Datasets:

bzantium

/

LongPerplexity

like 0

in recent years , density functional theory ( dft ) has emerged as a powerful tool for describing the structure of nuclei . based on the seminal works of kohn , hohenberg , and sham @xcite , dft was originally developed for use in condensed matter physics ; interestingly , it was discovered that it also provides a systematic framework for modeling the low - energy sector of the strong interaction @xcite . central to dft is the notion that , if the energy functional is considered a functional of the density , then by minimizing the energy one can acquire the exact ground - state density @xcite . furthermore , if the quantum many - body equations are replaced by single - particle equations with local , classical fields , one can exactly reproduce the total energy , scalar and vector densities , and chemical potential @xcite . thus the many - body problem is reduced to finding the proper form of the energy functional . as it turns out , the lagrangian of quantum hadrodynamics determines just such a functional . quantum hadrodynamics ( qhd ) is a low - energy effective theory of the strong interaction that uses the observed degrees of freedom at this energy scale ( hadrons ) . it has evolved from the early model of walecka @xcite into more modern incarnations that are based on chiral effective field theory ( eft ) and dft @xcite . the qhd energy functional is constructed as an expansion in powers of the mean meson fields , which are in fact kohn sham potentials , over a heavy mass , such as the chiral symmetry breaking scale , and includes all possible terms consistent with the underlying symmetries of qcd . field redefinitions are used to move the complexity of the problem ( e.g. , many - body forces and nucleon substructure ) into the meson field self - interactions . as this is an effective theory , the terms are characterized by unknown coefficients . once all the dimensional and combinatorial factors have been removed , however , the remaining dimensionless coefficients are assumed to be _ natural _ , or of order unity @xcite . the long - range dynamics are included explicitly , and short - range physics is contained in the parametrization ; therefore , this theory incorporates the natural separation of length scales . while qhd in principle contains all possible terms consistent with the underlying symmetries of the system , in practice it is a perturbative expansion for the energy functional that can be truncated at a manageable level ( since the ratio of the mean fields to the heavy mass is small ) . the now finite number of coefficients are fixed by experimental data , and this theory can then be used for predictive purposes @xcite . qhd is a strong - coupling theory . unlike qed , there is no obvious asymptotic expansion to use to obtain results and refine them systematically . it is thus unknown whether qhd permits any expansion for systematic computation and refinement of theoretical results . one possibility is the loop expansion , which was partially explored in @xcite . why a loop expansion ? the loop expansion is a simple and well - developed expansion scheme in powers of @xmath3 that is derived from the path integral @xcite . the mean meson fields are included non - perturbatively , and the correlations are included perturbatively . therefore , one can analyze the many - body effects order by order . previous work in qhd has shown that the mean - field terms dominate the nuclear energy , and exchange and correlation effects do not significantly modify the energy or nucleon self - energies , at least for states in the fermi sea @xcite . we stress that we are not certain that the loop expansion is practical ; the answer to this question is left for future consideration . however , the loop expansion has the advantage that it is fairly easy to separate the short - range and long - range dynamics and to analyze their structures . the loop expansion was investigated in the simplest case at the two - loop level @xcite ( only the most rudimentary couplings were retained ) . this was a severe truncation of the eft lagrangian and implies that we were interested only in the lowest level of accuracy . here we are also concerned with the loop expansion only up to two - loop order ( consideration of higher - loop effects is left for the future ) , but now nonlinear meson self - couplings and tensor terms are retained . note that the nonlinear terms incorporate short - range many - body nuclear forces , vacuum dynamics , and the effects of nucleon substructure . the extension of the level of truncation to include these terms allows us to describe nuclear matter ( and presumably nuclei ) with greater accuracy . the result of the loop expansion up to the two - loop level is the mean field energy density @xcite plus a series of integrals . we can categorize these integrals in two groups : baryon loops and purely mesonic loops . the baryon loops are integrals at the two - loop level that contain two factors of the baryon propagator and one factor of a meson propagator . each nucleon propagator can be separated into a feynman and a density part @xcite . this is accomplished by incorporating the proper pole structure of the propagator . the feynman part describes the propagation of a baryon or an antibaryon ; the density part involves only on - shell propagation in the fermi sea and accounts for the exclusion principle . as before @xcite , the two - loop baryon integrals can each be separated into three distinct parts : the exchange , lamb - shift , and vacuum - fluctuation contributions . the exchange term has two factors of the density portion of the nucleon propagator . the lamb - shift term contains both feynman and density parts of the nucleon propagator . the vacuum - fluctuation term contains two factors of the feynman propagator . the lamb - shift and vacuum - fluctuation parts are short - range physics , and it was shown in the simple case of @xcite that they can be expressed as a sum of terms that are already present in the eft lagrangian . this is an effective theory , so the coefficients of these terms are not yet determined ; thus , the two - loop terms are just absorbed into terms already present in the lagrangian and are not to be calculated explicitly @xcite . as a result , only the exchange portion of the two - loop integrals must be calculated explicitly . in addition , a number of purely mesonic loops arise when nonlinear meson field self - interactions are included . these terms , however , can be expressed as a power series in the meson fields with undetermined coefficients . as before , these terms are just absorbed into the lagrangian parameters and should not be calculated explicitly . it is interesting to note that the natural separation of the length scales in qhd and the parametrization of the short - distance dynamics is similar to infrared regularization @xcite . in infrared regularization , loop integrals are separated into infrared regular and singular contributions ( corresponding to short- and long - range physics ) . the regular portion is expanded as a power series in the momentum and is absorbed into the unknown coefficients of the underlying lagrangian ; the singular part must be calculated explicitly . this is directly analogous to what was done for the loop expansion in qhd @xcite . one question that remains unanswered is how the addition of these many - body correlations to the energy density affects the naturalness assumption , which states that once all the dimensional and combinatorial factors in a given term are accounted for , what remains is a dimensionless coefficient of order unity @xcite . once one decides on a level of truncation in the underlying lagrangian , the now finite number of unknown constants can be parametrized by the data . thus , one purpose of this work is to prove that sets of parameters exist at different levels of truncation in the two - loop energy density in which the parameters are all natural and thereby show that the naturalness assumption holds . for the purposes of this work , we will consider only uniform , symmetric nuclear matter , as in @xcite . here , however , we also include nonlinear meson self - interactions and tensor terms . these additional terms not only increase the number of two - loop integrals , but also complicate the meson propagators , due to the short - range and many - body physics mentioned earlier . the loop expansion is carried out to the two - loop level , and the resulting two - loop integrals are separated into long - range and short - range physics . the short - range contributions are expressed in forms that already appear in the lagrangian . since the coefficients of these terms have yet to be determined , they are just redefined , thus incorporating these new contributions into the mean field energy ( as in @xcite ) . as a result , they are already present in the one - loop qhd calculation . the long - range physics must be calculated explicitly . in this work , we fit the two - loop energy to empirical properties of nuclear matter . then we compare parameter sets constructed at the two - loop level with those of previous works obtained at the mean field level and consider the naturalness of the parameters . in this section , we investigate the effects of nonlinear meson self - interactions ( including isoscalar meson mixing ) on the loop expansion . here we will keep terms up to order @xmath4 in the isoscalar meson sector of the qhd lagrangian @xcite ( i.e. , @xmath5 , @xmath6 , @xmath7 , @xmath8 , and @xmath9 ) , and let @xmath10 . since we will calibrate the parameters using nuclear matter properties only ( and not finite nuclei ) , we want to keep the number of parameters at a manageable level . therefore , the baryon and pion parts of the lagrangian are unchanged from @xcite ( except for the addition of the two - pion nucleon vertex ) : @xmath11 + \left(m - g_{s}\phi\right)\right]\psi \nonumber \\ & & % - \frac{1}{2}\left(\partial_{\mu}\phi\right)^{2 } - \frac{1}{2}m_{s}^{2}\phi^{2 } % - \frac{1}{4}v_{\mu\nu}v_{\mu\nu } - \frac{1}{2}m_{v}^{2}v_{\mu}v_{\mu } - \frac{1}{2}\left(\partial_{\mu}\pi_{a}\right)^{2 } - \frac{1}{2}m_{\pi}^{2}\pi_{a}^{2 } % \nonumber \ , \label{eqn : lagrangian}\end{aligned}\ ] ] which follows from the _ chirally invariant _ lagrangian in @xcite by retaining only the relevant ( lowest - order ) terms in the pion fields . , we discuss the role of the tensor coupling to the neutral vector meson at the two - loop level . ] the neutral scalar and vector meson lagrangian is now @xmath12 where @xmath13 and @xmath14 . here @xmath15 are the fermion fields and @xmath16 , @xmath17 , and @xmath18 are the meson fields ( isoscalar - scalar , isoscalar - vector , and isovector - pseudoscalar , respectively ) . the heavy meson fields are also chiral scalars . note that in this work , the conventions of @xcite are used . as in @xcite , we are not making a chiral expansion in powers of the pion mass and include it for kinematical purposes only . the generating functional is defined in the usual way by @xmath19 & \equiv & \exp\left\{iw[j , j_{\mu}]/\hbar\right\ } \nonumber \\ & = & { \cal n}^{-1}\int d({\psibar})d(\psi)d(\phi)d(v_{\mu})d(\pi_{a } ) \nonumber \\ & & \times \exp\left\{\frac{i}{\hbar } \int d^{4}x\left[{\cal l}(x ) + j(x)\phi(x ) + j_{\mu}(x)v_{\mu}(x)\right]\right\}\ , \label{eqn:3}\end{aligned}\ ] ] where @xmath20 is the normalization factor ( in effect , the vacuum subtraction , which is also an expansion in loops ) , @xmath21 and @xmath22 are the external sources corresponding to the meson fields @xmath16 and @xmath17 , respectively , and the connected generating functional is @xmath23 $ ] . the path integrals of interest here are those over the _ isoscalar quantum fluctuation fields _ @xmath24 and @xmath25 ( the others are discussed in @xcite ) . with @xmath7 , @xmath26 , the integral is bilinear in form , or @xmath27 & & \left.\left . { } - m_{v}^{2}\left(\frac{\eta_{1}g_{s}}{m}v_{\mu}^{0 } + \frac{\eta_{2}g_{s}^{2}}{m^{2}}\phi_{0}v_{\mu}^{0}\right)\sigma\tilde{\eta}_{\mu } + u\sigma + u_{\mu}\tilde{\eta}_{\mu } \right]\vphantom{\int } \right\ } \ , \end{aligned}\ ] ] where the meson propagators in momentum space are @xmath28 { \cal d}_{\mu\nu}^{-1}(k ) & = & \left[k^{2 } + m_{v}^{2}\left(1 + \eta_{1}\frac{g_{s}\phi_0}{m } + \frac{\eta_{2}}{2}\frac{g_{s}^{2}\phi^{2}_{0}}{m^{2}}\right ) - \frac{\zeta_{0}}{6}g_{v}^{2}v_{\lambda}^{0}v_{\lambda}^{0}\right]\delta_{\mu\nu } - k_{\mu}k_{\nu } - \frac{\zeta_{0}}{3}g_{v}^{2}v_{\mu}^{0}v_{\nu}^{0 } \ . \nonumber \\\end{aligned}\ ] ] to simplify the equations , we introduce a five - component notation for the meson fields and their corresponding auxiliary sources @xcite : @xmath29 where @xmath30 and @xmath31 have @xmath32 . we can then rewrite the path integral as @xmath33\right\ } } & & \nonumber \\[5pt ] & = & \exp\left\{-\frac{1}{2}\tr\ ; \ln\left[a^{-1}\right]\right\ } \exp\left\{-\frac{i}{2}\ , t_{\alpha}a_{\alpha\beta}t_{\beta}\right\ } \ , \end{aligned}\ ] ] where the matrix @xmath34 is @xmath35 \eta v_{\mu}^{0 } & { \cal d}_{\mu\nu}^{-1}(k ) \end{array}\right ) \ , \ ] ] and @xmath36 if @xmath37 , then @xmath34 is diagonal , there is no mixing , and the system reduces to the non - mixing case , a simple example of which was discussed in @xcite . now we want to solve for the components of @xmath38 using @xmath39 : @xmath40 a_{\nu ' 0 } & a_{\nu'\mu } \end{array}\right ) \left(\begin{array}{cc } \delta_{s}^{-1}(k ) & \eta v_{\nu}^{0 } \\[5pt ] \eta v_{\mu}^{0 } & { \cal d}_{\mu\nu}^{-1}(k ) \end{array}\right ) \ .\end{aligned}\ ] ] this leads to the following four equations : @xmath41 + a_{0\mu}\eta v_{\mu}^{0 } & = & 1 \label{eqn : ab } \ , \\[5pt ] a_{00}\eta v_{\nu}^{0 } + a_{0\mu}\left[{\cal d}_{\mu\nu}^{-1}(k)\right ] & = & 0 \label{eqn : aa } \ , \\[5pt ] a_{\nu ' 0}\left[\delta_{s}^{-1}(k)\right ] + a_{\nu'\mu}\eta v_{\mu}^{0 } & = & 0 \label{eqn : ae } \ , \\[5pt ] a_{\nu ' 0}\eta v_{\nu}^{0 } + a_{\nu'\mu}\left[{\cal d}_{\mu\nu}^{-1}(k)\right ] & = & \delta_{\nu\nu ' } \ . \label{eqn : af}\end{aligned}\ ] ] solving these , we get ( @xmath42 ) @xmath43 & & \times \left . \left.\left(m_{v}^{2}-\frac{\zeta_{0}}{6}g_{v}^{2}v^{2}\right ) -\frac{\zeta_{0}}{3}g_{v}^{2}\left(k \cdot v\right)^{2}\right ] \right/ c \ , \label{eqn : ak}\end{aligned}\ ] ] @xmath44 \right/ c \ , \label{eqn : al}\ ] ] @xmath45 / c \ , \ ] ] @xmath46\delta_{\mu\nu'}\right.\right . \nonumber \\[5pt ] & & + \left(k^{2}+m_{v}^{2}-\frac{\zeta_{0}}{2}g_{v}^{2}v^{2}\right)k_{\mu}k_{\nu ' } + \frac{\zeta_{0}}{3}g_{v}^{2}\left(m_{v}^{2 } -\frac{\zeta_{0}}{6}g_{v}^{2}v^{2}\right)v_{\mu}^{0}v_{\nu'}^{0 } \nonumber \\[5pt ] & & \left.{}+\frac{\zeta_{0}}{3}g_{v}^{2}\left(k \cdot v\right ) \left(k_{\mu}v_{\nu'}^{0}+k_{\nu'}v_{\mu}^{0}\right)\right\ } \nonumber \\[5pt ] & & -\eta^{2}\left\{\left[\left(m_{v}^{2}-\frac{\zeta_{0}}{6}g_{v}^{2}v^{2}\right)v^{2 } + \left(k \cdot v\right)^{2}\right]\delta_{\mu\nu'}+v^{2}k_{\mu}k_{\nu ' } \right . \nonumber \\[5pt ] & & \left.\left . { } -\left(m_{v}^{2}-\frac{\zeta_{0}}{6}g_{v}^{2}v^{2}\right)v_{\mu}^{0}v_{\nu'}^{0 } -\left(k \cdot v\right)\left(k_{\mu}v_{\nu'}^{0}+k_{\nu'}v_{\mu}^{0}\right)\right\}\right ) / c \ , \label{eqn : ah}\end{aligned}\ ] ] where @xmath47 & & \times \left\{\left(k^{2}+m_{s}^{2}\right ) \left[\left(k^{2}+m_{v}^{2}-\frac{\zeta_{0}}{2}g_{v}^{2}v^{2}\right ) \left(m_{v}^{2}-\frac{\zeta_{0}}{6}g_{v}^{2}v^{2}\right ) -\frac{\zeta_{0}}{3}g_{v}^{2}\left(k \cdot v\right)^{2}\right]\right . \nonumber \\[5pt ] & & \left . { } - \eta^{2}\left[\left(m_{v}^{2}-\frac{\zeta_{0}}{6}g_{v}^{2}v^{2}\right)v^{2 } + \left(k \cdot v\right)^{2}\right]\right\ } \ , \label{eqn : ad}\end{aligned}\ ] ] and @xmath48 m_{v}^{2 } & \equiv & m_{v}^{2}\left(1 + \eta_{1}\frac{g_{s}\phi_{0}}{m } + \frac{\eta_{2}}{2}\frac{g_{s}^{2}\phi_{0}^{2}}{m^{2}}\right ) \ . \label{eqn : aj}\end{aligned}\ ] ] the corresponding path integral in the vacuum subtraction has a similar structure , except with vanishing classical meson fields ; therefore , @xmath49 is replaced by @xmath50 , or @xmath51 where @xmath52 and @xmath53 are the free propagators for the scalar and neutral vector mesons , respectively : @xmath54 the pion self - couplings do not contribute to the energy density if there is no pion condensate . therefore , we simply ignore the pure pion terms in the following calculation . the remaining generating functional is @xmath55 = { \cal n}^{-1 } \exp\left\{\frac{i}{\hbar}\int d^{4}x\left[{\cal l}_{0}(x ) + j(x)\phi_{0}(x ) + j_{\mu}(x)v_{\mu}^{0}(x)\right]\right\ } } & & \nonumber \\[5pt ] & & { } \times \exp\left\{\int d^{4}x\int\frac{d^{4}k}{(2\pi)^{4 } } \left(\tr \ ; \ln \left[g_{f}^{0}(k)g_{h}^{-1}(k)\right ] - \frac{1}{2}\,\tr \ ; \ln \left[a_{0}(k)a^{-1}(k)\right]\right)\right\ } \nonumber \\[5pt ] & & \times \left[\!\!\left [ \exp\left\{i\hbar^{1/2}\int d^{4}x \left[\frac{i\delta}{\delta \xi(x)}\right ] \left(g_{s}\left[\frac{-i\delta}{\delta u(x)}\right ] + ig_{v}\gamma_{\mu}\left[\frac{-i\delta}{\delta u_{\mu}(x)}\right ] \right.\right.\right.\right . \nonumber \\[5pt ] & & \left . \quad { } + i\frac{g_{a}}{2f_{\pi}}\gamma_{\mu}\gamma_{5}\partial_{\mu}\left [ \frac{-i\delta}{\delta \zeta_{a}(x)}\right ] \cdot\tau_{a}\right)\left[\frac{-i\delta}{\delta \bar{\xi}(x)}\right ] \nonumber \\[5pt ] & & { } -i\hbar^{1/2}\int d^{4}x\left(m_{s}^{2 } \left(\frac{\kappa_{3}g_{s}}{6m}+\frac{\kappa_{4}g_{s}^{2}}{6m^{2}}\phi_{0}\right ) \left[\frac{-i\delta}{\delta u(x)}\right]^{3 } \right . \nonumber \\ & & \left . { } + m_{v}^{2}\left(\frac{\eta_{1}g_{s}}{2m}+\frac{\eta_{2}g_{s}^{2}}{2m^{2}}\phi_{0}\right ) \left[\frac{-i\delta}{\delta u(x)}\right]\left[\frac{-i\delta}{\delta u_{\mu}(x)}\right]\left [ \frac{-i\delta}{\delta u_{\mu}(x)}\right]\right . \nonumber \\[5pt ] & & \left . \quad { } -\frac{\zeta_{0}}{6}g_{v}^{2}v_{\mu}^{0}\left [ \frac{-i\delta}{\delta u_{\mu}(x)}\right ] \left[\frac{-i\delta}{\delta u_{\nu}(x)}\right]\left [ \frac{-i\delta}{\delta u_{\nu}(x)}\right ] + m_{v}^{2}\frac{\eta_{2}g_{s}^{2}}{2m^{2}}v_{\mu}^{0}\left[\frac{-i\delta}{\delta u(x)}\right]^{2 } \left[\frac{-i\delta}{\delta u_{\mu}(x)}\right]\right ) \nonumber \\[5pt ] & & { } + \hbar\int d^{4}x\left ( \frac{1}{24}\ , \zeta_{0}g_{v}^{2 } \left[\frac{-i\delta}{\delta u_{\mu}(x)}\right]\left[\frac{-i\delta}{\delta u_{\mu}(x)}\right ] \left[\frac{-i\delta}{\delta u_{\nu}(x)}\right]\left[\frac{-i\delta}{\delta u_{\nu}(x)}\right ] -m_{s}^{2}\frac{\kappa_{4}g_{s}^{2}}{24m^{2}}\left[\frac{-i\delta}{\delta u(x)}\right]^{4 } \right . \nonumber \\[5pt ] & & \left.\left . { } - m_{v}^{2}\frac{\eta_{2}g_{s}^{2}}{4m^{2}}\left[\frac{-i\delta}{\delta u(x)}\right]^{2 } \left[\frac{-i\delta}{\delta u_{\mu}(x)}\right]\left[\frac{-i\delta}{\delta u_{\mu}(x)}\right ] \right)\right\ } \nonumber \\ & & { } \times \exp\left\{i\int d^{4}x\int d^{4}y \left[\frac{1}{2}\ , t_{\alpha}a_{\alpha\beta}(x - y)t_{\beta } \right . \right . \nonumber \\[5pt ] & & \left.\left.\left.\left . \quad { } + \frac{1}{2}\ , \zeta_{a}(x)\delta_{\pi}^{ab}(x - y)\zeta_{b}(y ) - \bar{\xi}(x)g_{h}(x - y)\xi(y)\right]\right\ } \right]\!\!\right]_{\mathrm{sources}\ = \ 0 } \ .\end{aligned}\ ] ] here @xmath56 is the pion source and @xmath57 is the nucleon source . we define the contribution at the mean field level as @xmath58 & & { } + \frac{1}{4!}\ , \zeta_{0}g_{v}^{2}\left(v_{\mu}^{0}v_{\mu}^{0}\right)^{2 } \ , \end{aligned}\ ] ] and the pion and fermion propagators are given by @xmath59 g_{f}^{0}(k ) & = & \frac{-1}{i{\not\ ! k}+m - i\epsilon } \ , \\[3pt ] g_{h}(k ) & = & \frac{-1}{i{\not\ ! k}-ig_{v}\gamma_{\mu}v_{\mu}^{0}+ \left(m - g_{s}\phi_{0}\right)}\ .\end{aligned}\ ] ] it is straightforward to acquire the connected generating functional at the one - loop order @xmath60 - \frac{1}{2}\ , \tr \ ; \ln \left[a_{0}(k)a^{-1}(k)\right]\right\}\ , \ ] ] and therefore @xmath61 & = & \int d^{4}x\left\{{\cal l}_{0 } + i\hbar\int\frac{d^{4}k}{(2\pi)^{4 } } \left(\vphantom{\int } \tr \ ; \ln \left[g_{f}^{0}(k)g_{h}^{-1}(k)\right ] \right.\right . \nonumber \\[5pt ] & & \left . \left . { } - \frac{1}{2}\ , \tr \ ; \ln \left[a_{0}(k)a^{-1}(k)\right]\right)\right\ } \ .\end{aligned}\ ] ] the baryon term is identical to the simple case in @xcite . to treat the isoscalar terms , we first consider the identity @xmath62 \equiv \ln \det\left[a_{0}(k)a^{-1}(k)\right ] \ , \ ] ] since it is easy to acquire the determinants : @xmath63 & = & \left[m_{v}^{2}\left(k^{2 } + m_{s}^{2}\right)\left(k^{2 } + m_{v}^{2}\right)^{3}\right]^{-1 } \ , \\[5pt ] \det\left[a^{-1}(k)\right ] & = & \left(k^2+m_{v}^{2}-\frac{\zeta_{0}}{6}g_{v}^{2}v_{e}^{2}\right)^{2 } \left\{\left(k^{2}+m_{s}^{2}\right)\left[\left(k^{2 } + m_{v}^{2}-\frac{\zeta_{0}}{6}g_{v}^{2}v_{e}^{2}\right)\right.\right . \nonumber \\[5pt ] & & \left . \times \left(m_{v}^{2}-\frac{\zeta_{0}}{2}g_{v}^{2}v_{e}^{2}\right ) + \frac{\zeta_{0}}{3}g_{v}^{2}\left[k^{2}v_{e}^{2}-\left(k \cdot v_{e}\right)^{2}\right]\right ] \nonumber \\[5pt ] & & \left . \quad { } - \eta^{2}\left[\left(m_{v}^{2}-\frac{\zeta_{0}}{6}g_{v}^{2}v_{e}^{2}\right)v_{e}^{2 } + \left(k \cdot v_{e}\right)^{2}\right]\right\ } \ , \end{aligned}\ ] ] where @xmath64 and @xmath65 are defined by eqs . ( [ eqn : ai ] ) and ( [ eqn : aj ] ) , and @xmath66 . apparently , @xmath67 $ ] is an analytic function of @xmath68 and @xmath69 , and @xmath70\left(\phi_{e}=0,v_{e}=0\right ) = \det\left[a_{0}^{-1}(k)\right ] \ .\ ] ] therefore , the meson contribution to the effective action can be categorized by the general form @xmath71 where the prime indicates that @xmath72 . however , terms of this type already appear in the general qhd lagrangian ; since the coefficients of these terms have yet to be determined , these meson contributions can be absorbed into the preexisting terms . as a result , they do nt need to be separately considered . now we want to determine the connected generating functional at the two - loop level , @xmath73 . for this , we have to make a legendre transformation to get the effective action . the classical meson fields are approximated as @xcite @xmath74 v_{\mu}^{0}(x ) \end{array}\right ) = \left(\begin{array}{c } \phi_{e}(x ) \\[5pt ] v^{e}_{\mu}(x ) \end{array}\right ) + \int d^{4}y a(x - y)\left(\begin{array}{c } \left.\displaystyle{\frac{\delta w_{1}}{\delta\phi_{0}(y)}}\right|_{e } \\[12pt ] \left.\displaystyle{\frac{\delta w_{1}}{\delta v_{\mu}^{0}(y)}}\right|_{e } \end{array}\right ) + o(\hbar^{2 } ) \ .\ ] ] one can verify that the tadpole diagrams in @xmath73 eventually cancel , and so we are left with only the one - particle irreducible diagrams : @xmath75 = \int d^{4}x\,{\cal l}_{0}\left[\phi_{e}(x),v_{\mu}^{e}(x)\right ] } & & \nonumber \\[5pt ] & & { } - i\hbar \int d^{4}x\int\frac{d^{4}k}{(2\pi)^{4 } } \left\{\tr \ ; \ln \left[g_{f}^{0}(k)g_{h}^{-1}(k)\right ] - \frac{1}{2}\ , \tr \ ; \ln \left[a_{0}(k)a^{-1}(k)\right]\right\ } \nonumber \\[5pt ] & & { } + \hbar^{2}\int d^{4}x \int\frac{d^{4}k}{(2\pi)^{4}}\int\frac{d^{4}q}{(2\pi)^{4 } } \left\{\frac{g_{s}^{2}}{2}a_{00}(k - q ) \tr\left[g_{h}(k)g_{h}(q)\right ] \right . \nonumber \\[5pt ] & & \quad { } -\frac{g_{v}^{2}}{2}a_{\mu\nu}(k - q ) \tr\left[g_{h}(k)\gamma_{\mu}g_{h}(q)\gamma_{\nu}\right ] \nonumber \\[5pt ] & & \quad { } -\frac{g_{a}^{2}}{2f_{\pi}^{2}}\ , \delta^{ab}_{\pi}(k - q ) \tr\left[g_{h}(k)(\not\ ! k - \not\ ! q)\gamma_{5}\frac{\tau_a}{2 } g_{h}(q)(\not\ ! k - \not\ ! q)\gamma_{5}\frac{\tau_b}{2}\right ] \nonumber \\[5pt ] % & & \quad { } -\frac{g_{a}^{2}}{4f_{\pi}^{2}}\left(\frac{5\gamma-8}{8}\right)i\delta^{ab}_{\pi}(k - q ) % \tr\left[g_{h}(k)\gamma_{\mu}(k - q)_{\mu}\gamma_{5}\frac{\tau_a}{2 } % g_{h}(q)\gamma_{\nu}(k - q)_{\nu}\gamma_{5}\frac{\tau_b}{2}\right ] % \nonumber \\[5pt ] & & \quad { } + ig_{s}g_{v}a_{0\mu}(k - q ) \tr\left[\gamma_{\mu}g_{h}(k)g_{h}(q)\right ] \nonumber \\[5pt ] & & \quad { } + 3\left(\kappa_{3}\frac{g_{s}m_{s}^{2}}{6 m } + \kappa_{4}\frac{g^2_{s}m_{s}^{2}\phi_0}{6m^2}\right)^{2 } a_{00}(k)a_{00}(q)a_{00}(k - q ) \nonumber \\[5pt ] & & \quad { } -\kappa_{4}\frac{g^{2}_{s}m_{s}^{2}}{8m^2}a_{00}(k)a_{00}(q ) \nonumber \\[5pt ] & & \quad { } + \frac{1}{4}\ , \eta^{2}\left[a_{00}(k - q)a_{\mu\nu}(k)a_{\mu\nu}(q ) + 2a_{0\mu}(k)a_{0\nu}(q)a_{\mu\nu}(k - q)\right ] \nonumber \\[5pt ] & & \quad { } + \frac{1}{2}\ , \eta\left(\kappa_{3}\frac{g_{s}m_{s}^{2}}{m } + \kappa_{4}\frac{g^2_{s}m_{s}^{2}\phi_0}{m^2}\right)a_{00}(k - q)a_{0\mu}(k)a_{0\mu}(q ) \nonumber \\[5pt ] & & \quad { } -\eta_{2}\frac{g_{s}^{2}m_{v}^{2}}{4m^{2 } } \left[a_{00}(k)a_{\mu\mu}(q)+2a_{0\mu}(k)a_{0\mu}(q)\right ] \nonumber \\[5pt ] & & \quad { } + \frac{1}{24}\ , \zeta_{0}g_{v}^{2}\left[a_{\nu\nu}(k)a_{\mu\mu}(q ) + 2a_{\mu\nu}(k)a_{\mu\nu}(q)\right ] \nonumber \\[5pt ] & & \quad { } + \frac{\eta_{2}^{2}g_{s}^{4}}{2m^{4}}\ , m_{v}^{4 } \left[a_{00}(k)a_{00}(q)v_{\mu}^{e}a_{\mu\nu}(k - q)v_{\nu}^{e } + 2a_{00}(k)v_{\mu}^{e}a_{\mu 0}(q)a_{0\nu}(k - q)v_{\nu}^{e}\right ] \nonumber \\[5pt ] & & \quad { } -m_{v}^{2}\frac{\eta_{2}\zeta_{0}g_{s}^{2}g_{v}^{2}}{6m^{2 } } \left[a_{0\lambda}(k)a_{\lambda\mu}(q)v_{\mu}^{e}v_{\nu}^{e}a_{\nu 0}(k - q ) + 2a_{0\lambda}(k)a_{\lambda 0}(q)v_{\mu}^{e}a_{\mu\nu}(k - q)v_{\nu}^{e}\right ] \nonumber \\[5pt ] & & \quad { } + \frac{1}{36}\ , \zeta_{0}^{2}g_{v}^{4}\left[a_{\mu\nu}(k)a_{\mu\nu}(q ) v_{\lambda}^{e}a_{\lambda\lambda'}(k - q)v_{\lambda'}^{e } \right . \nonumber \\[5pt ] & & \left.\left . \qquad\qquad\qquad { } + 2v_{\mu}^{e}a_{\mu\nu}(k)a_{\nu\lambda}(k - q ) a_{\lambda\lambda'}(q)v_{\lambda'}^{e}\right ] - \mathrm{vev } \vphantom{\int } \right\ } \ , \end{aligned}\ ] ] where vev is the vacuum subtraction . notice that there are now four distinct fermion loops at the two - loop level . these reduce to the familiar three loops from the non - mixing case when @xmath37 . with dimensional regularization , we can shift momentum variables to remove the dependence on the vector field from the fermion propagators and thereby replace @xmath76 by @xmath77 , where @xmath78 & = & \left(i{\not\ ! k}-m^{*}\right)\left[\frac{1}{k^{2}+{m^{*}}^{2}-i\epsilon } - \frac{i\pi}{e^{*}(k)}\delta[k_{4 } - e^{*}(k)]\theta(k_{f}-|\vec{k}|)\right ] \nonumber \\[5pt ] & \equiv & g_{f}^{*}(k ) + g_{d}^{*}(k ) \ . \label{eqn : bb3}\end{aligned}\ ] ] here @xmath79 . we can then separate each of the fermion loops [ using @xmath80 into three contributions : an exchange term , a lamb - shift term , and a vacuum - fluctuation term . the exchange terms will be evaluated explicitly later . for now , let us consider the lamb - shift term for the fermion loop involving scalar vector mixing : @xmath81 } & & \nonumber \\[5pt ] & \propto & \int\frac{d^{4}k}{(2\pi)^{4 } } \int\frac{d^{4}q}{(2\pi)^{4}}\ , \delta\left(q_{4 } - e^{*}(q)\right ) \theta\left(k_{f } - |\mathbf{q}|\right)\frac{i\pi}{e^{*}(q ) } \left[\frac{m^{*}(k+q)_{\mu}}{k^{2}+{m^{*}}^{2}}\right ] \nonumber \\[5pt ] & & { } \times \left[\left(k - q\right)^{2}+{m_{v}^{*}}^{2}\right ] \left(\left[\left(k - q\right)\cdot v\right](k - q)_{\mu } + { m_{v}^{*}}^{2}v_{\mu}\right ) \nonumber \\[5pt ] & & \times \left\ { \vphantom{\int}\!\left[\left(k - q\right)^{2}+m_{s}^{2}\right ] \left[\left(k - q\right)^{2}+{m_{v}^{*}}^{2}\right]\right . \nonumber \\[5pt ] & & \times \left(\left[\left(k - q\right)^{2}+{m_{v}^{*}}^{2 } + \frac{\zeta_{0}}{3}g_{v}^{2}v^{2}\right]{m_{v}^{*}}^{2 } -\frac{\zeta_{0}}{3}g_{v}^{2}\left[\left(k - q\right)\cdot v\right]^{2}\right ) \nonumber \\[5pt ] & & \left . { } - \eta^{2}\left(\left[\left(k - q\right)^{2}+{m_{v}^{*}}^{2}\right]{m_{v}^{*}}^{2}v^{2 } + \left[\left(k - q\right)^{2}+{m_{v}^{*}}^{2}+\frac{\zeta_{0}}{3}g_{v}^{2}v^{2}\right ] \left[\left(k - q\right)\cdot v\right]^{2}\right)\right\}^{-1 } \nonumber \\[5pt ] & \propto & \int\frac{d^{4}q}{(2\pi)^{4 } } \int\frac{d^{4}p}{(2\pi)^{4 } } \frac{\theta\left(k_{f } - |\mathbf{q}|\right)}{e^{*}(q ) } \frac{\delta\left(q_{4 } - e^{*}(q)\right ) } { \left[p^{2}+2\left(p\cdot q\right)+q^{2}+{m^{*}}^{2}\right ] } \nonumber \\[5pt ] & & { } \times \left[p^{2}+{m_{v}^{*}}^{2}\right ] \left[\left(p\cdot v\right)\left(p^{2}+2p\cdot q+{m_{v}^{*}}^{2}\right ) + 2{m_{v}^{*}}^{2}\left(q\cdot v\right)\right ] \nonumber \\[5pt ] & & \times \left\{\left(p^{2}+m_{s}^{2}\right)\left(p^{2}+{m_{v}^{*}}^{2}\right ) \left[\left(p^{2}+{m_{v}^{*}}^{2}+\frac{\zeta_{0}}{3}g_{v}^{2}v^{2}\right){m_{v}^{*}}^{2 } -\frac{\zeta_{0}}{3}g_{v}^{2}\left(p\cdot v\right)^{2}\right]\right . \nonumber \\[5pt ] & & \left . { } - \eta^{2}\left[\left(p^{2}+{m_{v}^{*}}^{2}\right){m_{v}^{*}}^{2}v^{2 } + \left(p^{2}+{m_{v}^{*}}^{2}+\frac{\zeta_{0}}{3}g_{v}^{2}v^{2}\right ) \left(p\cdot v\right)^{2}\right]\right\}^{-1 } \nonumber \\[5pt ] & \propto & \int\frac{d^{4}q}{(2\pi)^{4 } } \int\frac{d^{4}p}{(2\pi)^{4}}\frac{\theta\left(k_{f } - |\mathbf{q}|\right)}{e^{*}(q ) } \,\delta\left(q_{4 } - e^{*}(q)\right)\sum_{i\ , = \ , 0}^{\infty}\sum_{j\ , = \ , 0}^{\infty } f_{ij}\left(p^{2},q^{2},v^{2},m^{*}\right ) \nonumber \\[5pt ] & & \times \left(p\cdot q\right)^{i}\left(p\cdot v\right)^{j } \left\{\left(p\cdot v\right)\left[p^{2}+2\left(p\cdot q\right ) + { m_{v}^{*}}^{2}\right]+2{m_{v}^{*}}^{2 } \left(q\cdot v\right)\right\ } \nonumber \\[5pt ] & \propto & \int\frac{d^{4}q}{(2\pi)^{4 } } \frac{\theta\left(k_{f } - |\mathbf{q}|\right)}{e^{*}(q ) } \,\delta\left(q_{4 } - e^{*}(q)\right ) \widetilde{f}\left(m^{*},v^{2},q\cdot v\right ) \ , \end{aligned}\ ] ] where @xmath82 @xmath83 is the fermi momentum , and @xmath84 . we have made the substitution @xmath85 , which is allowed if we keep @xmath86 fixed in the @xmath87 integral . the function @xmath88 comes from the fractional integrand and is an analytic function of the variables @xmath89 , @xmath90 , @xmath91 , and @xmath92 . the function @xmath93 comes from the integral over @xmath94 and is analytic in the variables @xmath92 , @xmath91 , and @xmath95 , since @xmath96 is always spacelike . the on - shell condition @xmath97 is used to eliminate the dependence on @xmath90 in the last line . remember that @xmath98 . thus , after integrating over @xmath99 , the remaining expression is an analytic function of @xmath100 . the final result will be complicated , but should be explicitly dependent on some types of baryon densities . these terms can be directly related to terms in the effective lagrangian that are some combination of the nucleon fields ( with some derivatives ) multiplied by some lorentz - invariant polynomials in the mean meson fields ; as a result , they can be parametrized by local terms in the effective lagrangian . a simple case was illustrated in @xcite . the term with the pion propagator is the same as in the non - mixing case . the fermion loops with @xmath101 and @xmath102 will , in the case of @xmath37 , reduce to the equivalent loops with the pure neutral scalar and vector mesons from the non - mixing case . we expect that they will behave similarly in both the mixing and non - mixing cases , with the former simply leading to more general local combinations of the fields . _ thus the lamb - shift terms can be parametrized by terms already present in the lagrangian ( before truncation ) . _ the vacuum - fluctuation terms and all the pure meson loops have no explicit density dependence , so all the conclusions of the previous analysis @xcite are still applicable here . the short - range physics which is contained in the lamb - shift and vacuum - fluctuation terms is absorbed in preexisting terms in the qhd lagrangian . the long - range physics arising from the exchange term must be explicitly calculated ; these integrals are considered in the following section . we first consider the integral with the propagator from eq . ( [ eqn : ak ] ) . therefore , one can write @xmath103 \nonumber \\[5pt ] & = & \frac{\gamma g_{s}^{2}}{32\pi^{4}}\int_{0}^{k_{f } } \frac{|\mathbf{k}|^{2}d|\mathbf{k}|}{e^{*}(k ) } \int_{0}^{k_{f}}\frac{|\mathbf{q}|^{2}d|\mathbf{q}|}{e^{*}(q ) } \int_{-1}^{1}d(\cos\theta ) \nonumber \\[5pt ] & & \times \left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}|\cos\theta+{m^{*}}^{2}\right ] \nonumber \\[5pt ] & & \times \left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}|\cos\theta -2{m^{*}}^{2}+{m_{v}^{*}}^{2}\right ] \nonumber \\[5pt ] & & \times \left\{\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}|\cos\theta -2{m^{*}}^{2}+{m_{v}^{*}}^{2 } { } + \frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\right]{m_{v}^{*}}^{2 } \right . \nonumber \\[5pt ] & & \left.\left . -\frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\left[e^{*}(k)-e^{*}(q)\right]^{2 } \right\ } \right/ b \ , \end{aligned}\ ] ] where @xmath104 \nonumber \\[5pt ] & & \times \left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+{m_{v}^{*}}^{2}\right ] \nonumber \\[5pt ] & & \times \left(\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+{m_{v}^{*}}^{2 } + \frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\right]{m_{v}^{*}}^{2 } \right . \nonumber \\[5pt ] & & \left.\qquad { } -\frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2 } \left[e^{*}(k)-e^{*}(q)\right]^{2 } \right ) \nonumber \\[5pt ] & & { } + \eta^{2}v_{0}^{2}\left(\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+{m_{v}^{*}}^{2}\right ] { m_{v}^{*}}^{2}\right . \nonumber \\[5pt ] & & { } -\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}|\cos\theta-2{m^{*}}^{2}+{m_{v}^{*}}^{2 } + \frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\right ] \nonumber \\[5pt ] & & \quad \left . { } \times \left[e^{*}(k)-e^{*}(q)\right]^{2}\right ) \ . \label{eqn : den}\end{aligned}\ ] ] now we consider the integral with the propagator from eq . ( [ eqn : ah ] ) ; as a result , we get @xmath105 \nonumber \\[5pt ] & = & \frac{\gamma g_{v}^{2}}{32\pi^{4}}\int_{0}^{k_{f } } \frac{|\mathbf{k}|^{2}d|\mathbf{k}|}{e^{*}(k ) } \int_{0}^{k_{f}}\frac{|\mathbf{q}|^{2}d|\mathbf{q}|}{e^{*}(q ) } \int_{-1}^{1}d(\cos\theta ) \nonumber \\[5pt ] & & \times \left[\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+m_{s}^{2}\right ] \right . \nonumber \\[5pt ] & & \times \left\{2\left(\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+{m_{v}^{*}}^{2 } + \frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\right]{m_{v}^{*}}^{2}\right.\right . \nonumber \\[5pt ] & & \left . -\frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2 } \left[e^{*}(k)-e^{*}(q)\right]^{2}\right ) \nonumber \\[5pt ] & & \times \left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}|\cos\theta-2{m^{*}}^{2}\right ] \nonumber \\[5pt ] & & \left . { } + \frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}{m_{v}^{*}}^{2 } \left[e^{*}(k)e^{*}(q)+|\mathbf{k}||\mathbf{q}|\cos\theta+{m^{*}}^{2}\right]\right\ } \nonumber \\ & & -\eta^{2}v_{0}^{2}\left(2\left\{\left[e^{*}(k)-e^{*}(q)\right]^{2}-{m_{v}^{*}}^{2}\right\ } \left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}|\cos\theta-2{m^{*}}^{2}\right ] \right . \nonumber \\[5pt ] & & \left.\left.\left . -{m_{v}^{*}}^{2}\left[e^{*}(k)e^{*}(q)+|\mathbf{k}||\mathbf{q}|\cos\theta+{m^{*}}^{2}\right]\right)\right ] \right/ b \ . \end{aligned}\ ] ] lastly , we consider the integral with the mixed propagator from eq . ( [ eqn : al ] ) , which becomes @xmath106 \nonumber \\[5pt ] & = & -\frac{\gamma g_{s}g_{v}\eta m^{*}v_{0}}{16\pi^{4}}{m_{v}^{*}}^{2 } \int_{0}^{k_{f}}\frac{|\mathbf{k}|^{2}d|\mathbf{k}|}{e^{*}(k ) } \int_{0}^{k_{f}}\frac{|\mathbf{q}|^{2}d|\mathbf{q}|}{e^{*}(q ) } \int_{-1}^{1}d(\cos\theta ) \nonumber \\[7pt ] & & \times \left[e^{*}(k)+e^{*}(q)\right ] % \nonumber \\ % & & \times \left.\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+{m_{v}^{*}}^{2}\right ] \right/ b \ .\end{aligned}\ ] ] the formula for the pion contribution to the two - loop energy density is the same as in @xcite . in the previous section , we showed that the short- and long - range dynamics of the two - loop contributions can be separated even in the presence of various nonlinearities in the isoscalar meson fields . the short - range physics can be expressed as a series of terms that are already present in the underlying lagrangian ; thus , they are implicitly contained in the undetermined coefficients . the long - range exchange contributions must be calculated explicitly . the question remains what effect these new contributions have on the naturalness assumption . naturalness @xcite states that once all the dimensional and combinatorial factors in a given term of the lagrangian are removed , what is left is a dimensionless constant of order unity . to show that naturalness is not violated by the inclusion of the two - loop exchange integrals , we will show that sets of natural parameters exist that adequately reproduce the empirical properties of symmetric nuclear matter at equilibrium @xcite . .parameter sets used in this work compared with one - loop parameter sets at the same level of truncation . [ cols="^,^,^,^,^,^,^",options="header " , ] in _ conclusion _ , we have shown that the results of @xcite are still valid when various nonlinearities in the isoscalar meson fields are included in the lagrangian . the short- and long - range physics at the two - loop level in the loop expansion can be separated . the short - range ( local ) dynamics is parametrized by the undetermined coefficients of the lagrangian , like those in the nonlinear meson self - interactions , and is either removed by field redefinitions or fitted to empirical data . the long - range ( nonlocal ) physics must be calculated explicitly . the assumption of naturalness is shown to hold in the presence of the two - loop exchange integrals . the size of the two - loop contributions is roughly equal to third order in the mean field power counting . the inclusion of tensor contributions was also considered and found to be small . we thank our colleague j. d. walecka for valuable comments on the manuscript . this work was supported in part by the department of energy under contract no . de fg0287er40365 . in this section , we present the corrections to the theory arising from the tensor coupling to the neutral vector meson at the two - loop level , where @xmath107 after working out the variational derivatives , we acquire the energy density . now we can replace the hartree propagator by @xmath108 , as before . the lamb - shift and vacuum fluctuation terms are absorbed by the same process as before . ( the only difference here is more gamma matrices and they do nt influence the result . ) now we are left with only the exchange terms @xmath109 \nonumber \\ & = & -\frac{f_{v}^{2}g_{v}^{2}}{64\pi^{4}m^{2 } } \int_{0}^{k_{f}}\frac{|\mathbf{k}|^{2}d|\mathbf{k}|}{e^{*}(k ) } \int_{0}^{k_{f}}\frac{|\mathbf{q}|^{2}d|\mathbf{q}|}{e^{*}(q ) } \int_{-1}^{1}d(\cos\theta ) \nonumber \\[7pt ] & & \times \left(\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+m_{s}^{2}\right]\right . \nonumber \\ & & \times \left\{2\left(\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+m_{v}^{2}+\frac{\zeta_{0}}{2}g_{v}^{2}v_{0}^{2}\right]\right.\right . \nonumber \\ & & \times \left . \left[m_{v}^{2}+\frac{\zeta_{0}}{6}g_{v}^{2}v_{0}^{2}\right ] -\frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\left[e^{*}(k)-e^{*}(q)\right]^{2}\right ) \nonumber \\ & & \times \left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}| \cos\theta-{m^{*}}^{2}\right]^{2 } \nonumber \\ & & + \frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\left[m_{v}^{2}+\frac{\zeta_{0}}{6}g_{v}^{2}v_{0}^{2}\right ] \nonumber \\ & & \times \left(\left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}|\cos\theta-{m^{*}}^{2}\right ] \left[2e^{*}(k)e^{*}(q)-2{m^{*}}^{2}\right ] \right . \nonumber \\ & & \left.\left . -{m^{*}}^{2}\left[e^{*}(k)-e^{*}(q)\right]^{2}\right)\right\ } \nonumber \\ & & -\eta^{2}v_{0}^{2}\left[2\left\{\left[e^{*}(k)-e^{*}(q)\right]^{2 } -\left(m_{v}^{2}+\frac{\zeta_{0}}{6}g_{v}^{2}v_{0}^{2}\right)\right\}\right . \nonumber \\ & & \times \left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}|\cos\theta-{m^{*}}^{2}\right]^{2 } \nonumber \\ & & + \left(m_{v}^{2}+\frac{\zeta_{0}}{6}g_{v}^{2}v_{0}^{2}\right ) \left\{2\left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}|\cos\theta-{m^{*}}^{2}\right]\right . \nonumber \\ & & \left.\left.\left . \times \left[e^{*}(k)e^{*}(q)-{m^{*}}^{2}\right]-{m^{*}}^{2}\left[e^{*}(k)-e^{*}(q)\right]^{2}\right\}\right]\right ) / b \ , \nonumber \\ { \cal e}^{(2)}_{xt - ex } & = & \frac{f_{v}g_{v}^{2}}{2m}\int\frac{d^{4}k}{(2\pi)^{4 } } \int\frac{d^{4}q}{(2\pi)^{4 } } a_{\mu\nu}(k - q ) \nonumber \\ & & { } \times \tr\left[\gamma_{\mu}g_{d}^{*}(k)\sigma_{\nu\rho}\left(k - q\right)_{\rho}g_{d}^{*}(q)\right ] \nonumber \\ & = & -\frac{f_{v}g_{v}^{2}m^{*}}{32\pi^{4}m } \int_{0}^{k_{f}}\frac{|\mathbf{k}|^{2}d|\mathbf{k}|}{e^{*}(k ) } \int_{0}^{k_{f}}\frac{|\mathbf{q}|^{2}d|\mathbf{q}|}{e^{*}(q ) } \int_{-1}^{1}d(\cos\theta ) \nonumber \\[7pt ] & & \times \left(\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+m_{s}^{2}\right]\right . \nonumber \\ & & \times \left\{8\left(\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}| \cos\theta-2{m^{*}}^{2}+m_{v}^{2}+\frac{\zeta_{0}}{2}g_{v}^{2}v_{0}^{2}\right]\right.\right . \nonumber \\ & & \times \left . \left[m_{v}^{2}+\frac{\zeta_{0}}{6}g_{v}^{2}v_{0}^{2}\right ] -\frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\left[e^{*}(k)-e^{*}(q)\right]^{2}\right ) \nonumber \\ & & \times \left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}| \cos\theta-{m^{*}}^{2}\right ] \nonumber \\ & & + \frac{\zeta_{0}}{3}g_{v}^{2}v_{0}^{2}\left[m_{v}^{2}+\frac{\zeta_{0}}{6}g_{v}^{2}v_{0}^{2}\right ] \nonumber \\ & & \left . \times \left(\left[2e^{*}(k)e^{*}(q)-2|\mathbf{k}||\mathbf{q}|\cos\theta-2{m^{*}}^{2}\right ] + \left[e^{*}(k)-e^{*}(q)\right]^{2}\right)\right\ } \nonumber \\ & & -\eta^{2}v_{0}^{2}\left[8\left\{\left[e^{*}(k)-e^{*}(q)\right]^{2 } -\left(m_{v}^{2}+\frac{\zeta_{0}}{6}g_{v}^{2}v_{0}^{2}\right)\right\}\right . \nonumber \\ & & \times \left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}|\cos\theta-{m^{*}}^{2}\right ] \nonumber \\ & & + \left(m_{v}^{2}+\frac{\zeta_{0}}{6}g_{v}^{2}v_{0}^{2}\right ) \left\{2\left[e^{*}(k)e^{*}(q)-|\mathbf{k}||\mathbf{q}|\cos\theta-{m^{*}}^{2}\right]\right . \nonumber \\ & & \left.\left.\left . + \left[e^{*}(k)-e^{*}(q)\right]^{2}\right\}\right]\right ) / b \ , \end{aligned}\ ] ] p. hohenberg and w. kohn , phys . * 136b * ( 1964 ) 864 . w. kohn and l. j. sham , phys . * a140 * ( 1965 ) 1133 . w. kohn , rev . * 71 * ( 1999 ) 1253 . r. m. dreizler and e. k. u. gross , density functional theory , springer , berlin ( 1990 ) . c. speicher , r. m. dreizler , and e. engel , ann . * 213 * ( 1992 ) 312 . r. n. schmid , e. engel , and r. m. dreizler , phys . c * 52 * ( 1995 ) 164 . r. j. furnstahl , b. d. serot , and h .- b . tang , nucl . * a615 * ( 1997 ) 441 . r. j. furnstahl , b. d. serot , and h .- b . tang , nucl . * a640 * ( 1998 ) 505 ( e ) . b. d. serot and j. d. walecka , int . j. mod . e * 6 * ( 1997 ) 515 . j. d. walecka , theoretical nuclear and subnuclear physics , second ed . , world scientific , singapore ( 2004 ) . c. o. almbladh and u. von barth , phys . rev . b * 31 * ( 1985 ) 3231 . j. d. walecka , ann . * 83 * ( 1974 ) 491 . r. j. furnstahl and b. d. serot , nucl . phys . * a671 * ( 2000 ) 447 . r. j. furnstahl , int . j. mod . b * 17 * ( 2003 ) 5111 . b. d. serot , in : lecture notes phys . * 641 * , g. a. lalazissis , p. ring , and d. vretenar , eds . , springer , berlin , heidelberg ( 2004 ) , p. 31 . r. j. furnstahl , in : lecture notes phys . * 641 * , g. a. lalazissis , p. ring , and d. vretenar , eds . , springer , berlin , heidelberg ( 2004 ) , p. 1 . a. manohar and h. georgi , nucl . phys . * b234 * ( 1984 ) 189 . h. georgi , phys . * b298 * ( 1993 ) 187 . j. l. friar , d. g. madland , and b. w. lynn , phys . c * 53 * ( 1996 ) 3085 . j. l. friar , few body syst . * 22 * ( 1997 ) 161 . j. mcintire , phys . rev . c * 66 * ( 2002 ) 064319 . j. mcintire , acta phys . pol . b * 35 * ( 2004 ) 2261 . j. mcintire , ` arxiv : nuc - th/0507006 ` . m. a. huertas , phys . c * 66 * ( 2002 ) 024318 . m. a. huertas , phys . c * 67 * ( 2003 ) 019901 ( e ) . m. a. huertas , acta phys . b * 34 * ( 2003 ) 4269 . m. a. huertas , acta phys . pol . b * 35 * ( 2004 ) 837 . r. j. furntahl , r. j. perry , and b. d. serot , phys . c * 40 * ( 1989 ) 321 . k. wehrberger , r. wittman , and b. d. serot , phys . c * 42 * ( 1990 ) 2680 . b. d. serot and h .- b . tang , phys . c * 51 * ( 1995 ) 969 . y. hu , j. mcintire , and b. d. serot , ` arxiv:0706.1547 [ nucl - th ] ` . p. a. m. dirac , phys . z. sowjetunion * 3 * ( 1933 ) 64 . c. itzykson and j .- b . zuber , quantum field theory , mcgraw hill , new york ( 1980 ) . b. d. serot and j. d. walecka , adv . * 16 * ( 1986 ) 1 . s. coleman and e. weinberg , phys . d * 7 * ( 1973 ) 1888 . j. iliopoulos , c. itzykson , and a. martin , rev . * 47 * ( 1975 ) 165 . s. coleman , new phenomena in subnuclear physics , plenum , new york ( 1977 ) . tang , ` arxiv : hep - ph/9607436 ` . p. j. ellis and h .- b . tang , phys . c * 57 * ( 1998 ) 3356 . t. becher and h. leutwyler , eur . j. c * 9 * ( 1999 ) 643 . t. becher , ` arxiv : hep - ph/0011079 ` . s. scherer , adv . * 27 * ( 2003 ) 277 . t. fuchs , m. r. schindler , j. gegelia , and s. scherer , phys . * b575 * ( 2003 ) 11 . m. r. schindler , j. gegelia , and s. scherer , phys . * b586 * ( 2004 ) 258 . m. r. schindler , j. gegelia , and s. scherer , nucl . * b682 * ( 2004 ) 367 . s. scherer , int . j. mod . phys . * a21 * ( 2006 ) 881 . d. djukanovic , j. gegelia , and s. scherer , eur . j. * a29 * ( 2006 ) 337 . s. a. chin , ann . * 108 * ( 1977 ) 301 . r. j. furnstahl , nucl . * a706 * ( 2002 ) 85 . j. mcintire , ` arxiv:0708.2860 ` .

two - dimensional ( 2d ) superconducting arrays , weakly coupled by josephson junctions , have attracted much attention @xcite . in the classical array , where the charging energy is neglected , the relevant variable is the phase of the superconducting order parameter at each grain , and the logarithmic interaction between vortices is well known to lead to the berezinskii - kosterlitz - thouless ( bkt ) transition @xcite . recent development of the fabrication techniques , on the other hand , allows to make regular arrays composed of very small superconducting grains . in such arrays the charging energy can not be neglected any longer @xcite , and studies considering both the self - capacitance @xmath0 and the junction capacitance @xmath1 are required . in the presence of ohmic dissipation , the charges , which are conjugate to the phase variables , take continuous values and the corresponding phase diagrams have been obtained through the use of a variational method @xcite . this has revealed the existence of a low - temperature reentrant transition , which appears consistent with the recent quantum monte carlo study @xcite . without dissipation , in contrast , the charges change in discrete quanta @xmath2 with @xmath3 being the electron charge . in this case the phase diagrams have been investigated via the coarse - graining approximation , which is mean - field - like in nature @xcite . in general , the mean - field approximation does not give reliable results in two dimensions , where fluctuations are too strong to be neglected . it is thus desirable to study the 2d superconducting arrays , with general capacitance matrices and discrete charge states , beyond the mean - field approximation . the quantum phase transitions in the superconducting arrays have drawn much interest also in relation to the bose - hubbard model , which describes strongly interacting bosons under the competition between the kinetic - energy and the potential - energy effects . here the kinetic energy and the potential energy correspond to the josephson energy and the charging energy in the superconducting array , respectively . in the presence of the on - site potential , the system at zero temperature exhibits a transition between the mott insulating phase and the superfluid phase , displaying lobe - like structures in the resulting phase diagram @xcite . in addition , nearest - neighbor interactions and next nearest - neighbor interactions produce various phases such as the checker - board - type solid , the striped solid , and the supersolid . in particular , the supersolid phase , which is characterized by superfluidity in the presence of underlying commensurability , has been studied extensively @xcite . this paper investigates the phase transitions in 2d superconducting arrays with emphasis on the effects of charge frustration and on the competition between the self- and junction capacitances . at zero temperature , we use the perturbation expansion and the simulated annealing method to obtain various insulating lobes . here rational charge frustration introduces commensurability effects to the array , leading to the solid phases with various charge densities . the detailed phase boundaries depend crucially on the ratio of the junction to self - capacitances , @xmath4 : for example , the insulating lobes become narrower and their central positions approach the charge frustration values , as @xmath4 is increased . further , as the system size is increased , more and more insulating lobes are expected to be observed in the phase diagram , suggesting the interesting possibility of the lobe at every rational frustration in the thermodynamic limit . to check the robustness of the lobes against finite temperatures , we apply the monte carlo ( mc ) method to the system without the josephson term and find the persistence of the insulating lobes with low - order rational values of the charge density . we also use a variational method to obtain an effective classical hamiltonian describing the system at finite temperatures . the corresponding phase diagrams are found approximately in a self - consistent manner , which again display lobes with the half - filled charge density . in the absence of charge frustration , the effective classical hamiltonian has only real terms , which allows us to use the mc method : we obtain the critical strengths of the josephson coupling and compare them with ones obtained from the self - consistent approximation . in the self - charging limit , the possibility of a reentrant transition is also pointed out . this paper is organized as follows : section [ sec_model ] introduces the quantum mechanical hamiltonian for the superconducting arrays with general capacitance matrices and discrete charge states . such a superconducting array is described by quantum mechanical conjugate variables , charges and phases , and charge frustration is induced by applying an external voltage between the array and the substrate . in sec . [ sec_zero ] , we study the zero - temperature phase transitions in the system via the perturbation expansion together with the simulated annealing method . the phase transitions at vanishing charge density is investigated in sec . [ subsec_0 ] , where the results in the nearest - neighbor charging limit are also compared with those of the previous works . in sec . [ subsec_pq ] , phase transitions at finite charge densities are studied , and the nature as well as the existence of the supersolid phase is discussed . section [ sec_finite ] is devoted to the investigation of the finite - temperature phase transitions by means of a variational method and mc simulations . the system without the josephson coupling is described by only charge variables , and we present the corresponding mc results in sec . [ subsec_mc ] , where the charge densities are obtained as functions of the charge frustration at various temperatures . the resulting step structures at zero temperature are compared with those obtained in sec . [ subsec_pq ] . whereas the phase diagrams in the presence of the josephson coupling are obtained in a self - consistent manner in sec . [ subsec_var ] , the comparison with mc results reveals that the self - consistent approximation is accurate only at sufficiently high temperatures . finally , a summary of the paper is given in sec . [ sec_conclusion ] . we consider an @xmath5 ( @xmath6 ) square array of josephson junctions . the extra charge @xmath7 on the superconducting grain at site @xmath8 can be written as ' '' '' ' '' '' @xmath9 where @xmath0 and @xmath1 are the self- and the junction capacitances , respectively , and @xmath10 is the electric potential of the grain @xmath8 with respect to the substrate . if we apply an external voltage @xmath11 between the substrate and the ground , eq . ( [ eq_q ] ) is changed to @xmath12+@xmath13 , with the `` gauge charge '' @xmath14 . the hamiltonian describing such a superconducting array consists of two parts : @xmath15 ' '' '' 2 where @xmath16 represents the phase of the superconducting order parameter at site @xmath8 . note that the charge @xmath7 and the phase @xmath16 are quantum mechanical conjugate variables satisfying the commutation relation @xmath17 = 2ei$ ] ( @xmath18 ) . in the charging energy term @xmath19 , the charges at sites @xmath8 and @xmath20 interact via the inverse of the capacitance matrix @xmath21 whereas the summation in the josephson energy term @xmath22 is to be done over all the nearest neighboring pairs . when the lattice constant of the array approaches zero , we can write eq . ( [ eq_q ] ) in the continuum form @xmath23 and obtain @xmath24 due to a single positive charge at the origin @xcite : @xmath25 with the modified bessel function @xmath26 , where the screening length @xmath27 ( @xmath28 ) measures the interaction range between charges . the above hamiltonian in eq . ( [ eqh1 ] ) is closely related to the bose - hubbard hamiltonian and to the spin-1/2 @xmath29 hamiltonian @xcite . the charging energy term in eq . ( [ eqh1 ] ) drives the system to arrange its discrete charges such that the average charge per site is as close to @xmath13 as possible , and the josephson energy corresponds to the hopping energy in the bose - hubbard model @xcite . consequently , if the charging energy is sufficiently larger than the josephson energy , hopping is suppressed and all the charges are arranged to form a periodic lattice , leading to the insulating phase at zero temperature . in particular , when @xmath30 , the energy eigenstate can be chosen also as a simultaneous eigenstate of the charge operator since @xmath31=0 $ ] . the uncertainty relation between the charge and the phase then asserts that phase coherence and consequently , superconductivity can not be attained in this limit . in the opposite limit of the strong josephson coupling , all the charges are in the extended state owing to the dominant hopping term , and the superconducting phase is favored . for convenience , we write @xmath32 , @xmath33 , @xmath34 , and @xmath35 , and obtain eq . ( [ eqh1 ] ) in the form ' '' '' ' '' '' @xmath36 ' '' '' 2 where the inverse of the dimensionless capacitance matrix is given by @xmath37 in general the state which minimizes the energy determines the zero - temperature phase of the system . suppose that we have obtained the energy levels of the system which correspond to the ground state , the first excited state , etc . if we introduce a small change in the control parameters , the energy levels shift by a small amount , and it is possible that the ground state and the first excited state cross each other as we pass through a set of values of the control parameters ; this `` level crossing '' leads to a zero - temperature phase transition in the parameter space @xcite . it should be noted here that consideration of only the two low - lying states , the ground state and the first excited one , may not be sufficient . figure [ fig_sch ] shows schematic dependence of the energy levels @xmath38 on the control parameter @xmath39 in two possible cases . in fig . [ fig_sch ] ( a ) , the ground state and the first excited state cross at the critical value @xmath40 of the control parameter , which corresponds to the true transition point . in fig . [ fig_sch ] ( b ) , on the other hand , the true transition point is given by @xmath41 , where the ground state and the second excited one cross each other . in this case consideration of only the two low - lying states would lead to @xmath40 , which is not the true transition point . to find the true transition point , we thus need to consider all the energy levels , which is not possible in practice . despite this , considering the level crossing of the ground state and the first excited state is very useful and convenient to obtain the upper limit of the true transition point . in terms of charges the system can be either in the incommensurate conducting phase or in the commensurate insulating phase ; in the latter charges are expected to form a superlattice which covers periodically the whole system . if the commensurate state has energy lower / higher than that of the incommensurate state for given values of parameters , the system should be in the insulating / conducting phase . ( here @xmath42 , @xmath43 , and @xmath4 constitute the control parameters . ) accordingly , for given total number of charges , it is needed to find the charge configurations of the commensurate state and of the low - lying incommensurate states . the symmetry of the hamiltonian under the transformations @xmath44 allows us to consider @xmath43 only in the range @xmath45 . in this restricted range @xmath46 , the commensurate states of the system may be classified in terms of the average number of charges , @xmath47 @xmath48 . for @xmath49 with relatively prime integers @xmath50 and @xmath51 , the charges are expected to form a @xmath52 superlattice , on the unit cell of which @xmath53 charges , each with the unit charge , are located . we first investigate the phase transitions in the system with @xmath54 . the energy of the @xmath54 commensurate state is computed via the second - order perturbation expansion with the josephson term as a perturbation in eq . ( [ eqh2 ] ) , and is written as @xmath55 without the perturbation , the hamiltonian of the system contains only charge operators and yields the eigenfunction ' '' '' ' '' '' @xmath56 2 [ ] which is simply the plane - wave state of the usual free particle system , with the interpretation of @xmath16 and @xmath57 as the position and the momentum , respectively . for the commensurate state with @xmath58 , all the charge eigenvalues @xmath57 s are zero and it is easy to compute the zeroth - order contribution @xmath59 and the first - order contribution @xmath60 : @xmath59 = @xmath61@xmath62 = @xmath63 and @xmath60 = @xmath64 = @xmath65 @xmath66 the second - order contribution is given by @xmath67 where @xmath68 is the zeroth - order energy for the charge configuration @xmath69 , and ' '' '' ' '' '' @xmath70 \nonumber \\ & = & -\frac{e_j}{2}\sum_{\langle i , j\rangle}\left[\prod_{k\ne i , j } \delta(q'_k,0)\right ] \left[\delta(q'_i,1 ) \delta(q'_j,-1 ) + \delta(q'_j,1 ) \delta(q'_i,-1 ) \right ] .\end{aligned}\ ] ] ' '' '' 2 [ ] inserting eq . ( [ eq_v ] ) into eq . ( [ eq_ec2 ] ) , we obtain the energy of the @xmath54 commensurate state to the 2nd order in @xmath71 : @xmath72 we next consider the incommensurate state with @xmath73 . among the various excitations present in the system with the constraint @xmath54 , the lowest excitation is expected to be point - like and we consider two types of excitations : one is the creation of a single charge ( sc ) at one site @xcite , and the other is the creation of a charge dipole ( cd ) at one bond [ @xmath74 , @xmath75 with @xmath76 being the nearest - neighboring pair ] . both excitations do not change the value of @xmath77 in the thermodynamic limit ( @xmath78 ) . in the self - charging limit ( @xmath79 ) , the interaction matrix @xmath21 takes a diagonal form and the first excited state should be of the sc type , since the cd type excitation requires the creation energy of two charges . on the other hand , in the nearest - neighbor ( nn ) charging limit ( @xmath80 ) , only the excitations which does not change the total number of charges are allowed and the first excited state is expected to be of the cd type . accordingly , as the value of @xmath4 is increased , the first excited state should change from the sc type to the cd type at a critical value of @xmath4 . in particular , the system in the nn charging limit displays the charge - anticharge unbinding transition @xcite , the bkt nature of which is manifested by the square - root cusp in the resistance @xcite . this charge bkt transition is suppressed by the presence of the self - capacitance , making the screening length finite , and it seems likely that the change of the first excited state according to the value of @xmath4 in turn leads to the change in the nature of the transition . between the above two types of excitations , the lowest one is determined from the comparison of the zeroth - order energies of both types : if @xmath81 given by ' '' '' ' '' '' @xmath82 is negative / positive , the creation of a single charge / charge dipole is the lowest excitation . figure [ fig_1st ] shows the boundary for @xmath83 in the limit of @xmath84 . it is found that for @xmath85 the lowest excitation is of the sc type regardless of the value of @xmath43 , and that the region of the cd type excitation ( the region below the curve in fig . [ fig_1st ] ) is quite small , allowed only for @xmath86 . therefore we below focus on the sc type excitation for @xmath87 and @xmath88 , and then consider the cd type excitation only in the nn charging limit ( @xmath89 ) with @xmath90 . we write the energy of the first excited state of the sc type as @xmath91 and the eigenstate as @xmath92 which is expected to be a superconducting state since the extra charge at site @xmath93 can hop to any site without energy cost . inserting eq . ( [ eqql ] ) to @xmath19 in eq . ( [ eqh2 ] ) , we obtain the zeroth - order contribution : @xmath94 = @xmath95 , where we have used the translational symmetry @xmath96 and @xmath97 . since the above excited state in eq . ( [ eqql ] ) has an @xmath98-fold degeneracy ( @xmath93 can take values @xmath99 ) , we use the degenerate perturbation theory to calculate the first - order contribution @xmath100 . we thus need to diagonalize the matrix @xmath22 whose components are given by @xmath101 for that purpose , it is convenient to use the fourier transform : @xmath102 which gives straightforwardly the eigenvalues @xmath103 . the first - order contribution is then given by the lowest eigenvalue : @xmath104 . for this lowest - lying excited state with @xmath105 , it is also straightforward to obtain the second - order contribution , which leads to the energy of the incommensurate state @xmath106 here @xmath107 denotes the summation over @xmath8 and its four nearest - neighbors @xmath20 with the restriction @xmath108 . the zero - temperature transition occurs when the energy of the commensurate state given by eq . ( [ eq_ec0 ] ) and that of the incommensurate state given by eq . ( [ eqeic0 ] ) becomes equal . this reveals the zero - temperature phase transition in the system with @xmath54 , as @xmath43 is varied . namely , the system becomes superconducting as @xmath43 is increased beyond the critical value @xmath109 , which is given by @xmath110 the resulting phase boundary between the insulating phase and the superconducting phase is shown in fig . [ q0_bound ] , obtained for a @xmath111 array . the junction capacitance here tends to enhance superconductivity , which is in accord with the result in ref . . it is also obvious that as the system approaches the nn charging limit , the insulating region shrinks to zero ; _ this apparently cures the well - known artifact of the coarse - graining approximation _ , where the system remains in the insulating phase for arbitrarily large values of @xmath71 in the nn charging limit @xcite . to check the validity of the second - order perturbation expansion , we have computed the third - order contribution in the self - charging limit , and obtain the critical value @xmath109 : @xmath112 2 [ ] the three curves in fig . [ 3rd_bound ] represent the phase boundaries obtained from eq . ( [ eq_3rd ] ) up to the first- , the second- , and the third - order contributions , respectively . it is shown that the third - order contribution does not change significantly the boundary obtained from the second - order calculation , inducing 8.3% of the difference in the critical value of @xmath42 at @xmath90 . we next consider the nn charging limit in the absence of charge frustration ( @xmath90 ) . in this case , eq . ( [ eq_ec0 ] ) gives the energy of the commensurate state in the form @xmath113 where @xmath114 and @xmath115 in the thermodynamic limit , we replace the discrete sum by the integral and obtain ' '' '' ' '' '' @xmath116 ' '' '' 2 [ ] leading to @xmath117 . the first - excited incommensurate state in the nn charging limit is of the cd type ( see fig . [ fig_1st ] ) with the zeroth - order energy @xmath118=@xmath119-@xmath120 = @xmath121 and the null first - order energy @xmath122 . it is complicated but straightforward to calculate @xmath123 from the second - order degenerate perturbation theory , which leads to a @xmath124 matrix . the fourier transform then yields @xmath98 @xmath125 hermitian matrices , each of which can be diagonalized . the lowest eigenvalue then gives the energy of the cd type incommensurate state , up to the second order : @xmath126 where the number 62.000 has been obtained numerically for a sufficiently large array ( @xmath127 ) ; at this size finite - size effects have been confirmed to be negligible . the critical value of @xmath128 , below which the array is in the insulating phase , is obtained from the condition @xmath129 : @xmath130 note that this value is somewhat larger than the critical value @xmath131 obtained in refs . . one possible explanation of the above discrepancy is the failure of the second - order perturbation expansion . namely , consideration of higher - order terms may eliminate the discrepancy . the other possibility is that the cd type excitation does not give the true transition point : see fig . [ fig_sch ] ( b ) for a schematic picture , where @xmath40 may correspond to the obtained transition point @xmath132 and @xmath41 the true transition point . in experiment , the even larger critical value @xmath133 has been obtained , via the curve - fit to the square - root cusp form of the resistance @xcite . recent quantum monte carlo simulations , on the other hand , have yielded the result @xmath134 . in the nn charging limit , there exists an interesting duality between the charge and the vortex @xcite , which concludes that in the absence of the josephson term , the charges form a superlattice determined by the value of charge frustration @xmath43 . this is a counterpart of the vortex superlattice in the classical array without the charging energy , formed in the presence of magnetic frustration @xcite . in the system considered here , on the other hand , the non - vanishing self - capacitance gives the interaction range between charges @xmath136 in units of the lattice constant , thus making it finite @xcite . accordingly , we expect that the formation of the @xmath52 charge superlattice with @xmath137 is improbable in the system with the rational charge frustration @xmath138 . this implies that @xmath77 is not necessarily equal to @xmath43 in the array with non - vanishing @xmath0 , which reflects that a short - ranged potential in general makes the commensurate phase extend over a finite range of the chemical potential . ( note that the charge frustration in this work plays the role of the chemical potential . ) we now investigate the phase transitions for @xmath139 @xmath140 , using the canonical ensemble with the total number of charges fixed . to find the zero - temperature phase diagrams we need to know the configurations of the commensurate states and of the lowest - lying incommensurate states , which is a highly nontrivial problem in the 2d system considered here , especially for large @xmath51 . for the 1d josephson junction array with @xmath141 , the commensurate state , the first excited state , and the resulting zero - temperature phase boundaries have been obtained via the perturbation expansion @xcite . we write the energy of the commensurate state with @xmath142 in the form @xmath143 where the relations @xmath144 and @xmath97 give @xmath145 . in general , when @xmath146 , the commensurate ground states are degenerate , which makes it necessary to consider the matrix element : ' '' '' ' '' '' @xmath147 & & \left [ \delta\left(q_i^{(l)},q_i^{(m ) } - 1\right ) \delta\left(q_j^{(l)},q_j^{(m ) } + 1\right ) \right . \nonumber \\ & & + \left . \delta\left(q_i^{(l)},q_i^{(m ) } + 1\right ) \delta\left(q_j^{(l)},q_j^{(m ) } - 1\right ) \right ] \end{aligned}\ ] ] with @xmath148 being the @xmath93th degenerate commensurate state . note that the configuration of the superlattice unit cell in a degenerate commensurate state is in general different from that in a different ( degenerate ) state . therefore , for @xmath149 , @xmath150 differs from @xmath151 at the total of @xmath152 sites for finite @xmath51 , leading to @xmath153 . further , it is obvious that @xmath154 . it is thus concluded that all the components of the matrix @xmath22 vanish , which results in @xmath155 . the second - order contribution is given by @xmath156 where @xmath157 and @xmath158 is the zeroth - order energy for given charge configuration @xmath69 and @xmath159 , respectively , and the summation @xmath160 is over the states @xmath159 making the denominator nonzero . the superscript @xmath93 on @xmath69 has been dropped since all the degenerate commensurate states give the same contribution to @xmath161 . namely , the degeneracy is not removed in a finite order since @xmath150 and @xmath151 can not be connected by a finite number of charge hoppings . with eq . ( [ eq_h ] ) , eq . ( [ eq_i ] ) reads @xmath162 where @xmath163 @xmath164 @xmath165 is obtained from @xmath69 by a charge hopping between sites @xmath8 and @xmath20 . it is easy to compute the denominator in eq . ( [ eq_j ] ) : @xmath166 \nonumber \\ & = & 8e_0\left [ \sum_l\left ( { \widetilde c}_{jl}^{-1 } -{\widetilde c}_{il}^{-1 } \right)q_l - \left ( { \widetilde c}_{00}^{-1 } - { \widetilde c}_{10}^{-1}\right ) \right ] , \end{aligned}\ ] ] which leads to the expression for the energy of the commensurate state with @xmath167 : @xmath168 here @xmath57 is the charge at site @xmath8 in the commensurate state with @xmath169 . since the ( commensurate ) ground state configurations are not known for general @xmath170 , @xmath57 in this work is obtained via the simulated annealing method . to compute the energies of the incommensurate states with @xmath171 , we classify excitations according to the number @xmath172 of extra charges ( with the sign taken into account ) , and let @xmath173 denote the energy of this excited state which has the total number of charges @xmath174 . to the second order in @xmath71 , it is straightforward to obtain the energy of the @xmath172-charge excited state : @xmath175 ' '' '' 2 where @xmath176 is the charge at site @xmath8 in the @xmath172-charge excited state , and the first - order contribution @xmath177 has been observed to vanish in numerical calculations , regardless of the values of @xmath170 @xmath178 and @xmath172 considered in this work . further , in eq . ( [ eq_eic ] ) , we have assumed that degeneracy is not removed in the second order although there are some exceptions , e.g. , for @xmath170=1/2 and @xmath179 ( see below ) . to find the phase boundary , we need to know the charge configuration which corresponds to the first excited state for given values of @xmath4 and @xmath43 . for this purpose , we use the following numerical procedure : 1 . distribute @xmath180 charges randomly . 2 . use the simulated annealing method to find the charge configuration @xmath176 , which minimizes the zeroth - order energy @xmath181 . 3 . change the value of @xmath172 , and repeat the steps 1 - 3 . 4 . among the obtained charge configurations with their zeroth - order energies , find one corresponding to the lowest - lying incommensurate state . in the numerical simulations , we have used the conventional simulated annealing method adopting the metropolis algorithm for the arrays of sizes @xmath182 and @xmath183 . starting from high temperatures ( @xmath184 ) , we decrease the temperature slowly with the decrements @xmath185 ( at high temperatures ) and 0.002 ( at low temperatures ) , and perform 10000 mc steps per site at each temperature step . we have found that the charge configuration corresponding to the lowest energy depends crucially on the value of @xmath4 , and thus used three different values : @xmath4 = 0.1 , 1.0 , and 5.0 . through 20 independent runs , we have obtained the charge configurations corresponding to the lowest energy , which have also been checked via the entropic annealing algorithm @xcite . although we have allowed double occupancy ( @xmath186 ) as well , in the lowest - lying configurations @xmath57 is always found to have the value either 0 or 1 . once the charge configuration @xmath176 which minimizes the zeroth - order energy in eq . ( [ eq_eic ] ) is found , the phase boundary is determined from the condition @xmath187 . this leads to the critical value @xmath109 separating the superconducting phase from the insulating phase : ' '' '' ' '' '' @xmath188 \nonumber \\ & + & \frac{(e_j / e_0)^2}{256q_e}\left [ { \sum_{(i , j)}}^\prime \frac{1}{\sum_l\left ( { \widetilde c}_{jl}^{-1 } -{\widetilde c}_{il}^{-1 } \right)q^{(q_e)}_l - \left ( { \widetilde c}_{00}^{-1 } - { \widetilde c}_{10}^{-1}\right ) } \right . \nonumber \\ & & \left . -{\sum_{(i , j)}}^\prime \frac{1}{\sum_l\left ( { \widetilde c}_{jl}^{-1 } -{\widetilde c}_{il}^{-1 } \right)q_l - \left ( { \widetilde c}_{00}^{-1 } - { \widetilde c}_{10}^{-1}\right ) } \right ] , \end{aligned}\ ] ] ' '' '' 2 where @xmath189 again describes the charge configuration of the commensurate state . when @xmath190 , the system , displaying superconductivity for @xmath191 , becomes insulating as @xmath43 is increased beyond @xmath109 . this behavior according to whether @xmath43 is larger or smaller than @xmath109 is reversed when @xmath192 . table [ tab_qe ] shows the values of @xmath172 in the lowest excitations , breaking the commensurability of the states with @xmath77= 1/2 , 1/3 , 1/4 , and 1/5 , in the arrays of sizes @xmath193 10 and 12 . except for @xmath194 , there are two values of @xmath172 in each case , one for the positive - charge excitation ( @xmath195 ) and the other for the negative - charge excitation ( @xmath196 ) . since the value of the charge frustration has been restricted in the range @xmath45 , we need to consider only the negative - charge excitations for @xmath197 . it displays that as @xmath1 is increased , the absolute value of @xmath172 decreases , manifesting that the excitations become more point - like . as an example , we present the charge configurations for @xmath198 in fig . [ q2_conf ] , where ( a ) , ( b ) , and ( c ) correspond to the commensurate state , the excited state with @xmath179 , and the excited state with @xmath199 , respectively , in a @xmath200 array . it is of interest to note that the configuration ( c ) , which has 60 charges , is precisely the same as the ground state configuration of vortices in the presence of the magnetic frustration 5/12 @xcite . when @xmath179 is the lowest excitation , there exists @xmath201-fold degeneracy since the vacant site , represented by a small square in ( b ) , can be located on any of the total @xmath201 sites . in this case , the degenerate states are coupled in @xmath202 , and we should use the second - order degenerate perturbation theory to calculate the energy . it is again straightforward to diagonalize the second - order matrix by means of the fourier transform , which yields ' '' '' ' '' '' @xmath203 , \end{aligned}\ ] ] ' '' '' 2 in place of eq . ( [ eq_eic ] ) . here @xmath204 denotes the position of the vacant site . figure [ fig_zerot ] shows the resulting phase diagrams for @xmath4= ( a ) 0.1 , ( b ) 1.0 , and ( c ) 5.0 . the various lobe - like structures , in which the array is in insulating phase , are obtained . since the screening length @xmath136 , beyond which the magnitude of the interaction between charges decreases exponentially , is much smaller than the system size , the finite - size effects on the phase boundaries are presumably negligible . we have checked the phase boundary for @xmath197 , and indeed found that it hardly changes with the system size considered here ( see below ) . accordingly , the phase boundaries in fig . [ fig_zerot ] are expected to be qualitatively correct even in the thermodynamic limit . as we increase the array size , more insulating - lobes should be observed , since the array with @xmath30 should be insulating , regardless of the value of @xmath43 . at this stage , however , it is still unclear whether the insulating lobe exists at every rational value of @xmath43 . in the nn charging limit , the charging - energy term takes the form @xmath205 with @xmath206 given by eq . ( [ eq_cbar ] ) . here the divergence of @xmath207 leads to @xmath208 , which is simply the counterpart of the vortex - neutrality condition in the classical array . indeed fig . [ fig_zerot ] shows that the position of the insulating region approaches @xmath208 , as @xmath4 is increased . to study the effects of the junction capacitance versus the self - capacitance in detail , we now concentrate on the @xmath197 insulating phase , whose commensurate charge configuration displays the @xmath209 superlattice structure . [ see fig . [ q2_conf ] ( a ) . ] similarly to the vortex superlattice in the fully - frustrated classical array @xcite , the commensurate state has the two - fold degeneracy . if we remove a single positive charge from the system , commensurability is destroyed near the resulting point defect . the charge configuration of the corresponding lowest - lying incommensurate state has been shown in fig . [ q2_conf ] ( b ) , which describes the first excited state for @xmath4 = 1.0 or 5.0 . when @xmath4 = 0.1 , in contrast , the first excited state has the configuration shown in fig . [ q2_conf ] ( c ) , with the number of extra charge @xmath199 . this observation that the nature of the lowest excited state changes with @xmath4 has an important implication with regard to the supersolid phase , where a commensurate charge - density wave is believed to coexist with superfluidity @xcite : the charge configuration in fig . [ q2_conf ] ( b ) regarded as a snapshot of the charge configuration in the supersolid phase demonstrates that the underlying global commensurability still holds despite of the local point defect . furthermore , since the point defect can hop to any site without energy cost , it is delocalized in the energy eigenstate , leading to superfluidity . accordingly , we conclude that the existence of the supersolid phase depends on the value of @xmath4 . in particular , when @xmath4 is sufficiently small , the supersolid phase does not exist . it is of interest to compare this argument pointing out the importance of the vacant defect in supersolidity with that of ref . , emphasizing the role of the double occupancy . in the numerical investigations to find the lowest - energy charge configurations , we have found that all the charges have the value either 0 or 1 at zero temperature , and that those configurations with @xmath210 or 2 at some sites have much higher energies . the @xmath197 phase boundaries in fig . [ fig_zerot ] ( b ) and ( c ) may be considered to divide the insulating commensurate phase and the supersolid phase . as we go further into the supersolid phase starting from the phase boundary , the more point defects are generated ; the commensurability should be completely destroyed when we cross another phase boundary separating the supersolid phase from the superfluid phase , although within the formalism adopted here , we are unable to find this new boundary . to obtain the phase diagram for @xmath197 , we have considered @xmath5 arrays up to @xmath211 , and found that the finite - size effects are negligible for @xmath212 , which corresponds to the interaction range @xmath213 , and that the qualitative features of the phase diagram do not change even when @xmath214 . we have further found that the difference between the phase boundary obtained from the excitation @xmath179 and that from @xmath199 is insignificant in the @xmath200 array with @xmath215 . therefore , for simplicity , we consider only the excitation @xmath179 and compute the phase boundary from eqs . ( [ eq_o ] ) and ( [ eq_deg ] ) . the resulting phase diagram obtained for the array of size @xmath211 is shown in fig . [ q2_bound ] . the insulating region is shown to expand with @xmath4 when @xmath216 and then to shrink as @xmath4 is increased further . this is to be compared with the case @xmath54 ( shown in fig . [ q0_bound ] ) , where superconductivity is enhanced monotonically with @xmath4 , even for small @xmath4 . in this section , we investigate the phase transitions of the superconducting arrays at finite temperatures , where thermal fluctuations as well as quantum fluctuations suppress the ordering of phases . while at high temperatures the system should be in the normal ( disordered ) state due to large thermal fluctuations , strong charging effects at low temperatures produce large quantum fluctuations , tending to destroy superconductivity of the system . we thus expect rich behaviors resulting from the interplay of thermal and quantum fluctuations , which may be controlled by the temperature and charging energy . we first study the system in the absence of the josephson coupling via mc simulations , and then investigate the phase transitions in the presence of the josephson coupling via a variational method together with mc simulations . in the absence of the josephson coupling ( @xmath217 ) , the hamiltonian is diagonal in the charge representation and written in the form @xmath218 where @xmath57 is the eigenvalue of the charge operator at site @xmath8 . we perform mc simulations on the above hamiltonian to investigate how the insulating phase corresponding to the regions inside the lobes in fig . [ fig_zerot ] changes with the temperature . using the conventional mc method with the metropolis algorithm , we start from a sufficiently high temperature ( @xmath219 ) , and decrease the temperature slowly with the decrements @xmath185 ( at high temperatures ) and 0.002 ( at low temperatures ) , during the simulations , we allow creation and annihilation of a charge at a site , as well as hopping to other sites . at each temperature step , we compute the thermal average @xmath220 using 10000 mc steps per site , after 1000 mc steps of equilibration . figure [ fig_mc ] displays the results for arrays of sizes @xmath221 ( left ) and 12 ( right ) , with @xmath222 ( a ) 0.1 , ( b ) 1.0 , and ( c ) 5.0 . at zero temperature , the curves representing @xmath77 versus @xmath43 exhibit steps at rational values giving the ( commensurate ) insulating phase . such steps structures , which correspond to the lobe structures in fig . [ fig_zerot ] , are common in systems displaying commensurate - incommensurate transitions , and signal the commensurate locking @xcite . it is shown that , as expected , the width of a step decreases with @xmath4 , and the curve approaches the straight line @xmath223 in the limit @xmath224 , manifesting the suppression of the insulating phase . comparing the results for the @xmath221 array with those for the @xmath225 array , we also expect more rational steps in larger arrays , which is apparently suggestive of the devil s staircase structure @xcite in the thermodynamic limit . for conclusive results , however , more detailed study is needed . table [ tab_mc ] presents the widths of the rational steps of @xmath226 1/3 , and 1/2 in the @xmath225 array at zero temperature . one can see here that the values in table [ tab_mc ] are in good agreement with the @xmath217 results in fig . [ fig_zerot ] , which supports the validity of the approach in sec . [ sec_zero ] . figure [ fig_mc ] also shows that steps tend to disappear at finite temperatures , reflecting that thermal fluctuations suppress the charge ordering due to quantum coherence as well as the phase ordering . accordingly , the insulating commensurate phase in general turns into the normal ( disordered ) phase as the temperature is increased from zero . note , however , that some steps survive weak thermal fluctuations . namely , whereas steps at higher - order rationals easily disappear , low - order steps such as @xmath54 , 2 , and 1/3 persist at low but nonzero temperatures . these features presumably can also be observed in the systems with nonzero josephson couplings . thus as we increase the temperature from zero , higher - order insulating lobes in fig . [ fig_zerot ] disappear , but some low - order lobes are expected to remain at low temperatures . in the presence of the josephson coupling , both the charge and the phase variables should be treated quantum - mechanically , which prohibits exact analytical treatment . here , we use the giachetti - tognetti - feynman - kleinert ( gtfk ) variational method @xcite to evaluate the path integral and to obtain the effective classical hamiltonian , which is convenient for studying the phase transitions at finite temperatures . the gtfk method , which remains reliable in known cases even at zero temperature , has been successfully applied to the superconducting arrays with continuous charge states in the presence of ohmic dissipation @xcite . we begin with the hamiltonian given by eq . ( [ eqh2 ] ) and write the partition function of the system in terms of the path integral @xcite : @xmath227 \right\}\ ] ] with the euclidean action @xmath228 where @xmath229 is the inverse temperature , and the planck constant has been set equal to unity ( @xmath230 ) . the euclidean lagrangian @xmath231 can be obtained from the hamiltonian via the wick rotation @xmath232 and the legendre transformation . we use the representation @xmath233 obeying the commutation relation @xmath234 = i$ ] , and get ' '' '' ' '' '' @xmath235 with @xmath236 , where the charge frustration enters through a purely imaginary term . in the presence of ohmic dissipation , the charge at each site takes continuous values , and only the paths satisfying @xmath237 contribute to the partition function , with the dissipative action term added @xcite . in this case , the charge frustration does not appear in the euclidean action since @xmath238 , and the resulting phase diagram does not depend on @xmath43 . if there is no ohmic dissipation present , in contrast , only discrete charge states are allowed , and all the paths satisfying @xmath239 with integer @xmath240 contribute to the path integral in eq . ( [ eqz ] ) @xcite . therefore the path integral should include the summation over the `` winding numbers '' @xmath241 , yielding interesting behaviors associated with the charge frustration @xcite . accordingly , the allowed charge states , i.e. , continuous or discrete , are crucial in the resulting phase diagrams . from the boundary condition @xmath242 , we decompose @xmath16 into the periodic variable @xmath243 satisfying @xmath244 and @xmath240 , according to @xmath245 , and obtain the euclidean action in the form @xmath246 where @xmath247 \right\ } .\ ] ] following the gtfk method , it is straightforward to obtain the effective classical hamiltonian @xmath248 where @xmath249 is determined by @xmath250^{-1 } \right\ } \nonumber\\ & = & \label{eq : g0 } \frac{1}{8 g_0 \beta e_j } ( x_1 \coth x_1 + x_2 \coth x_2 - 2)\end{aligned}\ ] ] with @xmath251 and @xmath252 . here we have used identity @xmath253 when eq . ( [ eq : g0 ] ) has more than one zeros , we should choose the largest one according to the extremal principle . the detailed procedure is entirely similar to that in ref . , and will not be repeated here . as a result , the quantum fluctuations renormalize the josephson coupling @xmath71 to @xmath254 . furthermore , in contrast to the continuous charge case @xcite , the charge frustration appears explicitly in the hamiltonian , and is expected to play an important role in the phase transition of the system . unfortunately , the kronecker delta @xmath255 between winding numbers of nearest neighboring pairs in eq . ( [ eqhcl ] ) makes further analysis very difficult . for simplicity , we thus replace it by its self - consistent average @xmath256 and obtain the approximate hamiltonian @xmath257 where @xmath258 is the ensemble average with respect to the approximate hamiltonian @xmath259 . the poisson summation formula then allows us to write @xmath260 } { \sum_{\{m_i\ } } \exp \left [ - 4\beta e_0 \sum_{i , j } ( m_i - q_x ) { \widetilde c}_{ij}^{-1 } ( m_j - q_x ) \right]}\ ] ] ' '' '' 2 with @xmath261 , which manifests the absence of dependence on @xmath71 . once @xmath256 is determined from eq . ( [ eqde ] ) , the partition function of the system obtains the form @xmath262 where the charge - part @xmath263\ ] ] has been obtained from the poisson summation formula and the phase - part is given by @xmath264.\ ] ] unless the ratio @xmath4 is infinite , the interaction @xmath265 between charges is short - ranged and does not have any singularity . it is then reasonable to assume that the criticality of the system is governed by the phase - part , which implies that the system exhibits a vortex - unbinding bkt transition at the critical temperature given by @xcite @xmath266 we first consider the self - charging limit ( @xmath79 ) . in this limit , @xmath256 can easily be evaluated since the terms in eq . ( [ eqde ] ) factorize . the resulting phase diagrams determined from eq . ( [ eqcritical ] ) at temperatures @xmath267 , and @xmath268 ( in units of @xmath269 ) are shown in fig . [ fig_selfbd ] . the region in the right - hand side of each line corresponds to the superconducting phase whereas the left region represents the normal ( disordered ) phase . in this phase diagram , we observe that the superconducting region expands as @xmath43 is increased at all temperatures . ( recall that the symmetry of the system allows to restrict within the range [ 0,1/2 ] . ) therefore it is concluded that the charge frustration in general enhances superconductivity . the qualitative features are largely similar to those of the results from the coarse - graining approach @xcite at temperatures @xmath270 . unlike the latter , however , fig . [ fig_selfbd ] also shows reentrant behavior at low temperatures , for small @xmath43 . at zero temperature , unfortunately , we obtain the unphysical result that the system is in the superconducting state for @xmath271 , even in the limit of the vanishing josephson coupling . this is presumably an artifact of the self - consistent approximation ( sca ) replacing of the kronecker delta by its average , with correlations neglected . when @xmath272 , on the other hand , the terms in eq . ( [ eqde ] ) do not factorize , and the average @xmath256 can not be evaluated analytically , which makes it inevitable to resort to a numerical method . at extremely low temperatures , @xmath273 in eq . ( [ eqde ] ) can have the value either 0 or 1 , which allows us to calculate @xmath256 by direct summation of @xmath274 terms for small @xmath275 . at high temperatures , other values of @xmath273 are allowed ; this makes the direct summation impractical and we use mc simulations . figure [ fig_1.0bd ] shows the resulting phase diagram obtained for a @xmath125 array with @xmath276 at temperature @xmath277 , 0.2 , 0.3 , and 0.4 . at @xmath278 , @xmath256 has been calculated by means of both the direct summation with @xmath279 and @xmath280 and mc simulations , which give results in good agreement with each other . at higher temperatures , @xmath256 has been obtained via mc simulations , yielding the enhancement of superconductivity by charge frustration in a similar manner to that in the self - charging limit . at low temperatures ( @xmath277 ) , on the other hand , there exists a small lobe near @xmath281 , which indicates the expansion of the insulating region due to the commensurability effects on the ordering of charges . to check the validity of the sca adopted here , replacement of @xmath282 to its self - consistent value @xmath256 , we have performed mc simulations with @xmath90 , at which the hamiltonian in eq . ( [ eqhcl ] ) has only real terms : ' '' '' ' '' '' @xmath283 we have considered a @xmath284 array with various values of @xmath42 , and measured the helicity modulus @xmath285 with 100000 mc steps : @xmath286 , \end{aligned}\ ] ] ' '' '' 2 where @xmath287 with @xmath288 denoting the @xmath289 coordinate of the @xmath8th grain in units of the lattice constant . the universal jump condition of the helicity modulus @xcite @xmath290 then determines the critical coupling @xmath291 , the values of which are displayed in table [ tab_j ] , for @xmath292 and 1 at various temperatures . table [ tab_j ] also displays the results from the sca replacing the kronecker delta by its average , which has yielded figs . [ fig_selfbd ] and [ fig_1.0bd ] : the two results indeed exhibit behaviors qualitatively in accord with each other . in particular , for @xmath293 , both results give the value of @xmath291 larger at @xmath277 than at @xmath294 , strongly suggesting the reentrant transition . from a quantitative viewpoint , on the other hand , the agreement between the two results is not satisfactory , especially at low temperatures , which seems to indicate that the sca fails at low temperatures . to examine this , we have also measured in the mc simulations the average value of @xmath255 for @xmath79 and display the results in fig . [ fig_delta ] . at high temperatures , the mc data indeed agree well with the results from eq . ( [ eqde ] ) which is represented by the solid line in fig . [ fig_delta ] . however , at low temperatures the difference is apparent unless @xmath42 is sufficiently small . in particular , for @xmath295 , the mc simulations give the value approaching unity as the temperature is decreased to zero , whereas the value from eq . ( [ eqde ] ) approaches zero . this is the case that the second term in eq . ( [ eq_hcl ] ) is dominant over the first term , which leads to the ordering of both @xmath296 and @xmath297 . this reflects strong correlations between the two , which have not been taken into account in the sca . it is thus concluded that the _ quantitative _ behaviors of the phase diagrams in fig . [ fig_selfbd ] are accurate only at sufficiently high temperatures as already shown in table [ tab_j ] . we have investigated the quantum phase transitions in two - dimensional superconducting arrays with various values of external charge frustration at both zero and finite temperatures . we have considered general capacitance matrices , allowing both the self- and the junction capacitances . in the absence of dissipation , the charge variable in the array changes in discrete quanta of @xmath2 , and the phase variable becomes compact in the interval @xmath298 . we have studied in detail zero - temperature phase transitions at various values of the charge density and also considered the system at finite temperatures with and without the josephson energy . the states of the system at zero temperature have been determined from the investigation of the crossing of the ground - state level and the first - excited - state level . at vanishing charge density ( @xmath54 ) , we have considered the commensurate insulating state , where @xmath299 at all sites , and two types of the incommensurate superconducting state , i.e. , the single - charge ( sc ) type and the charge - dipole ( cd ) type . unless the self - capacitance vanishes , the sc type incommensurate state has been found to have lower energy than the cd type one except at very small values of the charge frustration . from the comparison of the energies of the commensurate insulating state and the sc type superconducting state , computed via the second - order perturbation expansions , the phase boundary has been obtained . the validity of the second - order expansion has been checked for the array in the self - charging limit , which reveals that the inclusion of the third - order term does not change significantly the phase diagram . at finite charge densities , identification of the commensurate state and of the first - excited incommensurate state is a highly nontrivial problem . we have thus adopted the simulated annealing method and applied the second - order perturbation theory to obtain a number of lobe - like insulating phases such as @xmath197 , 1/3 , 1/4 , and 1/5 in @xmath182 and @xmath183 arrays . as the system size is increased , more insulating lobes are expected to be observed . the effects of the junction capacitance have been investigated in detail in the insulating phases of @xmath54 and 1/2 : as @xmath4 is increased , the insulating region shrinks monotonically for @xmath54 , while the insulating region for @xmath197 expands with @xmath1 and then shrinks , with its maximum area at @xmath300 . the mc simulations have displayed that the lowest excited state changes as the junction capacitance is increased , which in turn indicates that the existence of the supersolid phase with @xmath197 depends on the interaction range of charges : the supersolid phase exists for sufficiently large values of the ratio of the junction capacitance to the self - capacitance . in the supersolid phase , it has also been shown that the point defect in the charge configuration plays an important role . at finite temperatures , thermal fluctuations as well as quantum fluctuations play an important role in the phase transitions of the system . in the limit of vanishing josephson coupling strength , we have performed the mc simulations to obtain @xmath77 as functions of the charge frustration at various temperatures , and pointed out the possibility of the devil s staircase structure in the thermodynamic limit . as the temperature is increased , higher - order insulating states disappear but some lower - order ones still remain at sufficiently low temperatures . in the presence of the josephson coupling , we have obtained the effective classical hamiltonian via a variational method in the path integral formalism . a self - consistent approximation has been applied to the resulting hamiltonian and phase boundaries have been computed for @xmath293 and 1 . a reentrant transition at @xmath301 has been observed for the array in the self - charging limit ( @xmath302 ) , while an insulating lobe - like structure has been found near @xmath303 in the case @xmath304 . the validity of the formalism used here , the self - consistent approximation ( sca ) on the effective classical hamiltonian , has been checked using mc simulations in the absence of charge frustration . although the reentrant behavior of the array in the self - charging limit has also been observed in mc simulations , the sca appears to be accurate only at sufficiently high temperatures . the quantum phase transitions of the superconducting arrays have been studied theoretically by many authors via mean - field approximations such as the coarse - graining approximation @xcite and via the semiclassical methods including the variational methods @xcite and the wkb approximation @xcite . although the mean - field approximations are in general expected to give better results at zero temperature , where the dimension of the system is effectively increased by one , than at finite temperatures , it is well known that the coarse - graining approximation gives unphysical results even at zero temperature , for the arrays without self - capacitances . on the other hand , the semiclassical methods should be applied with great care at zero temperature , where quantum fluctuations are strong . in comparison with these methods , the formalism adopted here to study the zero - temperature phase transitions , perturbation expansion combined with simulated annealing , appears to be more reliable and can be applied to other systems systematically . we thank g.s . jeon for useful discussions . this work was supported in part by the basic science research institute program , ministry of education , and in part by the korea science and engineering foundation through the src program . 99 for a list of references , see , e.g. , _ proceedings of the 2nd ctp workshop on statistical physics : kt transition and superconducting arrays _ , edited by d. kim , j.s . chung , and m.y . choi ( min - eum sa , seoul , 1993 ) ; physica b * 222 * 253 - 406 ( 1996 ) . berezinskii , zh . . fiz . * 59 * , 907 ( 1970 ) [ sov . jetp * 32 * , 493 ( 1971 ) ] ; j.m . kosterlitz and d.j . thouless , j. phys . c * 6 * , 1181 ( 1973 ) ; j.m . kosterlitz , j. phys . c * 7 * , 1046 ( 1974 ) ; j.v . jos , l.p . kadanoff , s. kirkpatrick , and d.r . nelson , phys . b * 16 * , 1217 ( 1977 ) . geerligs , m. peters , l.e.m . de groot , a. verbruggen , and j.e . mooij , phys . lett . * 63 * , 326 ( 1989 ) ; h.s.j . van der zant , l.j . geerligs , and j. e. mooij , europhys . * 19 * , 541 ( 1992 ) . kim and m.y . choi , phys . rev . b * 52 * , 3624 ( 1995 ) . c. rojas and v. jos , phys . rev . b * 54 * , 12361 ( 1996 ) . c. bruder , r. fazio , a. kampf , a. van otterlo , and g. schn , phys . scr . t * 42 * , 159 ( 1992 ) ; a. van otterlo , k .- h . wagenblast , r. fazio , and g. schn , phys . b * 48 * , 3316 ( 1993 ) . m.p.a . fisher , p.b . weichman , g. grinstein , and d.s . fisher , phys . b * 40 * , 546 ( 1989 ) ; g.g . batrouni , r.t . scalettar , and g.t . zimanyi , phys . lett . * 65 * , 1765 ( 1990 ) . freericks and h. monien , europhys . 26 * , 545 ( 1994 ) . meisel , physica b * 178 * , 121 ( 1992 ) ; g.g . batrouni , r.t . scalettar , g.t . zimanyi , and a.p . kampf , phys . lett . * 74 * , 2527 ( 1995 ) ; a. van otterlo and k .- h . wagenblast , _ ibid . _ * 72 * , 3598 ( 1994 ) ; e. roddick and d. stroud , phys . b * 51 * , 8672 ( 1995 ) . a. van otterlo , k .- h . wagenblast , r. baltin , c. bruder , r. fazio , and g. schn , phys . b * 52 * , 16176 ( 1995 ) . keldysh , pisma zh . * 29 * , 716 ( 1979 ) [ jetp lett . * 29 * , 658 ( 1979 ) ] ; r. fazio and g. schn , phys . b * 43 * , 5307 ( 1991 ) ; j.e . mooij , b.j . van wees , l.j . geerligs , m. peters , r. fazio , and g. schn , phys . rev . lett . * 65 * , 645 ( 1990 ) . cha , m.p.a . fisher , s.m . girvin , m. wallin , and a.p . young , phys . b * 44 * , 6883 ( 1991 ) . n. goldenfeld , _ lectures on phase transitions and the renormalization group _ ( addison - wesley , 1992 ) , p. 44 . h.s.j . van der zant , l.j . geerligs , and j.e . mooij , europhys . lett . * 19 * , 541 ( 1992 ) . a. kanda and s. kobayashi , j. phys . . jpn . * 64 * , 3172 ( 1995 ) ; a. kanda , s. katsumoto , and s. kobayashi , _ ibid . _ * 63 * , 4306 ( 1994 ) ; r. yagi , t. yamaguchi , h. kazawa , and s. kobayashi , czech . * 46 * , suppl . s2 , 693 ( 1996 ) . lobb , d.w . abraham , and m. tinkham , phys . b * 27 * , 150 ( 1983 ) . straley and g.m . barnett , phys . b * 48 * , 3309 ( 1993 ) . a.a . odintsov , phys . b * 54 * , 1228 ( 1996 ) . j. lee , phys . lett . * 71 * , 211 ( 1993 ) ; j. lee and m.y . choi , phys . e * 50 * , r651 ( 1994 ) . choi and s. doniach , phys . b * 31 * , 4516 ( 1985 ) . p. bak , rep . . phys . * 45 * , 587 ( 1982 ) . r. giachetti and v. tognetti , phys . lett . * 55 * , 912 ( 1985 ) ; r.p . feynman and h. kleinert , a * 34 * , 5080 ( 1986 ) . s. kim and m.y . choi , phys . b * 41 * , 111 ( 1990 ) ; * 42 * , 80 ( 1990 ) . r.p . feynman and a.r . hibbs , _ quantum mechanics and path integrals _ ( mcgraw - hill , new york , 1965 ) . g. schn and a.d . zaikin , phys . rep . * 198 * , 237 ( 1990 ) . see , e.g. , s. oh , s. lee , and k .- c . lee , j. korean phys . * 27 * , 365 ( 1994 ) and references therein . nelson and j.m . kosterlitz , phys . lett . * 39 * , 1201 ( 1977 ) . .values of @xmath172 in the lowest excited states of the arrays with size ( a ) @xmath221 and ( b ) @xmath225 . it is shown that @xmath305 tends to decrease as @xmath1 is increased . [ cols=">,^,^,^",options="header " , ]

photoionization models for chemically homogeneous gaseous nebulae of constant density predict an almost constant temperature ; consequently , observers often assume a constant temperature to determine chemical abundances . the spatial temperature inhomogeneities in regions are usually characterized by @xmath1 the mean square temperature variation in three dimensions . recent reviews on the temperature structure of gaseous nebulae are those by peimbert ( 1995 , 2002 ) , esteban ( 2002 ) , stasiska ( 2002 ) , liu ( 2002 ) , torres - peimbert & peimbert ( 2003 ) . from uniform - density chemically - homogeneous photoionization models it has been found that @xmath1 is in the 0.002 to 0.025 range , for objects with solar or subsolar chemical abundances , with typical values around 0.005 ( e.g. , gruenwald & viegas 1992 , garnett 1992 ) . baldwin et al . ( 1991 , henceforth b91 ) produced a photoionization model for the orion nebula , in which grains and gas are well mixed , that included photoelectric heating and cooling of the gas by grain ionization and grain collisions respectively ; for this model peimbert , storey , & torres - peimbert ( 1993 ) found that @xmath1 = 0.004 . alternatively from observations of ngc 346 in the smc ( peimbert , peimbert , & ruiz 2000 ) , 30 doradus ( peimbert 2003 ) , and the orion nebula ( peimbert et al . 1993 , esteban et al . 1998 , henceforth epte98 ) it has been found that @xmath1 is in the 0.02 to 0.04 range . the @xmath1values were derived by combining the temperature derived from the ratio of the @xmath3 4363 , 5007 lines [ @xmath4 with the temperature derived from the ratio of the balmer continuum to @xmath5 [ @xmath6 and by combining @xmath7 with the temperatures derived from the ratio of @xmath8 5007 to the recombination lines of multiplet 1 of o ii and the ratio of the of @xmath3 1906 , 1909 [ c iii ] lines to the @xmath8 4267 c ii recombination line . the differences between the observed @xmath1 values and those predicted by models have two profound implications for the study of ionized gaseous nebulae : they significantly affect the derived abundances and they imply the presence of one or several physical mechanisms not considered by the models . to study this problem further we decided to obtain an independent measurement of @xmath1 based on over a million temperatures determined across the face of the orion nebula . the subarcsecond resolution of the hst makes it ideal for this purpose . each of these temperatures represents an average along a given line of sight . since the columnar temperatures already represent an average temperature it is not possible to obtain directly from them the full @xmath1 value . instead it is possible to obtain the mean square temperature variation across the plane of the sky and from this value it is possible to estimate @xmath1 . in 2 we discuss the observations . in 3 and 4 we present the method for deriving the columnar temperature ( @xmath9 ) image together with its analysis . in 5 we formally define the columnar temperature , and the mean square temperature variation measured in the plane of the sky ( @xmath10 ) , then determine @xmath10 from the data . we then estimate @xmath1 based on @xmath10 and additional data from the @xmath9 image and temperatures . in 6 we compare our results with those of previous work . a new model for the structure of the nebula near the ionization front is presented in 7 . finally , the conclusions are presented in the intent of this observational program was to establish accurate 4363 and 5007 line ratios across the orion nebula using the wfpc2 camera ( holtzman et al . 1995 ) of the hubble space telescope ( hst ) and its narrow - band interference filters . this required imaging the nebula in the filters intended to isolate these lines ( f437n and f502n , respectively ) . since the extinction varies at small scales across the face of the nebula ( odell & yusef - zadeh 2000 ) , the magnitude of the extinction was determined on the same size scale as the line observations . this required determination of the h@xmath11 and h@xmath12 line ratios , which demanded observations with the f656n and f487n filters . since the former filter allows some emission from the nearby doublet to contaminate the h@xmath11 signal , it was also necessary to observe with the f658n filter , which is dominated by the 6583 line . there was the possibility of the h@xmath13 line contaminating the f437n signal . the anticipation was that this could be handled by scaling from the h@xmath12 observations in f487n . each emission line filter also passes the nebular continuum radiation . in the case of the emission line signal being very strong as compared with the underlying continuum it is sufficient to measure the visual continuum with the f547 m filter and extrapolate to the wavelength of the emission line filter ( odell & yusef - zadeh 2000 , odell 1998a ) . however , the continuum of the orion nebula is quite strong ( odell & hubbard 1965 , b91 ) because the large optical depth of dust particles lying just beyond the main ionization front ( mif ) back - scatter a significant fraction of the visual starlight reaching it . this back - scatter produces a continuum much stronger than that arising from atomic processes . the observations most sensitive to continuum contamination are those of the intrinsically weak 4363 line . fortunately , the f469n filter provided a good measure of the strength of the local continuum because it is free of strong emission lines when used to observe a low ionization object like the orion nebula . it is known that the he ii 4686 line ( which the f469n filter was designed to isolate in high ionization objects ) is undetected in the orion nebula even in spectroscopic studies reaching nearly 10@xmath14 the intensity of h@xmath12 ( b91 , epte98 ) . combining the f547 m and f469n observations allowed a linear extrapolation or interpolation of the wavelength dependence of the continuum flux to the wavelength of each filter . the filters employed and their pre - launch characteristics ( wfpc2 instrument handbook , version 6.1 available at http://www.stsci.edu ) are summarized in table 1 . this table also gives the wfpc2+hst system throughput as recently determined in orbit . because of the host of filters that had to be used and the length of the observing times necessary for producing an adequate signal in the fainter lines , only one field of view was covered . this was selected to include regions that had been spectroscopically studied recently using high dynamic range detectors ( b91 and epte98 ) , in order to allow a precise calibration of the filters , and to avoid the presence of the bright trapezium stars . as will be noted in 3.3 , the brightest trapezium star ( @xmath0 ori c ) was so close ( 7 ) to the ccd1 detector that low - level scattered light problems affected the long exposure images with that detector , rendering them useless for this program . the image is shown in figure 1 . the four ccd detectors of the wfpc2 are numbered counter - clockwise , with the high - angular resolution , but smaller field of view , ccd being designated as ccd1 . all exposures were made with exactly the same pointing on 2002 january 21 , 22 , and 26 and at the same gain ( electrons / adu ) values . multiple exposures of the same duration were made with each filter in order to allow correction for cosmic ray induced events on the detectors . the individual exposure times and number of exposures were : f437n(1100 s , 6x ) , f469n(1100 s , 6x ) , f487n(350 s , 3x ) , f502n(160 s , 3x ) , f547m(230 s , 3x ) , f656n(140 s , 2x ) , f658n(200 , 2x ) . all images were `` on - the - fly '' calibrated by the space telescope science institute ( stsci ) and then processed using their stsdas and the national optical astronomy observatories iraf tasks . the final product of the routine processing was a combined image such as shown in figure 1 . a considerable number of steps were necessary to convert the photometric observations with wfpc2 into 4363 and 5007 line ratios and from these line ratios into values . in this section we first describe the method of calibration of the filter properties and other components of the photometric system and their application to obtaining line ratios . we then describe the method of correcting for interstellar reddening . finally , we derive the pixel by pixel values of and estimate the statistical noise of the values . the method of calibration adopted was identical to that employed by odell & doi ( 1999 , henceforth od99 ) . however , in the present case we are only interested in the relative line fluxes , so that their equations for an absolute calibration can be simplified . the reader interested in absolute calibrations should refer to od99 , but note that an editing error occurs in their equation ( 5 ) . the term that reads @xmath15 should read @xmath16 , this error does not affect any of the numerical results in that paper . the basic approach is that one uses the wfpc2 observations of spectroscopically well studied regions to determine the calibration constants , which can then be applied to all the pixels of the image since the filter profile and system sensitivity does not vary across the image ( after the appropriate flat - field corrections have been applied ) . the most important calibration constants are determined directly from the observations of the reference regions and only the secondary constants draw on filter characteristics that were last determined before launch of the wfpcs . since the filter and detector characteristics can change with time , this method of calibration permits accurate determination of the most important characteristics . our method of deriving relative calibrations relates n@xmath17 ( the recorded number of adu counts per unit time ) to the characteristics of the source and the filter and camera system through the equation @xmath18 where we have used the following nomenclature : @xmath19=system throughput of the telescope - camera - detector system for a particular line ( excluding the filter ) @xmath20=transmission at a particular line of a filter @xmath21=surface brightness of a particular line in photons @xmath22 @xmath23=system throughput of the telescope - camera - detector system for the continuum ( excluding the filter ) @xmath24=effective width of the filter in @xmath25=surface brightness of the continuum in photons @xmath26 . it is important to note that this calibration is done in units of the number of photons . when necessary , the resulting numbers can be converted to the relative flux . in this program multiple lines contributed to the signal from a single filter only in the case of f656n ( h@xmath11 - 6563 , -6548 +6583 ) and f658n ( -6583 , h@xmath11 - 6563 ) . we considered the contribution of h@xmath13 4340 to the f437n filter , but found it to be negligible . the reference regions were all taken from b91 . they observed with a long slit of east - west orientation located to the east of @xmath0 ori c. the exact location was incorrectly expressed in b91 , but is correctly described in od99 . b91 gives line ratios and continuum to h@xmath12 ratios for 21 samples along their slit . we used the six samples that overlapped with the ccd4 portion of our images ( shown in figure 1 ) that included 4363 relative intensities . each sample was 2x 15.76 in size . the results of this calibration gave the following equations for the two important line ratios , h@xmath11/h@xmath12 and 4363/5007 . @xmath27}{(1 - 1.28\frac{n_{469}}{n_{487}})},\ ] ] and @xmath28}{[1 - 1.62\frac{n_{469}}{n_{502}}(33.4\frac{n_{469}}{n_{547}}+0.423)]}.\ ] ] equation ( 2 ) uses a full - image average value for the color of the continuum and scales the continuum intensity from the closer continuum filter . equation ( 3 ) derives the color of the continuum at each pixel and then extrapolates or interpolates from the two continuum filters , this step being necessary as the continuum correction is quite important in the f437n measurements . the continuum typically contributes about 73% of the total signal in the f437n filter . in a similar investigation of planetary nebulae rubin et al . ( 2002 , henceforth r2002 ) argued that there should be a small correction in the f437n filter for contamination in the blueward wing by h@xmath13 4340 emission . if one takes the pre - launch transmission profiles in the instrument handbook at face value , this would seem to be the case . however , it must be expected that the profiles of the filters change as they age in space . these changes are why we wanted to calibrate our observations within the data set . one would expect that the transmission near the principal lines to not change much with time , since they are located near the peak of the filters transmissions , whereas the h@xmath13 line falls far out on the wing of the filter curve . the instrument handbook indicates that the ccd sensitivities and throughput of the system have been remarkably constant . this combination of factors allowed us to empirically search for evidence of h@xmath13 contamination . we could determine the ratio @xmath29 with and without a correction for h@xmath13 contamination . with the correction the ratio was 0.42 and without the correction the ratio was 0.68 . the instrument handbook values indicate an expected value for the ratio of 0.62 . this indicates that the correction to the 4363 line surface brightness for h@xmath13 is about 10% . since the method of calibration essentially forces the calibration constants to be such as to give the right intensity ratios for the lines assumed to be present , this means that the effects of h@xmath13 correction will only enter in regions where the line and the h@xmath13 line vary considerably in relative intensity . this is not the case for the area covered by the present images ( b91 , epte98 ) . the scatter of results from the six different samples in the calibration region indicates that the uncertainty of the balmer line ratios will be 7% , while that of the line ratios will be about 12% . since the reddening corrections between 4363 and 5007 are small , the effects of the absolute uncertainty of the balmer line ratios are negligible . a 12% systematic uncertainty of the line ratios would make the derived values be off systematically by @xmath30 k. this will not significantly affect our search for fine scale variations in . as we will see in 3.3 , there is excellent agreement between the electron temperatures derived in b91 and those derived in this study . the extinction corrections are based on determining the balmer line ratio for each pixel and then calculating the total extinction from comparison with the intrinsic ratio . this total extinction was then used to correct the line ratios . we assumed with epte98 that the intrinsic balmer line flux ratio was 2.91 . using the same extinction law as epte98 means that the extinction corrected ( @xmath31 ) flux ratios could be calculated from the observed flux ratios ( @xmath32 ) by the relation @xmath33 the extinction values derived from the balmer line ratios are in good agreement with those of odell & yusef - zadeh ( 2000 ) , this being the case for both their results obtained from radio continuum to h@xmath11 ratios and those obtained from their balmer line ratios . the reddening corrected line ratios were then used to derive local values of . a convenient formulation of the relation of line ratio and is given in r2002 . their equation ( 8) can be restated as @xmath34 where we have restated the relation using the common - usage designation of the wavelengths in air ( rather than the more physically meaningful vacuum wavelengths employed by r2002 ) . since we are dealing with large numbers that are not usually determined to high precision , we sometimes will express by the frequent nomenclature @xmath35@xmath36/10,000 . the resulting image is shown in figure 2 . figure 2 shows multiple features . the most common are the `` scars '' that result from stars . these continuum sources are often saturated on some of the images used to derive . even when they are not saturated , the formal value of at the star positions is subject to big uncertainties owing to the multiple subtractions that occur in the derivation . within ccd1 there are multiple features that are related to scattered light from @xmath0 ori c and the more westward members of the trapezium . any apparent fluctuation in within ccd1 is suspect as being an artifact . there are a few features characterized by very low values of , which are discussed in 4.3 . this derivation gives for the six sample regions from the b91 study @xmath37@xmath35@xmath38=0.92@xmath390.01 , while values given in b91 for the same samples gives @xmath37@xmath35@xmath38=0.91@xmath390.02 . a comparison with the epte98 study is not possible because only their position 2 fell within a part of our image that was not useful as it fell within ccd1 and their position 1 was north of our field . it is important to assess the effect of the expected poisson noise on the determination of because of the low signal level in some of the filters . in the region that was used for calibration the total number of recorded electrons ( 7 times the number of adu s recorded per pixel ) for the key filters were f437n(1260 ) , f469n(1260 ) , f547m(8400 ) , and f502n(10500 ) . in the following discussion , we ll assume that the root mean square deviation of a signal n is n@xmath40 . this is what we will mean as the probable error ( p.e . ) . we will assume that for multiplicative factors the p.e.s add quadratically . unfortunately , one can not determine the p.e . directly from the observations since the nebula is intrinsically variable in surface brightness . an assessment of the p.e . per pixel was done by creating noisy images ( using iraf task _ mknoise _ ) of the appropriate noise level for each of the observationally determined components of equation ( 3 ) , then determining the p.e . of the resulting image . the p.e . was 20% . in most of the following discussions we will assume this uncertainty . equation ( 5 ) was then used to determined the effect of uncertainties of the line ratio on the derived value of . a 20% photometric p.e . produces a p.e . in of 5.7% . throughout this paper we will present uncertainties as probable errors ( p.e . ) calculated as the reciprocal of the root - mean square deviation of the recorded signal . the assumptions used in 3.1 for the derivation of eq . 3 are not fully justified since there are two weak lines that affect the signal in the more important continuum filter , f469n . the line appears at 4702 and the line at 4713 , where the transmissions are 0.51 and 0.15 , respectively . as we will demonstrate in this section , the effect of these lines is small in determination of in the nebula . the presence of these weak lines will make the continuum at f469n appear brighter than it is . if we call the enhanced signal in the f469n n@xmath41 and the signal that would be observed if there were only continuum radiation n@xmath42 , then their ratio will be given by @xmath43 the strength of the continuum is usually measured with respect to the h@xmath12 line at 4861 . this was done in b91 , where their figure 2 presents a quantity which is related to the equivalent width ( @xmath44 ) by @xmath8/@xmath44 . the average value of their plotted quantity is 9.5 , which gives an average @xmath44= 510 , which is in good agreement with earlier filter photometry over a wide range of the nebula ( odell & hubbard 1965 ) which showed that @xmath44 is nearly constant in the inner region of the nebula , then decreases outward . having values of @xmath44 allows us to express the observed - quantity ratios of equation ( 6 ) as s@xmath45/s@xmath46=@xmath44s@xmath47/s@xmath48 and s@xmath49/s@xmath46=@xmath44s@xmath50/s@xmath48 . making this substitution and inserting the known values allows the restating of equation ( 6 ) as @xmath51 if the ratio of line intensities with respect to h@xmath12 are constant across the region observed , then there is no net effect , since they will be constant in the calibration region and their affects will cancel out . however , if the line ratios vary , there will be an effect on the derived line ratios . anticipating that this effect is small , we address only the effect on equation ( 3 ) . since the correction term in the denominator of that equation is only about 0.015 , we can demonstrate adequately the effect of varying contaminating line strengths by considering only the terms in the numerator . if ( n@xmath41 / n@xmath42)@xmath52 is the value of the ratio of these quantities in the region used for calibration , we can express the local effect by multiplying the n@xmath53 rates by ( n@xmath41 / n@xmath42)/ ( n@xmath41 / n@xmath42)@xmath52 . the global average values ( taken from a section of ccd2 free of the peculiar effects due to shocks that are discussed in 4.2 are n@xmath54/n@xmath55= 0.0094 , n@xmath53/n@xmath56=1.04 , and n@xmath53/n@xmath57=0.016 . inserting these values into the numerator of equation ( 3 ) and applying a constant correction of 0.015 for the subtractive term in the denominator , gives @xmath58\}.\ ] ] the weak contaminating lines are seen only at high spectral resolution . in the study of osterbrock , tran , & veilleux ( 1992 ) , fluxes relative to h@xmath12 were 0.00245 and 0.00686 for 4702 and 4713 respectively ( for an east - west region 47 long centered 58 north of @xmath0 ori c ) . epte98 found ratios of 0.00179 and 0.00645 for a region 45 north of @xmath0 ori c and 0.00186 and 0.00621 for a region25 south and 10 west of @xmath0 ori c. these three samples would indicate values of n@xmath41 / n@xmath42 of 1.066 , 1.056 , and 1.055 respectively . the contaminating lines were not measured in b91 due to their lower spectra resolution . there is another set of data that provide even more information . co - authors antonio and manuel peimbert together with mara teresa ruiz obtained a long slit spectrum with the vlt3 telescope ( melipal ) and the fors1 spectrograph at the paranal observatory in chile on 2002 september 10 . the slit was oriented at a position angle of 340 and passed through the feature called hh 202-s in odell et al.1997 ( henceforth o97)and designated in figure 2 . to estimate the intensity of the 4702 line , they measured the intrinsically stronger line at 4658 and we scaled it to the 4702 intensity by using the line ratio of 0.32 from epte98 . likewise , they also measured the strength of the 4713 feature , a blend of [ ar iv ] line at 4711 and the at 4713 the [ ar iv ] line contributing about 15% of the blend intensity . the average of a 2section , which is close to the east - west calibration samples from b91 , gave ratios relative to h@xmath12 of 0.0061 and 0.0060 for 4702 and 4713 respectively , giving a value of n@xmath41 / n@xmath59=1.0551 . the scatter of this value , taken from six 20subsamples was @xmath390.0064 . a similar scatter was found from an analysis of both 20 and 2 samples . recalling that the effect of this 5% level contamination cancels out in our calibration procedure leaves us only to consider the effect of the variation in the value of due to the varying level of contamination . the variations in contamination indicated by the paranal data indicate that the corresponding variations in the derived is + 80 and -120 k. the `` temperature image '' can be used to look for variations in at large and fine scales . in this section we describe the results of a search for systematic changes , and explain the regions of apparently very low and high values of . we also present the evidence for small scale changes , and give the results of a fourier analysis of the temperature fluctuations . we looked for variations of with distance from the dominant ionizing star @xmath0 ori c. nine samples of 10x10 were taken from each of ccd2 - 4 . in each case there were no scars from bright stars and none of the low regions discussed in the next section . the average gave @xmath37@xmath608890@xmath39496 k. no statistically significant correlation was found when trying to fit a linear relation to the observed values of and distance from @xmath0 ori c. we conclude that there are real variations in across the region sampled in this study , but there is no obvious relation to the position in the nebula . the scatter of 496 k is significant since the scatter due to poisson statistics within each sample would be only 5 k and we estimated in 3.5 that the scatter due to variations in the contaminating weak lines would be about @xmath39100 k. there are several regions within ccd2-ccd3-ccd4 that have anomalously low values of the derived , as seen in figure 2 . these all fall within the province of known herbig - haro ( hh ) objects hh 202 ( cant et al . 1980 , o97 ) and hh 269 ( walter et al . 1995 , bally , odell , & mccaughrean 2000 ) . the former is probably the result of a pre - main sequence star s blueshifted jet striking the foreground veil of neutral material(o97 ) . the latter is probably the result of blueshifted jet outflow from a pre - main sequence star that lies behind the mif and has yet to be identified ( odell & doi 2003 ) . it is almost certain that hh 269 has the same source as that producing the series of shocks extending to the east and designated as hh 529 ( bally , odell , & mccaughrean 2000 ) and it is likely ( odell & doi 2003 ) that this is also the source for the jets producing hh 202 and the double object hh 203+hh 204 that lies to the southeast , near @xmath61 ori a. these anomalously low features are probably not real , rather , they are the result of contamination of the f469n filter . our method of calculation of the line ratios assumed that the f469n filter was free of any emission lines , i.e. it was used as a measure of the continuum . in all of the nebula for which spectra exist this assumption is good since the star producing most of the ionization of the nebula ( @xmath0 ori c ) is too cool to doubly ionize helium and there are only the two weak contaminating lines ( c.f . hartigan ( 1999 ) has shown that if a fast jet forms a bow shock in material that is already photoionized then one can expect he ii to be present . he used ground based images to look for evidence for emission in the he ii 4686 line . the only region where he thought that this line was present was in hh 202 , but with his low resolution images he could not conclude that 4686 emission was present . the present high resolution images show that there are anomalies at several of the shocks known to exist in this field of view ( but certainly not in all the known shocks ) . the anomalous feature to the southwest from hh 202-s corresponds to a faint shock visible only in h@xmath11 and , but not in ( o97 ) . the paranal spectra described in 3.5 were centered on hh 202-s . the spectra show no 4686 emission . however , there is a considerable enhancement of emission . in a sample centered on hh 202-s , the flux ratio of 4658 to h@xmath12 was 0.031 , which predicts a 4702 to h@xmath12 ratio of 0.010 ( using the flux ratio from epte98 ) . together with a normal 4713 contribution , this would raise the ratio n@xmath41 / n@xmath42 to 1.18 , which would artificially reduce the 4363 ratio by a factor of 4.6 , and decrease the derived value of to 6300 k. since the paranal sample on hh 202-s was 20 long and the object is much smaller , the contamination would be much larger on the object itself and the artificially derived correspondingly lower . direct confirmation of contamination by 4702 was obtained by high velocity resolution spectra obtained by co - author odell and will henney with the manchester echelle spectrometer ( meaburn , et al . 1984 ) on the 2.1-m telescope at san pedro martir observatory on 2002 october 23 , where one sees significant enhancement of that line , peaking at the optically brightest portion of hh 202-s , with no indication of 4686 emission . this means that enhanced emission is a common feature of the anomalously low spots seen in figure 2 , rendering useless the values there . why is enhanced in these shocks is in itself an interesting problem that should be addressed . several proplyds occur within our sample , the brightest within ccd2 - 4 are 141 - 301 and 182 - 413 . these proplyds and the fainter , less well resolved ones all show enhanced values of . in this case the anomaly is probably due to the mis - application of equation ( 5 ) , which is applicable only in the limit of electron densities ( @xmath62 ) low as compared with the critical density ( @xmath63 ) for . @xmath63 is the density at which collisional de - excitation of the upper state giving rise to an emission line is equal in probability to a spontaneous emission . this is only of concern for the 5007 line as its @xmath63value is much lower than that for the 4363 line . rubin et al . ( 2003 , henceforth r2003 ) give @xmath63(5007 ) @xmath64 6.4x10@xmath65 @xmath66 , which is much higher than densities encountered in the nebula , but comparable to characteristic values for the proplyds ( odell 1998a ) . this means that in the local ionization fronts of the proplyds the 5007 emission is being suppressed by collisional de - excitation and the value derived from equation ( 5 ) will be an upper limit . this same interpretation is applied by r2003 to their high for the proplyd 159 - 350 , which lies in the center of our ccd1 image and fell in their long slit spectrum . in their models of the proplyds close to @xmath0 ori c henney et al . ( 2002 ) argue that the formal temperatures derived without consideration of collisional de - excitation would be about 20,000 k. if the densities of the more distant proplyds are much lower , as indicated by their surface brightnesses , then the effects of collisional de - excitation are less and some of the temperature increase we find is real . it is not clear if the similar local increase in at the jet feature just south of the dark rim ( odell & yusef - zadeh 2000 ) in the southeast corner of ccd4 is due to this same effect . the jet has a known high radial velocity ( odell et al . 1997 ) and proper motion ( odell & doi 2003 ) , so it may be that is enhanced where this flow shocks against ambient nebular gas . figure 2 indicates that there are variations in even though we ve shown in 4.1 that these variations are not a function of distance from @xmath0 ori c. the smaller scale changes have been illuminated by looking at the three sample regions depicted in figure 2 , these samples being selected to be free of the [ fe iii ] contamination . the results are shown in figure 3 . in each case the range of temperature is @xmath39 1000 k. for reference , random noise at the level of 5.7 % probable error is shown , this being the level expected for the bright portions . the southwest sample is about one fourth the surface brightness of the other samples . the noise level was brought down to that of the other samples by taking samples of 2x2 pixels in this region . examination of figure 3 shows that the face of the nebula is mottled with small scale variations of with angular dimensions of about 10 and amplitudes of about 400 k. in order to gain a more quantitative evaluation of the data , we performed a fourier analysis of two regions of the @xmath35 image . the first was a 400x400 pixel section in the northwest portion of ccd2 and the second was 250x250 pixels in the northeast section of ccd4 , both being free of camera scars and the local apparent drops that are due to shocks , as discussed in the analysis was done using stsdas tasks with iraf . the procedure was the same for each sample . the first step was to subtract the average value of the sample from each pixel and then to make a power spectrum of the resulting image , using the task _ powerspec_. a radial profile was taken of the power spectrum image using the center as the origin . the result was a profile that dropped down to a constant value at large spatial frequency values and represents the noise level due to the poisson noise in the data . this constant value was then subtracted from the radial profile of the power spectrum . the resulting noise - subtracted power spectrum was then converted to the physically more meaningful length space by multiplying the power spectrum in spatial frequency by the square of the frequency and binning the result in the spatial dimension . the results are shown in figure 4 and figure 5 . these plots show that there is a large amount of power at the finest scales ( sub - arcsec ) but that power is found at all scales . this analysis is similar to that of vannier et al . ( 2001 ) of h@xmath67 emission arising near the imbedded bn - kl sources to the north of ccd4 . the aims of this section are to estimate representative values of the temperature variations [ @xmath1(o@xmath68 ) , @xmath1(o@xmath69 ) , and @xmath1(h@xmath69 ) ] in the orion nebula in order to compare them with predictions of photoionization models and to be able to study the physical conditions of the nebula including its chemical composition . we proceed in three steps : 1 . ) use of the @xmath9 map obtained in this paper to determine @xmath10(o@xmath68 ) across the plane of the sky ( this is the first detailed 2-dimensional set of temperatures across of the orion nebula ) , 2 . ) estimate the 3-dimensional value of @xmath1(o@xmath68 ) ; 3 . ) estimate @xmath1(h@xmath69 ) including both the o@xmath70 and the o@xmath68 regions . in order to obtain the observed mean square temperature variation in the plane of the sky , @xmath10(o@xmath68 ) , we will define a set of equations . the average temperature , @xmath71 , and the mean square temperature fluctuation , @xmath72 , are given by @xmath73 and@xmath74 respectively , where @xmath75 and @xmath76 are the local electron temperature and density , @xmath77 is the local ion density corresponding to the observed emission line , and @xmath78 is the observed volume ( peimbert 1967 ) . by recognizing that the observations correspond to averages over a given line of sight , it becomes necessary to define the columnar temperature , , as @xmath79 where @xmath11 is the right ascension and @xmath80 is the declination of a given line of sight ( corresponding to a given pixel ) . then , the average temperature can be written as @xmath81 and the mean square temperature variation over the plane of the sky , @xmath10 , can be defined as : @xmath82 in table 2 we list a set of @xmath10(o@xmath68 ) values obtained from equation ( 13 ) and from the @xmath83 map presented in figure 2 . to determine @xmath10(o@xmath68 ) we have excluded several features in order to diminish the noise of the samples : the pixels with very low flux , where we expect the temperature deviations to be mostly due to errors and not to physical inhomogeneities ; the shocks , where the destruction of dust could be enhancing the 4702 line and thus producing spurious temperatures ; the proplyds , whose density is not well represented by equation ( 5 ) ; the seams between the ccds , where there are border effects . keeping in mind these exclusions we defined the six regions from table 2 as follows : regions 1 and 2 include most of the pixels of the se and sw regions as defined in figure 3 ; region 3 includes most of the pixels in a 54 `` @xmath84 73 '' region in the eastern part of ccd4 ; region 4 is the largest sample , it includes most of the pixels from ccd2 , ccd3 , and ccd4 ; finally region 5 includes approximately 72% of the pixels of region 4 where , in order to enhance the s / n ratio , the southern 10% of ccd2 , the southwestern 50% of ccd3 , and the western 25% of ccd4 have been removed , the regions removed are the ones of relatively low emission measure . although the signal in ccd1 was high , the presence of scattered light artifacts ( c.f . 2 ) prevents its use in our analysis . column 2 , of table 2 , presents the number of independent temperatures used in the region ; column 3 presents the average temperature , weighted by @xmath85 ( defined in equation ( 12 ) ) ; while column 4 presents the measured @xmath10(o@xmath68 ) for each region . the @xmath10 of regions 1 , 2 , and 3 is in the 0.0050.016 range . the differences in @xmath10 are real since they are larger than the expected observational errors . this could have 2 explanations : a ) region 3 is more homogeneous than regions 1 and 2 ; b ) another possibility is that @xmath1 is similar in the 3 regions , and the variations along each line of sight in region 3 are larger , while there are not so many variations along the lines of sight in regions 1 and 2 , thus giving a set of more homogeneous values of @xmath9 for region 3 than for regions 1 and 2 . we added most of the pixels from ccd2 , ccd3 , and ccd4 to generate region 4 . figure 6 presents a histogram of the 1.5 million temperatures included in region 4 . the mean square temperature deviation obtained from the histogram is @xmath86 ; which is not @xmath10 . it is not possible to derive directly @xmath10 from the histogram since each temperature has to be weighted by the corresponding emission measure ( see equation ( 13 ) ) . after weighting each value for the emission measure we obtain a @xmath87 for region 4 . we still have to subtract a value associated with the noise . the average p.e . in @xmath88 for the whole sample is 5.7 % , however this value is for an average pixel , while the measured @xmath89 gives more weight to the brightest pixels , the ones with the smaller errors . by considering that the average pixel is 1.89 times fainter than the average pixel weighted by emission measure , we expect the representative p.e . to be 4.2 % , which corresponds to an error on @xmath10 of 0.0017 , that is a @xmath90 . if we take out the noisiest parts of region 4 ( the outermost parts ) , we have region 5 . here the measured @xmath91 . for region 5 we expect the noise to be about 2/3 the noise of the entire region , that is a p.e . of 3.8 % per pixel . in region 5 the average pixel is 1.31 times fainter than the average pixel weighted by emission measure ; thus we expect the representative p.e . to be 3.3% , which corresponds to an error on @xmath10 of 0.0011 , that is a @xmath92 . we expect the @xmath10 result for region 5 to be more accurate ; however , by excluding the outermost parts we seem to be ignoring a fraction of the object which is intrinsically cooler . therefore we will adopt the value derived from region 4 , @xmath93 , as representative for the whole object . this value is similar to the one obtained by r2003 where they find @xmath94 values of 0.0095 , for their slit 1 , and 0.0104 , for their slit 2 . unfortunately , their slits 1 and 2 are located in our unusable ccd1 , which precludes a direct comparison . up to this point we have presented the results in terms of @xmath94(o@xmath68 ) , but what is of physical interest are the variations in three dimensions @xmath1(@xmath96 ) . in order to obtain the total @xmath1(@xmath96 ) value we need to consider the variations along the line of sight . it can be shown that the relevant equation is : @xmath97 where @xmath98 , the variation along a given line of sight , is given by : @xmath99 and the average over all lines of sight is given by : @xmath100 there are two estimates that we can make of @xmath98 : the first based on geometrical considerations and the information contained in @xmath101 ; and the second based on the ratio of @xmath96 collisionally excited lines to @xmath102 recombination lines . from geometrical considerations one would expect the inhomogeneities of one additional dimension , @xmath98 , to be at least half as large as the inhomogeneities of two dimensions , @xmath94 ; this works for large scale inhomogeneities . the power included in @xmath98 increases in the presence of small scale inhomogeneities . this is because the many independent thermal elements along each line of sight will be masked in the averaging process , greatly reducing the total @xmath94 compared to @xmath95 . therefore , @xmath98 would be expected to be much larger than @xmath94 . from the previous considerations , we estimate that @xmath103 ; this result is very model dependent which is reflected by the large range in the estimated @xmath104 value . this estimate together with the @xmath94(o@xmath68 ) value derived in the previous subsection yield @xmath105 . from the ratio of o ii recombination lines to lines in two regions of the orion nebula epte98 find that @xmath95(o@xmath106 . this @xmath95 value is representative of a very small fraction of the nebula , so we expect it to be a lower bound to the total @xmath95(o@xmath68 ) . at the same time it must be bigger than @xmath98(o@xmath68 ) because it represents an average over more than one line of sight . this argument , together with the @xmath94(o@xmath68 ) value estimated in the previous subsection , yields @xmath107 . by combining both estimates we find @xmath108 , which implies that there is power in small scale inhomogeneities . our prefered value of @xmath98 together with the @xmath94 measured in the previous subsection and equation ( 14 ) gives us a @xmath109 value . we are interested in the global @xmath95 value that we will define as : @xmath110 in order to derive the physical condtions in the region . in an region , like the orion nebula , where oxygen is only once or twice ionized , it is useful to define a value @xmath13 , that characterizes the degree of oxygen ionization , as : @xmath111 dl da};\ ] ] for such an region @xmath1 ( ) is given by ( peimbert , peimbert , & luridiana 2002 ) : @xmath112 ^ 2 \over t_0({\rm h}^+)^2},\ ] ] the first term corresponding to the temperature variations in the o@xmath68 zone , the second term corresponds to the variations in the o@xmath70 zone , and the third term corresponds to the thermal variations introduced by having two zones with different average temperatures . from the observations of peimbert & torres - peimbert ( 1977 ) and epte98 it follows that @xmath13 decreases from about 0.85 near the center of the nebula to about 0.15 in the outer regions . we will adopt a value of @xmath113 as representative of the whole object . it is necessary to estimate @xmath1(o@xmath70 ) . since @xmath114 is rarely measured we will use @xmath115 as a surrogate since we expect both temperatures to be similar as they arise from the same narrow + zone . r2003 present values of @xmath116 ; their measured values are in very good agreement with their measured values of @xmath89 ; therefore we will assume @xmath117 . it is impossible , however , to estimate @xmath118 from any set of observations available . from the geometrical arguments presented in the previous section and the similarities between the @xmath89 and @xmath119 presented by r2003 we will assume also that @xmath120 , and consequently that @xmath121 . to compute the final term in equation ( 19 ) we need to estimate the differences in temperatures between the o@xmath70 and o@xmath68 zones ( again we will use @xmath115 as a surrogate ) . our observations do not yield @xmath115 values , therefore we are forced to look for them in the literature . we are more interested in finding @xmath122 than @xmath115 to avoid possible biases , since the calibration is uncertain by a few hundred degrees ( see final paragraphs of sections 3.1 and 3.5 ) ; an additional advantage of such comparison is that papers in the literature cover regions considerably smaller than the one presented in this paper . epte98 have measured @xmath123 k and @xmath124 k ; while r2003 , for their slits 1 and 2 and assuming @xmath125 , obtain @xmath126 k and @xmath127 k. by averaging the innermost 9 regions of b91 , we find @xmath123 k and @xmath128 k. these three results give us an average @xmath129 k. as mentioned before , we have adopted a @xmath113 , giving a @xmath130 approximately 700 k higher than @xmath131 . the inputs presented above yield values of 0.011 , 0.010 , and 0.007 for the three terms in equation ( 19 ) . therefore the total magnitude of the temperature fluctuations becomes @xmath132 . notice that for values of @xmath13 in the 0.3 to 0.7 range the change in total @xmath1 is very small since @xmath121 , thus changes in @xmath13 will make very small changes in the sum of the first 2 terms of equation ( 19 ) , while the third term will change by less than 15% for that range of @xmath13 . the total @xmath133 value derived in this paper is considerably larger than the 0.004 value derived by peimbert et al . ( 1993 ) from the one dimensional model of the orion nebula presented by b91 . the type of temperature variations reported here have not been seen before because previous slit spectra determinations usually employed integrations over long areas of the nebula . the best slit spectrum sub - sampling until r2003 is in b91 , where the sample sizes were as short as 16 and one could see sample to sample changes in of about 300 k ( b91 , table 5 ) . their changes were comparable to the uncertainties in their derivations of from the line ratios , so this variation did not attract attention . however , the b91 variations are consistent with the small scale variations we see in our figure 3 . a much more valuable study is that of r2003 , where they employed the hst s stis ( space telescope imaging spectrometer ) in a long slit mode to determine in samples of 0.5x 0.5 in both and . they presented results for about 400 samples from each of four slit settings , two within our ccd1 field and the others near or off the southeast corner of our ccd2 field . it is possible to use this data to estimate a lower limit to @xmath95(h ii ) using the same procedure as in the previous section . from their slits 1 and 2 , those included in our map , it is found that @xmath134 k and @xmath135 k ; notice that these @xmath136 s corresponds to two rectangular columns of the nebula and not to a large volume . from these slits we also find @xmath137 and @xmath138 . from r2003 observations it is not possible to obtain direct estimates of the o@xmath68 and n@xmath69 temperature fluctuations along any given line of sight ; so we will use @xmath89 and @xmath116 as lower bounds to @xmath139 and @xmath140 . to estimate the two zone temperature term , the term due to the difference between @xmath141 and @xmath142 ( see equation ( 19 ) , we will adopt @xmath143 . this , together with @xmath141 and @xmath142 , yields a contribution of 0.009 . finally , adding these 3 terms , we obtain @xmath95(h ii)@xmath144 . this lower limit to @xmath95(h ii ) is consistent with our determination and with the @xmath95 determination by epte98 . moreover it is already a factor of 5 higher than the values predicted by one - dimensional photoionization models of the orion nebula , like the one by b91 . in this section we will review the large scale structure of the orion nebula ( 7.1 ) , present a conundrum arising from the presence of fine scale structure in images ( 7.2 ) , demonstrate that high resolution spectra and line ratio changes argue for a three dimensional fine scale structure ( 7.3 ) , argue that the ionization front is advancing into a highly clumpy pdr ( 7.4 ) , show how these clumps can lead to low shadows within the ionized layer that can help to explain the @xmath1 phenomenon in the region of low ionization regions ( 7.5 ) , and present a preliminary discussion of the effect of these clumps on the anomalous line - broadening in orion ( 7.6 ) . the accepted model for the orion nebula is that of a blister . the visible nebula is the result of a thin layer of photoionized gas on the front of a giant molecular cloud , with a dense photon - dominated region ( pdr ) ( van der werf et al . 1996 ) , also known ( young owl et al . 2000 ) as a photodissociation region , on the neutral side of the ionization front . gas is accelerated away from this ionization . the dominant ionizing star is @xmath0 ori c which lies at about 0.25 pc in front of the ionization front . in the foreground is an irregular veil of neutral material that produces most of the extinction ( odell 2002 ) . this model is reviewed in detail in a recent review article ( odell 2001b , henceforth od2001 . the basic physics is discussed in a review paper by ferland ( 2001 ) . the shape of the ionization front has been derived ( wen & odell 1995 , henceforth wo95 ) using a basic process originally proposed by gary ferland in b91 . this draws on the fact that at a substellar point , the surface brightness in a recombination line ( for example h@xmath11 ) is directly proportional to the local flux of ionizing photons . wo95 solved the more general problem , where the line of sight to a point in the nebula does not pass through the ionizing star . they showed that the ionization front is an irregular concave form , with an escarpment that explains the bright bar feature running near @xmath61 ori a and has a dominant hill to the southwest of @xmath0 ori c , where the orion - s radio and infrared sources are located . odell & doi ( 2003 ) have shown that the latter region is the source of multiple optical and radio outflows and probably contains a secondary source of star formation . this model was derived using a resolution of 2 . because of the progressive nature of the calculation of the shape of the surface , it is most accurate near the direction of @xmath0 ori c. the only diagnostics that we have along the line of sight into the nebula are the increasingly blue - shifted emission from the higher ionization states that are found further from the mif and forbidden line doublet intensities that indicate a decreasing density with increasing distance from the mif ( od2001 ) . these few tests that we have are consistent with the model . a detailed hydrodynamic model of the orion nebula has not been calculated , however , it is expected that the high density in the pdr region is the result of a shock moving into the molecular cloud and that the ionization front is slowly progressing into the neutral region . the hst images of the orion nebula are filled with detail , especially in the inner regions . at first blush , this appears to present a fundamental quandary if the standard model for the nebula is correct . one can derive a formal thickness ( l ) for the ionized layer , using the assumption that the density is constant and knowing the emissivity in the recombination line being observed . wo95 show that this value is characteristically about 0.13 pc for the central part of the nebula , which corresponds to an angle of 58 if the distance to the nebula is 460 pc ( bally , odell , & mccaughrean 2000 ) . they also show that if the density has an exponential decrease , its scale - height will be l/2 ( 29 ) and the emissivity would decrease with a scale - height of l/4 ( 15 ) . this means that in the standard model one should not be able to see variations in the structure of the nebula with size scales smaller than about 15 . one certainly see lots of fine - scale ( less than 15 ) structure in ground - based and hst images . some of this can be attributed to variations in the amount of extinction ( odell & yusef - zadeh 2000 ) , most of which is caused by the foreground neutral veil of material ( od2001 ) . however , the extinction corrected images in h@xmath11 show significant structure right down to the limit of the extinction corrections , which was about 1.7 . this is shown in figure 7 , which is a profile of the extinction corrected h@xmath11 image of the nebula ( odell & yusef - zadeh 2000 ) derived with a resolution of 1.7 and taken across the velocity sample region designated in figure 1 . there are even smaller features , which appear to be a continuation of the structure , as shown in the hst images , while the smallest features are due to the outflows from young stellar objects and the shocks they create in the low density ambient gas ( odell & doi 2003 , od2001 ) and are seen in the hst images to be sub - arcsecond in extent . the presence of fine scale structure associated with the main emitting layer of ionized gas presents a fundamental challenge to the accepted model of the nebula . the structure is there , for we see it , but if the ionized gas was simply a smooth flow of gas away from a homogeneous ionization front , we should not see any small scale structure . we demonstrate in 7.5 and 7.6 , that the resolution of the quandary lies in the emitting layer having a much more complex structure than assumed in published models . there are two additional lines of evidence indicating that small scale structure exists within the ionized layer , these being the structure seen in the velocities and fine scale variations in the ionization . both argue for structure down to at least a few arcseconds . the internal velocities within the orion nebula have been the subject of investigation for nearly a century , beginning with a fabry - perot study by fabry ( buisson , fabry , & bourget 1914 ) , an extensive mapping using photographic detectors ( wilson et al . 1959 ) , and a series of studies in various emission lines using ccd detectors ( [ o iii ] , castaeda 1988 ; [ o ii ] , jones 1992 ; [ o i ] , odell & wen 1992 ; [ s iii ] , wen & odell , 1993 ) at the kitt peak national observatory coud feed spectrograph at selected position angles with respect to the trapezium stars . this work was used to compare the radial velocities with the predictions of turbulence within a thin ionized layer ( von hrner 1951 ) and the results are summarized in od2001 . a more ambitious and complete mapping of the radial velocities is now being created as part of the rice university phd thesis of takao doi , who is creating velocity images of the huygens region at resolutions of 10 km s@xmath145 and 2 in the h@xmath11 , 5007 , and 6583 lines , and a similar study at the instituto de astronoma , unam , morelia , mxico in the red doublet lines by w. j. henney & m. t. garca - daz that uses in part data obtained by co - author odell . with the permission of our colleagues we have used their north - south orientation slits to create figure 8 and figure 9 . these are spectra taken from the velocity sample region indicated in figure 1 and used for the creation of the surface brightness profiles of figure 7 . the instrumental full width at half maximum intensity ( fwhm ) was very close to 10 km s@xmath145 in each case . although the original samples along the slit were at intervals of 0.64 , the seeing was characteristically about 2 during the observations . the former figure shows a half tone depiction of the spectra in lines characterizing various parts of the ionized layer and the latter presents contour diagrams of the same data . the power of this approach lies in the fact that various emission lines are characteristic of different samples of the ionized layer . although this fact is implicit in the discussion of several texts , perhaps the best illustration is in a discussion of structure of the helix nebula ( odell 1998b ) . emission comes from throughout the ionized layer , its emissivity being weighted towards the lower regions along the line of sight . the forbidden line emissivity in , , and are all weighted towards the higher regions along the same line of sight . delineation of information about structure along the line of sight is also provided by the ionization differences . emission will arise exactly in the ionization front , where ionization of hydrogen rapidly transitions from nearly unity to nearly zero . a good fraction of the emission will occur in the adjacent thin + zone , while the emission will arise in the outermost ( closest to the observer ) + zone . we see in figure 8 and figure 9 that there is considerable velocity variation along the slit and that the velocity in one ion is not well correlated with that in the other ions . thermal broadening of the h@xmath11 line precludes a detailed analysis of its velocity structure , but is low enough in the other , more massive , ions to see the differences clearly . one sees that the narrower the expected emitting region , the more structure there is along the slit . emission gives evidence for velocity structure with a scale of the seeing values of 2 and emission has a characteristic size scale down to about 5 . these data argue that structure exists within the main ionized layer at sizes down to at least 2 and that gas of different states of ionization is moving at different velocities . it is the spatially unresolved average velocity along the slits which give the characteristic progression to more blueshifted radial velocities at higher states of ionization that is the earmark of a thin blister of gas expanding towards the observer . the second line of evidence for fine scale structure within the nebula is from comparison of emission from very different states of ionization or conditions of emissivity . for this comparison we used two pairs of images , whose closeness in wavelength essentially eliminate the effects of interstellar extinction . the first pair used the f487n images , which are primarily due to h@xmath12 emission whose emissivity is like that of h@xmath11 in that it selectively comes from the lower regions , and the f502n images , which are primarily emission from 5007 high collisionally excited regions . the second pair used the f656n h@xmath11 filter and the f658n filter . the considerations for the emissivity are similar in the second pair of filters , but the emitting layer of the emission should be thinner . the hydrogen emission arises from throughout the ionized zone , but the and emission arise from selected zones , as described earlier in this section . the results of these comparisons are shown pictorially in figure 10 . one sees that there are wide variations in these ratios . some of these changes are due to shocks caused by outflows from pre - main sequence stars ( odell & doi 2003 ) while others such as the bright bar region in the southwest corner of the images is due to large scale topographic features of the ionization front ( odell & yusef - zadeh 2000 ) . in addition to these , there are many additional features . a more quantitative depiction is shown in figure 11 and figure 12 , where profiles along the same region as the velocity samples are presented . we see fine scale structure in the ratios down to nearly the size of the 1x2 bins over which the data was averaged , although a characteristic value would be 3 to 4 . the probable errors of the ratios were derived from the total counts used in their calculation and are 1.8 % for the f658n / f656n ratio and 1.2 % for the f502n / f487n ratio . as expected , the peaks are sharper in the f658n / f656n ratio , indicating that the sources are small concentrated regions . the fact that there is a general anti - correlation of the ratios indicates that the variations are largely due to ionization structure , rather than temperature differences . in summary , we can say that both the spatially resolved high resolution velocity data and the line ratio data indicate structure in the main ionized layer of 3 to 5 . a comparison with a profile of the results of the determination of is also useful and is shown in figure 13 , where a profile along the same region as the velocity data is shown . we see that the proplyd 141 - 301 shows the common characteristic of the other proplyds , in having a higher than normal formal value of @xmath35 , but , as shown in 4.3 , this is probably due to collisional de - excitation of the electron state giving rise to the 5007 emission . the `` jet '' feature shows a similar enhancement , which may also be due to having densities comparable to the critical density for this nebular transition , but it could be real , being the result of shock heating at the interface with the ambient nebular gas . aside from these features , there are variations in all along the slit . in the well exposed region from 50 to 150 we calculate that the probable error of is 0.5% i.e. 45 k. , which is small as compared with the observed variations . there are no obvious general patterns of changes with the extinction corrected surface brightness in h@xmath11 , the velocity changes , nor the line ratios . there is considerable evidence that the pdr behind the ionization front of the orion nebula is very clumpy ( tauber , goldsmith , & dickman 1991 , hogerheijde , jansen , & van dishoeck 1995 ) . hcn and hco@xmath69 observations by young owl et al . ( 2000 ) demand that the portion of the pdr seen obliquely behind the bright bar feature possesses dense clumps ( 3 x 10@xmath146 @xmath66 ) embedded within an interclump medium of lower density ( 5 x 10@xmath147 @xmath66 ) . more direct observations come from imaging in h@xmath148 . van der werf et al . ( 1996 ) show that clumps exist from a scale of 40 down to their resolution limit of 1.5 . the highest angular resolution study in h@xmath148 is that of vannier et al . ( 2001 ) who studied a region northwest of @xmath0 ori c at a resolution of 0.15 . again , they saw clumpy structure , in this case going down to 1.5 - 2.0 . a clumpy structure in the pdr is not unexpected and there would be several instability mechanisms that operate there ( capriotti 1973 , brand 1981 , vishniac 1994 , garca - segura & franco 1996 , williams 1999 ) . the implications of these clumps for the nature of the ionized layer of gas in the nebula are very important since it is into these clumps and the interclump medium that the ionization front advances . where the density is lower , the advance will be more rapid . the expected rate of advance of the ionization front is not known as the hydrodynamic problem has not been solved . the orion nebula is old enough that it is in a state of quasi - equilibrium and the velocity will depend on many factors , including whether or not the photoionizing star @xmath0 ori c has a velocity component towards or away from the molecular cloud . this means that one would expect the ionization front advancing into a clumpy pdr to move beyond the dense clumps and be retarded at the clumps , so that a microscopic view of the ionization front would be like that of a stubble - field . the region behind a dense knot would then be shadowed from ionizing lyman continuum ( ) radiation coming directly from @xmath0 ori c and would be illuminated only by diffuse ( sometimes called scattered ) photons . where the density is insufficient to form a quasi - static ionization front , the clumps will simply retard the advance of the ionization front , while the dense clumps will form a small local ionization front that will exist until it has lost sufficient material through photoevaporation that the ionization front can penetrate , then totally disperse it . the shadowed regions behind the knots within the ionized layer can be sources of low temperature emission and of temperature variations , therefore they will increase the @xmath1 value . the photoionization equilibrium in such a shadowed region was considered 30 years ago ( van blerkom & arny 1972 , kirkpatrick , 1972 , capriotti 1973 ) , but it is only recently that the more general case has been considered ( cant et al . 1998 , henceforth c98 ) . the region illuminated only by diffuse lyc will see a lower temperature radiation field because most of the recombinations of that produce the diffuse lyc will result in a photon just above the ionization threshold , while the direct stellar field will have more high energy photons and produce higher regions . c98 have solved the general problem , rather than trying to apply their calculations to only a specific object , including the role of ambient gas density and the time variability of the shadowed zone structure . when the combination of the stellar radiation density ( the diffuse radiation field density is about 15 % of that ) and gas density is right an ionization front can form immediately at the edge of the shadowed cylinder of gas . as the radiation density increases and/or the density decreases , the ionization front can form within the shadowed cylinder and even produce a condition where the shadowed cylinder is completely photoionized . c98 argue that within the ionized shadow zone will be about 6000 k , considerably cooler than that of the directly illuminated gas . the radiation from the shadowed zone can be seen by the distant observer for two reasons . in some cases the angle between the line of sight from the observer to @xmath0 ori c can be quite different from the angle between @xmath0 ori c and a local spot on the ionization front , allowing direct view of the shadowed region . even when that is not the case , a significant fraction of the radiation from the shadowed region will be scattered by dust in the nearby pdr , in the same way that radiation from the ionized layer is scattered and produces the observed redward asymmetry of the emission lines ( odell , walter , & dufour 1992 , henney 1998 ) . shadowed regions have already been detected in the orion nebula . in a study of both the cometary knots in the helix nebula and the proplyds in the orion nebula , odell ( 2000 ) showed that the proplyds cast ionization shadows as long as 0.2 pc . the proplyds are optically thick to lyc radiation ( odell 1998a ) and in that sense resemble the optically thick clumps we argue exist at the main ionization front of the nebula . the proplyd ionization shadows are most visible when the geometry is such that the shadow falls into the ionized layer that provides most of the radiation identified as the nebula . odell ( 1998a ) showed that the shadows of the proplyds covered about 0.5% of the view from @xmath0 ori c. by assuming : pressure equilibrium , @xmath149 , @xmath150(n@xmath70 ) = 10,100 k ( b91 , epte98 , r2003 ) for the directly illuminated regions , and @xmath150(n@xmath70 ) = 6,000 k ( c98 ) for the shadowed regions we obtain a contribution of 0.0093 to @xmath1(o@xmath70 ) ( see equation ( 10 ) ) , that is a contribution of 0.0043 to the total @xmath1 ( due to second term of equation ( 19 ) ) . to reach the estimated 0.011 contribution to @xmath1 due to the o@xmath70 zone , we need an additional 0.7% of the volume of the nebula to be in a different type of ionized shadowed region . in what follows we will propose the presence of high density clumps inside the ionized layer left behind by the advancing ionization front . in the model we propose ( illustrated in cartoon form in figure 14 ) the clumps would not be permanent , rather , they would exist in the ionized layer only until they are destroyed by photoevaporation . the cool columns of gas behind them would then come to the higher equilibrium appropriate for illumination by @xmath0 ori c rather than diffuse lyc . one can approximate the results of an ionization front advancing into a clumpy pdr , as described in 7.4 . we dont know the velocity of this advance . it must be small , because the tracers of the ionization front ( [ o ii ] and [ s ii ] ) are blueshifted 2.5 km s@xmath145 with respect to the pdr tracers ( co , c ii ) ( odell 2001a ) . this blueshift is probably due to the hot gas tracers coming from a layer that has already been accelerated by the pressure gradient , rather than being the result of @xmath0 ori c moving away from the pdr . the radial velocity of @xmath0 ori c is highly uncertain because of variability probably arising in the extended atmosphere of the star . it is unlikely that this star is moving at a high velocity with respect to the other cluster stars and their velocity is the same as the pdr ( odell 2001a ) within the uncertainty of 3 km s@xmath145 . in the discussion below we will assume that the ionization front is advancing into the pdr at a rate of 2 km s@xmath145 , but it should be recognized that this could be off by a factor of two . in 7.5 we showed that the fluctuations in the surface brightness of the nebula indicate structure with a height of about 0.025 pc . this would mean that the characteristic clump will have been illuminated by lyc radiation for about 1.3x10@xmath147 years , which we will call the kinematic age . if we can determine the column density and rate of loss of atoms per unit time and area , we can calculate the time that the clumps can survive photoevaporation and have a check on our basic model . if the clump diameters are about 0.0044 pc ( corresponding to 2 ) and the densities are 3x10@xmath146 @xmath66 ( young owl et al . 2000 ) the average column density of a clump would be 2.8x10@xmath151 atoms @xmath152 . the corresponding mass would then be about 7x10@xmath153 g ) . the photoevaporation rate for these clumps will be very similar to those of the proplyds , which henney et al . ( 2002 ) have shown to be 8x10@xmath154 m@xmath155 yr@xmath145 . since these objects are comparable in size to the clumps at the ionization front , their average rate of loss of atoms is relevant , this being 2x10@xmath156 atoms @xmath152 s@xmath145 . in this case , the lifetime of the clumps would be about 5,000 yrs . however , the proplyd measured by henney et al . ( 2002 ) must be much closer to @xmath0 ori c than the ionization front , so that the rate of loss has been overestimated and the survival age of the clumps underestimated . this means that the survival age of the clumps and their kinematic ages are compatible . they are , however , close enough that the process involved must be a dynamic one , with any one clump surviving for only about 10@xmath147 years , but the number of clumps present at any time would be about constant as new clumps are revealed by the advancing ionization front . we can estimate the effect of these shadowed regions if we know their emissivities and volumes . if the shadowed regions exist long enough that dynamical equilibrium is established with their surroundings ( c98 ) , then their volume emissivity will be four times that of the ambient lower density and higher gas . if pressure equilibrium is not established , then the emissivity increase will be about two times that of the ambient gas . if one interprets the non - jet fluctuations in the extinction corrected h@xmath11 surface brightness ( figure 7 ) as being due to knots at various distances from @xmath0 ori c , then they lie in a range of about 0.025 pc or 11 , which is comparable to the emissivity scale - height of 15 derived in 7.2 . this means that the shadowed regions occur within the zone where most of the emission occurs . to calculate the contribution to @xmath1(o@xmath70 ) and @xmath1 due to the shadows behind these clumps we need an estimate of the fraction of the emission produced in the shadowed regions . assuming : a ) the typical clump density to be @xmath157 @xmath66 , b ) the average density of the ionized media to be @xmath158 @xmath66 , and c ) that half of the ionized matter originated from the photoevaporation of these clumps , it follows that the volume originally occupied by the clumps must be about 1/1200 of the total volume . however , the shadowed volume is larger than the volume occupied by the clumps ; we know that the shadows are approximately 0.025 pc long , while each clump is approximately 0.0044 pc across , so the shadowed volume can amount to approximately 1/200 of the volume of the region closest to the main ionization front , before these clumps are photoevaporated . since these clumps survive long enough to cast 0.025 pc shadows , and the emissivity scale - height of the orion nebula is of 0.034 pc the total volume of the shadowed regions is about 1/250 of the effective emitting volume . there are three possible physical conditions in the shadowed regions : a ) that they are completely neutral , in this case they will not contribute to @xmath1(o@xmath70 ) nor to @xmath1 ; b ) that they are completely ionized , but not in pressure equilibrium with their surroundings , then their contribution to @xmath1(o@xmath70 ) would be about 0.0038 and to @xmath1 of about 0.0016 ; and c ) that they are completely ionized and in pressure equilibrium with their surroundings , then their contribution to @xmath1(o@xmath70 ) would be about 0.0075 and to @xmath1of about 0.0032 . the contribution of the shadows , including the shadows of the clumps and of the proplyds , amounts to @xmath159 , that corresponds to @xmath160 . note that in the shadowed regions oxygen is either neutral or once ionized , and that therefore the shadowed regions do not contribute to the @xmath1(o@xmath68 ) value it can be shown that the ionized part of the proposed clumps has a very small effect on @xmath1 . the clumps are expected to be close to the pdr , and their ionized part will have a larger density than the interclump medium , so we expect the oxygen in the ionized fraction of the clumps to be singly ionized . additionally the clumps cover a total solid angle of about 0.03 steroradians , and therefore absorb less than 1/400 of the ionizing photons , about 1/800 . this implies that their emissivity amounts to about 1/800 of the total emissivity of the nebula and that the physical conditions in the high density photoevaporated region will produce a negligible effect in the increase of @xmath1(o@xmath70 ) or @xmath1 . there is an unexplained component of line broadening in the orion nebula that may be explained by the clumpy model for the ionized layer that we propose . this anomaly first surfaced with castaeda s ( 1988 ) study of , and remained with the inclusion of other massive ions ( , jones 1992 ; , odell & wen 1992 ; , wen & odell 1993 ) . after quadratic subtraction of the instrumental , thermal , and turbulent components of broadening , there remains a large and unexplained line width having fwhm@xmath6410 km s@xmath145 . the large thermal width of the lines makes it difficult to accurately determine the extra width there , however , wilson , et al.s ( 1997 ) study of the h64@xmath11 line indicates an unexplained broadening component ( called the turbulent velocity in their paper ) of 19.6@xmath390.9 km s@xmath145 . this extra broadening component has been re - addressed using spectra made with the keck telescope for the study of proplyds and their jets ( henney & odell 1999 , bally , odell , & mccaughrean 2000 . the keck data were very useful as they have the highest velocity resolution ( 6 km s@xmath145 ) of any of the optical wavelength studies and allow examination of all of the primary lines of , , , , he i , h@xmath11 , and h@xmath12 at exactly the same positions . samples were taken from 14 long spectra centered on 137 - 349 , 150 - 353 , 163 - 357 , 170 - 337 , 177 - 341 , 182 - 413 , and 244 - 440 ( the designation system is described in odell & wen 1994 ) and 28 long spectra taken at the positions 137 - 349 , 150 - 353 , and 163 - 357 . after correction for instrumental , thermal , and turbulent ( assumed to be 1 km s@xmath145 ) broadening , the remaining line widths ( fwhm in km s@xmath145 ) were h@xmath11 20.9@xmath391.3 , h@xmath12 18.8@xmath392.2 , he i 5876 18.4@xmath392.9 , 6583 10.6@xmath391.4 , 6300 9.0@xmath392.1 , 5007 13.0@xmath393.7 , 4959 11.3@xmath392.1 , 6731 11.3@xmath392.4 , and 6312 11.8@xmath391.9 . the only source of [ o ii ] data are those found in the thesis of michael jones ( 1992 ) who found for 10.5@xmath392.5 for the central 80x 80 . the longer keck slit spectra were subjected to a separate analysis using various fractions of the total slit length and no variation of the results were seen , i.e. there was not a relation between sample size and the residual line width . the radio ( h64@xmath11 , 19.6@xmath390.9 ) and optical ( h@xmath11 , 20.9@xmath391.3 and h@xmath12 , 18.8@xmath392.2 ) are in agreement within their individual uncertainties . the unexplained broadening in the emission lines that arise from the lower temperature component of the nebula is twice that of lines arising from the higher temperature component . we see that the recombination lines of h i and he i that preferentially come from cool regions have a clearly larger unexplained width ( average 19.4 km s@xmath145 ) than the collisionally excited forbidden lines that preferentially come from hot regions ( average 10.9 km s@xmath145 ) . this strongly argues against the broadening arising from alvn waves ( od2001 ) since they would have little dependence upon . the explanation certainly does not lie with there being a systematic difference in , because the sense of the correlation is opposite to that expected . the clearcut differences in the extra broadening in collisional and recombination lines is probably related to the the mechanism giving rise to the temperature inhomogeneities . can the broadening be due to the clumpy structure that we argue can explain the temperature inhomogeneities ? odell & wen ( 1992 ) argue that clumps having sizes of less than 2.3 could produce the extra broadening , but this argument was made before knowledge of the larger broadening of the hydrogen lines . the steady process of destruction of the clumps by photoevaporation would certainly give rise to flows away from the clumps . however , the line broadening due to this should selectively be a blueward distortion of the integrated line - of - sight line profile , except for the final stage of destruction , when material would be flowing in all directions . however , the initial phase of a shadowed region would be accompanied by material flowing inward towards the low pressure zone that would exist when the shadowed drops and the shadowed region is moving towards a state of hydrostatic equilibrium . it is obvious that these situations need to be theoretically modeled . although it is not now possible to quantitatively link them , it is likely that the explanation of the electron temperature fluctuations and the line widths are related . from our @xmath9(o@xmath68 ) map of the orion nebula , that includes 1.5 x 10@xmath146 independent temperature determinations , we have found that @xmath10(o@xmath68)@xmath161 . from our @xmath10(o@xmath68 ) value , together with geometrical considerations and other observations in the literature , we estimate that @xmath1(o@xmath68)@xmath162 . note that the total @xmath1(o@xmath68 ) is larger than @xmath10(o@xmath68 ) because in addition to the variations across the plane of the sky it includes the temperature variations along the line of sight . from our @xmath1(o@xmath68 ) value and comparisons between the temperatures in the low- and high - ionization zones , the o@xmath69 and o@xmath68 zones , we find that @xmath1(h ii)@xmath163 . @xmath1(h ii ) is different to @xmath1(o@xmath68 ) because in addition to the variations in the o@xmath68 zone includes the variations in the o@xmath69 zone and the difference in the average temperature between both zones . our derived @xmath1(h ii ) value is 7 times higher than those obtained from homogeneous one - dimensional photoionization models of the orion nebula . a combination of sources needs to be found to explain the large @xmath95 values observed in the orion nebula . many possible causes of temperature variations have been discussed in the literature like : density variations , deposition of mechanical energy , deposition of magnetic energy , presence of shadowed regions , chemical inhomogeneities , dust heating , and transient effects due to changes in the ionizing flux . from observations of variations in surface brightness , columnar electron temperature , radial velocity , and ionization , we have established the presence of significant small scale structure at the 1 - 5 arc sec level . the small scale velocity and ionization structures are probably strongly linked with the temperature structure . the presence of this small scale structure is compatible with a new model for the structure of the nebula near the ionization front . we posit that this front is advancing into the highly clumpy photodissociation region revealed by radio and infrared observations . these clumps possess ionized shadows that are of low . a quasi - static equilibrium exists in which any one clump survives the process of photoevaporation only for a short time . the process may be common to most regions which have a portion that is ionization bounded . there is an unexplained component of line broadening in the orion nebula that may be explained by the clumpy model for the ionized layer that we propose . we have discussed the presence of shadowed regions due to proplyds and to high density clumps near the pdr . we estimate that the shadowed regions can contribute @xmath160 to the total @xmath95 value , which together with the homogeneous one dimensional ionization structure provide only from about one third to about half of the observed @xmath95 value . additional sources of temperature variations are needed to explain the total @xmath95 observed value . moreover , since in the shadowed regions o is only once ionized , these regions could be responsible for temperature variations in the o@xmath70 zone , but not in the o@xmath68 zone , and since the @xmath1(o@xmath68 ) value is very large we need one or several additional sources of temperature variations to explain it . there is an anomalous broadening mechanism operating in the orion nebula which is unexplained . it broadens h and he recombination line emission by about 19 km s@xmath145and the forbidden line emission from heavy ions by about 11 km s@xmath145 . since these two types of lines selectively rise from the low and high temperature components along a line of sight , it appears likely that this process is related to the @xmath1 problem . michael richer and will henney provided valuable support in obtaining the high velocity resolution spectra at the mexican national observatory at san pedro martir . will henney also provided several useful comments on an earlier version of this paper . takao doi kindly shared before publication the results of his orion nebula radial velocity program , which allowed us to generate figure 8 and figure 9 . henney also provided observations on the expected rate of motion of the ionization front in the orion nebula . gary j. ferland provided estimates for the heating rates . christopher a. coco measured the keck spectra line widths as part of his senior thesis at rice university . the authors are grateful to patrick hartigan of rice university and robert rubin of the ames research center for discussions on some aspects of this work . cro s work was supported in part by grant go-9141 from the stsci to vanderbilt university . mp s work was supported in part by grant in 114601 from dgapa unam . baldwin , j. a. , ferland , g. j. , martin , p. g. , corbin , m. r. , cota , s. a. , peterson , b. m. , & slettebak , a. 1991 , apj , 374 , 580 ( b91 ) bally , j. , odell , c. r. , & mccaughrean , m. j. 2000 , aj , 119 , 2919 brand , p. w. j. l. 1981 , mnras , 197 , 217 buisson , h. , fabry , c. , & bourget , h. 1914 , apj , 40 , 241 cant , j. , goudis , c. , johnson , p. g. , & meaburn , j. 1980 , a&a , 85 , 128 cant , j. , raga , a. , steffen , w. , & shapiro , p. r. 1998 , apj , 502 , 695 ( c98 ) capriotti , e. r. 1973 , apj , 179 , 495 castaeda , h. o. 1988 , apjs , 67 , 93 esteban , c. 2002 , rev . mexicana astron . , 12 , 56 esteban , c. , peimbert , m. , torres - peimbert , s. , & escalante , v. 1998 , mnras , 295 , 401 ( epte98 ) ferland , g. j. 2001 , pasp , 113 , 41 garca - segura , g. , & franco , j. 1996 , apj , 469 , 171 garnett , d.r . 1992 , aj , 103 , 1330 . gruenwald , r. , & viegas , s.m . 1992 , apjs , 78 , 153 hartigan , p. 1999 , apj , 526 , 274 henney , w. j. 1998 , apj , 503 , 760 henney , w. j. , & odell , c. r. 1999 , aj , 118 , 2350 henney , w. j. , odell , c. r. , meaburn , j. , garrington , s. t. , & lpez , j. a. 2002 , apj , 566 , 315 hogerheijde , m. r. , jansen , d. j. , & van dishoeck , e. f. 1995 , a&a , 294 , 792 holtzman , j. a. , burrows , c. j. , castertano , s. , hester , j. j. , trauger , j. t. , watson , a. m. , & worthey , g. 1995 , pasp , 107 , 1065 jones , m. r. 1992 , phd thesis , rice univ . , houston , tx kirkpatrick , r. c. 1972 , apj , 176 , 381 liu , x .- w . 2002a , rev . mexicana astron . , 12 , 70 meaburn . j. , blundell , b. , carling , r. , gregory , d. f. , keir , d. , & wynne , c. g. 1984 , mnras , 210 , 463 odell , c. r. 1998a , aj , 115 , 263 odell , c. r. 1998b , aj , 116 , 1346 odell , c. r. 2000 , aj , 119 , 2311 odell , c. r. 2001a , pasp , 113 , 29 odell , c. r. 2001b , araa , 39 , 99 ( od2001 ) odell , c. r. 2002 , rev . mexicana astron . , 12 , 12 odell , c. r. , & doi , t. 1999 , pasp , 111 , 1316 ( od99 ) odell , c. r. , & doi , t. 2003 , aj , 125 , 277 odell , c. r. , hartigan , p. , lane , w. m. , wong , s. k. , burton , m. g. , raymond , j. , & axon , d. j. 1997 , aj , 114 , 730 ( o97 ) odell , c. r. , hartigan , p. , bally , j. , & morse , j. a. 1997 , aj , 114 , 2016 odell , c. r. , & hubbard , w. b. 1965 apj , 142 , 591 odell , c. r. , walter , d. k. , & dufour , r. j. 1992 , apj , 399 , l67 odell , c. r. , & wen , z. 1992 , apj , 387 , 229 odell , c. r. , & wen , z. 1994 , apj , 436 , 194 odell , c. r. , & yusef - zadeh , f. 2000 , aj , 120 , 382 osterbrock , d. e. , tran , h. d. , veilleux , s. 1992 , apj , 389 , 305 peimbert , a. 2003 , apj , 584 , 735 peimbert , a. , peimbert , m. , & luridiana , v. 2002 , apj , 565 , 668 peimbert , m. 1967 , apj , 150 , 825 peimbert , m. 1995 , in the analysis of emission lines , eds . r. e. williams , & m. livio ( cambridge ) , 165 peimbert , m. 2002 , rev . mexicana astron . , 12 , 275 peimbert , m. , peimbert , a. , & ruiz , m. t. 2000 , apj , 541 , 688 peimbert , m. , storey , p. j. , & torres - peimbert , s. 1993 , apj , 414 , 626 . peimbert , m. , & torres - peimbert , s. 1977 , mnras , 179 , 217 rubin , r. h. , bhatt , n. j. , dufour , r. j. , buckalew , b. a. , barlow , m. j. , liu , x .- w . , storey , p. j. , balick , b. , ferland , g. j. , harrington , j. p. , & martin , p. g. 2002 , mnras , 334 , 777 ( r2002 ) rubin , r. h. , martin , p. g. , dufour , r. j. , ferland , g. j. , blagrave , k. p. m. , liu , x .- w . , nguyen , j. f. , & baldwin , j. a. 2003 , mnras , in press , ( astro - ph/0212244 ) , ( r2003 ) stasiska , g. 2002 , rev . mexicana astron . , 12 , 62 tauber , j. a. , goldsmith , p. f. , & dickman , r. l. 1991 , apj , 375 , 635 torres - peimbert , s. , & peimbert , m. 2003 , in planetary nebulae and their role in the universe , iau symposium 209 , in press , ( astro - ph/0204087 ) van blerkom , d. , & arny , t. t. 1972 , mnras , 156 , 91 van der werf , p. p. , stutzki , j. , sternberg , a. , & krabbe , a. 1996 , a & a , 313 , 633 vannier , l. , lemaire , j. l. , field , d. , pineau des forts , g. pijpers , f. p. , & rouan , d. 2001 , a&a , 366 , 651 vishniac , e. t. 1994 , apj , 428 , 186 von hrner , s. 1951 , zap , 30 , 17 walter , d. k. , odell , c. r. , hu , x. , & dufour , r. j. 1995 , pasp , 107 , 686 wen , z. , & odell , c. r. 1993 , apj , 409 , 262 wen , z. , & odell , c. r. 1995 , apj , 438 , 784 ( wo95 ) williams , r. j. r. 1999 , mnras , 310 , 789 wilson , o. c. , mnch , g. , flather , e. m. , & coffeen , m. f. 1959 , apjs , 4 , 199 wilson , t. l. , filges , l , codella , c. , reich , w. , & reich , p. 1997 , a&a , 327 , 1177 young owl , r. c. , meixner , m. m. , wolfire , m. , tielens , a. g. g. m. , & tauber , j. 2000 , 540 , 886 lccccccl e@xmath17 & 16.56 & 17.40 & 19.87 & 22.79 & 601.9 & 22.0 4 & 31.19 + @xmath164 & 0.50 & 0.52 & 0.59 & 0.64 & & 0.78 & 0.79 + @xmath165 & 0.005 & & & & & 0.24 & 0.045 + s & 0.061 & 0.074 & 0.084 & 0.091 & 0.125 & 0.145 & 0.145 +

high energy physics experiments in recent years have confirmed the predictions of the standard model ( sm ) , a renormalizable , chiral gauge theory based on the group @xmath3 . in particular , the low energy dynamics of fermions and gauge bosons of the theory have been tested with great accuracy . the origin of masses of these fundamental particles , however , remains a mystery . in the sm , masses arise through the vacuum expectation value ( vev ) of a scalar field doublet , which spontaneously breaks the electroweak ( ew ) symmetry @xmath4 . a physical , neutral scalar field , the so called higgs field , appears in the spectrum . information on the mass of this scalar particle may be obtained through the quantum corrections that it induces to the masses and couplings of the ew gauge bosons . consistency of the sm predictions with experimental observations is improved for small values of the higgs mass , @xmath5 , close to the current experimental bound , @xmath6 gev @xcite . there are several aspects of the mechanism of the origin of mass that demand explanation . on one hand , the scale of the spontaneous symmetry breaking is governed by the size of the scalar mass parameter appearing in the higgs field effective potential , and is much smaller than the planck scale , the only known mass scale in nature , besides the dynamical generated qcd scale @xmath7 . moreover , this scale is associated with a negative value of the squared higgs mass . there is , however , no dynamical explanation for the origin of the higgs effective potential , or for the associated breakdown of the ew symmetry . finally , the hierarchy of the fermion masses of the different generations remains unexplained . warped extra dimensions provide a theoretically attractive framework for the solution of the hierarchy problem of the sm . for one extra spatial dimension as in rs1 @xcite , the curvature @xmath8 in the extra dimension induces a warp factor on the four dimensional metric , which depends on the position in the extra dimension . the coordinate point , @xmath9 , is associated with a trivial scale factor , leading to natural scales of the order of the planck scale . the other boundary in the compact dimension , @xmath10 , is associated with natural scales that are smaller by an exponential factor @xmath11 . for natural values of @xmath8 of the order of the planck scale and values of @xmath12 , a higgs field , located at @xmath13 will naturally acquire a vev of the order of the weak scale . neither the origin of the breakdown of the ew symmetry , nor the hierarchy of fermion masses , however , are explained by these considerations . the propagation of fermions and gauge fields in the extra dimension enables a possible explanation for the fermion mass hierarchy . indeed , chirality of the fermion zero modes is ensured by imposing an orbifold symmetry @xmath14 . even fields under this orbifold symmetry acquire zero modes whose masses arise from the vev of the higgs field . for fermion fields , the equations of motion demand that fields of opposite chirality to those having zero modes are odd under the orbifold symmetry and therefore have no zero modes . finally , localization in the fifth dimension is controlled by a mass parameter @xmath15 . the variation of @xmath15 by factors of order one induces exponential variations in the overlap of the fermion wave functions with the higgs field wave function and therefore on the induced masses for the fermions @xcite,@xcite . third generation left handed chiral quark fields which couple strongly to the higgs are associated with a mass parameter @xmath16 , and therefore their zero - modes are localized towards the so - called infra - red ( ir ) brane at @xmath10 . on the contrary , first and second generation left handed chiral fermions have mass parameters @xmath17 and the zero modes are localized towards the so - called ultraviolet ( uv ) brane at @xmath9 . gauge - higgs unification models @xcite provide a solution to the last mysterious aspect of the sm , while leading to a dynamical origin for the higgs field effective potential . the gauge symmetry of the sm is extended in the bulk and broken at the boundaries . the higgs is associated with the fifth component of gauge fields in the direction of the broken gauge symmetry . the gauge bosons associated with the broken symmetry are odd under the orbifold symmetry and the fifth component is guaranteed to be even , leading to the presence of massless scalar fields in the theory , with no potential at tree - level . the coleman - weinberg higgs potential is then determined by quantum corrections . for certain values of the fermion mass parameters , this effective potential leads to the higgs field acquiring a non - zero vev and to the proper generation of gauge boson and fermion masses . warped extra dimensions have a rich spectrum of four dimensional kaluza - klein ( kk ) excitations . in particular , after ew symmetry breaking , the kk modes of the weak gauge fields mix with the would - be zero modes , inducing important modifications to the masses and couplings of gauge fields at tree - level . such large modifications would render the theory inconsistent for values of the gauge field s masses at the reach of the lhc @xcite . recently , it was realized that these large corrections may be prevented by extending the weak gauge symmetry in the bulk of the extra dimension to incorporate the custodial symmetry : @xmath18 @xcite . moreover , a left - right parity symmetry , @xmath19 is necessary to prevent large corrections to the bottom - quark couplings @xcite . a gauge extension of the ew sector of the model to @xmath20 provides a framework in which gauge - higgs unification and custodial symmetry are possible . the higgs potential is fully calculable at one - loop level and a prediction for the higgs boson mass may be obtained . consistency of the theory at the quantum level , however , demands that the possible large corrections to precision electroweak parameters induced by the fermions of the theory be computed . it was recently shown that complete consistency may be obtained for certain values of the mass parameters associated with the fermions of the third generation @xcite , provided the higgs remains light , as is expected in gauge - higgs unification models . in this work , we compute the dynamically generated higgs potential in models similar to the ones previously analyzed in refs . the higgs potential is calculated by means of a generalization of the framework discussed in ref . @xcite and the specifics of our model are complementary to the one discussed in ref . we find that a consistent breakdown of the ew symmetry , with a dynamical generation of masses of the gauge bosons and the third generation quarks may only be obtained for certain values of the fermion mass bulk parameters . interestingly enough , these mass parameters are the same as those demanded by consistency with experimental observations at the quantum level @xcite . moreover , the higgs boson is predicted to be light , with mass @xmath21 in the range to be probed by the tevatron and the lhc in the near future . this article is organized as follows . in section 2 we describe our 5-dimensional ( 5d ) model and derive the gauge boson spectral functions . furthermore , we analyze the @xmath22 gauge boson form factor at low energy , defining its coupling to the higgs field . in section 3 we derive the fermion spectral functions , necessary to compute the loop - induced higgs effective potential and the fermion mass spectrum . we also compute the top quark form factor at low energy and the associated yukawa coupling . section 4 deals with the explicit computation of the effective potential . in section 5 we present the numerical analysis of the effective potential for the higgs and the kk spectrum of the theory . we reserve section 6 for our conclusions . we are interested in a 5d gauge theory with gauge group @xmath0 . the geometry of our space - time will be that of rs1 @xcite , with an orbifolded extra spatial dimension in the interval @xmath23 $ ] . the metric for such a geometry is given by @xmath24 where @xmath25 . the space spanning the fifth dimension corresponds to a slice of , with branes attached at the two boundary points : @xmath9 ( uv brane ) and @xmath10 ( ir brane ) . we place our gauge fields , @xmath26 and @xmath27 , in the bulk , where @xmath28 are the hermitian generators of the fundamental representation of @xmath29 and generically @xmath30=c(5)\delta^{\alpha,\beta}$ ] . the explicit form of the generators @xcite are given in appendix a. our fermions @xmath31 also live in the bulk , and they transform under a representation @xmath32 of @xmath29 . however , for the moment , we concentrate only on the gauge content of our model . the phenomenologically required fermionic content will be discussed in detail in section 3 . the 5d action is @xmath33 where @xmath34 and @xmath35 and @xmath36 are the 5d dimensionful gauge couplings . the choice @xmath37 is a convenient choice , since it allows us to identify the eigenvalues of our generators as the weak isospin , with the four dimensional coupling given by @xmath38 . any other choice for @xmath39 may be absorbed into a redefinition of the gauge fields or the gauge coupling leaving the physics unchanged . to construct a realistic 4d low energy theory , we will break the 5d @xmath40 gauge symmetry down to the subgroup @xmath41 on the ir brane and to @xmath42 on the uv brane , where @xmath43 is the hypercharge and @xmath44 is the @xmath45 associated charge which is accommodated to obtain the correct hypercharge . we divide the generators of @xmath29 as follows : the generators of @xmath46 are denoted by @xmath47 and @xmath48 , while the generators from the coset @xmath49 are denoted by @xmath50 and @xmath51 . in order to obtain the correct hypercharge and therefore the right weinberg angle @xmath52 , we need to rotate the fields @xmath53 and @xmath54 @xcite , @xmath55 @xmath56 where we will enforce @xmath57 to have even parity , corresponding to the hypercharge gauge boson in the 4d low energy limit . from now on we will drop the prime on @xmath58 , and it will be understood that @xmath59 refers to @xmath60 and @xmath61 . to implement the breaking of @xmath29 on the two branes as stated above , we impose the following boundary conditions on the gauge fields : @xmath62 these boundary conditions lead to 4d scalars @xmath63 originating from @xmath64 . at tree level , the @xmath63 s have a flat potential and therefore may be thought of as a dual description of goldstone bosons arising after spontaneous breaking of the global symmetry in 4d . furthermore , @xmath65 and @xmath66 also have zero modes which will become massive as soon as the @xmath63 s develop a vev , leading ultimately to electroweak symmetry breaking . part of our work consists of obtaining the mass spectrum of the theory in the presence of the vev of @xmath63 . the higgs forms a bidoublet under @xmath67 , whose doublet component under @xmath68 is given by @xmath69 , with charge 0 under the @xmath70 . the @xmath67 symmetry in principle implies that we can choose any one of the components to be the only non - zero one acquiring a vev . however , we want the scalar higgs to be a neutral particle , therefore , we choose @xmath71 . we kk - expand the fields @xcite , @xmath72 where we have put the explicit dependence on @xmath73 for the scalars @xmath74 . the normalization constant is chosen so that the scalars are canonically normalized , @xmath75 . we wish to solve for the kk profiles @xmath76 so that the 5d action may be rewritten in terms of a tower of 4d fields @xmath77 whose kinetic and mass terms , after proper diagonalization , take the form : @xmath78 ^ 2 + \frac{1}{2 } m_n^2(h ) a_{\mu , n}^2 \right ) + \dots \right \}\ ] ] this is done by solving the equations of motion in the presence of the scalar vev , and satisfying the boundary conditions implied by eqs . ( [ bc1 ] ) and ( [ bc2 ] ) . it turns out that solving the equations of motion in the presence of @xmath79 is complicated , as these mix the neumann and dirichlet modes . however , 5d gauge symmetry relates these solutions to solutions with @xmath80 @xcite . to that end , focusing on the gauge fields , we perform the following gauge transformation on them @xmath81 where @xmath82 is the gauge transformations that removes the vev of @xmath79 : @xmath83 .\ ] ] in the @xmath80 gauge , the gauge kk profiles satisfy the following equation of motion : @xmath84 which is the same for both dirichlet and neumann modes . we call the independent solutions @xmath85 and @xmath86 , which satisfy the following initial conditions : @xmath87 , @xmath88 , @xmath89 and @xmath90 . the explicit form of these solutions are given by @xcite : @xmath91\\ s(x_5,z ) & = & \frac{\pi z } { 2 k } a^{-1}(x_5 ) \left [ j_1 \left ( \frac{z } { k } \right ) y_1 \left ( \frac{z } { k a(x_5 ) } \right ) - y_1 \left ( \frac{z } { k } \right ) j_1 \left ( \frac{z } { k a(x_5 ) } \right ) \right ] \end{aligned}\ ] ] note from eq . ( [ om.tr ] ) that for @xmath9 , @xmath92 . this implies that the gauge kk profiles satisfying the uv boundary conditions given in eq . ( [ bc1 ] ) , can be directly written in terms of the basis functions defined above as @xmath93 where the coefficients @xmath94 are normalization constants . we can now calculate @xmath95 using eq . ( [ e.vrt ] ) . the ir boundary conditions give us a system of algebraic equations for the coefficients @xmath94 . the determinant of this system of equations must vanish to give us a nontrivial solution , and this condition in fact gives us the quantization condition for the masses @xmath96 . using this procedure we find for the gauge bosons : @xmath97^{2}\times\nonumber\\ & & \left[c'(l , m_n ) s(l , m_n)\left(-2+(1+{s_\phi}^{2})\sin^2\left(\frac{\lambda_g h}{f_h}\right)\right)-s'(l , m_n ) c(l , m_n)\left(1+{s_\phi}^{2}\right)\sin^2\left(\frac{\lambda_g h}{f_h}\right)\right ] = 0\nonumber\\ & & \end{aligned}\ ] ] where the `` higgs decay constant '' is defined as @xmath98 and @xmath99 . furthermore , using the wronskian relation @xmath100 we can rewrite the quantization condition in a more convenient form : @xmath101^{2}\times & \nonumber\\ & & \nonumber\\ & \left[2a^{2}_lc'(l , m_n ) s(l , m_n ) + ( 1+{s_\phi}^2)m_n\sin^2\left(\frac{\lambda_g h}{f_h}\right)\right]=0&\nonumber\label{gaugedet}\\ & & \end{aligned}\ ] ] therefore the kk mass spectrum for the @xmath22 and @xmath2 bosons , with the correct weinberg angle given by @xmath102 , is given by the zeroes of the following equations : @xmath103 we will identify the first zero of these equations with the masses of @xmath22 and z , respectively , and we shall denote the masses of the first excited states as @xmath104 and @xmath105 , associated with the second zeroes of both equations . we can gain some physical insight if we look at eq . ( [ gaugedet ] ) in the limit @xmath106 . in that case , this equation reduces to , @xmath107 we can identify the zeroes coming from @xmath108 with the kk mass spectrum of the 4 gauge bosons belonging to @xmath109 which are even on both the uv and the ir branes . in the same way , we identify the zeroes of @xmath110 with the kk spectrum of the 3 @xmath111 gauge bosons which are odd on the uv brane and even in the ir brane . @xmath112 is identified with the kk spectrum of the 4 @xmath49 gauge bosons which are odd on both the uv and ir branes . thus in eq . ( [ gaugedet ] ) we associate the photon kk spectrum with the zeroes of @xmath113 . at small momenta , below the scale @xmath114 , the form factor for the @xmath22 gauge bosons can be approximated by @xmath115 from the last equation , we can find an analytic expression for the w - mass : @xmath116 from this expression we can calculate the higgs - w - w coupling , @xmath117 , at linear order , by simply taking the derivative of @xmath118 with respect to the vev of @xmath79 : @xmath119 where we used @xmath120 and in the last step note that in the limit @xmath121 we recover the sm higgs - w - w coupling . we will discuss the physical implications of the above limiting case in section 5 . the sm fermions are embedded in full representations of the bulk gauge group . the presence of the @xmath122 subgroup of the full bulk gauge symmetry ensures the custodial protection of the @xmath1 parameter @xcite . in order to have a custodial protection of the @xmath123 coupling , the choice @xmath124 has to be enforced @xcite . an economical choice is to let the sm @xmath68 top - bottom doublet arise from a @xmath125 of @xmath0 , where the subscript refers to the @xmath70 charge . as discussed in @xcite , putting the sm @xmath68 singlet top in the same @xmath29 multiplet as the doublet , without further mixing , is disfavored since for the correct value of the top quark mass this leads to a large negative contribution to the @xmath1 parameter at one loop . hence we let the right - handed top quark arise from a second @xmath125 of @xmath0 . the right handed bottom can come from a @xmath126 that allows us to write the bottom yukawa coupling . for simplicity , and because it allows the generation of the ckm mixing matrix , we make the same choice for the first two quark generations . we therefore introduce in the quark sector three @xmath29 multiplets per generation as follows : @xmath127 @xmath128 where we show the decomposition under @xmath129 , and explicitly write the @xmath130 charges . the @xmath131s are bidoublets of @xmath129 , with @xmath132 acting vertically and @xmath122 acting horizontally . the @xmath133 s and @xmath134 s transform as @xmath135 and @xmath136 under @xmath129 , respectively , while @xmath137 and @xmath138 are @xmath129 singlets . the superscripts , @xmath139 , label the three generations . we also show the parities on the indicated 4d chirality , where @xmath140 and @xmath141 stands for odd and even parity conditions and the first and second entries in the bracket correspond to the parities in the uv and ir branes respectively . let us stress that while odd parity is equivalent to a dirichlet boundary condition , the even parity is a linear combination of neumann and dirichlet boundary conditions , that is determined via the fermion bulk equations of motion as discussed below . the boundary conditions for the opposite chirality fermion multiplet can be read off the ones above by a flip in both chirality and boundary condition , @xmath142 for example . in the absence of mixing among multiplets satisfying different boundary conditions , the sm fermions arise as the zero - modes of the fields obeying @xmath143 boundary conditions . the remaining boundary conditions are chosen so that @xmath144 is preserved on the ir brane and so that mass mixing terms , necessary to obtain the sm fermion masses after ew symmetry breaking , can be written on the ir brane . consistency of the above parity assignments with the original orbifold @xmath145 symmetry at the ir brane will be discussed in appendix b. as remarked above , the zero - mode fermions can acquire ew symmetry breaking masses through mixing effects . the most general @xmath146 invariant mass lagrangian at the ir brane compatible with the boundary conditions is , in the quark sector , @xmath147 ~ , \label{localizedmasses}\ ] ] where @xmath148 , @xmath149 and @xmath150 are dimensionless @xmath151 matrices , and matrix notation is employed . in appendix b we will show how these mass terms may be generated from an @xmath29 invariant lagrangian using the spurion formalism . to prevent negative corrections to the @xmath1 parameter @xcite , @xcite , we will take @xmath150 to be the null matrix . in appendix c however we briefly discuss the modifications of the effective higgs potential induced by a non - zero value of @xmath152 . moreover since the higgs effective potential is most sensitive to the third generation , we will concentrate on this one , discarding family indices . with the introduction of these brane mixing terms , the different multiplets are now related via the equations of motion . in the case of a mass term involving @xmath153 , with the fields having profile functions @xmath154 and @xmath155 , the odd parity profile @xmath156 on the ir brane is related to @xmath157 : @xmath158 the odd parity we demand at the ir brane implies a dirichlet boundary condition for the function @xmath159 . therefore , we can rewrite : @xmath160 and , similarly : @xmath161 eq . ( [ limit ] ) and ( [ limit2 ] ) can now be reinterpreted as the new boundary conditions for the profiles at the ir brane . the fermions , like the gauge bosons , can be expanded in their kk basis : @xmath162 in the @xmath106 gauge , we redefine @xmath163 and we write our vector - fermionic fields in terms of chiral fields . we can kk decompose the fermionic chiral components as , @xmath164 where @xmath165 is normalized by : @xmath166 therefore the profile function for the zero mode fermion corresponds to @xmath167 . from the 5d action , eq . ( [ 5daction ] ) , concentrating on the free fermionic fields , we can derive the following first order coupled equations of motion for @xmath168 , @xmath169 we see from eq . ( [ 1order ] ) that we can redefine @xmath170 and relate the opposite chiral component of the same vector - like field by @xmath171 . for the left handed field having dirichlet boundary conditions on the uv brane , we can derive a second order equation for the chiral component @xmath172 : @xmath173 \tilde{f}_{l , n } = 0\ ] ] the solution of which we shall call @xmath174 , with boundary conditions @xmath175 , @xmath176 . similarly , we can redefine @xmath177 and then relate the opposite chirality via @xmath178 . if the right - handed field fulfills dirichlet boundary conditions on the uv - brane , we can write the equation of motion for @xmath179 : @xmath180 \tilde{f}_{r , n } = 0.\ ] ] we shall correspondingly denote the solution to this equation with @xmath181 , fulfilling the boundary conditions @xmath182 , @xmath183 . as emphasized before , the solutions of the opposite chirality , are related to these solutions via the the first order equations of motion , eqs . ( [ 1order ] ) , which we rewrite as @xmath184 . here @xmath185 , and the @xmath186 is chosen depending on whether we want an odd ( - ) boundary condition on the uv brane for the left - handed or right - handed fermionic field . the solution to eq . ( [ feom ] ) is given by @xcite : @xmath187.\ ] ] the solution for eq . ( [ feom1 ] ) is given by eq . ( [ sm ] ) , with the replacement @xmath188 . before proceeding with the gauge transformations as was done with the gauge fields , we would like to discuss what is required of the gauge transformation for the fermions . the boundary conditions are explicitly given on the physical fermion which is an isospin eigenvector . hence , the first step is to identify the different components of the vector @xmath31 with the doublets , triplets and singlets as required from eq . ( [ multiplets ] ) . second , we need to assign functions to the different components of the multiplets consistent with the boundary conditions for the fermions given in eq . ( [ multiplets ] ) . the basis @xmath189 and @xmath190 are simply related by exponentials , therefore , the functions @xmath189 , given by @xmath191 and @xmath192 , are related to @xmath190 via @xmath193 \tilde{s}_{\pm m } ( \dot{\tilde{s}}_{\pm m})$ ] . odd and even solutions ( @xmath194 or @xmath195 ) can then be assigned to the component functions of the multiplets @xmath196 , @xmath197 and @xmath198 , depending on the boundary parities . for example , for a left handed component having boundary conditions @xmath199 or @xmath200 , since the function is odd at the uv brane , we assign a solution @xmath201 to that component of the left - handed multiplet . if the boundary condition is @xmath202 , or @xmath203 we need to look at the chiral companion @xmath204 or @xmath205 . in this case , our right handed fermion is odd on the uv brane . the relationship between the left and the right handed components are as defined above via derivatives . therefore , we assign @xmath206 to the left - handed fermion . following this procedure , we can assign either @xmath191 or @xmath207 to each component of the multiplets . this procedure defines our three fermion multiplets in the @xmath106 gauge as the vector functions : @xmath208\\ \\ \\ f_{2,r}(x_{5},0)=\left[\begin{array}{c } c_{6}s_{-m_{2}}\\ c_{7}s_{-m_{2}}\\ c_{8}s_{-m_{2}}\\ c_{9}s_{-m_{2}}\\ c_{10}\dot{s}_{m_{2}}\end{array}\right ] \end{array } & & f_{3,r}(x_{5},0)=\left[\begin{array}{c } c_{11}s_{-m_{3}}\\ c_{12}s_{-m_{3}}\\ c_{13}s_{-m_{3}}\\ c_{14}s_{-m_{3}}\\ c_{15}s_{-m_{3}}\\ c_{16}s_{-m_{3}}\\ c_{17}s_{-m_{3}}\\c_{18}s_{-m_{3}}\\ c_{19}s_{-m_{3}}\\ c_{20}\dot{s}_{m_{3}}\end{array}\right]\end{aligned}\ ] ] where , as for the gauge bosons , the @xmath209 are normalization constants . the opposite chiralities have opposite boundary conditions and can be read from these ones . the boundary conditions as written above , for example , for the vector components of @xmath210 are the boundary conditions on the isospin eigenvector with charges @xmath211 , @xmath212 , @xmath213 , @xmath214 , @xmath215 under @xmath129 , respectively . however , in the basis of our generators , the eigenvector with spin @xmath216 under @xmath217 , for example , is given by : @xmath218.\ ] ] to simplify matters , so that the boundary conditions can be taken to be on the eigenvector : @xmath219\ ] ] the transformation @xmath220 needs to be in the basis consistent with eq . ( [ f.can ] ) , or we can also think about this as changing the fermion vector functions given in eq . ( [ f.exp.bc ] ) to the basis of our generators . this means that we have to first understand how the eigenvectors of @xmath221 decompose in the canonical basis for both the @xmath125 and the @xmath126 . for the @xmath125 fermions , the basis transformation for @xmath196 , @xmath197 is given by @xmath222 for @xmath126 , the transformation matrix for @xmath198 is given by @xmath223 therefore , once the uv boundary conditions define the vectors as in eq . ( [ f.exp.bc ] ) , we change bases to be able to implement the gauge transformation and then transform back so that the ir boundary conditions can be imposed . using the transformations defined in eq . ( [ a.mat ] ) and ( [ b.mat ] ) , we can write the final fermion vector functions in the presence of the vev of @xmath79 as : @xmath224 applying the boundary conditions at @xmath225 , taking into account the mass mixing terms from eq . ( [ localizedmasses ] ) and using the procedure defined in eqs . ( [ limit ] ) and ( [ limit2 ] ) , we derive the conditions on @xmath226 : @xmath227 where the superscripts denote the vector components . asking that the determinant of this system of equations vanishes so that we get a non - trivial solution , we notice that all dependence on the prefactors @xmath228 $ ] relating @xmath229 and @xmath230 drops out , and thus we conclude that one of the following equations should vanish : @xmath231 ^ 2=0&\label{exotickk3}\\ & 2\tilde{s}_{m_{3}}\left[m_{b_2}^2\tilde{s}_{-m_{3}}\tilde{s}'_{-m_{1 } } + \tilde{s}_{-m_{1}}\tilde{s}'_{-m_{3}}\right ] -m_{b_2}^2\tilde{s}'_{-m_{1}}\sin^2\left(\frac{\lambda_f h}{f_h}\right)=0&\label{bottomkk}\end{aligned}\ ] ] @xmath232+&\nonumber\\ & \left[m_{b_2}^2\tilde{s}_{m_{2}}\tilde{s}_{-m_{3}}\tilde{s}'_{-m_{2 } } + m_{b_1}^2\left(2m_{b_2}^2\tilde{s}_{m_{1}}\tilde{s}_{-m_{3 } } \tilde{s}'_{-m_{1 } } + \tilde{s}'_{-m_{3 } } + 2\tilde{s}_{m_{1}}\tilde{s}_{-m_{1}}\tilde{s}'_{-m_{3}}-\right.\right . & \nonumber\\ & \left.\left.\tilde{s}_{m_{2}}\tilde{s}_{-m_{2}}\tilde{s}'_{-m_{3}}\right ) \right]\sin^2\left(\frac{\lambda_f h}{f_h}\right)-m_{b_1}^2\tilde{s}'_{-m_{3}}\sin^4\left(\frac{\lambda_f h}{f_h}\right)=0&\label{topkk}\end{aligned}\ ] ] where for simplicity we did not write the dependence on l and z and furthermore , we have used the crowian : [ e.cr ] -_m(x_5,z)_-m(x_5,z ) + _ m(x_5,z ) _ -m(x_5,z ) = 1 . in the following , and for simplicity , we shall omit the tildes and refer to the solutions @xmath233 by @xmath191 . the same value @xmath234 is found in the fermionic case , so from now on @xmath235 . the solutions to eqs . ( [ exotickk1 ] ) , ( [ exotickk2 ] ) and ( [ exotickk3 ] ) can be interpreted as the kk spectrum of exotic fermions . we shall name these states @xmath236 , @xmath237 and @xmath238 , respectively . eq . ( [ bottomkk ] ) is the kk mass spectrum for the bottom quark and eq . ( [ topkk ] ) gives the kk spectrum for the top quark . we can rewrite eqs . ( [ bottomkk ] ) and ( [ topkk ] ) in the same form as the gauge bosons : @xmath239 where @xmath240 @xmath241\\ f_{2 } ( z^2)&=&-m_{b_1}^2s'_{-m_{3}}\\ f_{d}(z^2)&= & 2\left[m_{b_1}^2s_{m_{1}}\left(-1 + s_{m_{2}}s_{-m_{2}}\right)\left(m_{b_2}^2s_{-m_{3}}s'_{-m_{1 } } + s_{-m_{1}}s'_{-m_{3}}\right)+\right.\nonumber\\ & & \left.s_{m_{2}}s'_{-m_{2}}\left(m_{b_2}^2\left(-1 + s_{m_{1}}s_{-m_{1}}\right)s_{-m_{3}}- \frac{a^2_{l}}{z^2}s_{-m_{1 } } s'_{m_{1 } } s'_{-m_{3}}\right)\right]\end{aligned}\ ] ] equations ( [ gsf ] ) , ( [ bottomform ] ) and ( [ topforms ] ) are the starting point for computing the one - loop effective potential for the scalars @xmath63 . it is physically very illuminating to interpret the origin of the fermionic resonances if we contemplate the fermionic determinant in the case @xmath106 : @xmath242 we recall that the way we defined the initial fermionic functions ( @xmath191 ) was via the boundary condition on the uv brane and the chirality . however , since we are in the @xmath106 gauge , we can use the boundary conditions on the ir brane to further give us the final equations the functions have to satisfy : @xmath243 we can now interpret eq . ( [ hzero1 ] ) in a simple way : the first equation in the first line provides the mass spectrum for the singlet right handed bottom . this with the first equation in the second line provide the mass spectrum of the triplet components @xmath244 , @xmath245 of @xmath198 . the second equation in the first line may be identified with the mass spectrum of the singlet @xmath246 and its modification from the mixing mass term with @xmath247 . it can be shown from this equation that the light top resonances will be mostly the kk modes of the right handed top . from the first equation in the third line we get the mass spectrum for @xmath248 . additionally we can also see here the mixing of the bidoublets @xmath249 and @xmath250 . the second equation in the second line is related to the mass spectrum for the @xmath251 doublet and its mixing with @xmath252 . the second equation in the third line may be identified as the spectrum of a new fermionic state coming from the doublet - doublet mixing between @xmath253 and @xmath254 . at small momenta , the solution to eq . ( [ sm ] ) takes the form : @xmath255 replacing this in eq . ( [ topkk ] ) , keeping only quadratic terms in z and for simplicity neglecting the small @xmath256 effects , we find that the top quark spectral equation reduces to : @xmath257\sin^2\left(\frac{\lambda h}{f_h}\right)-m_{b_{1}}^{2}\sin^4\left(\frac{\lambda h}{f_h}\right)=0,\ ] ] where a and b are coefficients given by : @xmath258 solving eq . ( [ top.mass ] ) for @xmath259 , we find that the top mass takes the form @xmath260 which can also be written as @xmath261}.\ ] ] this expression is similar to the one obtained in a related model in ref . @xcite . from this expression we can calculate the yukawa coupling of the top , @xmath262 , at linear order , by simply taking the derivative of @xmath263 with respect to @xmath79 : @xmath264 where we note that in the limit @xmath265 , we recover the sm top yukawa coupling . the coleman weinberg potential in 5d kk theories ( see for instance @xcite ) takes the form : @xmath267.\ ] ] here @xmath268 is the number of degrees of freedom of a given particle , @xmath186 depends on whether we are considering bosons or fermions and @xmath269 is the spectral function encoding the spectrum in the presence of @xmath79 . we showed in the previous section that in the 5d framework the spectral functions encoding the kk spectrum in the presence of the higgs vev are given by : @xmath270 thus , the scalar potential for the pseudo - goldstone can be written as : @xmath271+\right.\right.\nonumber\\ & & \left.\left . \log\big[1 + f_{b}(-p^2 ) \sin^2\left(\frac{\lambda h}{f_h}\right)\big]\right.\big\ } + \frac{6}{(4\pi)^2}\log\big[1 + f_{w}(-p^2 ) \sin^2\left(\frac{\lambda h}{f_h}\right)\big]+\right.\nonumber\\ & & \left.\frac{3}{(4\pi)^2}\log\big[1 + f_{z}(-p^2 ) \sin^2\left(\frac{\lambda h}{f_h}\right)\big]\right.\big)\end{aligned}\ ] ] the form factors @xmath272 and @xmath273 for the gauge bosons are given by eq . ( [ gsf ] ) , and @xmath274 , @xmath275 and @xmath276 for the fermions are given by eq . ( [ bottomform ] ) and ( [ topforms ] ) . we note here that the main contribution to the potential from the fermionic sector is due to the top quark . therefore , the first two fermion multiplets , @xmath277 and @xmath278 are the most important for fixing the higgs potential . [ 0.8 ] [ 0.8 ] [ 0.8 ] [ 0.8 ] [ 0.8 ] [ 0.8 ] [ 0.8 ] [ 0.8 ] [ 0.8 ] our goal is to find a non - trivial minimum of the higgs effective potential that will break @xmath279 down to @xmath130 . the relevant parameters which control the model are @xmath148 , @xmath149 , @xmath280 , @xmath281 , @xmath282 and @xmath283 , where @xmath284 , @xmath285 , @xmath286 and @xmath287 is the length of the extra dimension . looking at the potential we immediately see that @xmath288 is a non - trivial extremum . when @xmath289 corresponds to a maximum , this extremum can be a minimum . however this would lead to a non - standard effective low - energy theory , since as seen from the form factors , and from eq . ( [ hww ] ) , the higgs coupling to the gauge bosons vanishes at linear order . therefore , we shall concentrate only in the regions of parameter space such that @xmath290 with @xmath291 . after obtaining a non - trivial minimum , we calculate @xmath8 by finding the first zero of eq . ( [ gsf ] ) and setting this mass equal to the @xmath22 mass @xcite . this is the mass that the @xmath22 zero - mode gets from interacting with the higgs vev . since the relationship between the @xmath2 and the @xmath22 form factors remains the same as in the sm , we also recover a consistent mass for the @xmath2 boson . having obtained @xmath8 , we calculate the top - quark mass , given by the first zero of eq . ( [ topforms ] ) , and demand that its value is in the phenomenological range @xmath292 $ ] gev . we take such a wide range since we consider the result of our calculation to be associated with the running top - quark mass evaluated at a scale of the order of the weak scale . due to the strong gauge coupling , the running top - quark mass varies rapidly between values of @xmath293 gev , for a scale of the order of a few tev , to values of @xmath294 gev , for a scale of order of the top quark pole mass @xmath295 . furthermore , we also calculate the bottom quark mass , given by the first zero of eq . ( [ bottomform ] ) , and demand it to be in the phenomenological range @xmath296 $ ] gev . the higgs mass is given by the second derivative of the effective potential with respect to @xmath79 , evaluated at the vev . as we will show , somewhat smaller higgs boson mass values tend to be associated with smaller values of the top quark mass , a reflection of the power dependece of the low - energy higgs quartic coupling on the top quark mass . the higgs mass and the kk spectrum are seen to be only weakly dependent on variations of the bottom quark mass in the above defined range . although we will display all results consistent with the observed quark and gauge boson low energy spectrum , we will concentrate on points in parameter space in the linear regime , such that @xmath297 . by linear regime we mean the regime where higher order interactions involving the higgs field , @xmath298 with @xmath299 , can be neglected and , furthermore , the couplings of the higgs with gauge bosons and fermions are very close to the ones in the sm . therefore , in this regime the direct search lep bounds on the higgs mass apply , and @xmath300 gev . in addition , the low energy effective theory is well approximated by the sm , with renormalizable couplings and a light higgs boson . this allow us also to compare our results with those of ref . @xcite where they performed calculations in the linear regime as defined above . we can interpret @xmath301 as the associated mass scale of our theory and of the order of the natural uv cut off of the low - energy , sm - like effective theory . the scale @xmath301 is inversely proportional to @xmath302 and therefore as we move away from the linear regime , we are pushing the uv cut - off lower and additionally causing the linear couplings of the higgs to become smaller . therefore , the theory becomes increasingly non - renormalizable , which leads to the need of taking into account higher order kk - states ( beyond zero - modes ) in the calculation of low - energy processes , thus making connection with sm predictions harder . when @xmath303 tev , kk bosonic and fermionic resonances of exotic and sm fields have a chance to be detected in the next generation colliders such as the lhc ; hence , we will be specifically interested in those regions . we performed two major numerical parameter scans . in the first one we scanned the parameter space for values @xmath304 $ ] with @xmath139 and @xmath305 $ ] . we used @xmath306 in our numerical calculations for the masses since the results are invariant under variations of @xmath307 $ ] remains approximately constant ] . the first scan was done with a coarse grid and before we discuss details of our numerical results , we would like to point out that the effective potential is a completely well behaved function of all the parameters , therefore , we expect that the gaps in our scanned space are smoothly filled . we found that even though for some value of the other parameters , non - trivial minima existed for nearly all the regions of @xmath308 and @xmath309 , we only found phenomenologically consistent gauge bosons , top and bottom quark masses in the following regions of parameter space : @xmath310 , @xmath311 , @xmath312 , @xmath313 and @xmath314 . though the results show a skew symmetry between positive and negative values , we decided to concentrate on negative values of @xmath315 and @xmath316 , since interestingly enough , this is the region which is consistent with electroweak precision measurements @xcite . furthermore , positive values of @xmath317 lead to a smaller overlap of the left - handed top and bottom zero modes with physics at the ir brane , leading to a stronger suppression of potentially dangerous flavor changing operators @xcite . therefore , it is in this region where we performed a more thorough scan taking smaller steps for the @xmath318 . in figure [ figure : tophiggs ] we notice a clear relationship between the higgs and the top - quark masses , where on the plot we have included the bigger coarse scan of parameter space . focusing on the interesting phenomenological region , we see in figure [ figure : higgstop ] that in the linear regime we get higgs masses that tend to be above 115 gev , which is above the experimental bound from lep , and below 160 gev . furthermore , there is a somewhat weak dependence of the higgs mass values on the precise value of the running top quark mass . higgs mass values closer to the current experimental bound are obtained for the smallest values of the top quark mass in the phenomenologically allowed range . in figure [ figure : kahiggs ] , we manifestly see that in the linear regime we may obtain masses for the higgs which are compatible with @xmath319 . figure [ figure : hb1 ] provides us with information about the 5d parameter @xmath148 which as increased moves the minimum to higher values of @xmath302 , getting into the non - linear regime . we conclude that to remain in the linear regime , we can not have an arbitrarily high @xmath148 boundary mass parameter . the behavior @xmath320 is seen in figure [ figure : kah ] where , moreover , we notice that for the linear regime , values of @xmath321 tev are obtained . this low value of @xmath301 will lead to interesting phenomenology for the decays of light kk - fermions and kk - bosons . in figure [ figure : c1c2 ] we see that a selected region of the plane @xmath322 is suitable for minimizing the effective potential with the experimentally observed @xmath22 , @xmath2 , top and bottom masses , and that the linear regime basically applies through out this region . as emphasized before , @xmath323 and @xmath324 are the ones preferred by consistency with electroweak precision measurements @xcite . in figure [ figure : c2c3 ] we noticed that the same phenomenologically accepted region for @xmath315 in the @xmath325 plane with @xmath326 is chosen and it is also homogenously covered in the linear regime . figures [ figure : w1t1 ] and [ figure : w1e1 ] are very interesting for collider phenomenology and therefore deserve special attention . even though we havent included the plot , we calculated the mass of the @xmath22 first excited kk state numerically and verified the relation @xmath327 . if the masses of these fermions are below the black line denoting @xmath328 , the decay of @xmath329 and @xmath330 to pairs of these fermions is kinematically available . note here however , that for @xmath331 , @xmath332 is always greater than @xmath328 . moreover , these exotic fermions are unstable . the decay products of @xmath333 are analyzed in ref . the first top kk excited state , will mainly decay to zt , ht , wb . the charged 5/3 fermions of @xmath334 will necessarily decay to wt from charge conservation , leading to interesting phenomenology @xcite . in addition , the lightest kk quark state @xmath335 , the top quark ( via its right- and left - handed couplings ) and the bottom quark ( via its left - handed coupling ) will be included in the kk gluon decay products . this is a distinctive feature of this model when compared with the predominant decay into right - handed top quarks in the models considered in refs . in this article , we computed the one - loop coleman weinberg potential for the higgs field in a specific model of gauge - higgs unification in warped extra dimensions . we chose the specific group @xmath336 , which allows the introduction of custodial symmetries protecting the precision electroweak observables as well as a higgs field with the proper quantum numbers under the electroweak gauge groups . as a first step , we computed the spectral functions of the fermions and gauge bosons of the theory when the higgs field acquires a vev . these are then used to calculate the effective potential for the higgs . we demand non - trivial minima that lead to the proper values of the gauge boson and fermion masses in the low energy theory . this requirement leads to a selection of a restricted region of parameters . interestingly enough , the selected regions of parameters coincide with the ones previously selected in order to obtain good agreement with the precision electroweak observables . our main result is the computation of the higgs and kk mass spectra . demanding that the kk gauge bosons be accessible at the lhc and also that the effective potential minimum be in the linear regime associated with a sm - like low energy effective theory , we obtain higgs boson masses which are between the present experimental bound on this quantity and about 160 gev . this range of masses will be tested first at the tevatron and then at the lhc collider in the near future . the kk fermion spectrum also shows interesting features . we find that there are kk fermions which become much lighter than the kk gauge bosons . in particular , the lightest kk mode of the top becomes light enough so that the kk gauge bosons will decay into it . there is an additional light kk fermion , which may become light enough for the gauge kk bosons to decay into it , which has exotic quantum numbers under hypercharge , leading for instance to the presence of a light fermion with charge @xmath337 with interesting phenomenological properties . we would like to thank debajyoti choudhury , ben lillie , arjun menon , jing shu , tim tait and particularly marcela carena , eduardo ponton and jose santiago for useful discussions and comments . work at anl is supported in part by the us doe , div . of hep , contract de - ac02 - 06ch11357 . this work was also supported in part by the u.s . department of energy through grant no . de - fg02 - 90er40560 . the generators of the fundamental representation of @xmath29 , with @xmath338=c(5)\delta^{\alpha\beta}$ ] , are given by : @xmath339,\nonumber\\ t^{{\hat a}}_{i , j}&=&-i\sqrt{\frac{c(5)}{2}}(\delta^{{\hat a}}_i\delta^5_j-\delta^{{\hat a}}_j\delta^5_i),\end{aligned}\ ] ] where @xmath50 @xmath340 and @xmath341 @xmath342 are the generators of @xmath49 and @xmath67 respectively . since the higgs has positive parity , in order to preserve the @xmath145 orbifold symmetry , we need to ensure that the product of the parities of the fermions coupled via the higgs is always positive . the higgs profile is doubly exponentially suppressed on the uv brane , therefore , we will restrict ourselves to discussion of parities on the ir brane . concentrating on the parities of the first two multiplets given in eq . ( [ multiplets ] ) , we notice that the product of the fermion parities coupled by the higgs is not positive . one way to fix this would be to switch the parities for the @xmath67 singlet components : @xmath343\quad\leftrightarrow\quad [ + , + ] .\ ] ] due to the introduction of the brane mass terms as in eq . ( [ localizedmasses ] ) , and looking at eq . ( [ f.ir.bc ] ) we note that the switch in parity is exactly equivalent to taking @xmath344 . thus the physical results presented in this work are invariant under the switch of parity with just the above mentioned change in @xmath148 . another way to understand the realization of the @xmath145 symmetry is via the spurion formalism . we introduce a spurion field transforming in the fundamental representation of @xmath29 , and acquiring a very large vev in the @xmath67 singlet direction , breaking the @xmath29 gauge symmetry to @xmath67 on the ir brane . we also introduce local fermionic degrees of freedom on the brane and write large local mass terms involving the spurion , between these degrees of freedom and the ones propagating in the bulk . the fermion parities are chosen to be those preserving @xmath145 . by using a procedure similar to the one outlined in eqs . ( [ delta ] ) , ( [ limit ] ) and ( [ limit2 ] ) , we can now change the bulk fermion ir boundary conditions while preserving @xmath145 to give the model we are considering in this work . this mechanism would allow us to change the ir boundary conditions not just of the @xmath67 singlet fields in the @xmath345 s , but also of the bidoublets in the @xmath345 s or in the @xmath346 . for the case of the @xmath67 singlet fields in the @xmath345 s , we need to add @xmath29 singlet fermions on the ir brane , that couple to the @xmath347 s via the spurion field . for the case of the bidoublets in the @xmath346 or in the @xmath345 s , we need to add local @xmath345 s and singlet fermions on the ir brane , and couple them appropriately to the bulk @xmath346 or @xmath345 s via gauge invariant masses or couplings to the spurion field ( @xmath29 singlet fields are necessary to remove unwanted degrees of freedom ) . in this way , one could start with the opposite ir parities for the three bidoublets to the ones given in eq . ( [ multiplets ] ) , thus preserving the @xmath145 symmetry , and reach the choice of bulk fermion boundary conditions used in this work . the same spurion mechanism can be used to write the brane mass terms in eq . ( [ localizedmasses ] ) in an @xmath29 invariant way . a spurion field which is a @xmath345 under @xmath29 , can be used to write the bidoublet - bidoublet mixing brane mass term , @xmath149 , between the @xmath345 and the @xmath346 multiplets . the spurion field is coupled to these two via a yukawa type interaction which can be adjusted to give @xmath348 . this field can be the same as the one that is used to change the parities of the fermions as described in the last paragraph . the same procedure , but now using a @xmath349 of @xmath29 can be used to generate the mass mixing terms between the singlets , @xmath148 . the @xmath350 required can be thought of as arising from a tensor product of the @xmath345 spurion field : @xmath351 . note that this mechanism will not generate a localized bare mass for the higgs due to gauge invariance . we also studied the effects of flipping the parities of @xmath248 as was done in the context of ref . @xcite . in this case , the parity assignments for the rest of our fermionic content remains the same except for @xmath248 which now takes the form , @xmath352 if we keep the same boundary mass terms as in eq . ( [ localizedmasses ] ) and calculate the fermionic spectral functions we obtain , @xmath353 ^ 2=0&\\ & 2\tilde{s}_{m_{3}}\left[m_{b_2}^2\tilde{s}_{-m_{3}}\dot{\tilde{s}}_{-m_{1 } } + \tilde{s}_{-m_{1}}\dot{\tilde{s}}_{-m_{3}}\right ] -m_{b_2}^2\dot{\tilde{s}}_{-m_{1}}\sin^2\left(\frac{\lambda_f h}{f_h}\right)=0&\\ & 2\left[m_{b_2}^2\tilde{s}_{-m_{3}}\left(-\tilde{s}_{m_{2 } } + \tilde{s}_{m_{1}}\tilde{s}_{m_{2}}\tilde{s}_{-m_{1}}+m_{b_{1}}^{2}\tilde{s}_{m_{1}}\dot{\tilde{s}}_{m_{2}}\dot{\tilde{s}}_{-m_{1}}\right)+\right.&\nonumber\\ & \left.\tilde{s}_{-m_{1}}\left(\tilde{s}_{m_{2}}\dot{\tilde{s}}_{m_{1}}+m_{b_{1}}^{2}\tilde{s}_{m_{1}}\dot{\tilde{s}}_{m_{2}}\right)\dot{\tilde{s}}_{-m_{3}}\right]&\nonumber\\ & + \left(m_{b_{2}}^{2}\tilde{s}_{m_{2}}\tilde{s}_{-m_{3}}-m_{b_{1}}^{2}\dot{\tilde{s}}_{m_{2}}\dot{\tilde{s}}_{-m_{3}}\right)\sin^2\left(\frac{\lambda_f h}{f_h}\right)=0&\end{aligned}\ ] ] from these expressions we expect that the only non - trivial ew breaking vacua will be generated when @xmath354 which , as discussed in section 5 , leads to a theory which is highly non - renormalizable . we performed limited numerical analysis which confirmed this expectation . for completeness and as a reference for future work we write the expression for the determinant in the case of a non - zero mass mixing boundary term @xmath152 as in eq . ( [ localizedmasses ] ) . for simplicity , we turn off the singlet mass mixing @xmath355 . keeping parities as in our original model , we derive the following fermionic spectral functions , @xmath356=0&\\ & 2\tilde{s}_{m_{3}}\left[m_{b_2}^2\tilde{s}_{-m_{3}}\dot{\tilde{s}}_{-m_{1}}\dot{\tilde{s}}_{-m_{2}}+m_{b_3}^2\tilde{s}_{-m_{2}}\dot{\tilde{s}}_{-m_{1}}\dot{\tilde{s}}_{-m_{3 } } + \tilde{s}_{-m_{1}}\dot{\tilde{s}}_{-m_{2}}\dot{\tilde{s}}_{-m_{3}}\right]&\nonumber\\ & -m_{b_2}^2\dot{\tilde{s}}_{-m_{1}}\dot{\tilde{s}}_{-m_{2}}\sin^2\left(\frac{\lambda_f h}{f_h}\right)=0&\end{aligned}\ ] ] @xmath357&\nonumber\\ & + \left(m_{b_{2}}^{4}\tilde{s}_{m_{1}}\tilde{s}_{m_{2}}\tilde{s}_{-m_{3}}^2\dot{\tilde{s}}_{-m_{2}}^2+m_{b_{2}}^{2}\tilde{s}_{-m_{3}}\dot{\tilde{s}}_{-m_{2}}\left ( \tilde{s}_{m_{2}}\dot{\tilde{s}}_{m_{1}}\dot{\tilde{s}}_{-m_{2}}\right.\right.&\nonumber\\ & \left.\left.-2m_{b_{3}}^2\tilde{s}_{m_{1}}\left(\tilde{s}_{m_{2}}(-2+\tilde{s}_{m_{1}}\tilde{s}_{-m_{1 } } ) \tilde{s}_{-m_{2}}-\dot{\tilde{s}}_{-m_{1}}\dot{\tilde{s}}_{m_{1}}\dot{\tilde{s}}_{m_{2}}\dot{\tilde{s}}_{-m_{2}}\right)\right)\dot{\tilde{s}}_{-m_{3}}\right . & \nonumber\\ & \left.+ m_{b_{3}}^{2}\left(m_{b_{3}}^2\tilde{s}_{m_{1}}\tilde{s}_{-m_{2}}\left(2 - 2\tilde{s}_{m_{1}}\tilde{s}_{-m_{1}}+\tilde{s}_{m_{2}}\tilde{s}_{-m_{2}}\right)+ \dot{\tilde{s}}_{m_{1}}\dot{\tilde{s}}_{-m_{2}}\left(-2\tilde{s}_{m_{2}}\tilde{s}_{-m_{2}}\dot{\tilde{s}}_{m_{1}}\dot{\tilde{s}}_{-m_{1}}+\right.\right.\right . & \nonumber\\ & \left.\left.\left . \tilde{s}_{m_{1}}\tilde{s}_{-m_{1}}\dot{\tilde{s}}_{m_{2}}\dot{\tilde{s}}_{-m_{2}}+\dot{\tilde{s}}_{m_{1}}\dot{\tilde{s}}_{m_{2}}\dot{\tilde{s}}_{-m_{1 } } \dot{\tilde{s}}_{-m_{2}}\right)\right)\dot{\tilde{s}}_{-m_{3}}^{2}\right)\sin^2\left(\frac{\lambda_f h}{f_h}\right)&\nonumber\\ & -m_{b_{3}}^{2}\dot{\tilde{s}}_{-m_{3}}(m_{b_{2}}^{2}\tilde{s}_{m_{1}}\tilde{s}_{-m_{3}}\dot{\tilde{s}}_{-m_{2}}+(m_{b_{3}}^{2 } \tilde{s}_{m_{1}}\tilde{s}_{-m_{2}}+\dot{\tilde{s}}_{m_{1}}\dot{\tilde{s}}_{-m_{2}})\dot{\tilde{s}}_{-m_{3}})\sin^4\left(\frac{\lambda_f h}{f_h}\right)=0&\end{aligned}\ ] ] due to the term proportional to @xmath358 , we expect to have ewsb in this case , with non - trivial minima that for some region of parameter space could lie in the linear regime as in the model studied in this work . r. barate _ [ lep working group for higgs boson searches ] , phys . b * 565 * , 61 ( 2003 ) [ arxiv : hep - ex/0306033 ] . l. randall and r. sundrum , phys . lett . * 83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ] . n. s. manton , nucl . b * 158 * , 141 ( 1979 ) ; y. hosotani , phys . b * 126 * , 309 ( 1983 ) ; h. hatanaka , t. inami and c. s. lim , mod . a * 13 * , 2601 ( 1998 ) [ arxiv : hep - th/9805067 ] ; i. antoniadis , k. benakli and m. quiros , new j. phys . * 3 * , 20 ( 2001 ) [ arxiv : hep - th/0108005 ] ; m. kubo , c. s. lim and h. yamashita , mod . a * 17 * , 2249 ( 2002 ) [ arxiv : hep - ph/0111327 ] ; g. von gersdorff , n. irges and m. quiros , nucl . b * 635 * , 127 ( 2002 ) [ arxiv : hep - th/0204223 ] ; c. csaki , c. grojean and h. murayama , phys . d * 67 * , 085012 ( 2003 ) [ arxiv : hep - ph/0210133 ] ; n. haba , m. harada , y. hosotani and y. kawamura , nucl . b * 657 * , 169 ( 2003 ) [ erratum - ibid . b * 669 * , 381 ( 2003 ) ] [ arxiv : hep - ph/0212035 ] ; c. a. scrucca , m. serone and l. silvestrini , nucl . b * 669 * , 128 ( 2003 ) [ arxiv : hep - ph/0304220 ] ; c. a. scrucca , m. serone , l. silvestrini and a. wulzer , jhep * 0402 * , 049 ( 2004 ) [ arxiv : hep - th/0312267 ] ; n. haba , y. hosotani , y. kawamura and t. yamashita , phys . d * 70 * , 015010 ( 2004 ) [ arxiv : hep - ph/0401183 ] ; c. biggio and m. quiros , nucl . b * 703 * , 199 ( 2004 ) [ arxiv : hep - ph/0407348 ] ; y. hosotani , s. noda and k. takenaga , phys . b * 607 * , 276 ( 2005 ) [ arxiv : hep - ph/0410193 ] ; g. cacciapaglia , c. csaki and s. c. park , jhep * 0603 * , 099 ( 2006 ) [ arxiv : hep - ph/0510366 ] ; g. panico , m. serone and a. wulzer , nucl . phys . b * 739 * , 186 ( 2006 ) [ arxiv : hep - ph/0510373 ] ; g. panico , m. serone and a. wulzer , arxiv : hep - ph/0605292 ; r. contino , y. nomura and a. pomarol , nucl . b * 671 * , 148 ( 2003 ) [ arxiv : hep - ph/0306259 ] . h. davoudiasl , j. l. hewett and t. g. rizzo , phys . b * 473 * , 43 ( 2000 ) [ arxiv : hep - ph/9911262 ] . k. agashe , a. delgado , m. j. may and r. sundrum , jhep * 0308 * , 050 ( 2003 ) [ arxiv : hep - ph/0308036 ] . k. agashe , r. contino , l. da rold and a. pomarol , phys . b * 641 * , 62 ( 2006 ) [ arxiv : hep - ph/0605341 ] . m. carena , e. ponton , j. santiago and c. e. m. wagner , arxiv : hep - ph/0701055 . r. contino , l. da rold and a. pomarol , phys . d * 75 * , 055014 ( 2007 ) [ arxiv : hep - ph/0612048 ] . a. falkowski , phys . d * 75 * , 025017 ( 2007 ) [ arxiv : hep - ph/0610336 ] . k. agashe and r. contino , nucl . b * 742 * , 59 ( 2006 ) [ arxiv : hep - ph/0510164 ] . y. hosotani and m. mabe , phys . b * 615 * , 257 ( 2005 ) [ arxiv : hep - ph/0503020 ] . a. pomarol , phys . b * 486 * , 153 ( 2000 ) [ arxiv : hep - ph/9911294 ] . m. e. peskin and t. takeuchi , phys . d * 46 * , 381 ( 1992 ) .

today the universe appears spatially flat undergoing an accelerated expansion . there are many measurements proving this pictures @xcite . according to the successful cosmological model @xcite , there are two main ingredients in this scenario , namely dark matter ( dm ) and the cosmological constant @xmath1 ( dark energy ) . on the galactic scales , the evolution is driven by the usual newtonian gravitational potential , but it needs hypothesizing the existence of dm to obtain a good experimental agreement . a good model for the galactic distribution of dm , in the framework of general relativity ( gr ) , is the navarro - frenk - white model ( nfw model ) @xcite . however in recent years , the effort to give a physical explanation to the cosmic acceleration has attracted an amount of interest in so called fourth order gravity ( fog ) , and particularly the @xmath2-gravity , where @xmath3 is a generic function of ricci scalar @xmath4 . these alternative models have been considered as a viable mechanism to explain the cosmic acceleration . apart the cosmological dynamics , a systematic analysis of such theories were performed at short scale and in the low energy limit @xcite . in particular the paper _ most general fourth order theory of gravity at low energy _ @xcite analyzed the gravitational potential , induced by a @xmath5-gravity , where for sake of simplicity we set @xmath6 , @xmath7 and @xmath8 , generalizing the hilbert einstein lagrangian . the added quantities are the ricci tensor @xmath9 and the riemann tensor @xmath10 . as astrophysical application the modified potential has been used to build the rotation curves for the milky way and ngc 3198 @xcite . in this paper any galactic component ( bulge , disk and dm component ) required an onerous numerical computation since the gauss theorem is not applicable in the fog . the aim of the present paper is to point out the fundamental topics of the adopted strategy using the software _ _ mathematica__@xmath0 . our analysis is then organized as follows . in section [ theory ] we report the fundamental topics of the fourth order gravity : the field equations and their newtonian approximation , the solution for the gravitational potential and the mathematical models for the galactic componets . in section [ computation ] we build we build the code for the numerical simulation and in section [ data ] there is the data fit between our theoretical curves and the data of the rotation curve of the milky way and the galaxy ngc 3190 . finally in section [ conclusions ] we report the conclusions . let us start with a general class of fog given by the action @xmath11\end{aligned}\ ] ] where @xmath3 is an unspecified function of curvature invariants . the term @xmath12 is the minimally coupled ordinary matter contribution . in the metric approach , the field equations are obtained by varying ( [ fogaction ] ) with respect to @xmath13 . we get @xmath14_{;\nu)\alpha}\\\\\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,+\box[f_yr_{\mu\nu}]+[f_yr_{\alpha\beta}]^{;\alpha\beta}g_{\mu\nu } + 2f_zr_{\mu\alpha \beta\gamma}{r_{\nu}}^{\alpha\beta\gamma}-4[f_z{{r_\mu}^{\alpha\beta}}_\nu]_{;\alpha\beta}\,=\ , \mathcal{x}\,t_{\mu\nu } \\\\ f_xr+2f_yr_{\alpha\beta}r^{\alpha\beta}+2f_zr_{\alpha\beta\gamma\delta } r^{\alpha\beta\gamma\delta}-2f+\box[3 f_x+f_yr]+2[(f_y+2f_z)r^{\alpha\beta}]_{;\alpha\beta}\,=\,\mathcal{x}\,t \end{array}\end{aligned}\ ] ] where @xmath15 , @xmath16 , @xmath17 , @xmath18 and @xmath19 . @xmath20 is the energy - momentum tensor of matter and @xmath21 is its trace . the second line of ( [ fieldequationfog ] ) is the trace of the first one . in the case of weak field and slow motion we consider the field equation in the so called newtonian limit of theory . for our aim we can consider the metric tensor approximated as follows ( for details , see @xcite ) @xmath22 where @xmath23 is the gravitational potentials and @xmath24 is the kronecker delta . the set of coordinates adopted is @xmath25 . by introducing the quantities @xmath26 we get three differential equations for the curvature invariant @xmath27 and the gravitational potentials @xmath23 , @xmath28 @xmath29 x\,=\,-{m_2}^2\mathcal{x}\,\rho\\\\ ( \triangle-{m_1}^2)x\,=\,{m_1}^2\mathcal{x}\,\rho \end{array}\end{aligned}\ ] ] where @xmath30 is the laplacian in the flat space and @xmath31 is the matter density @xcite . further we assumes @xmath32 without loss of generality . by choosing @xmath33 and introducing @xmath34 the gravitational potential in the case of point - like source ( @xmath35 ) is given by @xmath36\end{aligned}\ ] ] while in the case of generic matter source distribution we perform the change @xmath37 . the passage from the pointlike source to extended one is correct only in the newtonian limit since a such limit corresponds also to the linearized version of theory . the motion of bodies is given by geodesic equation @xmath38 where @xmath39 is the relativistic distance and @xmath40 are the christoffel symbols . in the newtonian limit we obtain @xmath41 where the study of motion is very simple in particular cases of symmetry . for example the case of stationary motion on the circular orbit we get @xmath42 the distribution of mass can be modeled simply by introducing two sets of coordinates : the spherical coordinates @xmath43 and the cylindrical coordinates @xmath44 . an useful mathematical tool is the gauss flux theorem for gravity . since the newtonian mechanics satisfies this theorem and , by thinking to a spherical system of mass distribution , we get , from ( [ stationary_motion_2 ] ) , the equation @xmath45 where @xmath46 is the only mass enclosed in the sphere with radius @xmath47 . the green function of the @xmath5-gravity ( @xmath48 ) , instead , does not satisfy the theorem @xcite . in this case we must consider directly the gravitational potential @xmath49 \end{array}\end{aligned}\ ] ] apart the mathematical difficulties incoming from the research of gravitational potential for a given mass distribution , the non - validity of gauss theorem implies , for example , that a sphere can not be reduced to a point . in fact the gravitational potential generated by a ball ( also with constant density ) is depending also on the fourier transform of ball @xcite . only in the limit case where the radius of ball is small with respect to the distance we obtain the simple expression ( [ sol_pointlike ] ) . we remember that in the potential ( [ sol_gen ] ) we can distinguish the contributions of the bulge , the disk and the ( eventual ) dark matter . @xmath47 is the radial coordinate in the spherical system , while @xmath4 , @xmath50 are respectively the radial coordinate in the plane of disc and the distance from the plane then we have the geometric relation @xmath51 . the main item is the choice of models of matter distribution . the more simple model characterizing the shape of galaxy is the following @xmath52 where @xmath53 is the gamma function , @xmath54 is a free parameter and @xmath55 is the ratio of dark matter inside the sphere with radius @xmath56 with respect the total dark matter @xmath57 . moreover the couples @xmath58 , @xmath59 and @xmath60 , @xmath61 are the radius and the mass of the bulge and the disc . the computational analysis here described is referred to the study of the rotation curve ( [ stationary_motion_2 ] ) which can be replaced as follows @xmath62 where @xmath63 is the gravitational potential @xmath64 \nonumber\\\nonumber\\ & & + \frac{4\pi g}{3}\,\biggl[\frac{1}{r}\int_0^{\xi } dr'\,\rho_{dm}(r')\,r'\,\biggl(3\,\frac{|r - r'|-r - r'}{2 } -\frac{e^{-\mu_1|r - r'|}-e^{-\mu_1(r+r')}}{2\,\mu_1 } + 2\,\frac{e^{-\mu_2|r - r'|}-e^{-\mu_2(r+r ' ) } } { \mu_2}\biggr)\biggr ] \nonumber\\\nonumber\\ & & -2\,g\,\biggr\{\int_0^\infty dr'\,\sigma_{disc}(r')\,r'\,\biggl(\frac{\mathfrak{k}(\frac{4rr'}{(r+r')^2+z^2})}{\sqrt{(r+r')^2+z^2 } } + \frac{\mathfrak{k}(\frac{-4rr'}{(r - r')^2+z^2})}{\sqrt{(r - r')^2+z^2}}\biggr)+\int_0^\infty dr'\,\sigma_{disc}(r')\,r'\ , \\\nonumber\\ & & \times\int_0^{\pi } d\theta'\frac{1}{3\,\sqrt{(r+r')^2+z^2 - 4rr'\cos^2\theta ' } } \biggl[e^{-\mu_1\sqrt{(r+r')^2+z^2 - 4rr'\cos^2\theta ' } } -4\,e^{-\mu_2\sqrt{(r+r')^2+z^2 - 4rr'\cos^2\theta'}}\biggr]\biggr\}\nonumber\end{aligned}\ ] ] and @xmath65 is the elliptic function . the parameters @xmath66 and @xmath67 are the free parameters in the theory and only by fitting process can be fixed . a sensible item is the choice of distance @xmath68 on the which we are observing the rotation curve . in fact all models for the dark matter component are not limited and we need to cut the upper value of integration in ( [ potential ] ) . a further distinction are the contributions to the potential coming from terms of general relativity ( gr ) origin and terms of forth order gravity ( fog ) origin . finally our aim is the numerical evaluation of the rotation curve in the galactic plane @xmath69 the first step , after the definition of the numerical values for the parameters , has been the building of the velocity starting from the derivative of the potential . for this , we need to build the definitions of the contributions to the density coming from the bulge , the disk and the dark matter ( respectively ` \rhob[r _ ] ` , ` \sigmad[r_],\rhodm[r _ ] ` in the full code present in appendix ) , together with the already mentioned splitting in the gr contributions and fog contributions ( respectively ` tergr[x_,y _ ] ` and ` teryu[x_,y _ ] ` in the code ) . the derivative and integration operations commute , then we `` transport '' the derivative in the the integrand and then we make the integration . we found this computationally more rapid . we turning off the warning messages concerning the numerical integrations with the following commands : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ` off[nintegrate::inmur ] ` ` off[nintegrate::slowcon ] ` ` off[nintegrate::ncvb ] ` ` off[nintegrate::eincr ] ` _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the first ` off ` is justified since all the variables definitions are made with the `` setdelayed '' command ( : = ) that postpones the numerical evaluation of the integral making it not immediately numerical . the second turn off the message of slow convergence of the integration and making so , we avoid a long series of warning messages . the third avoid to the program to inform us of the need to use a larger number of recursive refinements in the computation . in effect , we increase the refinements with the command ` maxrecursion ` @xmath70 ` 20 ` for all the integrals , but this is not sufficient by itself to turn off the warnings . to do this , we need a bigger number than 20 , but the computation became excessively slow and the final result remains practically unchanged . since the computation is faced with oscillating error estimation , that explains the origin of the last warning message . we need to add to the numerical integration , the command ` method - > globaladaptive ` . still here , the warning message has to be manually closed since an improvement of ` method ` slows down the computation without an effective change in the results . an interesting thing to note , as it is possible to see in the complete code in appendix a and there noted with the comment ( * * ) , is that in the definition of the derivative by means of mute variables , it need not a `` setdelayed '' command , but a simple `` = '' command , otherwise the code is unable to make the computation . the next and more interesting step , is the comparison of the experimental data and what predicted by our model . from the literature cited in @xcite we can obtain the galactic speed values as function of the distance from the center and the corresponding errors . for instance , we show in some detail the manipulation of the data coming from the analysis of @xcite , concerning the external part of the milky way . we start copying the data listed in the table 1 of @xcite in a table called ` list1 ` . then we follow the prescriptions given by the authors with the introduction of new variables . as it is possible to see in the code present in appendix b , we preserve the same notations and with the command ` append ` we add to the initial ` list1 ` the new variables . for instance , for the @xmath4 variable , with the command ` mapthread[append,{list1,r } ] ` , we obtain a new table , here ` list2 ` , with one more column , the @xmath4 s valuer . and so on with the other variables . we introduce with the usual definitions , the errors on these derived quantities , here written as @xmath71 and added to the table . then @xmath72r and @xmath73 are , respectively , the error bars on the radius ( the distance from the galactic center ) and on the corresponding speeds and we process them together with the data in order to form the list with the values measured ( or derived ) and the errors on @xmath74 and @xmath75 . in order to obtai a plot with the error bars , we need to load the right package for this with the command ` need["errorbarplot"\ ` ] ` and this make us able to to make an ` errorlistplot ` . with the plot of ` list10 ` , we obtain a plot with the error bars only on y. a little bit more complicate procedure is need in order to obtain the bars on @xmath74 too . indeed we need to build a list of values like ` { { x , y},errorbar{err x , err y } } ` and this is done with the procedure shown in the last rows of appendix b. in figure [ stark ] it is shown the result . + similar procedures for the others two part of the milky way data and for the ngc 3190 data . at the end , the three experimental data for the milky way , each corresponding to a given range of distance from the center of the galaxy , are put together in order to obtain the complete galactic rotation curve . at this point we proceeded following two strategies . the first one , the faster , has been to overlap the theoretical graphs with the experimental one using the command ` show ` . in this case , the values of parameters in the densities ( bulge , dark matter and disk ) and of reduced masses , @xmath66 and @xmath67 , are chosen by a direct overlap of the graphs . the second strategy , more rigorous and slower , is the fit procedure . in appendix c is present the part of code of interest . in this case , in the code of the galactic rotation curve , we fix all other parameters except the `` masses '' @xmath66 and @xmath67 . these variables are the values that must be found whit the find fit procedure . we note that the ` findfit ` procedure uses the parameter constraints option . in this way , it is possible to eliminate all the solutions not physically allowed and to find the values obtained by the direct investigation , that is the first strategy , @xmath76 , @xmath77 where @xmath78 is the characteristic scale length fixed to the value of 1 kpc . obviously , an increase of the number of parameters to be found with this procedure , increase he time of the computation . in this paper we present a study of the galactic rotation curve when a fog is considered . the purely theoretical aspects have been fully exposed in @xcite . in the present work , after an obvious theoretical introduction , useful to remember the hypothesis used , we focus our attention on the salient points of the code that us permitted the computational analysis . with this program , we test the validity of our model of galactic rotation curve and the agreement of the experimental data of two galaxies , the milky way and the galaxy ngc 3190 , with our model . in order to make the explanation as complete as possible and to contextualize the several pieces of code examined , we show , in appendix a , the full code corresponding to the plot of the figure [ plot_1_prd ] , that is the code for a galaxy whose components are the bulge , the disk and the dark matter . the code referring also to the study of the galaxy ngc 3190 is exactly the same with the exclusion of the part of code referring to the bulge . as it is possible to see from figure [ plot_2_prd ] , the agreement of our model with the experimental data of the milky way is very good . only for very low values of the distance @xmath4 the agreement is not perfect . this suggest us that we only need an improvement of the parameters in the code , maintaining the code itself essentially unchanged . the complete code that refers to the data analysis is omitted since , apart the obvious introduction of the data coming from the cited literature , the complete program is a mere reply of what presented in appendix b. + dm ( dashed and dotted line ) , fog ( solid line ) , fog@xmath79 dm ( dotted line ) . the values of masses are @xmath80 and @xmath81 @xcite , title="fig : " ] + dm ( dashed and dotted line ) , fog ( solid line ) , fog@xmath79 dm ( dotted line ) on the experimental data for milky way . the values of masses are @xmath80 and @xmath81 @xcite.,title="fig : " ] + @xmath82;\text{msun}=1.98\times 10^{30};}\\ \pmb{\text{$\mu $ 1}=10^{-2};\text{$\mu $ 2}=10 ^ 2;\text{mb}=1.8\text{(*}\text{\textit{bulge } \text{mass}}\text{*)};\text{md}=6.5\text{(*}\text{\textit{disk } \text{mass}}\text{*)};\text{mdm}=4.2}\\ \pmb{\text{(*}\text{\textit{dark } \text{matter } \text{mass}}\text{*)};\text{$\xi $ b}=0.5\text{(*}\text{\textit{bulge } \text{radius}}\text{*)};\text{$\xi $ d}=3.5\text{(*}\text{\textit{galaxy } \text{radius}}\text{*)};}\\ \pmb{\text{$\xi $ dm}=5.5\text{(*}\text{d.m . radius}\text{*)};\chi = 20;\xi = \chi ; \text{stepr}=0.5;z=5\times 10^{-5};}\\ \pmb{\text{ab}\text{:=}\text{$\#$1 } ( \text{heavisidetheta}[\text{$\#$1}]-\text{heavisidetheta}[-\text{$\#$1}])\ & } \\ \pmb{k=10^{-3}\sqrt{\frac{\text{gr}\times 10^{10}\times \text{msun}\times \text{mb}}{10 ^ 3\times 3.08\times 10^{16}\times \text{$\xi $ b}}};}\\ \pmb{\text{$\rho $ b}[\text{r$\_$}]\text{:=}e^{-\left(\frac{r}{\text{$\xi $ b}}\right)^2 } \text{(*}\text{bulge density}\text{*)}}\\ \pmb{\text{$\rho $ dm}[\text{r$\_$}]\text{:=}\frac{1}{1+\left(\frac{r}{\text{$\xi $ dm}}\right)^2 } \text{(*}\text{dark matter density}\text{*)}}\\ \pmb{\text{$\sigma $ d}[\text{y$\_$}]\text{:=}e^{-\frac{y}{\text{$\xi $ d } } } \text{(*}\text{disk density}\text{*)}}\\ \pmb{\text{tergr}[\text{x$\_$},\text{y$\_$}]\text{:=}3\frac{\text{ab}[x - y]-x - y}{2}}\\ \pmb{\text{teryu}[\text{x$\_$},\text{y$\_$}]\text{:=}-\frac{e^{- \text{$\mu $ 1 } \text{ab}[x - y]}-e^{- \text{$\mu $ 1 } ( x+y)}}{2 \text{$\mu $ 1}}+2\frac{e^{- \text{$\mu $ 2 } \text{ab}[x - y]}-e^{-\text{$\mu $ 2 } ( x+y)}}{\text{$\mu $ 2}}}\\ \pmb{\text{f1}[\text{r$\_$},\text{rp$\_$}]\text{:=}\frac{\text{elliptick}\left[\frac{4r \text{rp}}{(r+\text{rp})^2+z^2}\right]}{\sqrt{(r+\text{rp})^2+z^2}}+\frac{\text{elliptick}\left[-\frac{4r \text{rp}}{(r-\text{rp})^2+z^2}\right]}{\sqrt{(r-\text{rp})^2+z^2}}}\\ \pmb{\text{f2}[\text{r$\_$},\text{rp$\_$},\text{$\theta $ p$\_$}]\text{:=}\frac{e^{-\text{$\mu $ 1 } \sqrt{(r+\text{rp})^2 - 4 r \text{rp}\text { } \text{cos}[\text{$\theta $ p}]^2+z^2}}-4e^{-\text{$\mu $ 2 } \sqrt{(r+\text{rp})^2 - 4 r \text{rp}\text { } \text{cos}[\text{$\theta $ p}]^2+z^2}}}{3\sqrt{(r+\text{rp})^2 - 4 r \text{rp}\text { } \text{cos}[\text{$\theta $ p}]^2+z^2}}}\\ \pmb{\text{integrandb1}[\text{r$\_$},\text{rp$\_$},\gamma \_]\text{:=}\frac{2}{3\text{$\xi $ b}^{2-\gamma } \text{gamma}\left[\frac{3-\gamma } { 2}\right]}\frac{1}{r } \text{$\rho $ b}[\text{rp}]\text{rp}^{1-\gamma } ( \text{tergr}[r,\text{rp}]+\text{teryu}[r,\text{rp}])}\\ \pmb{\text{derintegrandb1}[\text{r$\_$},\text{rp$\_$},\gamma \_]=d[\text{integrandb1}[r,\text{rp},\gamma ] , r ] ; \text{(*}\text{$\leftarrow $ } \text{*)}}\\ \pmb{\text{$\phi $ b1}[\text{r$\_$},\gamma \_]\text{:=}}\\ \pmb{\text{nintegrate}[\text{derintegrandb1}[r,\text{rp},\gamma ] , \{\text{rp},0,100\text{$\xi $ b}\},\text{maxrecursion}\to 20,}\\ \pmb{\text{method}\to \text{{``}globaladaptive{''}}]\text{/.}r\to \sqrt{r^2+z^2}}\\ \pmb{\text{integranddm1}[\text{r$\_$},\text{rp$\_$},\alpha \_]\text{:=}\frac{4 \alpha \text{mdm}/\text{mb}}{3(4-\pi ) \left.\text{$\xi $ dm}^3\right/\text{$\xi $ b}}\frac{1}{r } \text{$\rho $ dm}[\text{rp}]\text{rp}(\text{tergr}[r,\text{rp}]+\text{teryu}[r,\text{rp}])}\\ \pmb{\text{derintegranddm1}[\text{r$\_$},\text{rp$\_$},\alpha \_]=d[\text{integranddm1}[r,\text{rp},\alpha ] , r];\text{(*}\text{$\leftarrow $ } \text{*)}}\\ \pmb{\text{$\phi $ dm1}[\text{r$\_$},\alpha \_]\text{:=}}\\ \pmb{\text{nintegrate}[\text{derintegranddm1}[r,\text{rp},\alpha ] , \{\text{rp},0,\xi \},\text{maxrecursion}\to 20,}\\ \pmb{\text{method}\to \text{{``}globaladaptive{''}}]\text{/.}r\to \sqrt{r^2+z^2}}\\ \pmb{\text{integrandd11}[\text{r$\_$},\text{rp$\_$}]\text{:=}-\frac{(\text{md}/\text{mb})}{\pi \frac{\text{$\xi $ d}^2}{\text{$\xi $ b}}}\times ( \text{rp } \text{$\sigma $ d}[\text{rp } ] \text{f1}[r,\text{rp } ] ) } \\ \pmb{\text{integrandd12}[\text{r$\_$},\text{rp$\_$},\text{$\theta $ p$\_$}]\text{:=}-\frac{(\text{md}/\text{mb})}{\pi \frac{\text{$\xi $ d}^2}{\text{$\xi $ b}}}\times ( \text{rp } \text{$\sigma $ d}[\text{rp } ] \text{f2}[r,\text{rp},\text{$\theta $ p } ] ) } \\ \pmb{\text{derintegrandd11}[\text{r$\_$},\text{rp$\_$}]=d[\text{integrandd11}[r,\text{rp}],r];}\\ \pmb{\text{derintegrandd12}[\text{r$\_$},\text{rp$\_$},\text{$\theta $ p$\_$}]=d[\text{integrandd12}[r,\text{rp},\text{$\theta $ p}],r];\text{(*}\text{$\leftarrow $ } \text{*)}}\\ \pmb{\text{$\phi $ d1}[\text{r$\_$}]\text{:=}}\\ \pmb{(\text{nintegrate}[\text{derintegrandd11}[r,\text{rp}],\{\text{rp},0,50\text{$\xi $ d}\},\text{maxrecursion}\to 20,}\\ \pmb{\text{method}\to \text{{``}globaladaptive{''}}]+}\\ \pmb{\text{nintegrate}[\text{derintegrandd12}[r,\text{rp},\text{$\theta $ p}],\{\text{rp},0,50\text{$\xi $ d}\},\{\text{$\theta $ p},0,\pi \},\text{maxrecursion}\to 20,}\\ \pmb{\text{method}\to \text{{``}globaladaptive{''}}])}\\ \pmb{\text{integrandb2}[\text{r$\_$},\text{rp$\_$},\gamma \_]\text{:=}\frac{2}{3\text{$\xi $ b}^{2-\gamma } \text{gamma}\left[\frac{3-\gamma } { 2}\right]}\frac{1}{r } \text{$\rho $ b}[\text{rp}]\text{rp}^{1-\gamma } \text{tergr}[r,\text{rp}]}\\ \pmb{\text{derintegrandb2}[\text{r$\_$},\text{rp$\_$},\gamma \_]=d[\text{integrandb2}[r,\text{rp},\gamma ] , r];\text{(*}\text{$\leftarrow $ } \text{*)}}\\ \pmb{\text{$\psi $ b2}[\text{r$\_$},\gamma \_]\text{:=}}\\ \pmb{\text{nintegrate}[\text{derintegrandb2}[r,\text{rp},\gamma ] , \{\text{rp},0,100\text{$\xi $ b}\},\text{maxrecursion}\to 20,}\\ \pmb{\text{method}\to \text{{``}globaladaptive{''}}]\text{/.}r\to \sqrt{r^2+z^2}}\\ \pmb{\text{integranddm2}[\text{r$\_$},\text{rp$\_$},\alpha \_]\text{:=}\frac{4 \alpha \text{mdm}/\text{mb}}{3(4-\pi ) \left.\text{$\xi $ dm}^3\right/\text{$\xi $ b}}\frac{1}{r } \text{$\rho $ dm}[\text{rp}]\text{rp } \text{tergr}[r,\text{rp}]}\\ \pmb{\text{derintegranddm2}[\text{r$\_$},\text{rp$\_$},\alpha \_]=d[\text{integranddm2}[r,\text{rp},\alpha ] , r];\text{(*}\text{$\leftarrow $ } \text{*)}}\\ \pmb{\text{$\psi $ dm2}[\text{r$\_$},\alpha \_]\text{:=}}\\ \pmb{\text{nintegrate}[\text{derintegranddm2}[r,\text{rp},\alpha ] , \{\text{rp},0,\xi \},\text{maxrecursion}\to 20,}\\ \pmb{\text{method}\to \text{{``}globaladaptive{''}}]\text{/.}r\to \sqrt{r^2+z^2}}\\ \pmb{\text{derintegrandd2}[\text{r$\_$},\text{rp$\_$}]=d[\text{integrandd11}[r,\text{rp}],r];\text{(*}\text{$\leftarrow $ } \text{*)}}\\ \pmb{\text{$\psi $ d2}[\text{r$\_$}]\text{:=}\text{nintegrate}[\text{derintegrandd2}[r,\text{rp}],\{\text{rp},0,50\text{$\xi $ d}\},\text{maxrecursion}\to 20,}\\ \pmb{\text{method}\to \text{{``}globaladaptive{''}}]}\\ \pmb{\phi [ \text{r$\_$},\gamma \_]\text{:=}\text{$\phi $ b1}[r,\gamma ] + \text{$\phi $ d1}[r]}\\ \pmb{\psi [ \text{r$\_$},\gamma \_]\text{:=}\text{$\psi $ b2}[r,\gamma ] + \text{$\psi $ d2}[r]}\\ \pmb{\text{$\phi $ 1}[\text{r$\_$},\gamma \_,\alpha \_]\text{:=}\phi [ r,\gamma ] + \text{$\phi $ dm1}[r,\alpha ] } \\ \pmb{\text{$\psi $ 1}[\text{r$\_$},\gamma \_,\alpha \_]\text{:=}\psi [ r,\gamma ] + \text{$\psi $ dm2}[r,\alpha ] } \\ \pmb{\text{velfog}[\text{r$\_$},\gamma \_]\text{:=}k(r \phi [ r,\gamma ] ) ^{\frac{1}{2}}}\\ \pmb{\text{velgr}[\text{r$\_$},\gamma \_]\text{:=}k(r \psi [ r,\gamma ] ) ^{\frac{1}{2}}}\\ \pmb{\text{velfogdm}[\text{r$\_$},\gamma \_,\alpha \_]\text{:=}k(r \text{$\phi $ 1}[r,\gamma , \alpha ] ) ^{\frac{1}{2}}}\\ \pmb{\text{velgrdm}[\text{r$\_$},\gamma \_,\alpha \_]\text{:=}k(r \text{$\psi $ 1}[r,\gamma , \alpha ] ) ^{\frac{1}{2}}}\\ \pmb{\text{datavelfog}[\gamma \_]\text{:=}\text{table}\left[\text{velfog}[r,\gamma ] , \left\{r,10^{-7},\chi , \text{stepr}\right\}\right];}\\ \pmb{\text{datavelgr}[\gamma \_]\text{:=}\text{table}\left[\text{velgr}[r,\gamma ] , \left\{r,10^{-7},\chi , \text{stepr}\right\}\right];}\\ \pmb{\text{datavelfogdm}[\gamma \_,\alpha \_]\text{:=}\text{table}\left[\text{velfogdm}[r,\gamma , \alpha ] , \left\{r,10^{-7},\chi , \text{stepr}\right\}\right];}\\ \pmb{\text{datavelgrdm}[\gamma \_,\alpha \_]\text{:=}\text{table}\left[\text{velgrdm}[r,\gamma , \alpha ] , \left\{r,10^{-7},\chi , \text{stepr}\right\}\right];}\\ \pmb{\text{fig1}[\gamma \_]\text{:=}\text{listplot}[\text{datavelgr}[\gamma ] , \text{plotrange}\to \text{all},}\\ \pmb{\text{framelabel}\to \left\{\text{style}[\text{{``}r(kpc){''}},\text{large},\text{black}],\text{style}\left[\texttt{"}v_c\text{(r)(km / s)$\texttt{"}$},\text{large},\text{italic},\text{black}\right]\right\},}\\ \pmb{\text{datarange}\to \{0,\chi \},\text{plotstyle}\to \text{directive}[\text{black},\text{dashed},\text{thick}],\text{joined}\to \text{true } , } \\ \pmb{\text{axesorigin}\to \{0,0\},\text{frame}\to \text{true}]}\\ \pmb{\text{fig2}[\gamma \_]\text{:=}\text{listplot}[\text{datavelfog}[\gamma ] , \text{plotrange}\to \text{all},}\\ \pmb{\text{framelabel}\to \left\{\text{style}[\text{{``}r(kpc){''}},\text{large},\text{black}],\text{style}\left[\texttt{"}v_c\text{(r)(km / s)$\texttt{"}$},\text{large},\text{italic},\text{black}\right]\right\},}\\ \pmb{\text{datarange}\to \{0,\chi \},\text{plotstyle}\to \text{directive}[\text{black},\text{thick}],\text{joined}\to \text{true } , } \\ \pmb{\text{axesorigin}\to \{0,0\},\text{frame}\to \text{true}]}\\ \pmb{\text{fig3}[\gamma \_,\alpha \_]\text{:=}\text{listplot}[\text{datavelgrdm}[\gamma , \alpha ] , \text{plotrange}\to \text{all},}\\ \pmb{\text{framelabel}\to \left\{\text{style}[\text{{``}r(kpc){''}},\text{large},\text{black}],\text{style}\left[\texttt{"}v_c\text{(r)(km / s)$\texttt{"}$},\text{large},\text{italic},\text{black}\right]\right\},}\\ \pmb{\text{datarange}\to \{0,\chi \},\text{plotstyle}\to \text{directive}[\text{black},\text{dotdashed},\text{thick}],\text{joined}\to \text{true } , } \\ \pmb{\text{axesorigin}\to \{0,0\},\text{frame}\to \text{true}]}\\ \pmb{\text{fig4}[\gamma \_,\alpha \_]\text{:=}\text{listplot}[\text{datavelfogdm}[\gamma , \alpha ] , \text{plotrange}\to \text{all},}\\ \pmb{\text{framelabel}\to \left\{\text{style}[\text{{``}r(kpc){''}},\text{large},\text{black}],\text{style}\left[\texttt{"}v_c\text{(r)(km / s)$\texttt{"}$},\text{large},\text{italic},\text{black}\right]\right\},}\\ \pmb{\text{datarange}\to \{0,\chi \},\text{plotstyle}\to \text{directive}[\text{black},\text{dotted},\text{thick}],\text{joined}\to \text{true } , } \\ \pmb{\text{axesorigin}\to \{0,0\},\text{frame}\to \text{true}]}\\ \pmb{\alpha = 0.5;\gamma = 1.5;}\\ \pmb{\text{th}=\text{show}[\text{fig1}[\gamma ] , \text{fig2}[\gamma ] , \text{fig3}[\gamma , \alpha ] , \text{fig4}[\gamma , \alpha ] ] } $ ] @xmath83 ^ 2 - 2\text{r0 } \text{list1}[[\text{all},3]]\text{cos}[\text{list1}[[\text{all},1]]{}^{\circ}]\right)^{\frac{1}{2}};}\\ \pmb{\text{list2}=\text{mapthread}[\text{append},\{\text{list1},r\}];}\\ \pmb{\omega = \frac{\text{list2}[[\text{all},5]]}{\text{r0 } \text{sin}[\text{list2}[[\text{all},1]]{}^{\circ}]\text{cos}[\text{list2}[[\text{all},2]]{}^{\circ}]}+\text{$\omega $ 0}-}\\ \pmb{\frac{4.2 \text{cos}[\text{list2}[[\text{all},1]]{}^{\circ}]}{\text{r0 } \text{sin}[\text{list2}[[\text{all},1]]{}^{\circ}]};}\\ \pmb{\text{list3}=\text{mapthread}[\text{append},\{\text{list2},\omega \}];}\\ \pmb{\sigma \omega = \text{abs}\left[(\omega -\text{$\omega $ 0})\sqrt{\left(\frac{\text{list3}[[\text{all},6]]}{\text{list3}[[\text{all},5]]}\right)^2+\left(\frac{\text{$\sigma $ l}}{\text{tan}[\text{list3}[[\text{all},1]]{}^{\circ}]}\right)^2}\right];}\\ \pmb{\text{list4}=\text{mapthread}[\text{append},\{\text{list3},\sigma \omega \}];}\\ \pmb{\theta = r\times \omega ; } \\ \pmb{\text{list5}=\text{mapthread}[\text{append},\{\text{list4},\theta \}];}\\ \pmb{\text{$\sigma $ r}=}\\ \pmb{\text{abs}[}\\ \pmb{\frac{1}{r}\surd \left((\text{list4}[[\text{all},3]]-\text{r0 } \text{cos}[\text{list4}[[\text{all},1]]{}^{\circ}])^2\text{list4}[[\text{all},4]]^2+\right.}\\ \pmb{\left.\left.(\text{r0 } \text{list4}[[\text{all},3]]\text{sin}[\text{list4}[[\text{all},1]]{}^{\circ}])^2\text{$\sigma $ l}^2\right)\right];}\\ \pmb{\text{list6}=\text{mapthread}[\text{append},\{\text{list5},\text{$\sigma $ r}\}];}\\ \pmb{\text{list7}=\text{drop}[\text{list6},\{\},\{1,6\}];}\\ \pmb{\sigma \theta = \text{abs}\left[\theta \sqrt{\left(\frac{\sigma \omega } { \omega } \right)^2+\left(\frac{\text{$\sigma $ r}}{r}\right)^2}\right];}\\ \pmb{\text{list8}=\text{mapthread}[\text{append},\{\text{list7},\sigma \theta \}];}\\ \pmb{\text{list9}=\text{drop}[\text{list8},\{\},\{2,3\}]\text{(*errors on x and y*)};}\\ \pmb{\text{list10}=\text{drop}[\text{list9},\{\},\{3\}]\text{(*errors on y*)};}\\ \pmb{\text{needs}[\text{{``}errorbarplots$\grave { } $ { ' ' } } ] } \\ \pmb{\text{errorlistplot}[\text{list10},}\\ \pmb{\text{axeslabel}\to \{\text{style}[\text{{``}r(kpc){''}},\text{large},\text{black}],}\\ \pmb{\left.\text{style}\left[\texttt{"}v_c\text{(r)(km / s)$\texttt{"}$},\text{large},\text{italic},\text{black}\right]\right\},\text{plotstyle}\to \text{black},}\\ \pmb{\text{plotmarkers}\to \{\text{{``}$\blacksquare${''}}\ } , \text{plotrange}\to \text{all},\text{axesorigin}\to \{0,0\}]}$ ] + + + @xmath84}\\ \pmb{\text{errx}=\text{drop}[\text{drop}[\text{list9},\{\},\{1,2\}],\{\},\{2\}][[\text{all},1]]}\\ \pmb{\text{erry}=\text{drop}[\text{drop}[\text{list9},\{\},\{1,2\}],\{\},\{1\}][[\text{all},1]]}\\ \pmb{\text{listerr}=\{\text{errx},\text{erry}\}^{\mathsf{t}}}\\ \pmb{\text{error}=\text{cases}[\text{listerr},\{\text{x$\_$},\text{y$\_$}\}:\to \text{errorbar}[x , y]]}\\ \pmb{\text{valerr}=\text{map}[\{\#[[1]],\#[[2]]\}\&,\{\text{data},\text{error}\}^{\mathsf{t}}]}\\ \pmb{\text{errorlistplot}[\text{valerr},}\\ \pmb{\text{framelabel}\to \{\text{style}[\text{{``}r(kpc){''}},\text{large},\text{black}],}\\ \pmb{\left.\text{style}\left[\texttt{"}v_c\text{(r)(km / s)$\texttt{"}$},\text{large},\text{italic},\text{black}\right]\right\},\text{plotstyle}\to \text{black},}\\ \pmb{\text{plotmarkers}\to \{\text{{``}$\blacksquare${''}}\ } , \text{plotrange}\to \text{all},\text{axesorigin}\to \{0,0\},\text{frame}\to \text{true}]}$ ] + @xmath85\text{:=}(\text{$\phi $ b1}[r,\text{$\mu $ 1},\text{$\mu $ 2}]+\text{$\phi $ d1}[r,\text{$\mu $ 1},\text{$\mu $ 2}]+\text{$\phi $ dm1}[r,\text{$\mu $ 1},\text{$\mu $ 2}])}\\ \pmb{\text{velfogdm}[\text{r$\_$},\text{$\mu $ 1$\_$},\text{$\mu $ 2$\_$}]\text{:=}k(r \text{$\phi $ 1}[r,\gamma , \text{$\mu $ 1},\text{$\mu $ 2 } ] ) ^{\frac{1}{2}}}\\ \pmb{\text{model}= \text{velfogdm}[r,\text{$\mu $ 1},\text{$\mu $ 2}];}\\ \pmb{\text{findfit}\left[\text{val},\text{model},\left\{\left\{\text{$\mu $ 1},10^{-2}\right\},\left\{\text{$\mu $ 2},10 ^ 2\right\}\right\},r\right]}$ ]

vector auto regressive ( var ) models play an important role in contemporary macro economics , being an example of an approach called the `` dynamic stochastic general equilibrium '' ( dsge ) , which is superseding traditional large scale macro econometric forecasting methodologies @xcite . the motivation behind them is based on the assertion that more recent values of a variable are more likely to contain useful information about its future movements than the older ones . on the other hand , a standard tool in multivariate time series analysis is vector moving average ( vma ) models , which is really a linear regression of the present value of the time series w.r.t . the past values of a white noise . a broader class of stochastic processes used in macro economy comprises both these kinds together in the form of vector auto regressive moving average ( varma ) models . these methodologies can capture certain spatial and temporal structures of multidimensional variables which are often neglected in practice ; including them not only results in more accurate estimation , but also leads to models which are more interpretable . they are widely used by academia and central banks ( _ cf . _ the european central bank s smets wouters model for the euro zone @xcite ) , as they constitute quite a simple version of the dsge equations . varma models are constructed from a number of univariate arma ( box jenkins ; see for example @xcite ) processes , typically coupled with each other . in this paper , we investigate only a significantly simplified circumstance when there is no coupling between the many arma components . one may argue that this is too far fetched and will be of no use in describing an economic reality . however , one may also treat it as a `` zeroth order hypothesis , '' analogously to the idea of @xcite in finance , namely that the case with no cross covariances is considered theoretically , and subsequently compared to some real world data modeled by a varma process ; any discrepancy between the two will reflect nontrivial cross covariances present in the system , thus permitting their investigation . this latter route is taken in this communication . a challenging and yet increasingly important problem is the estimation of large covariance matrices generated by these stationary processes , since high dimensionality of the data as compared to the sample size is quite common in many statistical problems ( the `` dimensionality curse '' ) . therefore , an appropriate `` noise cleaning '' procedure has to be implemented , and random matrix theory ( rmt ) provides a natural and efficient outfit for doing that . in particular , the mean spectral densities ( a.k.a . `` limiting spectral distributions , '' lsd ) of the pearson estimators of the cross covariances for the @xmath7 and @xmath8 models , in the relevant high dimensionality sector and under the full decoupling , have been derived in @xcite by applying the framework proposed by @xcite . in this paper , we suggest that such calculations can be considerably simplified by resorting to a mathematical concept of the free random variables ( frv ) calculus @xcite , succinctly introduced in sec . [ s : doublycorrelatedwishartensemblesandfreerandomvariables ] . our general frv formula @xcite allows not only to rediscover , which much less strain , the two fourth order polynomial equations obtained in @xcite in the @xmath7 and @xmath8 cases , but also to derive a sixth order equation ( [ eq : varmaoneonemainequation ] ) which produces the mean spectral density for a more involved @xmath6 model . the results are verified by numerical simulations , which show a perfect agreement . this is all done in sec . [ s : varmafromfreerandomvariables ] . @xmath2 stochastic processes , as we will see below , fall within quite a general set up encountered in many areas of science where a probabilistic nature of multiple degrees of freedom evolving in time is relevant , for example , multivariate time series analysis in finance , applied macro econometrics and engineering . to describe this framework , consider a situation of @xmath3 time dependent random variables which are measured at @xmath4 consecutive time moments ( separated by some time interval @xmath9 ) ; let be the value od the @xmath10th ( @xmath11 ) random number at the @xmath12th time moment ( @xmath13 ) ; together , they make up a rectangular @xmath14 matrix @xmath15 . in what usually would be the first approximation , each is supposed to be drawn from a gaussian probability distribution . we will also assume that they have mean values zero , . these degrees of freedom may in principle display mutual correlations . a set of correlated zero mean gaussian numbers is fully characterized by the two point covariance function , if the underlying stochastic process generating these numbers is stationary . linear stochastic processes , including @xmath2 , belong to this category . we will restrict our attention to an even narrower class where the cross correlations between different variables and the auto correlations between different time moments are factorized , _ i.e. _ , @xmath16 in this setting , the inter variable covariances do not change in time ( and are described by an @xmath17 cross covariance matrix @xmath18 ) , and also the temporal covariances are identical for all the numbers ( and are included in a @xmath19 auto covariance matrix @xmath20 ; both these matrices are symmetric and positive definite ) . the gaussian probability measure with this structure of covariances is known from textbooks , @xmath21_{i j } y_{j b } \left [ \mathbf{a}^{- 1 } \right]_{b a } \right ) { \mathrm{d}}\mathbf{y } = \ ] ] @xmath22 where the normalization constant , and the integration measure , while the letters `` c.g . '' stand for `` correlated gaussian . '' now , a standard way to approach correlated gaussian random numbers is to recall that they can always be decomposed as linear combinations of uncorrelated gaussian degrees of freedom ; indeed , this is achieved through the transformation @xmath23 where the square roots of the covariance matrices , necessary to facilitate the transition , exist due to the positive definiteness of @xmath18 and @xmath20 ; the new normalization reads . an essential problem in multivariate analysis is to determine ( estimate ) the covariance matrices @xmath18 and @xmath20 from given @xmath3 time series of length @xmath4 of the realizations of our random variables . for simplicity , we do not distinguish in notation between random numbers , _ i.e. _ , the population , and their realizations in actual experiments , _ i.e. _ , the sample . since the realized cross covariance between degrees @xmath10 and @xmath24 at the same time @xmath12 is , the simplest method to estimate the today s cross covariance is to compute the time average , @xmath25 this is usually named the `` pearson estimator '' , up to the prefactor which depending on the context is @xmath26 or @xmath27 . other estimators might be introduced , such as between distinct degrees of freedom at separate time moments ( `` time delayed estimators '' ) , or with certain decreasing weights given to older measurements to reflect their growing obsolescence ( `` weighted estimators '' ) , but we will not investigate them in this article . furthermore , in the last equality in ( [ eq : estimatorcdefinition ] ) , we cast @xmath28 through the uncorrelated gaussian numbers contained in , the price to pay for this being that the covariance matrices now enter into the expression for @xmath28 , making it more complicated ; this will be the form used hereafter . the random matrix @xmath28 is called a `` doubly correlated wishart ensemble '' @xcite . let us also mention that the auto covariance matrix @xmath20 can be estimated through . however , it is verified that this object carries identical information to the one contained in @xmath28 ( it is `` dual '' to @xmath28 ) , and therefore may safely be discarded . indeed , these two estimators have same non zero eigenvalues ( modulo an overall rescaling by @xmath29 ) , and the larger one has @xmath30 additional zero modes . any historical estimator is inevitably marred by the measurement noise ; it will reflect the true covariances only to a certain degree , with a superimposed broadening due to the finiteness of the time series . more precisely , there are @xmath31 independent elements in @xmath18 , to be estimated from @xmath32 measured quantities @xmath15 , hence the estimation accuracy will depend on the `` rectangularity ratio , '' @xmath33 the closer @xmath29 to zero , the more truthful the estimate . this is a cornerstone of classical multivariate analysis . unfortunately , a practical situation will typically feature a large number of variables sampled over a comparably big number of time snapshots , so that we may approximately talk about the `` thermodynamical limit , '' @xmath34 on the other hand , it is exactly this limit in which the frv calculus ( see the subsection below for its brief elucidation ) can be applied ; hence , the challenge of de noising is somewhat counterbalanced by the computationally powerful frv techniques . any study of a ( real symmetric @xmath36 ) random matrix @xmath37 will most surely include a fundamental question about the average values of its ( real ) eigenvalues . they are concisely encoded in the `` mean spectral density , '' @xmath38 here the expectation map @xmath39 is understood to be taken w.r.t . the probability measure @xmath40 of the random matrix . we will always have this distribution rotationally ( _ i.e. _ , , with @xmath41 orthogonal ) invariant , and hence the full information about @xmath37 resides in its eigenvalues , distributed on average according to ( [ eq : meanspectraldensitydefinition ] ) . on the practical side , it is more convenient to work with either of the two equivalent objects , @xmath42 referred to as the `` green s function '' ( or the `` resolvent '' ) and the `` @xmath35transform '' of @xmath37 . the latter is also called the `` moments generating function , '' since if the `` moments '' of @xmath37 exist , it can be expanded into a power series around @xmath43 as . it should however be underlined that even for probability measures disallowing such an expansion ( heavy tailed distributions , preeminent in finance , being an example ) , the quantities ( [ eq : greenfunctionandmtransformdefinition ] ) still manage to entirely capture the spectral properties of @xmath37 ; hence the name `` @xmath35transform '' more appropriate , in addition to being more compact . we will show that for our purposes ( multiplication of random matrices ; see par . [ sss : thentransformandfreerandomvariables ] ) the @xmath35transform serves better than the green s function . however , it is customary to write the relationship between ( [ eq : meanspectraldensitydefinition ] ) and ( [ eq : greenfunctionandmtransformdefinition ] ) in terms of this latter , @xmath44 resulting from a well known formula for generalized functions , . the doubly correlated wishart ensemble @xmath28 ( [ eq : estimatorcdefinition ] ) may be viewed as a product of several random and non random matrices . the general problem of multiplying random matrices seems formidable . in classical probability theory , it can be effectively handled in the special situation when the random terms are independent : then , the exponential map reduces it to the addition problem of independent random numbers , solved by considering the logarithm of the characteristic functions of the respective pdfs , which proves to be additive . in matrix probability theory , a crucial insight came from d. voiculescu and coworkers and r. speicher @xcite , who showed how to parallel the commutative construction in the noncommutative world . it starts with the notion of `` freeness , '' which basically comprises probabilistic independence together with a lack of any directional correlation between two random matrices . this nontrivial new property happens to be the right extension of classical independence , as it allows for an efficient algorithm of multiplying free random variables ( frv ) , which we state below : step 1 : : : suppose we have two random matrices , and , mutually free . their spectral properties are best wrought into the @xmath35transforms ( [ eq : greenfunctionandmtransformdefinition ] ) , and . step 2 : : : the critical maneuver is to turn attention to the functional inverses of these @xmath35transforms , the so called `` @xmath3transforms , '' @xmath45 step 3 : : : the @xmath3transforms submit to a very straightforward rule upon multiplying free random matrices ( the `` frv multiplication law '' ) , @xmath46 step 4 : : : finally , it remains to functionally invert the resulting @xmath3transform to gain the @xmath35transform of the product , , and consequently , all the spectral properties via formula ( [ eq : sokhotskyformula ] ) . it is stunning that such a simple prescription ( relying on the choice of the @xmath35transform as the carrier of the mean spectral information , and the construction of its functional inverse , the @xmath3transform , which essentially multiplies under taking the free product ) resolves the multiplication problem for free random noncommutative objects . let us just mention that the addition problem may be tackled along similar lines : in this case , the green s function should be exploited , its functional inverse considered (; it is sometimes called the `` blue s function '' @xcite ) , which obeys the `` frv addition law , '' , for two free random matrices . in this paper , we do not resort to using this addition formula , even though our problem could be approached through it as well . let us also remark that in the original mathematical formulations @xcite of these frames , a slightly different language is employed : instead of the @xmath3transform , the `` @xmath47transform '' is found convenient , , while in place of the blue s function , one engages the `` @xmath48transform , '' . they fulfil simpler laws , and , respectively . the innate potential of the frv multiplication algorithm ( [ eq : frvmultiplicationlaw ] ) is surely revealed when inspecting the doubly correlated wishart random matrix ( [ eq : estimatorcdefinition ] ) . this has been done in detain in @xcite , so we will only accentuate the main results here , referring the reader to the original paper for a thorough explanation . the idea is that one uses twice the cyclic property of the trace ( which permits cyclic shifts in the order of the terms ) , and twice the frv multiplication law ( [ eq : frvmultiplicationlaw ] ) ( to break the @xmath3transforms of products of matrices down to their constituents ) , in order to reduce the problem to solving the uncorrelated wishart ensemble . this last model is further simplified , again by the cyclic property and the frv multiplication rule applied once , to the standard @xmath49 random matrix squared ( and the projector , designed to chip the rectangle off the square @xmath49 ) , whose properties are firmly established . let us sketch the derivation , @xmath50 @xmath51 @xmath52 this is the basic formula . since the spectral properties of @xmath28 are given by its @xmath35transform , , it is more pedagogical to recast ( [ eq : doublycorrelatedwishartensemblethemainequation1 ] ) as an equation for the unknown @xmath35 , @xmath53 it provides a means for computing the mean spectral density of a doubly correlated wishart random matrix once the `` true '' covariance matrices @xmath18 and @xmath20 are given . in this communication , only a particular instance of this fundamental formula is applied , namely with an arbitrary auto covariance matrix @xmath20 , but with trivial cross covariances , . using that , equation ( [ eq : doublycorrelatedwishartensemblethemainequation2 ] ) thins out to @xmath54 which will be strongly exploited below . let us mention that these equalities ( [ eq : doublycorrelatedwishartensemblethemainequation2 ] ) , ( [ eq : doublycorrelatedwishartensemblethemainequation3 ] ) have been derived through other , more tedious , techniques ( the planar feynman diagrammatic expansion , the replica trick ) in @xcite . in what follows , we will assume that the @xmath0 , @xmath1 , or @xmath2 stochastic processes are covariance ( weak ) stationary ; for details , we refer to @xcite . it implies certain restrictions on their parameters , but we will not bother with this issue in the current work . another consequence is that the processes display some interesting features , such as invertibility . for all this , we must in particular take both @xmath3 and @xmath4 large from the start , with their ratio @xmath55 fixed ( [ eq : thermodynamicallimit ] ) . more precisely , we stretch the range of the @xmath12index from minus to plus infinity . this means that all the finite size effects ( appearing at the ends of the time series ) are readily disregarded . in particular , there is no need to care about initial conditions for the processes , and all the recurrence relations are assumed to continue to the infinite past . we consider a situation when @xmath3 stochastic variables evolve according to identical independent @xmath0 ( vector moving average ) processes , which we sample over a time span of @xmath4 moments . this is a simple generalization of the standard univariate weak stationary moving average @xmath56 . in such a setting , the value of the @xmath10th ( @xmath11 ) random variable at time moment @xmath12 ( @xmath13 ) can be expressed as @xmath57 here all the s are iid standard ( mean zero , variance one ) gaussian random numbers ( white noise ) , . the s are some @xmath58 real constants ; importantly , they do not depend on the index @xmath10 , which reflects the fact that the processes are identical and independent ( no `` spatial '' covariances among the variables ) . the rank @xmath59 of the process is a positive integer . in order to handle such a process ( [ eq : vmaqdefinition ] ) , notice that the s , being linear combinations of uncorrelated gaussian numbers , must also be gaussian random variables , albeit displaying some correlations therefore , to fully characterize these variables , it is sufficient to calculate their two point covariance function ; this is straightforwardly done ( see appendix [ ss : theautocovariancematrixforvmaq ] for details ) , @xmath60 where @xmath61 in other words , the cross covariance matrix is trivial , ( no correlations between different variables ) , while the auto covariance matrix , responsible for temporal correlations , can be called `` @xmath62diagonal . '' in the course of this article , we will use several different auto covariance matrices , and for brevity , we decide to label them with superscripts ; their definitions are all collected in appendix [ ss : alistofthevariousautocovariancematricesused ] . for example , in the simplest case of @xmath7 , it is tri diagonal , @xmath63 such an infinite matrix ( [ eq : vmaqautocovariancematrix ] ) is translationally invariant ( as announced , it is one of the implications of the weak stationarity ) , _ i.e. _ , the value of any of its entries depends only on the distance between its indices , ; specifically , , for @xmath64 , and . hence , it is convenient to rewrite this matrix in the fourier space , @xmath65 in this representation , the @xmath35transform of is readily obtained @xcite , @xmath66 this integral can be evaluated by the method of residues for any value of @xmath59 , which we do in appendix [ ss : themtransformoftheautocovariancematrixforvmaq ] , where also we print the general result ( [ eq : vmaqgreenfunctionoftheautocovariancematrix ] ) . in particular , for @xmath67 , @xmath68 where the square roots are principal . we will be interested in investigating the spectral properties of the pearson estimator ( [ eq : estimatorcdefinition ] ) . the @xmath35transform of this correlated wishart random matrix , , can be retrieved from equation ( [ eq : doublycorrelatedwishartensemblethemainequation3 ] ) . we could write it for any @xmath59 using ( [ eq : vmaqgreenfunctionoftheautocovariancematrix ] ) , but we will restrict ourselves only to @xmath67 , in which case the substitution of ( [ eq : vmaonemomentsgeneratingfunctionfora ] ) leads to a fourth order polynomial ( ferrari ) equation for the unknown @xmath35 , @xmath69 @xmath70 @xmath71 the frv technique allowed us therefore to find this equation in a matter of a few lines of a simple algebraic computation . it has already been derived in @xcite , and ( [ eq : vmaonemainequation ] ) may be verified to coincide with the version given in that paper . in @xcite , the pertinent equation is printed before ( a.6 ) , and to compare the two , one needs to change their variables into ours according to @xmath72 , @xmath73 , and . the last equality means that and @xmath74 of @xcite correspond in our language to the green s functions and , respectively , where is the pearson estimator dual to @xmath28 . as mentioned , a quick extension to the case of arbitrary @xmath59 is possible , however the resulting equations for @xmath35 will be significantly more complicated ; for instance , for @xmath75 , a lengthy ninth order polynomial equation is discovered . a set up of @xmath3 identical and independent @xmath1 ( vector auto regressive ) processes is somewhat akin to ( [ eq : vmaqdefinition ] ) , _ i.e. _ , we consider @xmath3 decoupled copies of a standard univariate @xmath76 process , @xmath77 it is again described by the demeaned and standardized gaussian white noise ( which triggers the stochastic evolution ) , as well as @xmath58 real constants , , with @xmath78 . as announced before , the time stretches to the past infinity , so no initial condition is necessary . although at first sight ( [ eq : varqdefinition ] ) may appear to be a more involved recurrence relation for the s , it is actually easily reduced to the @xmath0 case : it remains to remark that if one exchanges the s with the s , one precisely arrives at the @xmath0 process with the constants , , @xmath78 . in other words , the auto covariance matrix of the @xmath1 process ( [ eq : varqdefinition ] ) is simply the inverse of the auto covariance matrix of the corresponding @xmath0 process with the described modification of the parameters , @xmath79 this inverse exists thanks to the weak stationarity supposition . the fourier transform of the auto covariance matrix of @xmath1 is therefore a ( number ) inverse of its counterpart for @xmath0 with its parameters appropriately changed , @xmath80 where @xmath81 and where we define . in order to find the @xmath35transform of the inverse matrix , , one employs a general result , true for any ( real symmetric ) random matrix @xmath37 , and obtainable through an easy algebra , @xmath82 since the quantity is known for any @xmath59 ( [ eq : vmaqgreenfunctionoftheautocovariancematrix ] ) , hence is via ( [ eq : mhinversefrommh ] ) , but we will not print it explicitly . let us just give it for @xmath67 , in which case ( [ eq : mhinversefrommh ] ) and ( [ eq : vmaonemomentsgeneratingfunctionfora ] ) yield @xmath83 despite being somewhat outside of the main line of thought of this article , an interesting question would be to search for an explicit expression for the auto covariance matrix from its fourier transform ( [ eq : varqautocovariancematrixfourier ] ) , @xmath84 where we exploited the fact that must be translationally invariant , . this computation would shed light on the structure of temporal correlations present in a var setting . this integral is evaluated by the method of residues in a very similar manner to the one shown in appendix [ ss : themtransformoftheautocovariancematrixforvmaq ] , and we do this in appendix [ ss : theautocovariancematrixforvarq ] . we discover that the auto covariance matrix is a sum of @xmath59 exponential decays , @xmath85 where are constants , and are the characteristic times ( [ eq : varqautocovariancematrixcharacteristictimes ] ) , @xmath86 ; these constituents are given explicitly in ( [ eq : varqautocovariancematrix2 ] ) . this is a well known fact , nevertheless we wanted to establish it again within our approach . for example , for @xmath67 , the auto covariance matrix of @xmath8 is one exponential decay , @xmath87 where we assumed for simplicity ( the formula can be easily extended to all values of ) . having found an expression for the @xmath35transform of the auto covariance matrix of a @xmath1 ( [ eq : mhinversefrommh ] ) , ( [ eq : vmaqgreenfunctionoftheautocovariancematrix ] ) , we may proceed to investigate the equation ( [ eq : doublycorrelatedwishartensemblethemainequation3 ] ) for the @xmath35transform of the correlated wishart random matrix ( [ eq : estimatorcdefinition ] ) . we will do this explicitly only for @xmath67 , when ( [ eq : varonemomentsgeneratingfunctionfora ] ) leads to a fourth order ( ferrari ) polynomial equation for the unknown @xmath35 , @xmath88 @xmath89 this equation has been derived by another method in @xcite , and our result confirms their equation ( a.8 ) , with the change in notation , @xmath72 , @xmath73 , @xmath90 . the two types of processes which we elaborated on above , and , can be combined into one stochastic process called , @xmath91 now it is a straightforward and well known observation ( which can be verified by a direct calculation ) that the auto covariance matrix of this process is simply the product ( in any order ) of the auto covariance matrices of the var and vma pieces ; more precisely , @xmath92 where corresponds to the generic model ( [ eq : vmaqautocovariancematrix ] ) , while denotes the auto covariance matrix of with a slightly different modification of the parameters compared to the previously used , namely , , for . we have thus already made use here of the fact that the auto covariance matrix of a var process is the inverse of the auto covariance matrix of a certain corresponding vma process ( [ eq : varqfromvmaq ] ) , but the new change in parameters necessary in moving from var to vma has effectively w.r.t . what we had before ( [ eq : varqfromvmaq ] ) ; it is understandable : this `` missing '' is now included in the matrix of the other process . of the cross covariances in the @xmath6 process computed numerically from the sixth order polynomial equation ( [ eq : varmaoneonemainequation ] ) , for various values of the process parameters . the scale of these parameters is determined by choosing everywhere . recall that the theoretical formula ( [ eq : varmaoneonemainequation ] ) is valid in the thermodynamical limit ( [ eq : thermodynamicallimit ] ) of @xmath93 , with @xmath94 kept finite . + up left : we set the remaining varma parameters to , , while the rectangularity ratio takes the values @xmath95 ( the purple line ) , @xmath96 ( red ) , @xmath97 ( magenta ) , @xmath98 ( pink ) ; each one is @xmath99 times smaller than the preceding one . we observe how the graphs become increasingly peaked ( narrower and taller ) around @xmath100 as @xmath29 decreases , which reflects the movement of the estimator @xmath28 toward its underlying value . + up right : we fix @xmath101 and consider the two varma parameters equal to each other , with the values ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath104 ( pink ) . + down left : we hold @xmath101 and , and modify ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath105 ( pink ) ; for this last value , the @xmath6 model reduces to @xmath8 . + down right : similarly , but this time we assign @xmath101 and , while changing ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath105 ( pink ) ; this last value corresponds to @xmath7.,title="fig:",width=321 ] of the cross covariances in the @xmath6 process computed numerically from the sixth order polynomial equation ( [ eq : varmaoneonemainequation ] ) , for various values of the process parameters . the scale of these parameters is determined by choosing everywhere . recall that the theoretical formula ( [ eq : varmaoneonemainequation ] ) is valid in the thermodynamical limit ( [ eq : thermodynamicallimit ] ) of @xmath93 , with @xmath94 kept finite . + up left : we set the remaining varma parameters to , , while the rectangularity ratio takes the values @xmath95 ( the purple line ) , @xmath96 ( red ) , @xmath97 ( magenta ) , @xmath98 ( pink ) ; each one is @xmath99 times smaller than the preceding one . we observe how the graphs become increasingly peaked ( narrower and taller ) around @xmath100 as @xmath29 decreases , which reflects the movement of the estimator @xmath28 toward its underlying value . + up right : we fix @xmath101 and consider the two varma parameters equal to each other , with the values ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath104 ( pink ) . + down left : we hold @xmath101 and , and modify ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath105 ( pink ) ; for this last value , the @xmath6 model reduces to @xmath8 . + down right : similarly , but this time we assign @xmath101 and , while changing ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath105 ( pink ) ; this last value corresponds to @xmath7.,title="fig:",width=321 ] of the cross covariances in the @xmath6 process computed numerically from the sixth order polynomial equation ( [ eq : varmaoneonemainequation ] ) , for various values of the process parameters . the scale of these parameters is determined by choosing everywhere . recall that the theoretical formula ( [ eq : varmaoneonemainequation ] ) is valid in the thermodynamical limit ( [ eq : thermodynamicallimit ] ) of @xmath93 , with @xmath94 kept finite . + up left : we set the remaining varma parameters to , , while the rectangularity ratio takes the values @xmath95 ( the purple line ) , @xmath96 ( red ) , @xmath97 ( magenta ) , @xmath98 ( pink ) ; each one is @xmath99 times smaller than the preceding one . we observe how the graphs become increasingly peaked ( narrower and taller ) around @xmath100 as @xmath29 decreases , which reflects the movement of the estimator @xmath28 toward its underlying value . + up right : we fix @xmath101 and consider the two varma parameters equal to each other , with the values ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath104 ( pink ) . + down left : we hold @xmath101 and , and modify ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath105 ( pink ) ; for this last value , the @xmath6 model reduces to @xmath8 . + down right : similarly , but this time we assign @xmath101 and , while changing ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath105 ( pink ) ; this last value corresponds to @xmath7.,title="fig:",width=321 ] of the cross covariances in the @xmath6 process computed numerically from the sixth order polynomial equation ( [ eq : varmaoneonemainequation ] ) , for various values of the process parameters . the scale of these parameters is determined by choosing everywhere . recall that the theoretical formula ( [ eq : varmaoneonemainequation ] ) is valid in the thermodynamical limit ( [ eq : thermodynamicallimit ] ) of @xmath93 , with @xmath94 kept finite . + up left : we set the remaining varma parameters to , , while the rectangularity ratio takes the values @xmath95 ( the purple line ) , @xmath96 ( red ) , @xmath97 ( magenta ) , @xmath98 ( pink ) ; each one is @xmath99 times smaller than the preceding one . we observe how the graphs become increasingly peaked ( narrower and taller ) around @xmath100 as @xmath29 decreases , which reflects the movement of the estimator @xmath28 toward its underlying value . + up right : we fix @xmath101 and consider the two varma parameters equal to each other , with the values ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath104 ( pink ) . + down left : we hold @xmath101 and , and modify ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath105 ( pink ) ; for this last value , the @xmath6 model reduces to @xmath8 . + down right : similarly , but this time we assign @xmath101 and , while changing ( purple ) , @xmath102 ( red ) , @xmath103 ( magenta ) , @xmath105 ( pink ) ; this last value corresponds to @xmath7.,title="fig:",width=321 ] ) ( the dashed red lines ) . the conformity is nearly perfect . we generate the matrices @xmath15 of sizes @xmath106 , @xmath107 ( _ i.e. _ , @xmath101 ) from the @xmath6 process with the parameters , , . the monte carlo simulation is repeated @xmath108 ( left ) or @xmath109 ( right ) times ; in this latter case , a significant improvement in the quality of the agreement is seen . one notices finite size effects at the edges of the spectrum ( `` leaking out '' of eigenvalues ) : in the numerical simulations , @xmath3 and @xmath4 are obviously finite , while equation ( [ eq : varmaoneonemainequation ] ) is legitimate in the thermodynamical limit ( [ eq : thermodynamicallimit ] ) only , hence the small discrepancies ; by enlarging the chosen dimensions @xmath110 one would diminish this fallout.,title="fig:",width=321 ] ) ( the dashed red lines ) . the conformity is nearly perfect . we generate the matrices @xmath15 of sizes @xmath106 , @xmath107 ( _ i.e. _ , @xmath101 ) from the @xmath6 process with the parameters , , . the monte carlo simulation is repeated @xmath108 ( left ) or @xmath109 ( right ) times ; in this latter case , a significant improvement in the quality of the agreement is seen . one notices finite size effects at the edges of the spectrum ( `` leaking out '' of eigenvalues ) : in the numerical simulations , @xmath3 and @xmath4 are obviously finite , while equation ( [ eq : varmaoneonemainequation ] ) is legitimate in the thermodynamical limit ( [ eq : thermodynamicallimit ] ) only , hence the small discrepancies ; by enlarging the chosen dimensions @xmath110 one would diminish this fallout.,title="fig:",width=321 ] the fourier transform of the auto covariance matrix of ( [ eq : varmaq1q2fromvarq1vmaq2 ] ) is simply the product of the respective fourier transforms ( [ eq : vmaqautocovariancematrixfourier ] ) and ( [ eq : varqautocovariancematrixfourier ] ) , @xmath111 where @xmath112 where we recall . for instance , for @xmath6 ( it is described by three constants , , , ) , one explicitly has @xmath113 the @xmath35transform of can consequently be derived from the general formula ( [ eq : momentsgeneratingfunctionforatranslationallyinvariant ] ) . we will evaluate here the pertinent integral only for the simplest @xmath6 process , even though an arbitrary case may be handled by the technique of residues , @xmath114 one might again attempt to track the structure of temporal covariances in a varma process . this can be done either by the inverse fourier transform of ( [ eq : varmaq1q2autocovariancematrixfourier ] ) , or through a direct computation based on the recurrence relation ( [ eq : varmaq1q2definition ] ) ( importantly , adhering to the assumption that it stretches to the past infinity ) . let us print the result just for @xmath6 , @xmath115 where for simplicity . this is an exponential decay , with the characteristic time of the var piece , with an additional term on the diagonal . expression ( [ eq : varmaoneonemomentsgeneratingfunctionforafive ] ) , along with the fundamental frv formula ( [ eq : doublycorrelatedwishartensemblethemainequation3 ] ) , allow us to write the equation satisfied by the @xmath35transform of the pearson estimator ( [ eq : estimatorcdefinition ] ) of the cross covariances in the @xmath6 process ; it happens to be polynomial of order six , and we print it ( [ eq : varmaoneonemainequation ] ) in appendix [ ss : theequationforthemtransformofthepearsonestimatorofthecovariancesforvarmaoneone ] . it may be solved numerically , a proper solution chosen ( the one which leads to a sensible density : real , positive definite , normalized to unity ) , and finally , the mean spectral density derived from ( [ eq : sokhotskyformula ] ) . we show the shapes of this density for a variety of the values of the parameters @xmath29 , , , in fig . [ fig : varmaoneonetheory ] . moreover , in order to test the result ( [ eq : varmaoneonemainequation ] ) , and more broadly , to further establish our frv framework in the guise of formula ( [ eq : doublycorrelatedwishartensemblethemainequation3 ] ) , the theoretical form of the density is compared to monte carlo simulations in fig . [ fig : varmaoneonetheoryplusmc ] ; they remain in excellent concord . these are the main findings of this article . in this paper we attempted to advertise the power and flexibility of the free random variables calculus for multivariate stochastic processes of the varma type . the frv calculus is ideally suited for multidimensional time series problems , provided the dimensions of the underlying matrices are large . the operational procedures are simple , algebraic and transparent . the structure of the final formula which relates the moments generating function of the population covariance and the sample covariance allows one to easily derive eigenvalue density of the sample covariance . we in detail illustrated how this procedure works for @xmath6 , confronted the theoretical prediction with numerical data obtained by monte carlo simulations of the varma process and observed a perfect agreement . the frv calculus is not restricted to gaussian variables . it also works for non gaussian processes , including those with heavy tailed increments belonging to the lvy basin of attraction , where the moments do not exist . since the majority of data collected nowadays is naturally stored in the form of huge matrices , we believe that the frv technique is the most natural candidate for the matrix valued `` probability calculus '' that can provide efficient algorithms for cleaning ( de noising ) large sets of data and unraveling essential but hidden correlations . this work has been supported by the polish ministry of science grant no . n n202 229137 ( 20092012 ) . aj acknowledges the support of clico ltd . in this appendix , we sketch a proof of the formula ( [ eq : vmaqautocovariancematrix ] ) for the auto covariance matrix of the @xmath0 process . as mentioned , since the random variables are centered gaussian , this matrix alone suffices to completely capture all their properties . we set @xmath116 ; the dependence on this index may be dropped as there are no correlations here . we use the definition ( [ eq : vmaqdefinition ] ) of @xmath0 , as well as the auto covariance structure of the white noise , . this leads to @xmath117 the double sum is symmetrized , the index @xmath118 replaced by @xmath119 , @xmath120 and the order of the sums interchanged ( an elegant method for this is explained in @xcite ) , @xmath121 which , upon splitting the sum over @xmath122 into three pieces ( from @xmath123 to @xmath124 , @xmath125 , and from @xmath126 to @xmath59 ) , is quickly seen to coincide with ( [ eq : vmaqautocovariancematrix ] ) . * by we denote the auto covariance matrix of the @xmath0 process with the generic constants , with @xmath127 , as defined in ( [ eq : vmaqdefinition ] ) . * by we denote the auto covariance matrix of the @xmath0 process with the constants , , where @xmath78 . * by we denote the auto covariance matrix of the @xmath1 process with the generic constants , , with @xmath78 , as defined in ( [ eq : varqdefinition ] ) . there holds ( [ eq : varqfromvmaq ] ) . * by we denote the auto covariance matrix of the @xmath128 process with the constants , , where . * by we denote the auto covariance matrix of the @xmath2 process with the generic constants , , with and , according to the definition ( [ eq : varmaq1q2definition ] ) . there is ( [ eq : varmaq1q2fromvarq1vmaq2 ] ) , where in the latter piece . [ [ ss : themtransformoftheautocovariancematrixforvmaq ] ] the @xmath35transform of the auto covariance matrix for @xmath0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ we will derive here the @xmath35transform ( [ eq : momentsgeneratingfunctionforatranslationallyinvariant ] ) of the auto covariance matrix of an arbitrary @xmath0 process , using the expression for its fourier transform ( [ eq : vmaqautocovariancematrixfourier ] ) . it is a little simpler to consider the green s function , @xmath129 where the integration range has been halved due to the evenness of the integrand . this integral is performed with help of the change of variables @xmath130 . the measure , when @xmath131 $ ] , reads . a basic observation is that the denominator of the integrand is a linear combination of @xmath132 , for @xmath133 , and each such a cosine can be cast as a polynomial of order @xmath122 in @xmath134 through the de moivre formula . hence , the denominator is a polynomial of order @xmath59 in @xmath134 , @xmath135 where the s are the @xmath59 roots ( which we assume to be single ) , and @xmath136 is the coefficient at . using the method of residues , one readily finds @xmath137 where the two square roots on the r.h.s . are principal . this is an explicit formula for the green s function of , provided one has factorized the order@xmath59 polynomial ( [ eq : vmaqgreenfunctionoftheautocovariancematrixdenominator ] ) . let us argue now that the fourier transform ( [ eq : varqautocovariancematrixfourier ] ) leads to the auto covariance matrix of @xmath1 ( [ eq : varqautocovariancematrixfromitsfouriertransform ] ) of the form of a sum of exponential decays ( [ eq : varqautocovariancematrix1 ] ) , and let us give precise expressions for the constants and the characteristic times , @xmath86 . we proceed by the technique of residues , analogously to appendix [ ss : themtransformoftheautocovariancematrixforvmaq ] , however this time with aid of another variable , , related to the previously used through @xmath138 . the integration measure is @xmath139 , and the integration path is counterclockwise around the centered unit circle . the denominator of the integrand is a polynomial of order @xmath59 in the variable @xmath134 , having thus some @xmath59 roots , @xmath78 . therefore , there are @xmath140 corresponding solutions for the variable @xmath141 , with a half of them inside the integration path and a half outside ; let be the solutions to with the absolute values less than @xmath126 . only them contribute to the integral , and their residues straightforwardly give @xmath142 this is indeed @xmath59 exponents , @xmath86 . remark that the solutions may be complex , hence this is really @xmath59 different exponential decays , with the characteristic times @xmath143 ( these times are positive as the roots have the absolute values less than @xmath126 ) , possibly modulated by sinusoidal oscillations when a root has an imaginary part . for example , for @xmath67 there is one exponential decay ( [ eq : varoneautocovariancematrix ] ) , while for @xmath75 , one obtains either two exponential decays ( the two roots are real and different ) , or one exponential decay modulated by oscillations ( the two roots are complex and mutually conjugate ) , _ etc . _ [ [ ss : theequationforthemtransformofthepearsonestimatorofthecovariancesforvarmaoneone ] ] the equation for the @xmath35transform of the pearson estimator of the covariances for @xmath6 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ the sixth order polynomial equation obeyed by in the case of @xmath6 reads , @xmath144 @xmath145 @xmath146 @xmath147 @xmath148 @xmath149 @xmath150 @xmath151 @xmath152 @xmath153 @xmath154 @xmath155 @xmath156 @xmath157 @xmath158 @xmath159 @xmath160 @xmath161 @xmath162 @xmath163 this equation in a ` mathematica ` file can be obtained from the authors upon request . 99 sims c. a. , _ macroeconomics and reality _ , econometrica * 48 * ( 1980 ) 1 . smets f. , wouters r. , _ an estimated stochastic dynamic general equilibrium model of the euro area _ , european central bank working paper series , working paper * 171 * , august 2002 [ ` http://www.ecb.int ` ] . box g. , jenkins g. m. , reinsel g. , _ time series analysis : forecasting and control _ , prentice hall , englewood cliffs , nj , 1994 . laloux l. , cizeau p. , bouchaud j.p . , potters m. , _ noise dressing of financial correlation matrices _ , phys . * 83 * ( 1999 ) 1467 [ ` arxiv : cond - mat/9810255 ` ] . plerou v. , gopikrishnan p. , rosenow b. , amaral l. a. n. , stanley h. e. , _ universal and non universal properties of cross correlations in financial time series _ , phys . * 83 * ( 1999 ) 1471 [ ` arxiv : cond - mat/9902283 ` ] . jin b. , wang c. , miao b. , huang m.n . l. , _ limiting spectral distribution of large dimensional sample covariance matrices generated by varma _ , journal of multivariate analysis * 100 * ( 2009 ) 2112 . bai z. d. , silverstein j. w. , _ spectral analysis of large dimensional random matrices _ , science press , beijing , 2006 . voiculescu d. v. , dykema k. j. , nica a. , _ free random variables _ , crm monograph series , vol . 1 , am . math . providence , 1992 . speicher r. , _ multiplicative functions on the lattice of non crossing partitions and free convolution _ , math . * 298 * ( 1994 ) 611 . burda z. , jarosz a. , jurkiewicz j. , nowak m. a. , papp g. , zahed i. , _ applying free random variables to random matrix analysis of financial data _ , submitted to quantitative finance . wishart j. , _ the generalized product moment distribution in samples from a normal multivariate population _ , biometrika * a 20 * ( 1928 ) 32 . zee a. , _ law of addition in random matrix theory _ , nucl . phys . * b 474 * ( 1996 ) 726 [ ` arxiv : cond - mat/9602146 ` ] . janik r. a. , nowak m. a. , papp g. , zahed i. , _ various shades of blue s functions _ , acta phys * b 28 * ( 1997 ) 2949 [ ` arxiv : hep - th/9710103 ` ] . burda z. , grlich a. , jarosz a. , jurkiewicz j. , _ signal and noise in correlation matrix _ , physica * a 343 * ( 2004 ) 295 [ ` arxiv : cond - mat/0305627 ` ] . burda z. , jurkiewicz j. , _ signal and noise in financial correlation matrices _ , physica * a 344 * ( 2004 ) 67 [ ` arxiv : cond - mat/0312496 ` ] . burda z. , jurkiewicz j. , wacaw b. , _ spectral moments of correlated wishart matrices _ , phys . e 71 * ( 2005 ) 026111 [ ` arxiv : cond - mat/0405263 ` ] . burda z. , grlich a. , jurkiewicz j. , wacaw b. , _ correlated wishart matrices and critical horizons _ , eur . j. * b 49 * ( 2006 ) 319 [ ` arxiv : cond - mat/0508341 ` ] . burda z. , jurkiewicz j. , wacaw b. , _ eigenvalue density of empirical covariance matrix for correlated samples _ , acta phys . pol . * b 36 * ( 2005 ) 2641 [ ` arxiv : cond - mat/0508451 ` ] . ltkepohl h. , _ new introduction to multiple time series analysis _ , springer verlag , berlin , 2005 . graham r. , knuth d. , patashnik o. , _ concrete mathematics : a foundation for computer science _ , addison wesley , 1994 .

dust emission gives us crucial information to understand both the history of cosmic star formation in galaxies and that of supermassive black hole ( smbh ) growth in galactic centers . intense star formation produces a great amount of dust , which makes its activity invisible in the ultraviolet band but visible in the infrared band . similarly , the central engines of active galactic nuclei ( agns ) are surrounded by dusty `` tori '' @xcite . since optical and ultraviolet lights are very easily obscured by the torus , the most complete search for agns can be made by detecting hard x - rays from the central engine or infrared emission from the heated dust . utilizing x - ray data of the chandra deep field south , @xcite reported that the fraction of `` compton - thick '' agns ( those with absorption column densities of @xmath6@xmath7 ) increases with redshift from @xmath8 to @xmath9 . this fact implies that the level of dust - obscuration of agn may be correlated to the star formation activity , which has a peak at @xmath10 @xcite . indeed , @xcite gathered all available _ spitzer _ spectra of local compton - thick agns and found strong starburst ( sb ) features in their average infrared spectrum , suggesting that sb activity and highly obscured tori are somehow coupled each other . @xcite also reported the same trend for buried agns with small opening angles of tori ( new type agns ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) detected in the _ swift_/bat survey . although hard x - rays are useful to search for obscured agns , extremely heavily buried agns with @xmath11@xmath7 are difficult to be detected due to flux attenuation by repeated comptonization even at energies above 10 kev ( e.g. , * ? ? ? * ; * ? ? ? candidates of galaxies that host such buried agns are infrared galaxies , defined as those having large infrared luminosities ( ) with @xmath12 . among infrared galaxies , those in the range @xmath13 are called luminous infrared galaxies ( lirgs ; * ? ? ? * ) and more luminous ones in the range @xmath14 are called ultraluminous infrared galaxies ( ulirgs ; * ? ? ? their bolometric luminosities are dominated by the infrared emission , suggesting that very luminous heating sources are surrounded by dust and then the heated dust re - emits in the infrared band . the hidden energy sources are believed to be either sb , agn , or both . disentangling the energy sources of infrared galaxies is crucial to unveil the history of dust obscured star formation and smbh growth in the universe @xcite . @xcite studied agn activity in u / lirgs on the basis of optical observations . they showed that the fraction of agns increases with infrared luminosities in the range of @xmath15 . one caveat of their optical diagnostics is that one can not identify buried agns embedded in dust - tori with very small opening angles where no significant narrow line regions can be formed . unlike the agns which are optically - classified type-2 seyfert , such buried agns are elusive to conventional optical spectroscopy @xcite . after the launch of _ spitzer _ , the agn activity in local u / lirgs has been also addressed in the infrared band . @xcite observed ulirgs with _ spitzer_/irs 58 @xmath0 m spectroscopy , and investigated both the sb and agn contributions by fitting sb / agn templates to the spectra . they found the increasing agn significance with infrared luminosity , confirming the same trend reported from the optical studies . @xcite observed the 538 @xmath0 m spectra of lirgs with _ spitzer_. they decomposed the spectra into sb and agn components using a sb galaxy template and clumpy torus models @xcite , which are supported by observations ( e.g. , * ? ? ? * ; * ? ? ? they confirmed that the agn energetic contribution to the total power increases with infrared luminosity from @xmath16 . infrared 2.55 @xmath0 m spectroscopy is another powerful tool to study optically - elusive buried agns in u / lirgs . one advantage of this band is that dust extinction is much lower than in the optical band @xcite and is similar to that in the 513 @xmath0 m band @xcite . another great advantage is that sb and agn activity can be distinguished based on the spectral features . first , strong polycyclic aromatic hydrocarbon ( pah ) emission , a pure star formation tracer , is located at 3.3 @xmath0 m in sb galaxies , while pure agns exhibit pah - feature free spectra due to the effects that 1 ) the x - ray emission destroying the pahs , and 2 ) the strong continuum emission originating from agn hot - dust diminishes this feature @xcite . sb galaxies generally show large 3.3 @xmath0 m pah equivalent widths of @xmath17 nm , which never go down below 40 nm . thus , objects with @xmath18 nm may be classified as galaxies that harbor buried agns . in sb / agn composite galaxies , strong pah emission can be observed because the agn can not destroy the pah molecules located in the outer ( @xmath19 pc ) region from the central engine due to shielding by dust and gas . however , strong agn continuum makes ew@xmath20 smaller , well below @xmath21 nm . using this method , many authors reported the strong sign of buried agns in ulirgs @xcite and lirgs @xcite . another key feature in the 2.55 @xmath0 m band is the continuum slope @xmath22 ( @xmath23 ) . sb galaxies have blue continuum slopes ( @xmath24 ) in this band due to the contribution of the stellar photospheric continuum . conversely , galaxies with agns have hot - dust emission heated by agns , which produces a red continuum ( @xmath25 ) . this method was first reported by @xcite and successfully applied to ulirgs @xcite and lirgs @xcite . @xcite combined the two methods ( @xmath26 nm and/or @xmath27 ) to find buried agns from a sample of u / lirgs showing no agn signatures in the optical spectra . they found that the fraction of buried agns increases with infrared luminosity in the range of @xmath28 . in this paper , we expand the studies of u / lirgs by @xcite into a lower infrared luminosity range , by including normal infrared galaxies ( irgs ) with @xmath29 , where comprehensive studies are still missing due to the limited sample . this is because l / irgs have spatially more extended infrared emission ( @xmath30 kpc , corresponding to @xmath31several arcseconds at @xmath32 ) compared with ulirgs @xcite , and therefore slit spectroscopy with ground based telescopes could miss their extended emission . slit - less spectroscopy of infrared camera ( irc ; * ? ? ? * ) on board _ akari _ @xcite with 1@xmath33 aperture opened a new window to probe such extended emission of l / irgs . throughout this paper , we adopt @xmath34 km s@xmath35mpc@xmath35 , @xmath36 , and @xmath37 for consistency with the previous publications @xcite . our motivation is to search for buried agns from optically non - seyfert infrared galaxies and investigate their properties over a wider range of infrared luminosity than in previous studies . an ideal sample can be selected from unbiased catalogs of infrared galaxies with various infrared luminosities at redshifts of @xmath38 , where 3.3 @xmath0 m pah is detectable in the _ akari_/irc 2.55.0 @xmath0 m spectra . to this end , we first gather u / lirgs with high far - infrared fluxes from _ iras _ catalogs . our ulirgs are mainly selected from the bright ulirg catalog by @xcite , consisting of 41 sources 25 out of which are optically non - seyfert galaxies . these ulirgs have _ iras _ 60 @xmath0 m fluxes @xmath39 jy and far - infrared ( 40120 @xmath0 m ) luminosities @xmath40 km s@xmath35mpc@xmath35 . therefore , in our study with @xmath41 km s@xmath35mpc@xmath35 , some galaxies reach @xmath42 . ] . in addition to these sources , we also gather 54 fainter , local ( @xmath38 ) optically non - seyfert ulirgs from the literature to increase the sample size . the lirgs are mainly taken from the lirg catalog of @xcite , which is based on the _ iras _ bright galaxy catalog . the catalog consists of 61 sources with the criteria of _ iras _ 60 @xmath0 m fluxes @xmath43 jy and far - infrared ( 40400 @xmath0 m ) luminosities @xmath44 . we select 41 optically non - seyfert galaxies out of this sample . to further increase the number of l / irg targets , we also utilize the catalog by @xcite . they compiled 76 _ iso _ detected sources in the _ iras _ 12 @xmath0 m galaxy catalog , which gathered all galaxies at galactic latitudes of @xmath45 down to the _ iras _ 12 @xmath0 m flux limit of @xmath46 jy @xcite . since agn tori have peak emission at the mir band , 12 @xmath0 m galaxy catalog could tend to include more agns than in far - infrared selected samples . indeed , a majority of the @xcite catalog ( 64 out of 76 sources ) are optically seyfert galaxies . therefore , we gather the 12 optically non - seyfert starburst galaxies from it . in summary , we select a sufficiently large number of non - seyfert infrared galaxies ( 132 sources ) as our `` parent '' sample to be followed - up with _ akari_. the redshifts of all the sources in the above catalogs are distributed within @xmath38 . , title="fig : " ] + observation was conducted with the irc infrared spectrograph @xcite on board _ akari _ @xcite . the spectra in the 2.55 @xmath0 m band were taken with the ng grism mode . this mode achieves an effective spectral resolution of @xmath47 at 3.6 @xmath0 m , which is sufficient for detecting and tracing the profile of @xmath48 @xmath0 m pah emission . the irc has a @xmath49 arcmin@xmath50 window with a pixel scale of @xmath51 arcsec@xmath50 . all the data were taken as a part of the _ akari _ mission program called `` agnul '' ( pi : t. nakagawa ) . the observation settings were same as those described in chapter 3 of @xcite . one to seven pointings were assigned for each source according to the brightness . the total on - source exposure time was @xmath52 minutes per each pointing . we used ircz4 ( phase 3 ; post liquid - he mission ) observing mode , where one pointing is composed of eight or nine independent frames @xcite . this mode successfully removes cosmic - ray contaminations even for sources observed by a single pointing . for sources observed with multiple pointings , we combined all the data to achieve the best signal - to - noise ( s / n ) ratio after excluding the data sets with particularly bad quality . we often had to discard data observed in later epochs of _ akari _ s phase 3 operation due to the increasing background signal . spectral analysis was conducted in a standard manner by using the idl data reduction package for _ akari _ irc spectra . the dark - subtraction , linearity correction , and flat - field correction were performed using this idl package . some infrared galaxies with low luminosities ( @xmath53 ) have spatially extended emission . hence , we varied the aperture size for spectral extraction according to the actual signal profile of each source . the background was estimated from both sides of the spectral dispersion direction of the target and was subtracted from the source . the idl package for _ akari _ also performed wavelength and absolute flux calibration . the accuracy of the wavelength calibration is @xmath54 pixel ( @xmath55 @xmath0 m ) , and that of the flux calibration is @xmath56% around the central wavelength of the spectra and up to @xmath57% at the edges close to 2.5 @xmath0 m and 5.0 @xmath0 m . out of the 132 sources in our parent sample , we observed 37 objects in total from 2008 june until 2010 february with _ akari_. after the epoch , _ akari _ fell into the stage where the data highly suffered from the significant background emission . as the result , we could not observe the remaining 95 targets . also , due to _ akari _ s sun - synchronous polar orbit , higher visibilities were achieved for objects at higher ecliptic latitudes . these _ akari _ s observational constraints make it difficult for us to construct a uniformly flux - limited sample over the entire sky with _ akari_. however , the observable targets are essentially randomly selected from the parent sample with no selection biases regarding the physical nature of the galaxies . hence , we regard that our sample has no obvious bias . next , we also set the criterion that the averaged s / n ratio for each spectral channel must be higher than 5 in order to apply our spectral deconvolution analysis . among the 37 targets observed with _ , five sources were discarded due to their low s / n ratio and the other 10 sources were too faint to be detected . 14 out of 15 non - detected sources are fainter ulirgs gathered from the literature . these sources have 60 @xmath0 m flux @xmath58 jy , which is roughly 10 times fainter than that of the detected sources ( @xmath59 jy ) . supposing that their infrared spectral energy distribution does not change drastically from that of the detected sources , we can regard that the most of the non - detected sources are intrinsically too faint to be observed with _ figure 1 shows the redshift distribution of the finally selected 22 infrared galaxies . lr*7c eso 286-ig19 & 1122176 - 001 & 0.043 & 0.28 & 1.90 & 12.71 & 10.58 & 12.00 & 2comp + ic 5135 & 1120236 - 001 & 0.016 & 0.63 & 2.14 & 16.67 & 26.27 & 11.33 & 2comp + iras 03068 - 5346 & 1122105 - 001 & 0.074 & 0.25 & 0.20 & 3.43 & 3.92 & 11.98 & 2comp + iras 03538 - 6432 & 1122106 - 001 & 0.30 & 0.25 & 0.25 & 0.96 & 1.54 & 13.01 & 2comp + iras 10494 + 4424 & 1122096 - 001 & 0.092 & 0.25 & 0.25 & 3.36 & 5.60 & 12.22 & 2comp + iras 17028 + 5817 & 1122101 - 001 & 0.11 & 0.25 & 0.08 & 2.49 & 4.05 & 12.22 & 2comp + irasf 07353 + 2903 & 1120200 - 001 & 0.33 & 0.08 & 0.10 & 0.22 & 0.64 & 12.63 & 2comp + mcg + 02 - 04 - 025 & 1122123 - 001 & 0.031 & 0.37 & 1.46 & 11.13 & 10.29 & 11.67 & 2comp + mcg + 08 - 23 - 097 & 1122130 - 001 & 0.029 & 0.25 & 0.53 & 5.08 & 8.10 & 11.34 & 2comp + mcg + 10 - 19 - 057 & 1122132 - 001 & 0.031 & 0.40 & 1.92 & 11.35 & 10.81 & 11.70 & 2comp + mrk 1490 & 1122134 - 001 & 0.026 & 0.25 & 0.86 & 6.21 & 8.38 & 11.30 & 3comp + mrk 551 & 1122171 - 001 & 0.050 & 0.26 & 0.80 & 4.63 & 6.13 & 11.80 & 2comp + mrk 848 & 1122136 - 001 & 0.040 & 0.32 & 1.52 & 9.38 & 10.26 & 11.86 & 2comp + ngc 2339 & 1120230 - 001 & 0.0074 & 0.53 & 2.11 & 18.96 & 32.24 & 10.69 & 3comp + ngc 2388 & 1120231 - 001 & 0.014 & 0.49 & 1.99 & 16.21 & 23.09 & 11.15 & 3comp + ngc 4102 & 1120232 - 001 & 0.0028 & 1.45 & 6.83 & 47.37 & 68.50 & 10.25 & 2comp + ngc 4194 & 1122091 - 001 & 0.0083 & 0.86 & 4.36 & 22.79 & 25.94 & 10.88 & 3comp + ngc 4818 & 1120234 - 001 & 0.0036 & 0.85 & 3.88 & 20.26 & 26.55 & 10.10 & 3comp + ngc 520 & 1120229 - 001 & 0.0076 & 0.78 & 2.83 & 31.52 & 48.40 & 10.91 & 3comp + ngc 6285 & 1122138 - 001 & 0.019 & 0.30 & 0.44 & 7.43 & 23.58 & 11.21 & 3comp + ngc 838 & 1122173 - 001 & 0.013 & 0.59 & 1.83 & 0.40 & 17.94 & 10.74 & 3comp + zw 453.062 & 1122140 - 001 & 0.025 & 0.29 & 0.57 & 7.59 & 11.04 & 11.34 & 2comp + + for understanding the properties of buried agns in infrared galaxies , it is crucial to analyze the infrared spectra by decomposing sb and agn components . the 2.55.0 @xmath0 m band covered by _ akari _ irc has unique advantages for detecting agn signatures . as introduced in section 1 , one promising method for finding buried agns is the strength ( equivalent width ) of the 3.3 @xmath0 m pah emission as proposed by @xcite . pure sb galaxies show generally strong 3.3 @xmath0 m pah emission with ew@xmath60 nm because plenty of sb - produced uv photons excite pah molecules @xcite . by contrast , if an agn resides in the galaxy , the pah feature becomes weaker , ew@xmath61 nm @xcite . @xcite also used the red continuum slope ( @xmath27 for @xmath62 ) for finding buried agns . in this paper , we develop the previous works not only to find buried agns but also to quantitatively estimate their energy contribution by performing detailed spectral fitting with a continuum model consisting of a stellar component and dust components from agn and/or sb . emission and absorption line features are also included in the model . the agn emission in the near- to mid - infrared originates from the radiation reprocessed by hot dust in the torus . recent studies of dusty torus models reproduce the observed nuclear mid - infrared emission well by assuming that the torus media have a clumpy structure ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . however , some authors mentioned that there are difficulties in fitting their near - infrared spectral energy distribution ( sed ) @xcite . this is due to the complicated degeneracies including a possible existence of an extra hot - dust ( @xmath63 k ) component originating from the vicinity of the agn @xcite , and/or contamination from the host galaxies , and/or the extinction of the torus emission by interstellar matter in the host galaxies . in this work , we take the simplest approach to approximate the torus dust emission by a single blackbody component , considering our limited wavelength range . presumably , the dusty torus has a continuous dust temperature distribution and produces emission peaked around @xmath510 @xmath0 m . hence , our single blackbody fit should trace its peak temperature at @xmath5300 k. accordingly , we set a conservative upper limit on its temperature as @xmath64 k @xcite . the stellar photospheric emission in the host galaxy contributes to the total radiation in the 2.55.0 @xmath0 m spectral range . this component produces a decreasing continuum flux at @xmath65 @xmath0 m @xcite . we first determine the temperature @xmath66 by fitting the 2mass @xmath67 and @xmath68 photometric data with a blackbody model , and fix the temperature at its best - fit value when analyzing the _ akari _ 2.55.0 @xmath0 m spectra . for mrk 551 and mrk 848 , we can not find any 2mass fluxes . therefore , we use the @xmath67 and @xmath69 photometric data from the literature @xcite instead of the 2mass data . sb produces a large amount of dust , which generally has lower temperatures than those of the agn - heated dust , typically @xmath70 k @xcite . as already suggested by @xcite and many other studies , such cold dust with @xmath71 k does not significantly contribute to the 2.55.0 @xmath0 m spectra , and therefore we ignore the contribution from the sb dust emission . however , some sb galaxies show excess emission from very hot dust with a temparature of @xmath72 k , requring at least two blackbody components . this very hot dust component of @xmath72 k would not originate from the dusty torus , from which blackbody radiation of @xmath5300 k is generally expected ( see section 4.1 . ) . one possible origin is the emission mainly heated by massive stars in hii regions @xcite . if such emission is required from the data , we model it with a single blackbody with a temperature of @xmath73 k in a range of @xmath74 k as suggested by @xcite . lc*8c eso 286-ig19 & @xmath75 & @xmath76 & @xmath77 & @xmath78 & n & y + ic 5135 & @xmath79 & @xmath80 & @xmath77 & @xmath81 & n & y + iras 03068 - 5346 & @xmath82 & @xmath83 & @xmath77 & @xmath84 & n & y + iras 03538 - 6432 & @xmath85 & @xmath86 & @xmath77 & @xmath87 & y & y + iras 10494 + 4424 & @xmath88 & @xmath89 & @xmath77 & @xmath90 & n & y + iras 17028 + 5817 & @xmath91 & @xmath92 & @xmath77 & @xmath93 & n & y + irasf 07353 + 2903 & @xmath94 & @xmath95 & @xmath77 & @xmath96 & y & y + mcg + 02 - 04 - 025 & @xmath97 & @xmath98 & @xmath77 & @xmath99 & n & y + mcg + 08 - 23 - 097 & @xmath100 & @xmath101 & @xmath77 & @xmath102 & n & y + mcg + 10 - 19 - 057 & @xmath103 & @xmath104 & @xmath77 & @xmath105 & n & y + mrk 1490 & @xmath106 & @xmath107 & @xmath108 & @xmath109 & n & n + mrk 551 & @xmath110 & @xmath111 & @xmath77 & @xmath112 & n & y + mrk 848 & @xmath113 & @xmath114 & @xmath77 & @xmath115 & n & y + ngc 2339 & @xmath116 & @xmath117 & @xmath118 & @xmath109 & n & n + ngc 2388 & @xmath119 & @xmath120 & @xmath121 & @xmath109 & n & n + ngc 4102 & @xmath122 & @xmath123 & @xmath77 & @xmath124 & y & y + ngc 4194 & @xmath125 & @xmath126 & @xmath127 & @xmath109 & n & n + ngc 4818 & @xmath128 & @xmath129 & @xmath130 & @xmath109 & n & n + ngc 520 & @xmath131 & @xmath132 & @xmath133 & @xmath109 & n & n + ngc 6285 & @xmath134 & @xmath135 & @xmath136 & @xmath109 & n & n + ngc 838 & @xmath137 & @xmath138 & @xmath139 & @xmath109 & n & n + zw 453.062 & @xmath140 & @xmath141 & @xmath77 & @xmath142 & n & y + + the 2.55.0 @xmath0 m spectra of infrared galaxies contain various emission / absorption line features . the previous studies of the _ akari _ `` agnul '' program @xcite reported three pah lines at 3.3 , 3.4 , and 3.5 @xmath0 m , hydrogen brakett series at 4.05 @xmath0 m br@xmath143 and 2.63 @xmath0 m br@xmath144 , and pfund series at 4.65 @xmath0 m pf@xmath144 and 3.74 @xmath0 m pf@xmath145 as emission lines . they also reported absorption features at 3.1 @xmath0 m by h@xmath146o ice , at 4.29 @xmath0 m by co@xmath146 , and at 4.67 @xmath0 m by co. because pah emission is believed to originate from solid state molecules , the line profile becomes lorentzian @xcite . for simplicity , however , we model all the lines including those of pah by gaussian profiles , following @xcite . the central wavelength and line width are fixed within a spectral resolution at appropriate values reported in the literature . lc*10c eso 286-ig19 & @xmath147 & @xmath148 & @xmath149 & @xmath150 & @xmath151 & @xmath152 & @xmath153 & @xmath154 & @xmath155 & @xmath156 + ic 5135 & @xmath157 & @xmath148 & @xmath158 & @xmath159 & @xmath160 & @xmath161 & @xmath153 & @xmath162 & @xmath163 & @xmath164 + iras 03068 - 5346 & @xmath165 & @xmath148 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath153 & @xmath170 & @xmath171 & @xmath172 + iras 03538 - 6432 & @xmath173 & @xmath148 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & @xmath153 & @xmath178 & @xmath179 & @xmath180 + iras 10494 + 4424 & @xmath181 & @xmath148 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath153 & @xmath186 & @xmath187 & @xmath188 + iras 17028 + 5817 & @xmath189 & @xmath148 & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath153 & @xmath194 & @xmath195 & @xmath196 + irasf 07353 + 2903 & @xmath197 & @xmath148 & @xmath198 & @xmath199 & @xmath200 & @xmath201 & @xmath153 & @xmath202 & @xmath203 & @xmath204 + mcg + 02 - 04 - 025 & @xmath205 & @xmath148 & @xmath206 & @xmath207 & @xmath208 & @xmath209 & @xmath153 & @xmath210 & @xmath211 & @xmath212 + mcg + 08 - 23 - 097 & @xmath213 & @xmath148 & @xmath214 & @xmath215 & @xmath216 & @xmath217 & @xmath153 & @xmath218 & @xmath219 & @xmath220 + mcg + 10 - 19 - 057 & @xmath221 & @xmath148 & @xmath222 & @xmath223 & @xmath224 & @xmath225 & @xmath153 & @xmath226 & @xmath227 & @xmath228 + mrk 1490 & @xmath229 & @xmath230 & @xmath148 & @xmath231 & @xmath232 & @xmath233 & @xmath234 & @xmath153 & @xmath235 & @xmath77 + mrk 551 & @xmath236 & @xmath148 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath153 & @xmath241 & @xmath242 & @xmath243 + mrk 848 & @xmath244 & @xmath148 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath153 & @xmath249 & @xmath250 & @xmath251 + ngc 2339 & @xmath252 & @xmath253 & @xmath148 & @xmath254 & @xmath232 & @xmath255 & @xmath256 & @xmath153 & @xmath257 & @xmath77 + ngc 2388 & @xmath258 & @xmath259 & @xmath148 & @xmath260 & @xmath232 & @xmath261 & @xmath262 & @xmath153 & @xmath263 & @xmath77 + ngc 4102 & @xmath264 & @xmath148 & @xmath265 & @xmath266 & @xmath267 & @xmath268 & @xmath153 & @xmath269 & @xmath270 & @xmath271 + ngc 4194 & @xmath272 & @xmath273 & @xmath148 & @xmath274 & @xmath232 & @xmath275 & @xmath276 & @xmath153 & @xmath277 & @xmath77 + ngc 4818 & @xmath278 & @xmath279 & @xmath148 & @xmath280 & @xmath232 & @xmath281 & @xmath282 & @xmath153 & @xmath283 & @xmath77 + ngc 520 & @xmath284 & @xmath285 & @xmath148 & @xmath286 & @xmath232 & @xmath287 & @xmath288 & @xmath153 & @xmath289 & @xmath77 + ngc 6285 & @xmath259 & @xmath290 & @xmath148 & @xmath291 & @xmath232 & @xmath292 & @xmath293 & @xmath153 & @xmath294 & @xmath77 + ngc 838 & @xmath295 & @xmath296 & @xmath148 & @xmath297 & @xmath232 & @xmath298 & @xmath299 & @xmath153 & @xmath300 & @xmath77 + zw 453.062 & @xmath301 & @xmath148 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath153 & @xmath306 & @xmath307 & @xmath308 + + we thus construct a spectral model consisting of the multiple continuum components over which the emission / absorption lines are superposed . for the continuum we consider three blackbody components , ( 1 ) a direct stellar component @xmath309 , ( 2 ) a dust component from agn with a temperature below 800 k @xmath310 , and ( 3 ) a very hot ( 800 k1200 k ) dust component from hii regions @xmath311 . each blackbody component has two parameters , normalization ( @xmath312 ) and temperature ( @xmath313 ) , which , except for @xmath314 , are left as free parameters in the spectral fit . previous studies suggest that emission from warm dust may be partly absorbed by outer cold dust ( e.g. , * ? ? ? * ; * ? ? ? * ) . to take this into account , we apply an extinction correction by incorporating @xmath315 into the blackbody components . we adopt the dust extinction index of @xmath316 @xcite as @xmath317 the normalization value @xmath318 can be estimated from the optical thickness of 3.1 @xmath0 m ice ( @xmath319 ) , which is calculated by using the deconvolved continuum from the dust ( @xmath320 ) . @xcite estimated the relationship between the dust extinction @xmath321 and @xmath319 with @xmath322 where @xmath323 is the fraction of dust that is covered with an ice mantle . the maximum value of @xmath323 is derived to be 0.3 in the core of u / lirgs in @xcite . from the relation between the optical thickness and extinction with @xmath324 and eq . ( 2 ) with @xmath325 , @xmath326 is derived . by combining eq . ( 1 ) and ( 3 ) , at the @xmath327 band ( 0.55 @xmath0 m ) , @xmath328 accordingly , we model ths observed component as @xmath329 , where @xmath330 . first , we adopt only two continuum components of direct stellar emission and agn / sb dust emission to fit the observed spectra , because the contribution of dust emission from hii regions is not always required from the data . this model ( 2 component model ) is written as : @xmath331 where @xmath332 represents each line s profile as described in section 4.3 . in some cases , mainly for sources showing little contribution from @xmath333 , the 2-component model given by equation ( 6 ) does not fit the data well . then , we add the @xmath334 component to the above model : @xmath335 we adopt this model ( 3 component model ) only when the fitting result is significantly improved from the 2-component model with a @xmath336% confidence level on the basis of an f - test . the fitting model we adopt for each source is summarized in table 1 . as shown in figure 2 , all sources that require the three components are galaxies showing blue contina in their spectra ( e.g. , ngc 4818 ) . as many authors ( e.g. , * ? ? ? * ; * ? ? ? * ) suggested , these galaxies do not show buried agn signs but are more likely normal sb galaxies . this is consistent with our assumption that the extra hot ( @xmath337 k ) component should originate from hii regions , not from the agn torus . + [ tbp ] figure 2 displays the observed _ akari _ spectra of all the 22 sources ( red dots with flux error bars ) overplotted with their best - fit models ( blue line ) . each line / continuum component is also plotted in the figure . as noticed , a majority of sources show 3.3 @xmath0 m pah emission features . table 2 and 3 summarize the fitting results . for finding buried agns , we apply two agn diagnostics : ( 1 ) pah emission line and ( 2 ) torus - dust continuum . the pah diagnostic is based on ew@xmath20 as discussed in section 4 ; we identify a buried agn if it shows a small 3.3 @xmath0 m pah equivalent width with ew@xmath61 nm . the torus - dust diagnostic is based on whether or not there is a contribution from a hot dust component from the torus ; we set the criterion of identifying buried agn as @xmath338 k. in the left panel of figure 3 , we present the two agn diagnostic diagram applied for our sample , where ulirgs / lirgs / irgs are plotted as pink circles / red triangles / brown squares , respectively . there is a clear boundary of distribution of the dust temperature around @xmath339 k. this supports our criterion that agns should have high dust temperatures ( @xmath338 k ) and are well distinguishable from sbs . interestingly , many infrared galaxies with large pah equivalent widths ( ew@xmath340 nm ) show high dust temperatures ( @xmath341 k ) . one reason could be that because of the large aperture of _ akari_/irc and less luminous nucleus emission in these galaxies , the spectra contain relatively large contribution of 3.3 @xmath0 m pah emission in the host galaxies where pah molecules are not destroyed by agn x - ray photons . this effect works to increase the pah equivalent width , and hence the pah diagnostic could miss buried agns . by contrast , the hot - dust diagnostic is not subject to the aperture effect because the continuum emission can be properly decomposed by the spectral fit . in fact , the torus - dust agn diagnostic can recover all the buried agns identified by the pah diagnostic ; in other words , there is no source with a low dust temperature ( @xmath342 k ) _ and _ a small pah equivalent width ( ew@xmath61 nm ) in our sample . this supports the superiority of using the hot - dust diagnostic for finding agns completely . + the right panel of figure 3 shows the plot of ew@xmath20 versus dust luminosity ( @xmath343 ) tabulated in table 2 and table 3 . in the figure , we plot only the sources with @xmath344 k. this is because the 2.55.0 @xmath0 m spectral range is not sensitive to the dust emission with low temperature @xmath345 k. in this case , only an upper limit of the dust luminosity can be derived . this is the reason why the dust luminosity with @xmath346 k is very small , @xmath347 erg s@xmath35 . the figure shows that all the sources with @xmath348 k have significant dust emission with @xmath349 erg s@xmath35 . thus , the alternative criterion for detecting agns can be @xmath350 erg s@xmath35 instead of @xmath348 k . in summary , we conclude that the hot - dust agn diagnostic is a better method for finding buried agns in the _ akari _ irc band . + one motivation of our study is to reveal how frequently buried agns exist in infrared galaxies . figure 4 shows the fraction of buried agn as a function of infrared luminosity . the left and right panels show the results based on the pah agn diagnostic and on the torus - dust agn diagnostic , respectively . as discussed above , the latter diagnostic is more complete than the former for finding buried agns from infrared galaxies . therefore , we discuss the result of the right panel hereafter . in total , we detect 14 buried agns out of the 22 infrared galaxies without any agn signs in the optical band . we also find a clear trend that the fraction of buried agn increases with infrared luminosity . while only 17% ( 1 out of 6 sources ) of the irgs contain buried agns , the u / lirgs contain them almost ubiquitously ( @xmath351% for lirgs and @xmath352% for ulirgs ) . our results for u / lirgs are consistent with those obtained by @xcite within the statistical uncertainties . although most of the irgs are selected from the 12 @xmath0 m galaxy catalog , which is sensitive to torus - dust emission , an agn is detected only from one source out of the five targets . the high agn fraction in u / lirgs may be related with their high merger rates . there is strong observational evidence that most of u / lirgs have experienced merging or strong galaxy - galaxy interactions @xcite . the mergers can induce strong sb within the galaxies , which also make galactic gas fall into the galactic center and trigger agn activity obscured by dust @xcite . @xcite show that infrared galaxies with signs of recent merger drastically increase at @xmath353 toward a higher luminosity range . this result is in accordance with our finding that the buried agn fraction drastically increases from @xmath354 . as a function of infrared luminosity . open circles / triangles / squares represent ulirgs / lirgs / irgs with agn signs , respectively . filled triangle / square represents lirgs / irgs without agn signs , respectively . note that all ulirgs in our sample are diagnosed to have buried agns . [ fig-5],title="fig : " ] + we confirm the dependence of the agn fraction on luminosity by investigating the ratio between the 3.3 @xmath0 m pah luminosity ( @xmath355 ) and total infrared luminosity @xmath356 . pure sb galaxies generally have a constant ratio @xmath357 . if an agn exists inside the galaxy , the ratio becomes smaller because x - rays from the agn destroy pah molecules and the infrared continuum luminosity is increased by the torus - dust emission . figure 5 plots @xmath358 as a function of infrared luminosity for our sample . as noted , it decreases with infrared luminosity , especially at @xmath359 . we also find that the sources with small @xmath360 ratios match to those with buried agn signs , supporting the above idea . note that there is a possibility that this trend may be caused by the destruction of small pah molecules by major mergers @xcite rather than by agns . nevertheless , since all the sources with @xmath361 show buried agn signs , the conservative criterion @xmath362 could be phenomenologically used as another diagnostic for finding buried agns . our second motivation is to determine the energetic importance of agns in these infrared galaxies . as discussed in section 5.1 , the luminosity contribution from dust blackbody component ( @xmath363 ) predominantly originates from agn - heated dust , and the luminosity contribution of sb - heated dust to @xmath363 is negligible . thus , using the best - fit blackbody parameters obtained by the spectral fit , we can estimate the total luminosity from agn - heated dust and hence its ratio to the total infrared luminosity . figure 6 shows the averaged fraction of agn power to the total infrared luminosity for irgs , lirgs , and ulirgs . as noticed , the agn contribution increases with infrared luminosity , while the values are quite small , @xmath364% in irgs , @xmath365% in lirgs , and @xmath366% in ulirgs . this suggests that the bulk of the infrared emission originates from sb , not from agns in these galaxies . one caveat for this result is that we can not completely exclude the possibility that the estimated luminosity of @xmath367 might be partially originated from the torus - dust if it has an extreme temperature distribution . we thus calculate the averaged @xmath368 ratio to check if it could have a significant contribution . we find @xmath369% , suggesting that the main results above would not be affected , even if the observed very hot ( @xmath370 k ) dust component were totally due to the agn emission . another caveat for interpreting this result is the selection bias for our sample , which consists of optically non - seyfert galaxies . previous studies based on samples including optically seyfert agns found larger average agn contribution to the total infrared luminosity , 1015% in lirgs @xcite and 1540% in ulirgs @xcite . our results are in good agreement with those by @xcite within the errors , who obtained the agn to total luminosity ratio of 68% in lirgs and 1119% in ulirgs , using a similar sample to ours that preferentially includes optically non - seyfert infrared galaxies . they calculated the agn contribution by performing sed fit with the decompir package @xcite . this method requires infrared sed data covering the far - infrared band , while our method only uses the _ akari _ 2.55 @xmath0 m spectra . this supports that our simple assumption in the spectral model ( ie , single blackbody for dust emission ) works well in estimating the agn luminosity . this simple method will have a great advantage in the era of _ jwst _ , because we can apply our method to high-@xmath371 galaxies by observing the rest 2.55.0 @xmath0 m band ( 5.010.0 @xmath0 m at @xmath372 and 7.515.0 @xmath0 m at @xmath373 ) with _ jwst_/miri ( 528 @xmath0 m ) . other methods based on the _ spitzer _ bandpass ( 535 @xmath0 m ) would have a difficulty in studying high-@xmath371 objects ; for instance , the observed spectral range corresponds to 15105 @xmath0 m at @xmath373 , which only the _ spica _ mission can cover . thus , our method is achievable for studying infrared galaxies in the distant universe even before the launch of _ spica _ ( from 2025 ) . , title="fig : " ] + combining our result with the total infrared luminosity density produced by u / lirgs in the local universe ( @xmath374 @xmath375/mpc@xmath376 and @xmath377 @xmath375/mpc@xmath376 ) @xcite , we estimate the local agn luminosity density in the infrared band ( @xmath378 ) . we find @xmath379/mpc@xmath380 for lirgs and @xmath381/mpc@xmath380 for ulirgs . these values are about an order of magnitude lower than the estimate by @xcite . this is because @xcite simply assumed that the infrared luminosity of galaxies classified as agns entirely originates from agns , while we quantitatively take into account the fraction of agn contribution to the total infrared luminosity of infrared galaxies . figure 7 shows the infrared luminosity density as a function of redshift derived by @xcite together with our results of agn the infrared luminosity density . in the era of _ jwst _ , we can fill in the agn infrared luminosity density at high redshifts thanks to their improved sensitivity in the mid - infrared band . ) of all galaxies , pink filled circles that of ulirgs ( @xmath382 ) , and red filled triangles that of lirgs ( @xmath383 ) , taken from @xcite . pale - pink filled circles and brown filled triangles represent the agn infrared luminosity density of ulirgs and that of lirgs estimated by our work . [ fig-7],title="fig : " ] + comparison between the starburst and agn activity gives crucial information to understand the process of galaxy - smbh co - evolution . for optically selected seyfert galaxies and qsos , many authors found that the luminosity of star formation ( @xmath384 ) and that of agn ( @xmath385 ) are well correlated with each other @xcite . we can derive a sf luminosity as @xmath386 . the agn - heated dust luminosity ( @xmath363 ) summarized in table 3 can be converted to an agn bolometric luminosity ( @xmath387 ) by using two relations given in @xcite and @xcite , respectively , @xmath388 here @xmath389 represents the intrinsic luminosity of the accretion disk integrated in the optical to x - ray band . assuming @xmath390 , we obtain @xmath391 noted that the estimated bolometric agn luminosity ( @xmath392 ) derived in equation ( 10 ) can be overestimated . the relations from equation ( 8) to ( 10 ) are derived from the sample of non - hidden seyfert agns , while our sources are infrared galaxies with highly obscured ( = hidden ) agns . therefore , our sources in this study could have higher covering fraction of dusty - torus than those of seyfert agns . this means the ratio of mid - infrared over bolometric luminosity in our sources could be larger than those of seyfert agns . figure 8 shows the @xmath384 versus @xmath392 relations obtained from our sample . the average luminosities of u / lirg are also plotted in the same figure . the dotted line corresponds to the linear correlation obtained from optically selected seyfert galaxies obtained by @xcite . interestingly , all the sources except two ( iras f07353 + 2903 and eso 286-ig19 ) are located above the seyfert line . this result may reflect the difference in the evolutional stage of galaxies , from a pure sb phase ( u / lirgs ) to an unburied agn phase ( seyferts ) . as discussed in section 5.2 , u / lirgs have recently experienced or are facing mergers , which result in strong sb activity from the epoch @xmath393 to @xmath394 in figure 8 . the sb also triggers agn activity , which continues from @xmath394 to @xmath395 . finally , the sb activity is gradually weakened due to the shortage of gas and feedback from the agn , and the supply of matter from the host galaxy onto the central engine also decreases , from @xmath395 to @xmath396 . in this framework of galaxy evolution , the evolutional stage of ulirgs should be a later phase of mergers than that of lirgs . the picture is in good agreement with observations that the morphology of ulirgs are more compact than that of lirgs , suggesting that ulirgs have just finished the merger process while that of lirgs is still on - going . this scenario suggests that infrared galaxies containing buried agns could be an earlier evolutional stage of agn , which will be evolved to normal ( unburied ) seyferts / qsos . this scenario becomes much more reasonable if our bolometric agn luminosity ( @xmath392 ) is overestimated as discussed in the previous paragraph because all the points are shifted to the left in the figure . while our sample size is still small , the new agn diagnostics developed in this paper can be applied to other sources . further studies using a well - defined complete sample of infrared galaxies should be urged to complete our view of agn evolution . we obtained the _ akari _ 2.55.0 @xmath0 m spectra of 22 infrared galaxies at @xmath397 selected mainly from three infrared catalogs : bright ulirg catalog @xcite , _ iras _ 12 @xmath0 m galaxy catalog @xcite , and _ iras _ bright galaxy catalog @xcite . this band includes unique spectral features for finding `` buried '' agns , the 3.3 @xmath0 m pah emission line and red continuum spectra originated from agn torus emission . we perform detailed spectral fitting to these data by properly modeling the continuum and emission / absorption features . this enables us to decompose the continuum spectra into sb and agn torus - dust components . the results are summarized as follows . 1 . using the agn diagnostics based on the torus - dust continuum ( @xmath398 k or @xmath399 erg s@xmath35 ) and 3.3 @xmath0 m pah emission strength ( ew@xmath61 nm ) , we find that 14 out of the 22 infrared galaxies have buried agns . we confirm the trend that the buried agn fraction increases with infrared luminosity : @xmath400% for normal irgs , @xmath401% for lirgs , and @xmath402% for ulirgs in our sample . this suggests that the presence buried agns is ubiquitous in u / lirgs . the ratio @xmath358 decreases with infrared luminosity . this is opposed to the previous reports that @xmath403 have a constant value @xcite . this decrease is possibly due to three effects : ( 1 ) destruction of pah molecules by x - rays emitted from the agn , ( 2 ) pah destruction by merger processes , and ( 3 ) increase of agn contribution to the total infrared luminosity . the energy contribution from the agn torus - dust emission to the total infrared luminosity also increases with infrared luminosity , but it only reaches @xmath57% in lirgs and @xmath520% even in ulirgs . this suggests that majority of the total infrared luminosity in u / lirgs originates from sb , _ not _ from agn . 4 . combining the above results with the luminosity function of infrared galaxies derived by @xcite , we estimate the agn luminosity density of @xmath404/mpc@xmath380 for lirgs and @xmath405/mpc@xmath380 for ulirgs in the local universe . @xmath384 versus @xmath392 luminosity plot for our sample shows that infrared galaxies with buried agns are located above the luminosity correlation line derived from optically - selected seyfert galaxies . we interpret that infrared galaxies could be an early phase of the evolutional track of agns . we are grateful to kenichi yano and rika yamada for their help with the data reduction of _ we also thank claudio ricci and kenta matsuoka for discussions . the research is based on observations with _ akari _ , a jaxa project with the participation of esa . this work was partly supported by the grant - in - aid for jsps fellows for young researchers ( k.i . ) and for scientific research 23540273 ( m.i . ) and 26400228 ( y.u . ) .

since the first optical counterpart of gamma - ray burst ( grb ) was identified on 1997 february 28 , more and more grbs have been optically detected , especially after _ swift _ was launched successfully , which allows early follow - up observations with ground - based optical telescopes . however , about 40 per cent of _ swift _ grbs still failed to be identified with optical or infrared counterparts ( burrows et al . 2008 ) . for the nature of the dark burst , several models have been proposed : extinction from dusty opaque star - forming regions ( e.g. groot et al . 1998 ; reichart & price 2002 ) ; high redshift , which diminishes the grb optical luminosity ( lamb & reichart 2000 ) , as the ly@xmath6 forest is shifted into the observational energy range ; intrinsically dim bursts ( fynbo et al . 2001 ; rol et al . 2005 ) ; the observational effect , as a result of the late response to the bursts . jakobsson et al . ( 2004 ) have proposed an operational definition of dark burst ( i.e. optical - to - x - ray spectral index @xmath7 ) , according to the fireball model . however , grbs with an optical afterglow have been studied widely and in depth(e.g . grb 060218 , grb 080319b ) . this has resulted in an increasing number of grbs with known redshift , which has allowed statistical studies to be carried out . the optical luminosity distribution of grbs varies widely up to several magnitudes ( kann et al . 2006 ; 2008 ) . after being transformed into a common redshift ( e.g. , @xmath8 ) , the optical luminosity shows a bi - model phenomenon ( nardini et al . 2006 ; liang et al . 2006 ; kann et al . 2006 ; 2008 ) : optically luminous and optically dim . the bi - model distribution is also found to exist in other energy bands ( gendre et al . 2008a , b ) . liang & zhang ( 2006 ) found that the apparent bimodality can not be interpreted as a manifestation of the extinction effect , and they suggested that there might be two types of progenitors or two types of explosion mechanisms in operation . considering the dark burst is usually a type of grbs with a faint optical afterglow , optically dim bursts might be related to dark bursts ( nardini et al . however , the number of well - studied optically dim bursts is very limited . thus , increasing the number of known optically dim bursts may help us to understand the origin of the bi - model distribution , the nature of optically dim bursts and dark bursts . one such interesting case is grb 070518 . it was detected by gamma - ray burst alert telescope ( bat ) onboard _ swift _ at 14:26:21 ( ut ) on 2007 may 18 . the x - ray telescope ( xrt ) and ultraviolent / optical telescope ( uvot ) on board _ swift _ observed the counterpart beginning at 70 and 100 sec after the trigger , respectively . t@xmath9 ( 15 - 350 kev ) is 5.5@xmath100.2 s , t@xmath11 is 2.9s ( sakamoto et al . the values of duration mean the burst belongs to the short tail of the long group ( kouveliotou et al . thus , it might be a classical long - duration grb or an intermediate - long grb ( horvath et al . 2008 ) . the ratio of fluence between the 25 - 50 kev and 50 - 100 kev band ( sakamoto et al . 2008a ) is about 1.06 , which makes grb 070518 belong to an x - ray rich burst ( xrr ) , according to the criterion0.72 c - grb ; + 0.72@xmath12s(25 - 50kev)/s(50 - 100kev)@xmath131.31 xrr ; + s(25 - 50kev)/s(50 - 100kev)@xmath141.32 xrf . ] given by sakamoto et al . ( 2008b ) . the first given magnitude and coordinates of the counterpart are about 18 mag in @xmath15 band reported by uvot and ra(j2000 ) = 16h 56 m 47.7s , dec(j2000 ) = 55d 17 m 42.3s ( radius , 90 pen cent containment ) , respectively ( guidorzi et al . 2007 ) . the afterglow was also observed by other ground - based telescopes later . besides , the redshift of grb 070518 was reported to be lower than 0.7 ( cucchiara et al . 2007 ) based on the photometry of the _ swift _ uvot . no other report about redshift of grb 070518 has been presented in the literature up to now . in our analysis , the redshift of 0.7 are used to investigate the properties of grb 070518 . in this paper , we report on optical photometric follow - up observations of grb 070518 at xinglong observatory of national astronomical observatories , chinese academy of sciences . afterglow observations and results are described in @xmath162 . an anaysis and discussion are presented in @xmath163 , and a summary and conclusion are given in @xmath164 . in our calculations , a flat universe is assumed with matter density @xmath17 , cosmological constant @xmath18 , and hubble constant @xmath19 km s@xmath20 mpc@xmath20 . the formula @xmath21@xmath22 is used for the afterglow decay analysis . we carried out a follow - up observation programme of _ swift _ grbs with the 0.8-m tsinghua university - national astronomical observatory of china optical telescope ( tnt ) at xinglong observatory , under the framework of the east - asia follow - up observation network ( eafon ; urata et al . 2003;urata et al . the tnt telescope is equipped with a pi 1300@xmath231340 ccd and filters in the standard johnson - bessel system . the field of view of the tnt is 11.4@xmath2311.4 arcmin , yielding 0.5 arcsec of the pixel scale . a custom - designed automation system has been developed for the grb follow - up observations ( zheng et al . 2008 ) . the follow - up observations of grb 070518 were performed with the tnt from 512s after the initial _ swift _ trigger , and lasted for about 5.5 hours until the dawn time . a sequence of @xmath15 , @xmath5 , @xmath24 , and @xmath25 bands images was obtained . after preliminary analysis , we first confirmed the counterpart of grb 070518 ( xin et al . 2007 ) after the report of _ swift _ uvot . data reduction was carried out following standard routine in iraf package , including bias and flat - field corrections . dark correction was not performed , as the temperature of our ccd was cooled down to @xmath26 . point spread function ( psf ) photometry was applied via daophot task in iraf package to obtain the instrumental magnitudes . during the reduction , some frames were combined in order to increase signal - to - noise ( s / n ) ratio . flux calibration was performed by re - observing the field of grb 070518 , together with landolt photometric standard stars with the tnt telescope on 2007 november 14 and december 12 . considering that the @xmath15 band is not a standard band , we treated the @xmath15-band filter as a very broad filter . when we calibrated the @xmath15 band data to @xmath5 band magnitude , no significant difference was found in the colour terms , and the response functions of the two filters were very similar with about 0.07 mag uncertainty . since the redshift of grb 070518 is no larger than 1 , the @xmath15 band data was calibrated to @xmath5 band magnitude , for simplicity . the transformation uncertainty of 0.07 mag is added to the final results . in addition , other optical photometries for grb 070518 are collected from the grb coordinate network ( gcn ) , in order to complete the optical light curves . all of these optical data are presented in table 1 . ccccccc t - t0 & exposure & band & mag & telescope + ( mid day ) & ( sec ) & & & or reference + 0.00636 & 2@xmath2320 & @xmath27 & 19.03@xmath100.10 & tnt + 0.00699 & 2@xmath2320 & @xmath27 & 19.23@xmath100.10 & tnt + 0.00804 & 3@xmath2320 & @xmath27 & 19.37@xmath100.12 & tnt + 0.00883 & 3@xmath2320 & @xmath27 & 19.27@xmath100.08 & tnt + 0.00961 & 3@xmath2320 & @xmath27 & 19.68@xmath100.14 & tnt + 0.01053 & 4@xmath2320 & @xmath27 & 19.56@xmath100.12 & tnt + 0.01157 & 4@xmath2320 & @xmath27 & 19.74@xmath100.13 & tnt + 0.01275 & 5@xmath2320 & @xmath27 & 20.08@xmath100.15 & tnt + 0.01406 & 5@xmath2320 & @xmath27 & 20.06@xmath100.18 & tnt + 0.02446 & 1@xmath23300 & @xmath5 & 20.20@xmath100.10 & tnt + 0.02814 & 1@xmath23300 & @xmath5 & 20.76@xmath100.16 & tnt + 0.03723 & 3@xmath23300 & @xmath5 & 20.48@xmath100.10 & tnt + 0.01237 & - & @xmath5 & 19.5 & gcn6443 + 0.23337 & - & @xmath5 & 22 & gcn6483 + 0.35948 & - & @xmath5 & 22.2 & gcn6418 + 0.37753 & - & @xmath28 & 22.3@xmath100.2 & gcn6458 + 0.71837 & - & @xmath5 & 23.03@xmath100.05 & gcn6462 + 1.45833 & - & @xmath5 & 23.3 & gcn6426 + 1.70837 & - & @xmath5 & 23.36@xmath100.05 & gcn6462 + 3.60837 & - & @xmath5 & 23.56@xmath100.05 & gcn6462 + 0.02065 & 1@xmath23300 & @xmath24 & 20.74@xmath100.15 & tnt + 0.06383 & 10@xmath23300 & @xmath25 & 20.47@xmath100.17 & tnt + 0.10197 & 10@xmath23300 & @xmath25 & 20.92@xmath100.13 & tnt + 0.15271 & 14@xmath23300 & @xmath25 & 21.37@xmath100.14 & tnt + 0.20660 & 15@xmath23300 & @xmath25 & 21.23@xmath100.24 & tnt + the multiband light curves of grb 070518 are shown in fig . galactic extinction ( schlegel et al . 1998 ) is corrected for all @xmath24,@xmath5 and @xmath25 bands data , which corresponds to a reddening of @xmath29 mag in the direction of the burst . the @xmath5- band data could be corrected for the flux contamination of its host galaxy , whose magnitude is estimated to be @xmath30 mag ( garnavich et al . 2007b ) . the contamination is also interpreted as a combination of its host galaxy and supernova bump ( dai et al . 2008 ) . as shown in fig . 1 , the @xmath5 band light curve corrected for the flux contamination is labeled with `` after '' , while the @xmath5 band light curve without the correction is labeled with `` before '' . the tnt data before 10@xmath31 s are not affected by the contamination very much . moreover , the `` after '' @xmath5 band light curve of whole data ( including tnt data and gcn data ) could be well fitted with a simple power law with a decay index of @xmath32 @xmath33 . the result is consistent with the fitting index of @xmath00.81@xmath100.08 @xmath34 only for tnt data . the general trends of light curves for the @xmath5 ( labeled `` before '' ) and @xmath25 bands are similar ; however , the fluctuations exist in both light curves , which might be a result of a low s / n ratio . all @xmath5 , @xmath24 and @xmath25 band data have been corrected for the galactic extinction . solid and open triangles present the @xmath5 band data before and after the correction of the flux contamination of the host galaxy , respectively . the insert shows the flares in the x - ray light curve at the beginning . , scaledwidth=50.0% ] in order to investigate the decay properties between optical and x - ray light curves , we use the x - ray afterglow data downloaded from the website of the uk _ swift _ science data center(labeled as uss ) ( evans et al . 2007 ) , where some analysis results are given , including the spectral indices for each segments in the x - ray light curve . these spectral indices are used directly for our later analysis . the x - ray light curve is also plotted by squares in fig . 1 , showing several drastic flares from the beginning up to about 250s after the burst trigger . considering the sparse data between 2000 and @xmath35 sec , we fit the light curve after 250 sec by separating it into two parts . one is between 250 sec and 2000 sec , and the other is from @xmath35 sec to over @xmath36 sec . the first part is well fitted by a single power law with an index of @xmath37 ( @xmath38 ) , while the second is well fitted by a single power law with a slope of @xmath39 ( @xmath40 ) . as shown in fig . 1 , the extrapolation of the first fitting is below the data detected after 1000 sec , suggesting that there is energy injection ( zhang et al . 2006 ) between the two segments , shown as a shallow decay . therefore , the x - ray light curve displays a `` steep - shallow - normal '' decays ( zhang et al . 2006 ) . the first steep decay in the x - ray afterglow light curve is usually explained by the curvature effect ( e.g. kumar & panaitescu 2000 ; zhang et al . according to their model , temporal index @xmath6 and spectral index @xmath41 have a tight relation @xmath42 . for grb 070518 , the first slope @xmath43 ( 1.78 ) of the x - ray light curve is definitely lower than the value ( 3.22 , from @xmath44 ) expected for high - latitude emission ( here , @xmath45 from uss ) . however , the high - latitude effect only provides an upper limit to the decay slope , as it assumes that the shell emission stops abruptly after the initial pulse ( mangano et al . the decline could be made in two cases : one is that residual emission from the shocked shells is still present , in which , the high - latitude emission would contribute only a small fraction of the overall radiation ; the other is that the x - ray band is below the cooling frequency , in which case , the expected decay slope of high - latitude radiation would naturally be shallower than 2 . for grb 070518 , the first possibility might be more reasonable , from the analysis for other segments of the x - ray light curve in the later sections . the x - ray afterglow is analyzed by applying the closure relations given by zhang et al . ( 2006 ) to the x - ray spectral and temporal indices . these relations help us to determine the location of the observed x - ray band relative to the synchrotron frequencies ( @xmath46,@xmath47,@xmath48 ) and the environment where the burst occurred . the normal decay segment with @xmath49 and @xmath50 , agrees with the non - injected closure relations : @xmath51 and @xmath52 . these relations are consistent with two cases : when @xmath53 in fast cooling phase or when @xmath54 in slow cooling phase . also , the environment condition [ interstellar medium ( ism ) , wind ] could not be distinguished for the above cases . the application of the closure relations to the x - ray temporal and spectral index in normal decay segment , combining the analysis in @xmath16 3.1 for the x - ray afterglow during 2000@xmath02@xmath55sec , suggests that there is energy injection before the start of the normal decay phase . the x - ray light curve of grb 070518 shows drastic flares between 80 and 200s since the trigger , as shown in the insert in the fig . 1 . several grbs were reported to have flares in x - ray afterglows ( e.g. bernardini et al . , 2009 ; krimm et al . , 2007b ) . with the method of krimm et al . ( 2007b ) , the ratios of raise time duration and the middle time for two big flares ( peak times of about 106.9 s and 189 s , respectively ) are calculated . @xmath56t / t for two flares was about 0.2 and 0.1 , respectively , and the @xmath56f / f was about 1 . such steep and large amplitude rises in the observed flux , as in grb 070518 , could not be explained by a sudden increase in the external density ( nakar & granot 2007 ) . another model , a patchy shell model , can produce the flare with @xmath56f / f of about 1 ( nakar & oren 2004 ) , athough it can not interpret the flares with @xmath56t / t @xmath57 0.1 . for two big flares of grb 070518 , @xmath56t / t are about 0.2 and 0.1,respectively . therefore , the model might be possible . moreover , such dramatic flares are also related to refreshed shocks or the central engine activity ( e.g. zhang et al . this is more reasonable for the case of grb 070518 , if we consider the above analysis that energy injection was still occurred after these flares . chromatic decay could be seen in the case of grb 070518 , as shown in fig . 1 . the x - ray afterglow shows a `` steep - shallow - normal '' decay ( zhang et al . 2005 ) , while optical afterglow is decaying with a constant index of 0.87 ( if the correct of the late contamination was right ) . the decay index of the optical light curve ( @xmath58=0.87 ) is slightly smaller than that of the normal decay segment ( 0.99 ) of the x - ray afterglow . the normal decay segment is usually explained the external shock model ( sari et al . 1998 ) , as expected for the optical afterglow . however , the origin of the normal phase might also be related to the internal energy dissipation ( see as the discussion by urata et al . thus , the classical external shock model will tested with the burst . according to the model , the difference between the decay indices of the optical and x - ray afterglow should be between -0.25 for the uniform ism case and 0.25 for the wind medium case ( urata et al . 2007 ) . for grb 070518 , @xmath59 , ignoring the uncertainties of those values . the case of grb 070518 agrees with the prediction of the classical external shock model . moreover , with the relation derived by urata et al . ( 2007 ) : @xmath60 , where s shows the condition of ambient matter density as a formate of @xmath61 , the value of s for grb 070518 could be deduced to be 0.84 . this suggests that the surrounding matter of the burst is a mix of ism and wind - type medium . the time - averaged spectrum of grb 070518 from t-1.8s to t+4.5s is best fitted by a simple power - law model with a photon index @xmath62 ( krimm et al . 2007a ; guidorzi et al . 2007 ) , corresponding to the energy spectral index of 1.11@xmath100.25 ( @xmath63 ) . the fluence in the 15 - 150 kev band is 1.6@xmath100.2@xmath2310@xmath64 erg cm @xmath65 ( krimm et al . 2007a ) , which is lower than the average value of the _ swift _ burst ( sakamoto et al . 2008a ) . 2 shows the relation between @xmath66 in the rest frame and @xmath67 ( i.e. amati relation ; amati et al . the data shown by open symbols and the best fitting parameters are obtained from amati et al . ( 2008 ) . for grb 070518 , spikes in the bat light curve reported by guidorzi et al . ( 2007 ) were only clearly seen under 50 kev , which implies that @xmath68 might be less than 50 kev . besides , if @xmath68 is within the bat energy range , the photon index @xmath69 of a simple power - law fit and @xmath68 are well correlated with a relationship ( liang & zhang 2005 ) : @xmath70 + sakamoto et al.(2009 ) also obtained a similar conclusion by combining the simulation : @xmath71 + with @xmath72 and the two relations above , @xmath73 and 30.8 are obtained , respectively . the mean value of @xmath0 35 kev is taken from the above calculations as its peak energy in the observational frame . thus @xmath74 kev in the rest frame . as a result , grb 070518 is consistent with amati relation within 3 @xmath75 uncertainties , as shown in fig . 2 . and e@xmath76 ( the amati relation ) . the solid triangle is shown for grb 070518 when @xmath77 and e@xmath78 kev . e@xmath76 is calculated using the band function ( band et al . 1993 ) , assuming photon index @xmath79 of the simple power law is the higher photon index , and fixing @xmath80 as the lower photon index . the data shown as open symbols are taken from amati et al . ( 2008 ) . , scaledwidth=50.0% ] ghirlanda , ghisellini & lazzati ( 2004 ) proposed a relation between peak energy @xmath66 in the rest frame and the collimation - corrected energy @xmath81 : @xmath82 kev , where @xmath83ergs and @xmath84 . in this relation , the jet angle @xmath85 is very important , which was given by sari , piran & halpern ( 1999 ) : + @xmath86 , + where z is the redshift , @xmath87 is the break time in days . for grb 070518 , the time of a possible jet break should be later than the last observation of x - ray afterglow , @xmath88 sec , according to the analysis of @xmath16 3.1 . thus , the opening angle of jet @xmath85 should be larger than 25 degree , under an assumption of the typical values : @xmath89 @xmath90 and @xmath91 ( ghirlanda et al . from the values of @xmath85 and e@xmath92 , @xmath81 is estimated to be larger than 6@xmath93 ergs . however , @xmath66 has been estimated to be about 60 kev in the observation frame , as analyzed in @xmath163.2.1 . the @xmath66 gives an @xmath81 of about 5.1@xmath93 , based on the ghirlanda relation . this value is slightly less than that from the above analysis ( 6@xmath93 ergs ) , but the values are consistent with each other , considering all the assumptions we have made . therefore , grb 070518 might agree with ghirlanda relation from above analysis . nardini et al.(2006 ) found that grb optical luminosity at 12 hours ( at rest - frame ) after the trigger shows a bimodal distribution . with a redshift of 0.7 for grb 070518 , 12 h at the rest frame corresponds to 20.4 hours in the observer frame . following the method of nardini et al . ( 2006 ) , and assuming @xmath94 , @xmath95 erg s@xmath20 is obtained based on the redshift of 0.7 , and @xmath96 . the low value of @xmath97 of grb 070518 means that the burst is one of the optically dimmest grb afterglows so far ( nardini et al . 2008 ) . we have also noted that the result might be affected by the subtraction of its contamination at late phase . in order to investigate this , we just calculate the optical luminosity at the rest frame with the data without correction of the contamination of its possible host galaxy . in this case , @xmath98 erg s@xmath20 is obtained with the assumptions as above , indicating that the burst is really an optically dim burst . moreover , if the real redshift of the burst was lower than 0.7 , its luminosity would become much lower than the value given above . another method to investigate the optically dim luminosity of grb 070518 is to compare its light curve with other long bursts directly . kann et al . ( 2007 ) have collected grbs in the pre-_swift _ and _ swift _ era to compare the luminosity of grbs in the common redshift of 1 . in this sample , grb 050416a , grb 060512 and grb 070419a are at the lower edge of the distribution of optical light curves ( seen in the figure 1 of kann et al . 2008 ) . in order to make the comparison , we have collected @xmath5 band data of the above three dim bursts from gcn circular and other literature . these data are transformed into a common redshift @xmath8 , with @xmath99 for grb 050416a ( soderberg et al . 2007 ) and fixed @xmath100 for other bursts , grb 070518 , grb 060512 and grb 070419a . in the transformation , all magnitudes are corrected for the galactic extinction , and without considering the extinction from their host galaxies . in addition , for grb 070518 , flux contribution of host galaxy is corrected . as shown in fig.3 , the luminosity of grb 070518 is close to that of the three bursts , based on the redshift of 0.7 . therefore , grb 070518 should belong to optically dim bursts , like the other three bursts ( e.g. grb 050416a , grb 060512 and grb 070419a ) . band luminosity of grb 070518 with three other long bursts after transformation into the same redshift of @xmath8 . the comparison long bursts are grb 050416a , grb 060512a and grb 070419a , as labeled in the plot . grb 070518 was shown as a solid line based on the redshift of @xmath77 . the reference for data are as follows : grb 050416a ( soderberg et al . 2007 ) ; grb 060512a ( mundell et al . gcn5118 ; malesani et al . gcn5122 ; cenko gcn5125 ; milne gcn5127 ) ; grb 070419a ( cenko et al . gcn6306 ; williams et al . gcn6328 ; milne et al . gcn6341 ; nishiura et al . gcn6308 ; updike et al . gcn6317 ; lizuka et al . gcn6316 ; fu et al.gcn6311;pozanenko et al . gcn6407 ; garnavich et al.gcn6406 ; hill et al . gcn6486).,scaledwidth=50.0% ] one of the interpretations for the optically dim luminosity is the extinction and absorption from circumburst medium in its host galaxy . here , we try to estimate the extinction along the line - of - sight to the burst . the observed neutral hydrogen column density along this line - of - sight to grb 070518 determined by xrt onboard _ swift _ is about @xmath101 cm@xmath65 ( guidorzi et al . 2007 ) , larger than the previously measured galactic value ( @xmath103 ; dickey & lockman , 1990 ) . the excess value is about @xmath104 cm@xmath65 . if the excess value is from a previously unobserved over - density located within or close to our galaxy , then the excess extinction @xmath105 would be obtained with the excess value of neutral hydrogen , based on the relation found by predehl & schmitt ( 1995 ) : @xmath106 + 0.23 $ ] mag . if this is the case , extinction should not be a prominent cause for the optically dim luminosity . it is likely that the excess column density is located within the host galaxy of grb 070518 . as in the discussion for grb 051022 ( nysewander et al . 2006 ) , the rest frame @xmath107 is given by the observed @xmath108 . the source frame is about @xmath109 with z=0.7 . from the relation of host extinction @xmath110 mag ( nysewander et al . 2006 ) , @xmath111 mag would be given in the rest frame based on a redshift of 0.7 . clearly , under these conditions , high extinction would be a dominate cause of the optically dim luminosity of grb 070518 . however , all above estimates are made with an assumption of the milky way - type extinction . the optical extinction from the measured @xmath107 absorption should depend on the assumed dust - to - gas ratio . some evidence shows that the extinction law in the grb host galaxy is likely to be a small magellanic clouds ( smc ) type . using the small dust - to - gas ratio of the smc ( pei 1992 ) , the host extinction @xmath112 would be 0.3 mag when @xmath77 . the difference of the extinction value estimated above is very large , about a factor of 10 . actually , the real extinction law in the grb host galaxy is uncertain in most cases , like grb 050401 , for example , the extinction @xmath2 is estimated to be @xmath113 by fitting the optical spectrum , which is lower than the value @xmath114 inferred from the x - ray column density ( watson et al . one reason for the case of grb 050401 was that high energy release of gamma - ray and hard x - ray photons might destroy the circumburst dust in the grb host galaxy ( watson et al . the reason might also be the case of grb 070518 . as shown in fig.3 , three _ swift _ optically dim bursts grb 050416 , grb 060512a and grb 070419a are labeled to compare with grb 070518 . in order to investigate the reason for the dim luminosity of grb 070518 , we try to compare them in different aspects . table 2 presented the main parameters for the three bursts along with those of grb 070518 . there are several similar properties , as follows . + ( 1 ) for all the bursts , t@xmath9 is lower than 10s , except for grb 070419a . this means that they belong to the shortest part of the `` long grb '' ( paciesas et al . 1999 ) , or intermediate long grb ( horvath et al . 2008 ) classification . + ( 2 ) the photon indices of these bursts are larger than 2 , which implies that the spectra are so soft that these bursts belong to xrrs and xrfs . + ( 3 ) the redshifts of all these bursts are lower than 1 . lcccccl grb & @xmath4 & @xmath69 & f & z & @xmath2 & xrr + name & ( s ) & & & & & or xrf + 070518 & 5.5 & 2.11 & 6.02 & 0.7@xmath01 & & yes + 050416a@xmath115 & 2.4 & 3.4 & 3.5 & 0.6535 & 0.2(0.32 ) & yes + 060512@xmath116 & 8.6 & 2.49 & 2.3 & 0.4428(2.1 ) & & yes + 070419a@xmath117 & 116 & 2.35 & 5.6 & 0.97 & & yes + @xmath115 the references for grb 050416a are sakamoto et al.(2005,2006 ) , holland et al . ( 2007 ) and kann et al . + @xmath116 the references for grb 060512 are cummings et al . ( 2006 ) , starling r. et al . ( 2006 ) , bloom et al . ( 2006 ) and kann et al.(2007 ) . + @xmath117 the references for grb 070419a are cenko et al . ( 2007 ) and stamatikos et al . + another aspect concerns the absorption and extinction of the grb host galaxy . of the bursts mentioned above , only grb 050416a has been studied extensively ( e.g. sakamoto et al . 2006 ; mangano et al . 2007 ; soderberg et al . 2007 ) , and the extinction from its host galaxy has been estimated to be @xmath118 ( holland et al . 2007 ) , or @xmath119 ( kann et al . 2007 ) . the two values are similar to the average value ( @xmath120=0.2 ) of grbs based on the study of kann et al . ( 2007 , 2008 ) . however , they are significantly smaller than the extinction ( a@xmath121 ) expected from the hydrogen column density inferred from x - ray observations of xrf 050416a , assuming a dust - to - gas ratio similar to that found for the milky way ( holland et al . the low value of extinction from host galaxy of grb 050416a implies that the extinction of host galaxy of grb 070518 might be not large , only about @xmath122 , as analyzed in @xmath163.4 . optically dim bursts might be related to dark bursts ( nardini et al . considering that grb 070518 is an optically dim burst , we wonder whether the burst is a dark burst or not . the index of optical to x - ray spectrum @xmath123 ( @xmath124 ) at 11 h after the burst trigger is used to define dark bursts ( jakobsson et al . cenko et al . ( 2009 ) argued that the x - ray flux at about 11 hours might be contaminated by the shallow decay for some bursts . as a result , the x - ray flux would be larger than the real x - ray flux radiated along with optical afterglow , which would make some normal bursts `` dark '' bursts . they calculated @xmath123 for the bursts in their catalogue at 1000s after the trigger time . however , it is too early for most of the _ swift _ bursts to carry out the calculation at 1000s after the trigger , because the x - ray light curves in some _ swift _ grbs might be at the steep or shallow decay segment about this time . for grb 070518 , @xmath123 at 1000s , 11 hours and @xmath125s after the burst are calculated by interpolation , as shown in fig.4 . @xmath123 for grb 070518 is 0.666 , 0.606 and 0.543 at the above three epochs , respectively . it seems that the burst is not a dark burst , for all of @xmath123 are slightly larger than 0.5 . however , spectral index @xmath1 depends on the energy distribution @xmath126 of electrons ( jakobsson et al . 2004 ) , for grb 070518 , @xmath126 should be 2.22@xmath127 , as @xmath128 ( from analysis of uss ) and @xmath129 . in this case , the criterion for dark bursts should become @xmath130 . consequently , grb 070518 might be a dark burst ( jakobsson et al . 2004 ) or a gray burst ( zheng et al . 2009 ) . band , shown as open hexagonal . the x - ray data at 3 kev is shown as solid squares . the spectral indices are calculated at different times ( 1000s , 11 hours and @xmath125s ) by the interpolation of their light curves . the optical to x - ray spectral indices @xmath1 are also shown . , scaledwidth=50.0% ] we have presented optical afterglow observations of grb 070518 from the tnt and an analysis of the optical and x - ray light curves . the optical afterglow shows a constant power law decay @xmath00.87 , while the x - ray shows a `` steep - shallow - normal '' decay behavior . based on the upper limit of redshift of 0.7 , the optical luminosity at 12 hours in the rest frame was estimated to be @xmath131 erg s@xmath20 , which means that the burst is an optically dim burst . the conclusion is also supported by comparing with three other optically dim bursts : grb 050416a , grb 060512 and grb 070419a . the spectral indices of @xmath123 at 1000 s , 12 hours,10@xmath132 are calculated . all @xmath123 were slightly larger than 0.5 , indicating that grb 070518 might be a dark burst or a gray burst , according to the definition of jakobsson et al . the optical extinction in its host galaxy inferred from x - ray hydrogen column density is 3.2 with galactic extinction law , and 0.3 with smc extinction law . afterglow model has been applied to the light curves . energy injection is found to occurr for the x - ray afterglow before the start of the normal decay segment . the behavior of the normal segment of the x - ray afterglow and the optical afterglow is consistent with the prediction of the classical external shock model ( urata et al . 2007 ) , indicating that they come from the same origin . as an xrr burst , grb 070518 agrees with amati relation . moreover , it fills the gape of high - luminosity and low - luminosity bursts in the amati relation , similar to grb 050416a ( sakamoto et al . it is also found that grb 070518 agrees with the ghalanda relation within reasonable assumptions . moreover , a comparison with three optically dim bursts ( grb 050416a , grb 060512 and grb 070419a and grb 070518 ) reveals that several similar properties were shared : burst duration t@xmath9@xmath1210s ( expect for grb 070419a , @xmath133 ) , soft spectrum at bat band @xmath69@xmath134 , low redshift @xmath1351 , etc . we are very grateful to the anonymous referee , who provided excellent advice and comments , and many suggestions for improving the paper . we wish to thank dr.xinyu dai for the discussion about the dark burst and the redshift of the burst . thanks go to dr.yufeng mao , xuhui han , huali li for their kindly help in correcting and improving the writing . this work made use of data supplied by the uk _ swift _ science data center at the university of leicester . this study has been supported by national basic research program of china-973 program 2009cb824800 and by the national natural science foundation of china - grant no.10673014 and 10803008 . 199 amati l. et al . , 2002 , a&a , 390 , 81 amati l. , guidorzi c. , frontera f. , della valle m. , finelli f. , landi r. , montanari e. , 2008 , mnras , 391 , 577 band l. et al . , 1993 , apj , 413,281 bernardini m. g. , dainotti m.g . , bianco c. l. , caito l. , guida r. , ruffini r. , 2009 , aipc , 1111 , 383 bloom j. , foley r. j. , koceveki d. , perley d. , 2006 , gcn circ . , 5217 burrows d. n. et al . , 2008 , preprint ( arxiv:0803.1844 ) cenko s. b. , 2006 , gcn circ . , 5125 cenko s. b. , fox d. b. , 2007 , gcn circ . , 6306 cenko s. b. , gezari s. , small t. , fox d.b . , chornock r. , 2007 , gcn circ . , 6322 cenko s. b. et al . , apj , 693 , 1484 covino s. , tagliaferri g. , chincarini g. , antonelli l. a. , israel g. l. , andreuzzi g. , garcia de gurtubai escudero a. , 2007 , gcn circ . , 6426 cucchiara a. , marshall f. e. , guidorzi c. , 2007 , gcn circ . , 6419 cummings j. et al . , 2006,gcn circ . , 5124 dai x. et al . , 2008 , apj , 682 , l77 dickey j. m. , & lockman , f. j. , 1990 , ara&a , 28 , 215 evans p. a. et al . , 2007 , a&a , 469,379 fu k.j . , lee y. h. , huang k. y. , urata y. , 2007 , gcn circ . , 6311 fynbo j.u . et al . , 2001 , a&a , 373 , 796 garnavich p. et al . , 2007a , gcn circ . , 6406 garnavich p. et al . , 2007b , gcn circ . , 6462 gendre b. , pelisson s. , boer m. , basa s. , mazure a. , 2008a , a&a , 492 , l1 gendre b. , galli a. , boer m. , 2008b , apj , 683 , 620 ghirlanda g. , ghisellini g. , lazzati d. , apj , 616 , 331 groot p. et al . , 1998 , apj , 493 , l27 guidorzi c. et al . , 2007 , gcn circ . , 6415 hill j. et al . 2007,gcn circ . , 6486 holland s. t. et al . , 2007 , aj , 133 , 122 horvath i. , balazs l. g. , bagoly z. , veres p. , 2008 , a&a , 489 , l1 jakobsson p. , hjorth j. , fynbo j. p. u. , watson d. , pedersen k. , bjornsson g. , gorosabel j. , 2004 , apj , 617 , l21 jelinek m. , castro - tirado a. j. , casanova v. , gorosabel j. , de ugarte postigo a. , vitek s. , cunniffe r. , kubanek p. , 2007 , gcn circ . , 6418 kann d. a. , klose s. , zeh a. , 2006 , apj , 641 , 993 kann d. a. et al . , 2007 , apj , preprint ( arxiv:0712.2186 ) kann d. a. et al . , 2008 , apj , preprint ( arxiv:0804.1959 ) kouveliotou c. , meegan c. a. , fishman g. j. , bhat n. p. , briggs m.s . , koshut t. m. , paciesas w. s. , pendleton g. n. , 1993 , apj , 413 , l101 krimm h. a. et al . , 2007a , gcn circ . , 6417 krimm h. a. et al . , 2007b , apj,665,554 kumar p. , panaitescu a. , 2000 , apj , 541 , l9 lamb d. q. , reichart d. , 2000 , apj , 536 , 1 liang e. w. , zhang b. , 2005 , apj , 633 , 611 liang e. w. , zhang b. , 2006 , apj , 638 , l67 malesani d. , piranomonte s. , covino s. , masetti n. , maiorano e. , magazzu a. , mainella g. , 2006 , gcn circ . , 5122 mangano v. et al . , 2007 , aj , 133 , 122 milne p. a. , 2006 , gcn circ . , 5127 milne p. a. , williams g. g. , 2007 , gcn circ . , 6341 mundell c. g. , steele i. a. , cummings j. r. , 2006 , gcn circ . , 5118 nakar e. , & oren y. , 2004 , apj , 602 , l97 nakar e. , & granot j. , 2007a , mnras , 380 , 1744 nardini m. , ghisellini g. , ghirlanda , g.,tavecchio f. , firmani c. , lazzati d. , 2006 , a&a , 451 , 821 nardini m. , ghisellini , g. , ghirlanda g. , 2008 , mnras , 386 , l87 nishiura s.,tornita h. , urata y. , huang k.y . , 2007,gcn circ . , 6308 nysewander m. c. , 2006 , ( phdt.2n ) publication number : aat 3239245 , isbn : 9780542940439 paciesas w. s. et al . , 1999 , apjs , 122 , 465 panaitescu a. , kumar p. , 2001a , apj , 554 , 667 panaitescu a. , kumar p. , 2001b , apj , 571,779 pei y. c. , 1992 , apj , 395 , 130 reichart d. , price p. , 2002 , apj , 565 , 174 pozanenko a. , shulga a. , volnova a. , hafizov b. , asfandiyarov i. , ibrabimov m. , 2007,gcn circ . , 6407 predehl p. , schmitt , j.h.m.m . , 1995 , a&a , 293 , 889 price p. a. minezaki t. , 2007 , gcn circ . , 6443 rol e. , wijers r.a.m . j. , kouveliotou c. , kaper l. , kaneko y. , 2005 , apj , 624 , 868 rumyantsev v. , biryukov v. , pozanenko a. , 2007 gcn circ . , 6483 sakamoto t. et al . , 2005 , gcn circ . , 3273 sakamoto t. et al . , 2006 , apj , 636 , l73 sakamoto t et al . , 2008a , apjs , 175 , 179 sakamoto t. et al . , 2008b , apj , 679 , 570 sakamoto t. et al . , 2009 , apj , 693 , 922 sari r. , piran t. , narayan r. , 1998 , apj , 497 , 17s sari r. , piran t. , halpern j. p. , 1999 , apj , 519,l17 schlegel e. m. , finkbeiner d. p. , davis m. , 1998 , apj , 500 , 525 soderberg a.m. et al . , 2007 , apj,661,982 stamatikos m. , parsons a. , perri m. , landsman w. , barthelmy s. d. , burrows d. n. , roming p. , gehrels n. , 2007 , gcn report , 47 starling r. et al . , 2006,gcn circ . , 5149 terra f. et al . , 2007,gcn circ . , 6458 updike a. c. , hartmann d. h. , rumstay k. , 2007 , gcn circ . , 6317 urata y. et al . , 2003 , apj , 595 , l21 urata y. et al . , 2005 , nuovo cimento c , 28 , 775 urata y. et al . , 2007 , apj , 668,95 watson d. et al . , 2006 , apj , 652 , 1011 williams g. g. , milne p. a. , 2007,gcn circ . , 6328 xin l. p. , zhai m. , qiu y. l. , wei j.y . , hu j. y. , deng j.s . , urata y. , zheng w.k . , 2007 , gcn circ . , 6416 zhang b. , fan y. z. , dyks j. , kobayashi s. , meszaros p. , burrows d. n. , nousek j. a. , gehrels n. , 2006 , apj , 642 , 354 zhang b. b. , zhang b. , liang e - w . , wang x - y . , 2009 , apj , 690 , l10 zheng w. k. et al.,2008 , chjaa , 8 , 693 zheng w. k. , deng j - s . , wang j. , 2009 , ra&a , 9 , 1103

it has been recognized for many years that observations of clusters of galaxies can be applied effectively to test theories of the evolution of the universe . as the largest gravitationally bound objects emerging from the growth and subsequent collapse of primordial density fluctuations , their spatial distribution at different epochs is a relic containing evidence which can test cosmological models of structure formation . x - ray astronomy has enhanced the results of such studies greatly , for a number of reasons . first , clusters are x - ray sources of relatively high luminosity , with a luminosity function sufficiently well defined to allow , in principle at least , cosmic structure to be studied out to epochs of @xmath6 and beyond . further , ambiguities in the identification of clusters , caused by chance line - of - sight superposition of objects , are fewer in x - ray than in optical sky surveys . this is a consequence of the sharply peaked x - ray brightness distribution in most clusters , and of the relatively low surface density of bright x - ray sources . second , the x - ray properties of clusters are defined by events reaching back to the earliest epoch of cluster formation , and therefore may be used to test cosmological models . the x - ray luminosity and temperature of the intracluster gas depend on the cluster mass , so that the shape and amplitude of the luminosity and temperature functions of the cluster population are dependent on the spectrum of the early density perturbations . further , the morphology of clusters in the whole population may be dependent on the characteristics of the expansion , and is a diagnostic of cluster merging processes . finally , the abundance of heavy elements in the intracluster gas is a measure of galaxy evolution and ram - pressure stripping , and of the proportion of primordial gas in the cluster . therefore much effort has been given to the production of statistically complete samples of x - ray emitting clusters . the first were produced using @xmath7 ( schwartz 1978 ) and @xmath8 ( mchardy 1978 ) observations , and the heao-1 all - sky surveys ( piccinotti et al . 1982 ; kowalski et al . 1984 ) , and contained up to 76 clusters . the flux limits of these surveys ( @xmath9 , expressed here in the rosat band @xmath1 ) , and their angular resolution were limited by the use of proportional counters , and improvements were possible only after the first missions using imaging x - ray instruments . edge et al . ( 1990 ) reported the use of einstein and exosat observations to refine the results of the earlier surveys , producing a sample of 55 clusters free of , for example , the effects of source confusion . in the einstein extended medium sensitivity survey ( emss ) it was possible , by examining the fields of einstein pointings covering a total of 740 @xmath10 , to create a sample with a flux limit of @xmath11 and containing clusters with redshifts out to @xmath12 ( henry et al . 1992 ) . the rosat x - ray telescope , equipped with an imaging proportional counter in the focal plane , allowed for the first time complete detailed surveys of large areas of sky ( trmper 1983 ) . the analysis of the rosat all - sky survey ( rass ) proceeded in two phases rass-1 and rass-2 , described below in section 3.1 , in which the second phase incorporated a number of improvements suggested by rass-1 ( voges et al . . the first x - ray flux - limited samples of clusters based on analysis of a large area of sky in the rosat survey were obtained by romer et al.(1994 ) , romer ( 1995 ) , ebeling et al.(1996 ) and ebeling et al . ( 1998 ) . the sgp sample described in romer et al.(1994 ) , based on analysis of the rass-1 data base , contained 161 cluster candidates with a limiting flux of @xmath13 . redshifts were obtained for 128 clusters in this sample , and this subset was used in a study of the cluster spatial correlation function ( romer et al . 1994 ) . the bcs survey by ebeling et al . ( 1998 ) addressed the whole extragalactic sky ( @xmath14 ) in the northern hemisphere , and reanalysed the fields in which clusters had been detected in the rass-1 analysis . the survey yielded a sample of 201 clusters at @xmath15 , having a limiting flux of @xmath16 , which is statistically complete at the 90 percent level . in a recent publication the bcs sample has been extended to an x - ray flux limit of @xmath17 , at which the completeness was 75 percent ( ebeling et al . 2000 ) . a recent survey of clusters in the northern sky ( noras ) , based on x - ray sources classified as extended in the rass analysis , is described by bhringer et al . deeper surveys have been possible in areas around the ecliptic poles ( henry et al . 2001 , gioia et al . 2001 , mullis et al . 2001 and voges et al . 2001 ) , where the rosat survey achieved unusually long exposure times , and in areas where deep pointings have been made ( romer et al . 2000 , jones et al . 1998 , rosati et al . 1998 , vikhlinen et al . 1998 and collins et al . the deepest survey was made by rosati et al . ( 1998 ) , in which the sample was complete down to a flux of @xmath18 and included clusters out a redshift of @xmath19 . in what follows we concentrate on complete large - area cluster surveys in the southern hemisphere . this has its roots in an agreement between the royal observatory in edinburgh ( roe ) , the naval research laboratory ( nrl ) and the max - planck - institut fr extraterrestrische physik ( mpe ) in garching , made early in the rosat program , to make the cosmos digitised uk schmidt iiia - j survey of the southern sky ( macgillivray & stobie 1985 , yentis et al . 1992 ) available for identification of x - ray sources . this data base , containing @xmath20 objects with a limiting magnitude of @xmath21 , was installed at mpe for use in x - ray source identification . in cluster studies it was used to derive unbiased lists of candidates through systematic correlations of x - ray source positions with concentrations of galaxies . a comprehensive search for clusters is being undertaken at mpe in the reflex southern hemisphere project ( bhringer et al . 2001 ) . the first results ( de grandi et al . 1999 ) described a sample of x - ray bright clusters , obtained by searching a region of area @xmath22 ( @xmath23 ) . the study reanalysed clusters identified in the rass-1 analysis and arrived at a sample of 130 clusters having a completeness of at least 90 percent and a limiting flux of @xmath24 in the @xmath1 band . the complete reflex study area covers @xmath25 in the southern sky , and the techniques used in constructing the cluster sample have been described by bhringer et al . ( 2001 ) . at the current stage of the study a sample of 452 clusters has been obtained , which is at least 90 percent complete , using a flux limit of @xmath26 . a power - law fit to the log n - log s distribution yielded an exponent of @xmath27 . collins et al . ( 2000 ) have analysed the spatial correlation function of clusters using a reflex sample of 452 clusters covering an area 4 times larger than the sgp , and derived a value of @xmath28 mpc for the correlation length , significantly greater than the value of @xmath29 mpc obtained by romer et al . however for the 109 reflex clusters in the sgp area collins et al . ( 2000 ) find a correlation length of @xmath30 mpc , consistent with the result of romer et al . the difference between the complete reflex and the sgp results is most likely due to sample fluctuations caused by cosmic variance ( see collins et al . a significant conclusion of the reflex study is that low - density cold dark matter ( cdm ) models ( @xmath31 ) provide a better fit to the observed correlation function , and that high - density cdm models ( @xmath32 ) yield a poorer fit because they provide insufficient power at large scales . schuecker et al . ( 2001 ) have analysed the power spectrum of the reflex sample , and the results are consistent with the results of collins et al . ( 2000 ) . a prominent aspect of the sgp and reflex survey projects has been optical observing programs to obtain cluster redshifts , which we describe briefly in section 2 . however , it should be borne in mind that while these two projects have a common root and have followed similar procedures for cluster selection , x - ray analysis and redshift determination , they do possess some differences in approach . following this introduction section 2 provides a summary of how the sgp survey was performed . section 3 describes the various procedures by which the cluster sample was extracted from the rass-2 data base and section 4 summarises how the x - ray energy flux and luminosity were calculated . section 5 presents the sgp cluster catalog in the form of a table and describes the archive of sgp cluster images in the form of overlays of the x - ray and optical data , which is accessible at mpe over the internet . section 6 examines the completeness of the cluster sample and section 7 discusses the large - scale structure traced by the clusters out to a redshift of @xmath33 . finally a summary of the results and conclusions of the paper are presented in section 8 . throughout the article we have assumed that the hubble constant @xmath34 , the deceleration parameter @xmath35 and the cosmological constant @xmath36 . this survey selected a region at high galactic latitude in the southern sky centered on the south galactic pole . rectangular in shape , it extends in right ascension from @xmath37 to @xmath38 and in declination from @xmath39 to @xmath40 ( see figure [ f18 ] ) , giving it an area of 1.013 steradian . the search for clusters in this region invoked no flux threshold , but intead worked to the limit of the rass data to maximise the completeness of the sample of cluster candidates . a flux limit is established subsequently to define a complete subsample . the first step in the search for clusters was to select all x - ray sources found in the region by the rass-2 analysis ( voges et al . 1999 ) . in this analysis the threshold values for the source detection algorithmns were lowered ( section 3.1 ) , with the aim of achieving as deep a survey as the rass would allow . therefore cluster searches using this data base have more sensitivity than has been available to previous southern hemisphere searches using the rass-1 data base ( romer 1995 , de grandi et al . 1999 ) . the rass-2 analysis was not equipped for detailed examination of the x - ray properties of extended sources such as clusters . to this end a number of techniques have been developed , including voronoi tesselation and percolation ( vtp , ebeling & wiedemann 1993 ) , the steepness ratio technique ( srt , degrandi et al . 1997 ) and the growth curve analysis ( gca , bhringer et al . the goal of these techniques is to obtain reliable fluxes for both extended and point sources , and to extract from the raw data as much information as possible about the x - ray source characteristics . in this study the gca technique has been used to analyse the list of candidates obtained from the rass-2 data base . cluster candidates were selected from this sample using the cosmos digitised iiia - j survey and a search procedure called csearch . this procedure , which is not designed to select clusters in an unambiguous manner , compares the galaxy density around the rass-2 position with the background galaxy density on the uk schmidt plate , and derives a probability that the result was a random coincidence . a sample of cluster candidates is obtained by setting a threshold to this probability , and this threshold is set low enough to ensure that few genuine clusters are lost and a high completeness is attained in the final sample . in our study a rass-2 list of 11981 sources is reduced by csearch to 3236 candidates ( section 3.3 ) . at this point much remains to be done to remove contaminating sources , which we estimate from internal statistical tests to comprise about half the candidates , in an objective process . at high redshifts ( @xmath41 ) the method reaches a limit of the cosmos data , when many cluster galaxies become so faint as to be classified by cosmos only as faint objects , which are not counteed by csearch . therefore some clusters at high redshift may be missed , which may be one source of incompleteness near the flux limit of the survey . the analysis of this sample of cluster candidates began by setting a limit of @xmath42 to the x - ray count rate in the hard band ( @xmath43 , section 3.1 , 3.4 ) , which reduced the number of candidates from 3236 to 477 . this was followed by the process of removing contaminating sources from this sub - sample , which proceeded through a number of stages . the first was the removal of cosmos artifacts , bright sources which to the eye were clearly not clusters , and redundant detections by rass-2 of the same source , particularly in extended sources such as clusters . this was done by examination of cosmos optical finding charts and of overlays of the optical and x - ray images . the second was to correlate rass-2 positions with the simbad catalog , in order to eliminate bright stars found in the rass-2 error circle , and then to use two x - ray properties of the source , hardness ratio and extent , as selection criteria . the hardness ratio was used to exclude those stars and active galactic nuclei ( agn ) whose spectra are sufficiently soft to exclude the possibility that the source is a cluster . the x - ray extent was used to classify candidates in three groups , namely clusters , agn candidates and a remainder of uncertain nature . the third and final stage , identifying the clusters and known agn among the remaining sample , proceeded along two parallel paths . the first comprised correlations of the candidate list with a variety of catalogs , in particular the nasa extragalactic data base ( ned ) , the simbad catalog , and the veron catalog of agn ( veron - cetty & veron 1998 ) . the second was an ongoing examination of the results of the supporting optical observations , described below , which identified agn in the sample . we summarise here briefly the sources of incompleteness in the final catalog obtained , which reduce the cluster count at low fluxes : + i ) the csearch analysis has a predetermined level of incompleteness in correlating + x - ray sources with galaxy concentrations ( section 3.3 ) . + ii ) the sharp count - rate limit ( @xmath42 ) rejects some sources at low energy + flux , due to the variation of interstellar column density in the field ( section 3.4 ) . + iii ) some distant clusters elude the tests applied in searching the list of candidates + yielded by csearch and the x - ray count - rate cut ( section 3.5 ) . + finally , a vital complementary part of the sgp survey and the reflex project has been a program of optical observations to obtain the redshifts of those clusters detected in the rosat survey , for which no redshifts were available in the literature . the sgp survey , which commenced not long after the launch of rosat in 1990 , was supported by three campaigns at the 4 m anglo - australian telescope ( aat ) and three at the 1.9 m telescope of the south african astronomical observatory . the results of these campaigns have been reported by romer ( 1995 ) . a larger program , which started in 1992 and was completed in 1995 , was installed as a key project at the european southern observatory ( eso ) . its purpose was to obtain redshifts for clusters in the reflex sample . an ongoing continuation of these observations is being pursued to support the extension of the reflex survey to lower x - ray flux limits . the rosat observatory was launched on 1990 june 1 , and saw first light on 1990 june 16 ( trmper et al . 1991 ) . after a two - month verification of the instrument operation it commenced its six - month long survey of the whole sky on 30 july 1990 . during the rass the observatory rotated about an axis through the sun , scanning on a given day a @xmath44 wide strip passing through the ecliptic poles the survey ended on 1991 february 18 . the first analysis of the survey ( rass-1 ) , which took place in 1991 - 1993 , used data collected into the same @xmath45 wide strips , and yielded a list of some 50,000 sources containing positions and such characteristics as count - rate , hardness ratio , angular extent and source detection likelihood . rass-1 was used to test the algorithms implemented in the analysis and to start scientific studies , and as a consequence a series of improvements was initiated ( voges et al . 1999 ) : + i ) sorting the data into 1378 sky fields of size @xmath46 , overlapping at least @xmath47 . + this improved the efficiency of source detection . + ii ) for operational reasons the rass-1 data base contained a few regions of low + exposure , which were filled in later by supplementary observations . this + later data was included in the rass-2 data base . + iii ) searching for sources in the soft @xmath48 , hard @xmath49 and broad + @xmath50 energy bands independently . + iv ) refining the pointing attitude solution and improving source positions . + v ) improvement of the spline - fit to the background , yielding more accurate source + count - rates . + vi ) the list of candidates selected by the sliding - window source detection algorithmns + was increased by lowering the detection threshold , and in the subsequent maximum + likelihood analysis the likelihood threshold for accepting sources was lowered + from 10 ( rass-1 analysis ) to 7 . + reanalysis of the survey data using the new software ( rass-2 ) was performed in 1994 - 1995 , yielding a list of 145060 sources . the sgp field contained 11981 sources above a likelihood threshold of 7 . the characteristics of the new software and results from the rass-2 analysis have been described by voges et al . the results described here relied solely on the rass-2 data base ( rass bright source catalog in voges et al . 1999 and rass faint source catalog in voges et al . 2000 ) . the @xmath45 wide rass scan strips overlap to an extent which increases with ecliptic latitude , and consequently the exposure time grows and reaches a maximum at the poles . this effect is visible in the map of the rass exposure time in the sgp survey region ( figure [ f1 ] ) , in which there is a systematic increase from the nw to the se corner . upon this change are superposed striations caused by periodic protective shutdown of the focal - plane detector as it traversed radiation belts . analysis of the exposure map of the sgp survey region yields the histogram shown in figure [ f2 ] , in which the mean exposure time is 300 seconds . there is an area of @xmath51 in the sgp field , in which the exposure time is less than @xmath52 . the roe cosmos digital data base contains the results of scanning 894 plates of the uk schmidt iiia - j ( blue ) southern hemisphere survey . the sgp study area contains 161 fields , each of size @xmath53 . cosmos analyses and classifies each object in a field , and records the @xmath54 magnitude , position ( r.a . and declination ( 2000 ) ) and object shape parameters . objects are classified primarily as stars , galaxies or faint objects , the last being objects whose low intensity on the plate defies a reliable determination of their class . heydon - dumbleton , collins , & macgillivray ( 1989 ) compared the cosmos galaxy number count - magnitude distribution with those derived from several independent optical surveys , and deduced that the galaxy sample was essentially complete ( @xmath55 percent completeness level ) to a magnitude limit @xmath56 . beyond this limit the star - galaxy discrimination gradually becomes less reliable as @xmath54 increases ( macgillivray & stobie 1985 , heydon - dumbleton et al . 1989 ) , leading to a significant incompleteness of the galaxy sample in the range @xmath57 . in the cosmos data base star / galaxy separation was performed down to a limit of @xmath58 . an example of the results is given in a figure , which shows the @xmath59 galaxies identified by cosmos in the sgp field ukj 411 ( this figure is in the published article , but was too large for this archived version ) . the galaxy magnitudes have been calibrated using photometric ccd observations of sequences of galaxies , which were used also in the calibration of the edinburgh / durham galaxy catalog ( heydon - dumbleton et al . 1989 ) . for any given plate cosmos yields a galaxy intensity , which depends on such factors as exposure time , emulsion characteristics and measures taken to keep the plate in a dry environment during exposure . this implies that the galaxy magnitudes for any given plate must be adjusted to match the surrounding plates . this is done by identifying galaxies in the overlapping border zone of adjacent plates , and deriving a shift which matches the galaxy magnitudes . no unique solution for the whole southern sky is possible , and instead a set of plate magnitude shifts is obtained by a fitting procedure which globally minimises the inconsistencies between adjacent plates . the resulting cosmos galaxy magnitudes are on an arbitrary machine scale , which now must be adjusted to match the ccd calibrations , in essence providing a global zero point solution to the matched plate data . this is done by fitting a correlation between the ccd and cosmos magnitudes of galaxies in the ccd observation sequences . to further ensure there were no calibration problems as a result of the iosophotal thresholding of the cosmos galaxies ( isophotal thresholding results in loss of light below the isophote level used and hence underestimation of the true brightness of a galaxy ) , we accepted for calibration purposes only galaxies in the cosmos data which were at least three magnitudes above the plate limit ( i.e. @xmath60 ) . this resulted in 207 sequence galaxies being accepted for calibration . the rms scatter of the correlation between the cosmos and the ccd magnitudes for these galaxies was found to be @xmath61 . as the intrinsic error in the ccd photometry is estimated to be @xmath62 , we estimate that the resulting error in the magnitudes derived for galaxies in the cosmos iiia - j southern hemisphere survey is @xmath63 . some independent corroboration of the cosmos galaxy photometry is provided by the study of caretta , maia , & willmer ( 2000 ) . they confirm after comparisons with their ccd data that the cosmos galaxy magnitudes are reliable at the @xmath63 level . their results indicate that galaxy samples are @xmath64 percent complete at magnitudes down to @xmath65 , in broad agreement with the cosmos estimates . however this agreement should be treated with caution , as caretta et al . ( 2000 ) studied a small area of sky having a lower galaxy count than is available in the uk schmidt southern hemisphere survey . the cluster search procedure ( csearch ) counts galaxies within 5 circles of radius @xmath66 = 1.5 , 3 , 5 , 7.5 and 10 arcmin ( circles 1 through 5 ) , centered at the rosat source position , and compares the results with those of random sampling in the ukj field . these radii correspond to the range of expected cluster core diameters ( e.g. at @xmath67 a radius of 300 kpc is equivalent to an angular scale of 1.1 arcmin and at @xmath68 to 11 arcmin ) . the random sampling was performed 1000 times for each radius , yielding for each field five histograms of the galaxy count . for each histogram a table is created , giving the probability @xmath69 that the galaxy count exceeds a value , @xmath70 : + @xmath71 where @xmath72 is the number of cases where the galaxy count was @xmath73 . selection of cluster candidates is made after setting a threshold @xmath74 , which implies that for each plate a threshold be set for @xmath70 . individual plate sensitivity and statistics make this galaxy count threshold vary from plate to plate and yield a significant uncertainty . the following argument describes how this is circumvented by analysing the ensemble of probability values @xmath69 for the whole sgp rass-2 source sample . the rosat source is accepted if the probability associated with one of the 5 galaxy counts ( circles 1 through 5 ) at the rosat position falls below a selected threshold @xmath75 . for convenience , we have defined and used the quantity @xmath76 , as this approaches a maximum value of 1 for cluster candidates . the range @xmath77 is divided into bins of equal width ( typically 400 in number ) , and for any given search radius @xmath66 we create a histogram of the source count ( @xmath78 ) per bin for all the rass-2 sources in the sgp field . if there were no correlation between the rosat position and the galaxy count , as would be close to the truth for an ensemble of stars and agn , the expectation value for the number of sources , @xmath78 , per bin would be independent of @xmath79 ( bhringer et al . figure [ f4 ] shows the histograms of the probability p obtained for the rosat sources detected in the sgp region . to save space we show results for @xmath66 = 1.5 , 3 , 5 and 10 arcmin only . the major characteristic of these histograms is indeed a flat distribution with statistical fluctuations , but in addition there is an enhancement at values of @xmath79 approaching @xmath80 , which is the signature of clusters . a comment on the data used to create these histograms is necessary . following the first csearch analysis of the sgp sources in the rass-2 data base , examination of images overlaying the x - ray and the optical fields revealed multiple detections by rass-2 of the more nearby clusters , due to their markedly extended x - ray emission ( e.g. a119 ) . this biases the csearch analysis , and therefore we removed 256 redundant multiple detections from the rass-2 sgp source list , and then repeated the analysis . this number is small compared to the total of 11981 sources , but is more significant in relation to the size of the eventual csearch sample of cluster candidates ( 3236 ) . the threshold @xmath81 must be chosen carefully in terms of the desired compromise between formal completeness @xmath82 and contamination @xmath83 , for inevitably the selected sample contains rosat sources having chance associations with concentrations of galaxies , which build the flat distribution , and so the price of higher completeness is an increased contamination . the probability threshold @xmath81 is related to @xmath82 and @xmath83 by the following equations : + @xmath84 @xmath85 where @xmath86 is the number of sources per bin in a specified flat portion of the probability histogram . before the results of csearch are discussed , two characteristics of the histograms shown in figure [ f4 ] should be clarified . first , the dip in @xmath79 for circle 1 in the range @xmath87 is a quantisation effect , due to the galaxy count jumping from 0 to 1 . the dip for circle 5 in the same region appears to be due to sampling limits , resulting from the circle diameter being a significant fraction of the field width . we have examined the number density of clusters found in our analyses as a function of position on the plate , and have found no sign of incompleteness near the cosmos digitisation boundaries on the plate . second , whereas the histograms for circles 1 and 2 decay rapidly from the peak at @xmath88 to a flat plateau , those for circles 3 and 5 decay even more rapidly to a minimum and then rise to a second enhancement in @xmath89 , covering the range @xmath90 for circle 3 and @xmath91 for circle 5 . to search for the cause of this second enhancement we examined how candidates found using circles 1 and 2 were registered using circles 3 through 5 . a typical result ( figure [ f5 ] ) shows how compact groupings of galaxies found by circle 2 , for which @xmath92 , are detected using circle 5 . the number of sources falls sharply from @xmath93 to a minimum , rises to a peak at @xmath94 and then falls again gradually , i.e. much of this second enhancement is due to relatively compact groups , for which the number of galaxies counted is not significantly enhanced when detected with circles 3 and 5 . as a consequence the effect of increasing the circle radius is to reduce @xmath79 . the five candidate lists ( @xmath66 = 1.5 , 3 , 5 , 7.5 and 10 arcmin ) are cross - correlated to remove redundant detections , and so this second enhancement is in principle not a problem . however , if for each circle @xmath81 is set so as to yield a high value of @xmath95 ( equation 2 ) , then for circles 3 through 5 the contamination @xmath83 is increased substantially , making the subsequent screening of the sample more difficult and time - consuming . therefore as a precaution , for circles * 4 and 5 only * we have set @xmath81 at @xmath96 , so as to include only the sharp peak . we show below that this is a reliable procedure . the sgp field contains 11981 rass-2 sources having a likelihood greater than or equal to @xmath97 . table 1 shows the results of analysing this sample with csearch , and summarises the assumptions made in selecting cluster candidates . the last entry @xmath98 gives the yield of new candidates as one starts with circle 1 , and then proceeds from circles 2 to 5 , * removing redundant detections * at each stage . the trends in these numbers confirm the expectation that the smaller circles should detect the majority of clusters . the decision to set @xmath81 to @xmath96 for circles 4 and 5 was supported later by a search of the final list of 186 cluster candidates , which revealed no case in which the csearch detection relied solely on circle 4 and/or circle 5 . the total sample extracted using csearch comprised 3236 candidates . table 1 indicates that the majority of clusters are detected using circles 1 , 2 and 3 , and we deduce from column 6 that the contamination of the sample should lie between 46 and 56 percent . + ccccccc 1 & 1.5 & 0.92 & @xmath99 & 0.925 & 0.46 & 1857 + 2 & 3 & 0.90 & @xmath99 & 0.900 & 0.50 & 884 + 3 & 5 & 0.90 & @xmath100 & 0.700 & 0.56 & 428 + 4 & 7.5 & 0.95 & @xmath101 & - & @xmath102 & 33 + 5 & 10 & 0.95 & @xmath101 & - & @xmath102 & 34 + & & & & & & + previous studies found that the energy flux from extended extended sources is underestimated by the standard rass source detection algorithm ( ebeling et al . 1996 , 1998 and de grandi et al . 1997 ) , and section 2 has stated the various analyses developed to correct this problem . the growth curve analysis ( gca ) method , described in detail by bhringer et al . ( 2000 ) , has been used in this study to analyse the rass-2 data base . the gca returns among other information the following important parameters : background subtracted source count - rate and its poisson ( photon statistical ) error , a locally redetermined position of the centroid of the source x - ray emission , the mean exposure for the source region , a significance of the source detection , the radius out to which source emission is significantly detected , a hardness ratio characterizing the source spectrum and its photon statistical error , a fitted source core radius , and the probability , obtained using a kolmogorov - smirnov test , that the object is consistent with being a point source . the source position is determined with an error which in most cases is less than @xmath103 , the exceptions being nearby clusters with irregular emission structure . errors in count - rate and hardness ratio are given in our catalog ( table 3 ) . taking a2726 as an example , the hard band count - rate is @xmath104 and the hardness ratio is @xmath105 . the basic parameters are derived from the photon distribution using the three energy bands ` hard ' ( 0.5 to 2.0 kev , channel 52 - 201 ) , ` broad ' ( 0.1 to 2.4 kev , channel 11 - 240 ) , and ` soft ' ( 0.1 to 0.4 kev , channel 11 - 40 ) . the band definitions are the same as those used in the standard analysis ( voges et al . most of the derived parameters are based on analyses of data in the hard energy band , as clusters are detected in this energy band with the highest signal - to - noise ratio . one of the exceptions is the hardness ratio , which requires also the results from the soft band . the gca method uses the photon event distributions in two circles ( figure [ f6 ] ) , of radii 20 and 41.3 arcmin , centered initially on the rass-2 source center . events within the inner circle are analysed to determine the source characteristics , and those in the outer annulus are used to determine the background level . in some cases special analyses were necessary , where either clusters had an unusually large angular extent , or where two or more sources lay in a confused region , requiring the use of deblending algorithms . the source count - rate is determined from the cumulative , radial source count rate profile ( growth curve ) after background subtraction . the construction of the growth curve is preceded by a redetermination of the source center and by the background measurement . a typical example for the resulting growth curve is displayed in figure [ f7 ] , showing the count - rate plotted as a function of integration radius ( solid curve ) . the dashed lines show the limits determined by applying the photon event statistical error ( including the error for the background subtraction ) . the count - rate is determined in two alternative ways . in the first a significant outer radius of the source is defined as the point where the increase in the @xmath106 error is larger than the increase of the source signal . the integrated count - rate is calculated using events within this cutoff radius . in the second method a horizontal level is fitted to the outer region of the plateau , and this value is adopted as the standard in quoting hard - band x - ray count - rates and in calculating cluster energy flux and luminosity ( section 4 ) . however , we use the first method also as a check on the count - rate and as a means of estimating systematic uncertainties in the count - rate determination , which must be combined with the pure photon statistical errors . careful examination of the overlays of x - ray and optical images ( section 3.5.1 ) of the cluster sample revealed 19 cases in which the source was very extended , or source confusion might cause errors in deriving the source count - rate . these 19 were subjected to the special analyses referred to above , in which the source count - rates were revised . six systems were identified as double clusters and not subjected to the deblending analysis . they are identified as double clusters in the survey catalog ( table 3 ) . the other two important source properties determined by the gca method are the spectral hardness ratio and the source extent . the hardness ratio , @xmath107 , is defined as @xmath108 where @xmath109 is the hard band and @xmath110 the soft band source count - rate ( both determined for the same outer radius limit ) . the source extent is investigated in two ways . in the first a @xmath111-model profile ( cavaliere & fusco - femiano 1976 ) convolved with the averaged survey psf is fitted to the differential count - rate profile ( using a fixed value of @xmath112 ) , and we derive from the results a core radius , @xmath113 . in the second a kolmogorov - smirnov ( ks ) test is used to estimate the probability that the object is a point source . it is taken as highly likely that the source is extended when the ks probability is less than 0.01 . tests with x - ray sources which have been identified with stars and agn showed that a false classification as an extended source occurred in less than 5% of the sample . the final step in the x - ray analysis was to establish a lower limit to the source count - rate , in order to set practical limits to source identification and redshift determination . the list of 3236 candidates from the csearch analysis ( section 3.3 ) was truncated using a threshold of @xmath42 for the cut - off radius count - rate in the hard band . this reduced the list of candidates to 477 . a sharp limit in count - rate corresponds to a range of energy flux , determined by the variation of interstellar column density @xmath114 in the field , which leads to some incompleteness at low fluxes . bhringer et al . ( 2001 ) estimate a range of @xmath115 in the band @xmath1 for @xmath116 . in the sgp field @xmath117 takes on relatively low values so that , for example , for only 8 percent of the clusters selected is @xmath118 . therefore we can be certain that this source of incompleteness is absent at fluxes greater than @xmath119 . the removal of contaminating sources from the list of 477 candidates was by necessity an iterative process , in which the decisions made were reviewed repeatedly to ensure that the criteria used had been applied consistently . the following sections describe the steps in this process . for each cluster candidate a finding chart @xmath120 in size was extracted from the cosmos data base . in addition an overlay was made of the x - ray surface brightness contours upon the cosmos optical field . for example figure [ f8 ] shows the overlay made for the abell cluster a3854 . examination of these finding charts and overlays during the selection process has been valuable for several reasons . first , visual scrutiny has removed cases in which csearch had been deceived by artifacts , for example satellite trails , the haloes and diffraction spikes of bright stars , and the attempt by cosmos to resolve the structure of bright nearby galaxies . these structures often were resolved by cosmos into objects identified as galaxies . second , the rass-2 analysis resulted in multiple detections of extended sources such as clusters , which may be removed after examination of the x - ray / optical overlays . as described in section 3.3 , a significant number of redundant detections was removed to minimise any bias of the csearch analysis . however this process was not perfect , and it was possible later to identify further cases by careful re - examination of the overlays . the latter are counted in the detection statistics shown later in table 2 . finally , scrutiny of the finding charts and overlays has identified occasional errors , which are simple to correct , where a csearch candidate is identified with a faint rass-2 x - ray source having a brighter x - ray neighbour . the gca technique then has preferred the latter and derived the higher count - rate . this problem was detected through visual examination of the overlays . the temperature and extent of the hot gas in clusters of galaxies make it possible to set limits to both the x - ray hardness ratio and the angular extent of a rass cluster source ( ebeling et al . 1996 , 1998 ) . this provides effective discriminants between clusters and the majority of contaminating sources in the csearch sample , made up of agn and stars . in this section we justify the following two criteria , which have been used in removing a large fraction of these contaminating sources : + 1 ) the * upper limit * of the x - ray hardness ratio ( section 3.4 ) is less than zero , + i.e. if @xmath121 , the source is too soft . + 2 ) the core radius @xmath113 obtained from the angular extent calculation is less than + @xmath122 , i.e. the object is a point source . + these criteria were applied independently in rejecting sources . in order to establish criteria using these parameters , we selected control samples describing the characteristics of clusters , agn and stars . the first contained @xmath123 abell clusters extracted from the total sample of @xmath124 candidates ( section 3.4 ) , using as a criterion a correlation between the x - ray position and the centroid of an abell cluster . conservatively we excluded from the sample clusters where the separation between the rosat and abell positions was greater than @xmath125 , where the survey exposure was low @xmath126 , or where x - ray maps revealed nearby components in the source region , one of which might not be a cluster . in each case the x - ray / optical overlay image ( section 3.5.1 ) and the cosmos finding chart were examined . the count - rates of these clusters covered the full range of the study reported here . the second and third control samples comprised agn and stars detected by the rosat survey and identified through spectroscopic optical observations . this project was a collaboration between mpe and eso , in which the error circles of all rass-1 sources in four relatively small regions of the southern hemisphere were searched using the cosmos data base to identify optical candidates . the nature of these candidates was examined through spectroscopic observations , after which the likely source of the x - ray emission was identified ( t. a. fleming 1994 , private communication ) . using three of these fields having a total area of @xmath127 , we selected samples of @xmath128 agn and @xmath129 stars , whose positions are within 20 arcsec of the rass-2 source positions . again , the x - ray count - rate of these sources covered the full range of the cluster survey reported here . the hardness ratio ( @xmath130 ) distributions for these samples are shown in figure [ f9 ] . whereas @xmath131 for @xmath132 percent of the abell clusters in our test sample , @xmath133 for @xmath134 and @xmath135 percent of the agn and star samples respectively . therefore a conservative threshold of zero for the * upper limit * to @xmath130 has been set , above which rass sources qualify for retention in the list of cluster candidates . this is an effective method of removing a large fraction of the stars and agn in our sample of 477 candidates . the angular extent , as measured by the derived core radius @xmath113 , provides an effective test for distinguishing clusters from agn , which however runs into two difficulties . the first is caused by the diminishing size of clusters with increasing redshift , an effect which is emphasised if the cluster contains a cooling flow . the second is the effect of source confusion in a rass-2 detection of an agn , when the gca analysis may yield a finite value for the source extent . these difficulties are constrained to a certain extent by the redshift limit ( @xmath67 ) and by the limit set to the x - ray count - rate in the hard band ( section 3.4 ) . the distributions of @xmath113 derived from the two samples are shown in figure [ f10 ] . the rise in these curves at a radius of @xmath136 is a result of setting a limit @xmath137 in the gca , and to a lesser extent of the influence of surrounding sources on the growth curve analysis ( section 3.4 ) . the important characteristic is that 58 percent of agns satisfy @xmath138 , whereas @xmath132 percent of abell clusters in our test sample satisfy the condition @xmath139 . therefore only those sources among the 477 csearch cluster candidates were retained which satisfied the condition @xmath139 . as discussed in section 3.5.5 , this results in some incompleteness near the survey limit , where the rass and cosmos data become marginal in the detection of distant clusters . we have examined four clusters in the abell control sample , which were rejected when these hardness ratio and core radius criteria were used . one , associated with a2800 , failed the hardness ratio test . this may be due to contamination , as it is in a confused region of x - ray emission . three sources , associated with a3866 , s181 and s1121 , failed the core radius test . in the case of s181 the galaxies are faint and displaced from the center of the x - ray source , which is comparatively strong ( @xmath140 in the hard band ) . the small rass-2 error circle contains a stellar object ( @xmath141 ) , and therefore we conclude that this source is an agn ( the emission from agn comes from a concentrated region , so that often cosmos classifies the object as a star ) . the source coincident with a3866 has a high count - rate ( @xmath142 ) yet shows only weak signs of extent . although the spectroscopic observations ( romer 1995 ) revealed evidence of an agn , overlaying an einstein hri image upon an optical dss image indicates that the central object is probably a cd galaxy and that the x - ray source is very compact but not a point source . thus the decision to reject this source is open to correction . s1121 appears to be an example of a distant cluster near the limit of the survey sensitivity . thus four clusters in the test sample of 118 abell clusters were rejected , from which we estimate that in applying the hardness ratio and core radius criteria to our cluster sample , the loss of clusters should be less than 4 percent . bright stars appear in the sample of cluster candidates either because of a chance association with a region of high galaxy density , or because cosmos has broken down the diffraction spikes and/or halo into objects classified as galaxies . they have been removed using two tests , one being the hardness ratio test described in section 3.5.2 and one being a statistical test . in the latter we have analysed @xmath143 regions of diameter @xmath144 in the sgp field , using the cosmos data to derive the mean density of stars as a function of @xmath54 magnitude . the following argument was used to reject rass-2 sources having stars brighter than @xmath145 in the error circle . the density of stars brighter than @xmath145 was found to be @xmath146 , which implies that the probability of finding such a star by chance in a rass-2 90-percent confidence error circle , which typically has a radius of about @xmath147 , is @xmath148 . therefore if we reject all candidates having a star brighter than @xmath145 within the error circle , the loss of clusters from the sample should be no more than @xmath149 . this procedure has been applied by correlating the candidate list with stars in the simbad and tycho ( hg et al . 1998 ) catalogs . a correlation of the @xmath150 magnitudes in these catalogs with the cosmos @xmath54 magnitudes shows that the latter are lower by about 0.5 magnitudes for stars fainter than @xmath151 , but that the difference widens steadily for brighter stars . therefore in examining bright simbad or tycho catalog stars in the rass-2 error circle we set a threshold of @xmath152 , below which the star was judged to be the cause of the x - ray emission . additional stars , not listed in the simbad and tycho catalogs , were identified later by inspection of the cosmos finding charts and the x - ray / optical overlay images for the 477 candidates . the x - ray sources removed by the hardness ratio @xmath153 and core radius @xmath154 tests have been reviewed carefully , using correlations of the list of 477 candidates with the veron catalog of active galaxies ( veron - cetty & veron 1998 ) and the nasa extragalactic data base ( ned ) to identify sources for which the 90 percent rass-2 error circle contained an identified agn , and examination of the x - ray / optical overlay images and the cosmos finding charts . however in the sample remaining after these tests , after removal of stars ( section 3.5.3 ) and after identification of detections resulting from cosmos artifacts , there remained a contaminant comprising 28 sources for which @xmath155 or @xmath139 , but whose identification as a cluster was suspect . these sources were reviewed using again the veron and ned correlations , overlay images and cosmos finding charts , but in addition using the results of the csearch analysis . of these 28 we rejected 9 as being agn and 1 as an x - ray source superposed on a nearby galaxy over 5 arcmin in extent . nine sources were excluded because the extent detection was due to source confusion , and 8 because csearch either made an identification at a large radius only , inconsistent with the small x - ray extent , or made an error due to cosmos artifacts . the one remaining source requires further study and was rejected tentatively . there is a possible association with a2813 , but the minimum found by gca in the core radius fit was very shallow and the evidence for a cluster in the x - ray / overlay image was indecisive . the total number of agn identified was 127 , and figure [ f11 ] shows the distribution of their @xmath54 magnitudes , as measured by cosmos . where the agn was not identified in the veron catalog or ned , @xmath54 for the brightest object inside the 90 percent error circle was used . the half - width of the distribution lies in the range @xmath156 . finally , we discovered five candidates showing large extended regions of x - ray emission @xmath157 in diameter , whose optical fields show no signs of a significant galaxy population and whose origin as the confusion of several sources is not obvious . after removal of rass-2 sources contaminating the sgp sample of 477 cluster candidates , using the procedures described above , there remained 186 candidates . the statistics of this selection process are summarised in table 2 . of the 61 redundant rass-2 detections in table 2 which eluded the earlier screening , 24 were associated with clusters . therefore of the 477 candidate sources @xmath158 are associated with clusters , which implies that the contamination of the sample of 477 was 0.56 . this is in marginal agreement with the results of the csearch analysis ( section 3.3 ) , which predicted @xmath159 . examination of table 1 shows that @xmath160 of the cluster candidates were identified by csearch using the smaller circles 1 and 2 , and consequently that formally the contamination of the sample should be 0.5 or less , and the completeness about 0.90 . therefore we can estimate that the sample of 477 candidates should contain @xmath161 clusters and that a complete sample obtained using csearch would have yielded @xmath162 clusters . therefore we can estimate that the overall completeness of the sample of selected clusters is @xmath163 . this incompleteness , which is the sum of the three effects summarised in section 2 , contributes to the flattening of the log n / log s distribution at fluxes below @xmath2 in the band @xmath1 ( section 5 ) . cc clusters of galaxies & 186 + active galactic nuclei ( agn ) & 127 + stars & 74 + extended x - ray emission regions & 5 + redundant rass-2 detections & 61 + sources below x - ray count - rate threshold & 24 + + total & 477 + the above argument implies that @xmath164 clusters have been missed in searching the 477 candidates . this has arisen in most cases while applying the criteria @xmath121 and @xmath138 to reject candidates , and in section 3.5.2 we have estimated that applying these criteria applied to a control sample of abell clusters should yield an incompleteness of @xmath165 . two examples of clusters with overly soft spectra are a2800 ( @xmath166 ) and the cluster @xmath167 ( @xmath168 ) discovered by tucker et al . ( 1995 ) in the emss . in neither case is the nature of the contaminating source clear . one limit of this study is that the criterion @xmath139 for selecting cluster candidates will inevitably discriminate against the more distant clusters of small angular extent , particularly those in which a cooling flow makes the x - ray source more compact . this loss is evident in the flattening of the log n - log s distribution at low flux ( section 5 ) . three aspects of the analysis bear out this trend . first , of the sources rejected because @xmath138 we identified at least 4 cases , where the x - ray / optical overlays showed suggestive but marginal evidence that they were clusters . second , there were 11 in our catalog of 186 clusters , for which the detection of angular extent appeared to be near its limits . the gca yielded a small but non - zero core radius , but simultaneously the ks - test yielded significant probabilities , between @xmath169 and @xmath170 percent , of the source being point - like . third , in the case of distant clusters accepted by csearch , cosmos classified some of the galaxies as faint objects , which the csearch analysis did not count . this is obvious in some of the x - ray / optical overlays of these objects , which show the dominant elliptical galaxy with halo surrounded by faint objects . therefore we conclude that some distant clusters may have been missed by csearch . in order to convert the rosat count - rate into an energy flux outside the galaxy , the spectrum of the hot intracluster medium ( icm ) has been convolved with the response functions of the interstellar medium ( ism ) and the instrument . the plasma emissivity was derived using the version of the raymond & smith ( 1977 ) code installed in 1994 in the mpe exsas software . the technique has been verified by comparing results with those obtained using the exspec analysis code ( bhringer et al . the cluster plasma was assumed to contain elements heavier than hydrogen having abundances 0.35 times the solar abundance . the hi column densities used in calculating the ism absorption have been taken from dickey & lockman ( 1990 ) . these calculations have been made over ranges of temperature between 0.05 and 10 kev ( @xmath171-@xmath172 k ) , and of hi column density between @xmath173 and @xmath174 . the resulting matrix was used to convert the count - rate in the hard band to an energy flux @xmath175 in units of @xmath176 in the 0.1 - 2.4 kev band . for each cluster whose redshift has been measured the x - ray luminosity has been calculated using the relations : + @xmath177 @xmath178 where @xmath179 is the luminosity in the 0.1 - 2.4 kev band , @xmath180 is the correction for redshifting of the radiation , and @xmath181 is the luminosity distance . k has been calculated for ranges of redshift between 0.025 and 0.5 , and of icm temperature between 0.05 and 10 kev . the resulting matrix was used to convert energy flux into luminosity . measured cluster temperatures have been taken from the compilation in table 1 of white et al . ( 1997 ) . where measured temperatures have not been found we have used the luminosity - temperature correlation derived by white et al . ( 1997 ) : + @xmath182 where @xmath183 is the x - ray luminosity in the 0.01 - 80 kev band and t is the temperature in kev . the assumed temperature was used to calculate a bolometric correction to the luminosity . for each source a cycle of iterations was performed in the flux and luminosity calculations , until the change in temperature fell below 1 percent . where neither redshift nor temperature has been measured , we have assumed a temperature of @xmath184 . the procedures described in sections 2 , 3 and 4 for searching the rass-2 data base for cluster candidates , selecting clusters from this sample and determining their x - ray characteristics , have resulted in the catalog of 186 clusters shown in table 3 . the following notes , listed by the column number in the table , augment the information given in the header to each column . + ( * 1 * ) rosat x - ray source name . + ( * 2 * ) right ascension ( h : m : s j2000 ) of the centroid ( gca ) of the cluster x - ray emission . + ( * 3 * ) declination ( d : m : s j2000 ) of the centroid ( gca ) of the cluster x - ray emission . + ( * 4 * ) count - rate in @xmath185 in the hard band ( @xmath43 ) , with error ( gca ) . + ( * 5 * ) x - ray hardness ratio , with error ( gca ) . + ( * 6 * ) total rosat survey exposure time for the cluster , in seconds . + ( * 7 * ) column density of interstellar atomic hydrogen at the cluster center , in units of @xmath186 . + ( * 8 * ) temperature of the cluster gas , derived by the procedure described in section 4 , in @xmath187 . the superscript @xmath188 indicates that no measurement was found , and the temperature was determined using an @xmath189 correlation . where no redshift was available a temperature of @xmath184 was assumed . + ( * 9 * ) energy flux from the cluster , with the absorption by galactic interstellar absorption removed . the units are @xmath190 in the @xmath1 band . + ( * 10 * ) x - ray luminosity of the cluster in its rest - frame , derived using the procedure described in section 4 . the units are @xmath191 in the energy band @xmath1 . we assume @xmath34 and @xmath35 . + ( * 11 * ) optical identification of the cluster . the meaning of the prefix to the identification and the appropriate reference are given in table 4 . for five sources identified as double clusters , the label dc has been appended to the identification . + ( * 12 * ) redshift of the cluster . + ( * 13 * ) label for the reference from which the redshift was obtained : + * 1 . * romer ( 1995 ) * 2 . * h. bhringer and l. guzzo 1999 , eso key project , private communication ) * 3 . * struble & rood ( 1999 ) * 4 . * katgert et al . ( 1996 ) * 5 . * quintana & ramirez ( 1995 ) * 6 . * fetisova et al . ( 1993 ) * 7 . * de grandi et al . ( 1999 ) * 8 . * h. andernach 1989,``a compilation of measured redshifts of abell clusters '' , unpublished , private communication * 9 . * crawford et al . ( 1995 ) * 10 . * huchra et al . ( 1999 ) * 11 . * shectman ( 1985 ) * 12 . * mazure et al . ( 1996 ) * 13 . * collins et al . ( 1995 ) * 14 . * dalton et al . ( 1997 ) * 15 . * de vaucouleurs et al . ( 1991 ) * 16 . * stocke et al . ( 1991 ) * 17 . * postman et al . ( 1992 ) * 18 . * abell , corwin , & olowin ( 1989 ) * 19 . * katgert et al . ( 1998 ) * 20 . * ebeling & maddox ( 1995 ) * 21 . * muriel , nicotra , & lambas ( 1995 ) * 22 . * shectman et al . ( 1996 ) * 23 . * lauberts & valentijn ( 1989 ) * 24 . * da costa et al . ( 1998 ) * 25 . * dalton , efstathiou , & sutherland ( 1994 ) * 26 . * di nella et al . ( 1996 ) * 27 . * ratcliffe et al . ( 1998 ) * 28 . * given in ned , reference uncertain . + ( * 14 * ) number of galaxy redshifts used in deriving the cluster redshift . + ccc a & abell catalog & abell et al . ( 1989 ) + s & abell catalog : southern supplementary clusters & abell et al . ( 1989 ) + am & cambridge apm catalog & ned : arp & madore ( 1987 ) + apmcc & cambridge apm cluster catalog & dalton et al . ( 1997 ) + cid & efar survey & barton et al . ( 1996 ) + edcc & edinburgh - durham cluster catalog & lumsden et al . ( 1992 ) + emss & einstein medium sensitivity survey & stocke et al . ( 1991 ) + eso & european southern observatory & no reference in ned + hcg & hickson compact group catalog & hickson ( 1982 ) , barton et al . ( 1996 ) + scg & catalog of cosmos southern compact groups & prandoni , iovino & macgillivray ( 1994 ) + ngc & cluster dominated by ngc galaxy & ned + sh & cluster catalog from shane - wirtanen galaxy counts & shectman ( 1985 ) + images of these clusters , showing overlays of the hard band x - ray contours upon optical images reconstructed from the cosmos digital data base , are accessible over the internet at ( http://wave.xray.mpe.mpg.de/publications/papers/2001/rass-sgp-clusters/ xray - optical - images ) . they comprise 134 abell clusters , 15 clusters found in other catalogs , and 37 newly discovered clusters . the number of abell clusters in the sgp survey area is 1215 , including 332 southern supplementary clusters , therefore the abell clusters in the sgp catalog ( table 3 ) comprise 11.0 percent of all abell clusters in the sgp survey area . the abell fraction of the sgp cluster sample ( 72% ) is in good agreement with that of the bcs ( ebeling et al . 1999 , 2000 ) . we have compared the catalog in table 3 with the reflex catalog of bright clusters published by de grandi et al . ( 1999 ) , which contains 54 clusters lying in the sgp region . applying their energy flux threshold of @xmath192 in the @xmath1 band to table 3 , we find all but one cluster from the list of de grandi et al . the exception is a3866 , which we have rejected tentatively as a source whose x - ray flux may be dominated by an agn ( section 3.5.2 ) . however table 3 contains 20 clusters above this threshold , which are not in the list of de grandi et al . the reason is that these sources were not in the rass-1 source list , but were found by the rass-2 analysis as a consequence of the improvements , in particular lowering source detection thresholds and searching for sources in the hard energy band , which are described in section 3.1 and by voges et al . these 20 sources are ngc 720 , ngc 1132 , five newly discovered clusters , sh518 , and the abell clusters 194 , 2410 , 2442 , 2474 , 2496 , 2554 , 2556 , 2566 , 3984 , 4010 , 4068 and s1136 . the sgp catalog contains 113 clusters having fluxes below the threshold of the de grandi et al . ( 1999 ) study . the log n - log s distribution for the sgp cluster sample is shown in figure [ f12 ] , where it is compared with the results of other surveys . to avoid any errors caused by converting hard band ( @xmath43 ) fluxes in these surveys to broad band ( @xmath1 ) fluxes , we have made this comparison in the hard band . the circular symbols ( @xmath193 ) are the result of the survey of clusters by de grandi et al . ( 1999 ) , who reanalysed clusters detected in the rass-1 data base over a @xmath23 area of the southern sky and made first use of redshifts obtained from the eso key project . figure [ f12 ] includes also one point from the heao-1 a-2 cluster survey ( piccinotti et al . 1982 ) , the fit to the log n - log s curve obtained by ebeling et al . ( 1998 ) from the rosat brightest cluster sample ( bcs ) , and a box setting limits at low fluxes , derived from an analysis of emss results ( rosati et al . 1995 ) . there are significant differences among these results at energy fluxes less than @xmath194 . there is a flattening of the sgp log n - log s distribution at energy fluxes less than about @xmath195 , equivalent to a flux of @xmath26 in the rosat broad band ( @xmath1 ) , which is the completeness limit derived in section 6 . applying this limit we obtain a sample of 112 clusters , of which 110 have measured redshifts . the sources of the incompleteness of the sgp catalog at low fluxes have been discussed in section 3.5.5 . in addition there is a steepening of the sgp log n - log s curve over a short range of hard band flux near @xmath196 , a characteristic evident to a lesser extent also in the results of de grandi et al . this results in the distribution being higher than the fit of ebeling et al . ( 1998 ) and being at the upper limit set by rosati et al . ( 1995 ) , within the range @xmath197 . this may be a statistical fluctuation , but as we show in section 7 , there is significant large - scale structure in the sgp field out to a redshift of @xmath33 , which may be one cause of this feature in the sgp log n - log s distribution ( figure [ f12 ] ) . the optical observing program and a search of the literature yielded redshifts for 157 of the 186 clusters in table 3 . the distribution of the clusters in redshift space is summarised in the wedge diagram shown in figure [ f13 ] , in which the clusters are projected onto a plane whose co - ordinates are cluster redshift and right ascension . in this figure the symbol size is a measure of the cluster x - ray luminosity in the @xmath1 band . table 3 contains five sources , labelled dc in column 11 , for which the x - ray and optical data provide compelling evidence that they are interacting double clusters . in contrast the pair of objects rxcj0108.8 - 1524 and rxcj0108.9 - 1537 , which in projection are close , do not constitute a double cluster as they are well separated in redshift space . the former is a151 , which has a redshift of 0.0537 ( h. andernach 1989 , unpublished , private communication ) , whereas spectroscopic observation of the latter yield two redshifts which tie the galaxies to a151 but nine having a mean of @xmath198 ( katgert et al . 1996 , 1998 ) . in section 3.5.5 it was estimated from a comparison of the number of clusters detected in the sgp with the csearch predictions , that the sample completeness was 0.79 , and it was concluded that the missing ones were mainly distant clusters of small angular extent . the effect is to flatten the log n - log s distribution at low fluxes ( figure [ f12 ] ) , and using the following arguments we have established that the sample is essentially complete at fluxes greater than @xmath2 . first , we compare our log n - log s distribution with the results of deeper surveys , and of the wider field reflex and bcs surveys . at @xmath2 in the sgp distribution ( figure [ f12 ] ) we obtain @xmath199 . for deeper surveys , which have smaller fields and therefore limited statistics at this flux , we obtain @xmath200 ( rosati et al . 1998 ) , @xmath201 ( vikhlinen et al . 1998 ) , @xmath202 ( henry et al . 2001 ) , and @xmath203 ( gioia et al . 2001 ) . for the reflex survey we obtain @xmath204 ( bhringer et al . 2001 ) and for the bcs survey @xmath205 ( ebeling et al . we find no evidence that the sgp result is inconsistent with these surveys at this flux threshold . second , we apply a standard measure of the formal completeness of a sample , namely the @xmath206 test ( e.g. , schmidt 1968 , avni & bahcall 1980 ) . for the computation of the maximum volume @xmath207 the effective area , the local variation of the x - ray flux limit ( both for a minimum of 10 source counts ) , and the @xmath180 correction ( see section 4 ) are taken into account . the test is performed for subsamples with different x - ray flux limits . the mean of the @xmath206 values and their @xmath208 standard deviations are plotted in figure [ f14 ] . note that the errors do not include cosmic variance . below @xmath209 the deviations from the ideal value of 0.5 are significant , indicating that the sample gets definitively incomplete below this flux limit . between @xmath210 and @xmath211 it is difficult to make a definite statement about significant deviations from the 0.5 line because the large scale inhomogeneities in the spatial distribution of the clusters ( see section 7 ) start to modulate the averaged @xmath206 values . note also that the measurements at different flux limits are not statistically independent . a good estimate of the flux completeness limit is @xmath212 , because for fluxes brighter than this limit the averaged @xmath206 values are all within the 90 percent confidence range ( note that cosmic variance increase the given formal errors ) . third , we examine the behavior of the radially averaged comoving cluster number density along the redshift direction . if we assume that within the tested redshift range no large evolutionary effects are present , one would expect an almost constant cluster number density profile with superposed large - scale inhomogeneities . figure [ f15 ] shows the superposition of a set of cluster number density profiles obtained from 10 volume - limited subsamples within the luminosity range @xmath213 to @xmath214 . outside this luminosity interval the sizes of the subsamples are smaller than 10 and are thus dominated by small - number statistics . the subsamples are restricted to clusters with at least 10 source counts and x - ray fluxes larger than @xmath215 , as obtained from the @xmath206 test mentioned above . the effective survey area covered by each subsample is taken into account . the individual density profiles are normalized using the average cluster number density estimated for each subsample . the bars in figure [ f15 ] represent the formal @xmath216 poisson errors , within which no systematic large - scale gradients of the number density are found for @xmath217 . the comparatively large high - density region found in the redshift interval @xmath218 is composed mainly of several superclusters . a more detailed analysis of the large - scale structure is given in section 7 . finally , we have verified that the variation of rass exposure time in the sgp field ( figure [ f1 ] ) has little effect on the effective area at a flux level of @xmath2 . at the detection threshold of the gca ( 10 source counts ) , this area is 98 percent of the total area of the field . the wedge diagram for the sgp cluster sample ( figure [ f13 ] ) reveals that the spatial density of clusters at redshifts greater than @xmath3 is too low for the purpose of tracing large - scale structure , and we concentrate the discussion on clusters within this limit . a detailed examination of large - scale structure in the southern sky was made by tully et al . ( 1992 ) , who analysed the abell , corwin & olowin ( 1989 ) catalog of clusters . in this study measured redshifts were complemented by estimates based on the magnitudes of the third and tenth brightest galaxies , having an accuracy of @xmath219 out to redshifts of about @xmath220 ( zamorani et al . we have condensed the results into a sky map of the centroids of superclusters identified in the sgp region ( figure [ f16 ] ) . another wedge diagram ( figure [ f17 ] ) shows the distribution of these superclusters in redshift space . . [ f16 ] the dotted line in figure [ f16 ] traces the supergalactic equator , defined by the plane of a disk - like structure , the local supercluster , identified by de vaucouleurs & de vaucouleurs ( 1964 ) using measurements of the radial velocities of a large sample of local galaxies . the maximum values of these velocities were around @xmath221 ( @xmath222 ) . tully et al . ( 1992 ) have presented persuasive evidence that the pisces - cetus supercluster is likewise a flattened structure extending along the supergalactic equator , which is an extension of the local supercluster out to redshifts of at least @xmath223 . this idea is supported by our x - ray survey of clusters in the sgp . figure [ f18 ] is a sky map of all clusters in the sgp catalog ( table 3 ) having a redshift of less than @xmath224 , which shows clearly a concentration of clusters around the supergalactic equator ( figure [ f16 ] ) . of the 36 clusters identified by tully ( 1987 ) as members of the pisces - cetus supercluster , the following seven appear in table 3 and have a redshift @xmath225 : a14 , a85 , a119 , a133 , a147 , a151 , and a168 . the six brightest clusters in figure [ f18 ] , having hard - band x - ray count rates greater then @xmath226 , lie close to the supergalactic equator . five of these clusters , a85 , a119 , a133 , a4059 and s1101 , have redshifts between @xmath227 and @xmath228 , consistent with their being members of the pisces - cetus supercluster . the fifth , a4038 , is nearer at a redshift of @xmath229 , placing it at the edge of the local supercluster . the pisces - cetus supercluster is evident also in the wedge diagram ( figure [ f13 ] ) , in which the group of x - ray clusters at @xmath230 and @xmath231 is close to the location of the supercluster center . more distant structures are difficult to identify in the x - ray survey . tully et al . ( 1992 ) identified a structure orthogonal to the supergalactic equator , which includes the superclusters aquarius a and b , aquarius - capricornus , and sculptor a and b. the location of this structure , which extends from @xmath231 to @xmath232 ( @xmath233 ) , may be traced in figures [ f16 ] and [ f17 ] . to search for evidence of this structure in the x - ray data , we have constructed a sky map ( figure [ f19 ] ) which includes only x - ray clusters in the range @xmath234 . the map shows an elongated feature in the nw quadrant , which is perpendicular to the supergalactic equator . the argument that this feature traces the structure evident in the findings of of tully et al . ( 1992 ) is supported by two observations . first , the group of x - ray clusters in the nw corner of figure [ f19 ] belongs to the aquarius - capricornus supercluster . a correlation in sky position is evident upon comparing figures [ f19 ] and [ f16 ] , and a correlation in redshift space is seen upon comparing figures [ f13 ] and [ f17 ] . second , the two bright clusters a2597 and a2670 , highlighted as black circles in the nw quadrant , were identified by tully ( 1987 ) as members of the aquarius supercluster . batuski et al . ( 1999 ) have performed redshift measurements recently which establish that aquarius a and b are essentially one filamentary structure almost aligned with the line - of - sight between @xmath235 and @xmath236 . this concentration of abell clusters on the sky near @xmath237 , @xmath238 ( batuski et al . 1999 ) is not evident in the x - ray data ( figure [ f19 ] ) . however , care is needed in interpreting this result as it may be biased by the relatively low exposure time in this part of the rass ( figure [ f1 ] ) . therefore we conclude that , at least for redshifts out to @xmath33 , there is significant large - scale structure in the sgp field . however , there is one part of this structure which is not strongly evident in the x - ray maps shown in figure [ f18 ] and figure [ f19 ] . this is the horologium - reticulum supercluster ( tully et al . 1992 ) , a rich association of clusters having its centroid at @xmath239 and located at the position shown in figure [ f16 ] . there is only one bright x - ray cluster in our survey , a3113 at ( @xmath240 , figure [ f19 ] ) and @xmath241 , which may be associated with this supercluster . we can not rule out the possibility that the supercluster is inconspicuous because of its location close to the boundary of the sgp survey region . a systematic search of the rosat all - sky survey for clusters of galaxies has been performed in an area of @xmath242 centered on the south galactic pole . the search was made by correlating the list of rass sources detected in the hard x - ray band ( @xmath43 ) with the projected surface density of galaxies derived from the cosmos digitised uk schmidt iiia - j survey of the southern sky . the analysis yielded a list of 3236 candidates , of which @xmath243 should be chance associations with non - cluster sources , mainly agn and stars , and which should be 90 percent complete in containing clusters . the study was bounded during x - ray analysis by applying a lower limit of @xmath42 to the count - rate in the hard band , yielding a list of 477 candidates . non - cluster sources were removed by applying tests using the source x - ray extent and hardness ratio , by scrutiny of cosmos finding charts and overlays of the rass and cosmos images , and by correlation with such catalogs as the nasa extragalactic ( ned ) , simbad , tycho and veron data bases . this was performed in as objective a manner as possible , requiring an iterative sequence of procedures . the search resulted in 186 clusters , comprising 134 abell clusters , 15 clusters found in other catalogs , and 37 newly discovered clusters . the catalog shown in table 3 , and the overlays of the rass hard x - ray and cosmos optical images , are accessible over the internet ( http://wave.xray.mpe.mpg.de / publications / papers/2001/rass - sgp - clusters / xray- optical - images ) . the catalog has a minimum energy flux of @xmath244 in the band @xmath1 . tests show that a sample of 112 clusters , obtained when a flux limit of @xmath2 is applied , is essentially complete . we estimate the completeness of the overall catalog at a flux limit of @xmath0 to be @xmath245 . in support of studies of cluster evolution and large - scale structure using the results of the survey , a series of optical observing campaigns at the aat and at saao was conducted to obtain cluster redshifts . in addition we have used a number of redshifts obtained in programs at eso in support of the reflex project . using these campaigns and an extensive search of the literature we have obtained redshifts for 157 of the clusters in the catalog ( table 3 ) . redshifts are available for 110 of the 112 clusters in the complete flux - limited sample . examination of the spatial distribution of the sgp x - ray cluster sample corroborates earlier studies of large - scale structure , which were based on optical observations ( de vaucouleurs & de vaucouleurs 1964 , tully et al . the basic features are a flattened structure in the supergalactic plane , reaching out to @xmath246 and embracing the pisces - cetus supercluster , and an orthogonal structure at @xmath247 which includes the aquarius , aquarius - capricornus and sculptor a and b superclusters . we thank prof . trmper and the rosat team for the unstinting support , the myriad valuable discussions , the access to the rass data base , the use of the rass and exsas software , and the use of the powerful computing system under joachim paul and his team , which have made this work possible . the rosat project is supported by the bildungsministerium fr bildung und forschung ( bmbf / dlr ) and the max - planck - gesellschaft ( mpg ) . in equal measure we thank the royal observatory edinburgh and the naval research laboratory ( nrl ) for the intensive effort that went into the creation of the cosmos digitised optical data base and its installation at mpe , which was an essential factor in identifying clusters in the rass data base . our gratitude goes also to the anglo - australian telescope , the south african astronomical observatory and the european southern observatory for the expert assistance provided to the optical observations crucial to this project . we thank the reflex project for its cooperation in exchanging redshifts , and in particular luigi guzzo of the osservatorio astronomica di brera for his analysis of those eso key project observations , which yielded 14 redshifts used in this study . finally , this research has been supported by the office of naval research and nrl , and by the national aeronautics and space administration through an astrophysics data program grant ( s57759-f ) . abell , g. o. , corwin jr . , h. g. , & olowin , r. p. 1989 , apjs , 70 , 1 + arp , h. c. , & madore , b. f. 1987 , `` a catalogue of peculiar galaxies and associations + ( cambridge : cambridge university press ) + avni , y. & bahcall , j.n . 1980 , apj , 235 , 694 + barton , e. , geller , m. j. , ramella , m. , marzke , r. o. , & da costa , l. n. 1996 , aj , 112 , + 871 + batuski , d. j. , miller , c. j. , slinglend , k. a. , balkowski , c. , maurogordato , s. , + cayatte , v. , felenbok , p. , & olowin , r. 1999 , apj , 520 , 491 + bhringer , h. , et al . 2000 , apjs , 129 , 435 + bhringer , h. , et al . 2001 , a&a , 369 , 826 + caretta , c. a. , maia , m. a. g. , & willmer , c. n. a. 2000 , aj , 119 , 524 + cavaliere , a. , & fusco - femiano , r. 1976 , a&a , 49 , 137 + collins , c. a. , guzzo , l. , nichol , r. c. , & lumsden , s l. 1995 , mnras , 274 , 1071 + collins , c. a. , burke , d. a. , romer , a.k . , sharples , r. m. , & nichol , r. c. 1997 , apj , + 479 , l117 + collins , c. a. , et al . 2000 , mnras , 319 , 939 + crawford , c. s. , edge a. c. , fabian , a. c. , allen , s. w. , boehringer , h. , ebeling , h. , + mcmahon , r. g. & voges , w. 1995 , mnras , 274 , 75 + da costa , l. n. , et al . 1998 , aj , 116 , 1 + dalton , g. b. , efstathiou , g. & sutherland , w. j. 1994 , mnras , 269 , 151 + dalton , g. b. , maddox , s. j. , sutherland , w. j. , & efstathiou , g. 1997 , mnras , 289 , 263 + de grandi , s. , molendi , s. , bhringer , h. , & voges , w. 1997 , apj , 486 , 738 + de grandi , s. , et al . 1999 , apj , 514 , 148 + de vaucouleurs , g. , & de vaucouleurs , a. 1964 , first reference catalogue of bright + galaxies ( austin , university of texas press ) + de vaucouleurs , g. , de vaucouleurs , a. , corwin , h. , buta , r. , paturel , g. & fouque , p. + 1991 , third reference catalogue of bright galaxies ( springer - verlag : new york ) + di nella , h. , couch , w. j. , paturel , g. , & parker , q. a. 1996 , mnras , 283 , 367 + dickey , j. m. , & lockman , f. j. 1990 , ann . astrophys . , 28 , 215 + ebeling , h. , & wiedemann , g. 1993 , phys . e , 47 , 704 + ebeling , h. , & maddox , s. j. 1995 , mnras , 275 , 1155 + ebeling , h. , voges , w. , bhringer , h. , edge , a. c. , huchra , j. p. , & briel , u. g. 1996 , + mnras , 281 , 799 + ebeling , h. , edge , a. c. , bhringer , h. , allen , s. w. , crawford , c. s. , fabian , a. c. , + voges , w. , & huchra , j. p. 1998 , mnras , 301 , 881 + ebeling , h. , edge , a. c. , allen , s. w. , crawford , c. s. , fabian , a. c. , & huchra , j. r. + 2000 , mnras , 318 , 333 + edge , a. c. , stewart , g. c. , fabian , a. c. , & arnaud , k. a. 1990 , mnras , 245 , 559 + fetisova , t. s. , kuznetsov , d. yu . , lipovetski , v. a. , starobinski , a. a. , & olovin , r. p. + 1993 , astronomy letters , 19 , 198 + gioia , i. m. , henry , j. p. , mullis , c. r. , voges , w. , briel , u. g. , bhringer , h. & + huchra , j. p. 2001 , apj , 553 , l105 + henry , j.p . , gioia , i. m. , maccacaro , t. , morris , s. l. , stocke , j. t. , & wolter , a. 1992 , + apj , 386 , 408 + henry , j. p , gioia , i. m. , mullis , c. r. , voges , w. , briel , u. g. , bhringer , h. & + huchra , j. p. 2001 , apj , 553 , l109 + heydon - dumbleton , n. h. , collins , c. a. , & macgillivray , h. t. 1989 , mnras , 238 , 379 + hickson , p. 1982 , apj , 255 , 382 + hg , e. , kuzmin , a. , bastian , u. , fabricius , c. , kuimov , k. , lindegren , l. , + makarov , v. v. , & rser , s. 1998 , a&a , 335 , l65 + huchra , j. p. , vogeley , m. s. & geller , m. j. 1999 , apjs . 121 , 287 + katgert , p. , et al . 1996 , a&a , 310 , 8 + katgert , p. , mazure , a. , den hartog , r. , adami , c. , biviano , a. , & perea , j. 1998 , a&as , + 129 , 399 + kowalski , m. p. , ulmer , m. p. , cruddace , r. g. , & wood , k. s. 1984 , apjs , 286 , 822 + jones , l. r. , scharf , c. , ebeling , h. , perlman , e. , wegner , g. , malkan , m. , & horner , d. + 1998 , apj , 495 , 100 + lauberts , a. , & valentijn , e. 1989 , eso catalog , ' ' the surface photometry of the + eso - uppsala galaxies `` ( european southern observatory : garching ) + lumsden , s. l. , nichol , r. c. , collins , c. a. , & guzzo , l. 1992 , mnras , 258 , 1 + macgillivray , h. t. , & stobie , r. s. 1985 , vistas in astronomy , 27 , 433 + mazure , a. , et al . 1996 , a&a , 310 , 31 + mchardy , i. 1978 , mnras , 184 , 783 + mullis , c.r . , henry , j. p , gioia , i. m. , bhringer , h. , briel , u. g. , voges , w. + & huchra , j. p. 2000 , apj , 553 , l115 + muriel , h. , nicotra , m. a. , & lambas , d. g. 1995 , aj , 110 , 1032 + piccinotti , g. , mushotzky , r. f. , boldt , e. a. , holt , s. s. , marshall , f. e. , + serlemitsos , p. j. , & shafer , r. a. 1982 , apj , 253 , 485 + postman , m. , huchra , j. p. , & geller , m. j. 1992 , apj , 384 , 404 + prandoni , i. , iovino , a. , & macgillivray , h. t. 1994 , aj , 107 , 1235 + quintana , h. , & ramirez , a. 1995 , apjs , 96 , 343 + ratcliffe , a. , shanks , t. , parker , q. a. , broadbent , a. , watson , f. g. , oates , a. p. , + collins , c. a. , & fong , r. 1998 , mnras , 300 , 417 + raymond , j. c. , & smith , b. w. 1977 , apjs , 35 , 419 + romer , a. k. , collins , c. a. , bhringer , h. , cruddace , r. g. , ebeling , h. , + macgillivray , h. t. , & voges , w. 1994 , nature , 372 , 75 + romer , a. k. 1995 , doctoral thesis , university of edinburgh + romer , a. k. , et al . 2000 , apjs , 126 , 209 + rosati , p. , della ceca , r. , burg , r. , norman , c. , & giacconi , r. 1995 , apj , 445 , l11 + rosati , p. , della ceca , r. , norman , c. , & giacconi , r. 1998 , apj , 492 , l21 + schmidt , m. 1968 , apj , 151 , 393 + schwartz , d. a. 1978 , apj , 220 , 8 + shectman , s. a. 1985 , apjs , 57 , 77 + shectman , s. a. , landy , s. d. , oemler , a. , tucker , d. l. , lin , h. , kirschner , r. p. , + & schecter , l. p. 1996 , apj , 470,172 + schuecker , p. , et al . 2001 , a&a , 368 , 86 + stocke , j. t. , morris , s. m. , gioia , i. m. , maccacaro , t. , schild , r. , wolter , a. , + fleming , t. a. , & henry , j.p . 1991 , apjs , 76 , 813 + struble , m. f. , & rood , h. j. 1999 , apjs , 125 , 35 + trmper , j. 1983 , adv . space res . , 27 , 1404 + trmper , j. , et al . 1991 , nature , 349 , 579 + tucker , w. h. , tananbaum , h. , & remillard , r. a. 1995 , apj , 444 , 532 + tully , r. b. 1987 , apj , 323 , 1 + tully , r. b. , scaramella , r. , vettolani , g. , & zamorani , g. 1992 , apj , 388 , 9 + veron - cetty , m. p. , & veron , p. 1998 , eso scientific report 18 + vikhlinen , a. , mcnamara , b. r. , forman , w. , & jones , c. 1998 , apj , 502 , 558 + voges , w. , et al . 1999 , a&a , 349 , 389 + voges , w. , et al . 2000 , iau circular 7432 + voges , w. , henry , j. p , briel , u. g. , bhringer , h. , mullis , c. r. , gioia , i. m. & + huchra , j. p. 2001 , apj , 553 , l119 + white , d. a. , jones , c. , & forman , w. 1997 , mnras , 292 , 419 + yentis , d. j. , cruddace , r. g. , gursky , h. , stuart , b. v. , wallin , j. f. , + macgillivray , h. t. , & collins , c. a. 1992 , ' ' digitised optical sky surveys `` , ed . + h. t. macgillivray & e. b. thomson , ( netherlands : kluwer academic publishers ) , 67 + zamorani , g. , scaramella , r. , vettolani , g. , & chincarini , g. 1992 , in ' ' traces of primordial + structure in the universe , ed . h. bhringer & r. a. treumann , mpe report 227 + cccccccccccccc rxcj2201.8 - 2226 & 22 01 50.5 & -22 26 37 & 0.286@xmath2480.043 & 0.82@xmath2480.30 & 237 & 2.61 & 3.554@xmath249 & 5.75 & 1.193 & s987 & 0.0693 & 8 & 20 + rxcj2202.1 - 0949 & 22 02 07.2 & -09 49 48 & 0.280@xmath2480.038 & 1.15@xmath2480.31 & 307 & 4.20 & 4.039@xmath249 & 5.89 & 1.629 & a2410 & 0.0801 & 17 & 1 + rxcj2203.8 - 2130 & 22 03 50.6 & -21 30 31 & 0.139@xmath2480.030 & 0.11@xmath2480.19 & 227 & 2.45 & 2.809@xmath249 & 2.77 & 0.644 & & 0.0732 & 1 & 3 + rxcj2205.6 - 0535 & 22 05 39.6 & -05 35 29 & 0.674@xmath2480.109 & 0.79@xmath2480.12 & 191 & 4.68 & 3.800@xmath250 & 14.40 & 2.103 & a2415 & 0.0582 & 2 & 3 + rxcj2210.3 - 1210 & 22 10 20.0 & -12 10 33 & 0.746@xmath2480.056 & 1.04@xmath2480.08 & 278 & 3.92 & 6.000@xmath250 & 15.70 & 4.642 & a2420 & 0.0831 & 1 & 3 + rxcj2213.0 - 2753 & 22 13 05.0 & -27 53 56 & 0.425@xmath2480.034 & 0.14@xmath2480.17 & 205 & 1.40 & 3.760@xmath249 & 8.28 & 1.373 & & 0.0620 & 1 & 2 + rxcj2214.4 - 1701 & 22 14 26.3 & -17 01 24 & 0.198@xmath2480.020 & 0.79@xmath2480.24 & 305 & 2.61 & 5.845@xmath249 & 4.02 & 4.016 & a3847 & 0.1530 & 3 & 1 + rxcj2214.5 - 1022 & 22 14 32.4 & -10 22 12 & 0.594@xmath2480.052 & 0.95@xmath2480.09 & 280 & 3.84 & 6.578@xmath249 & 12.50 & 5.343 & a2426 & 0.1001 & 1 & 4 + rxcj2216.2 - 0920 & 22 16 15.2 & -09 20 18 & 0.405@xmath2480.041 & 0.87@xmath2480.09 & 274 & 5.24 & 4.869@xmath249 & 8.78 & 2.567 & a2428 & 0.0825 & 1 & 2 + rxcj2216.9 - 1725 & 22 16 56.4 & -17 25 30 & 0.425@xmath2480.034 & 0.35@xmath2480.12 & 277 & 2.28 & 6.969@xmath249 & 8.56 & 6.163 & & 0.1301 & 1 & 2 + rxcj2217.7 - 3543 & 22 17 43.7 & -35 43 32 & 0.329@xmath2480.035 & 0.32@xmath2480.11 & 319 & 1.10 & 6.865@xmath249 & 6.42 & 5.929 & a3854 & 0.1474 & 13 & 9 + rxcj2218.2 - 0350 & 22 18 17.1 & -03 50 01 & 0.427@xmath2480.047 & 0.92@xmath2480.14 & 273 & 5.73 & 5.375@xmath249 & 9.39 & 3.267 & ms2215.7 - 0404 dc & 0.0901 & 1 & 3 + rxcj2218.6 - 3853 & 22 18 40.3 & -38 53 48 & 0.425@xmath2480.034 & 0.27@xmath2480.09 & 314 & 1.33 & 7.246@xmath249 & 8.35 & 6.746 & eso 344-g 019 & 0.1379 & 28 & ? + rxcj2218.8 - 0258 & 22 18 49.0 & -02 58 07 & 0.425@xmath2480.034 & 0.90@xmath2480.19 & 268 & 5.84 & 5.350@xmath249 & 9.37 & 3.246 & & 0.0899 & 2 & 7 + rxcj2223.9 - 0137 & 22 23 56.7 & -01 37 44 & 0.454@xmath2480.049 & 1.00@xmath2480.12 & 224 & 5.34 & 9.000@xmath250 & 9.92 & 3.509 & a2440 & 0.0912 & 8 & 24 + rxcj2225.8 - 0636 & 22 25 50.8 & -06 36 09 & 0.239@xmath2480.034 & 0.68@xmath2480.15 & 251 & 5.14 & 4.196@xmath249 & 5.16 & 1.787 & a2442 & 0.0897 & 2 & 12 + rxcj2226.1 - 4745 & 22 26 10.6 & -47 45 01 & 0.091@xmath2480.020 & 0.31@xmath2480.23 & 329 & 1.44 & 3.289@xmath249 & 1.77 & 0.978 & a3876 & 0.1130 & 14 & ? + rxcj2227.4 - 4852 & 22 27 28.3 & -48 52 54 & 0.126@xmath2480.022 & 0.24@xmath2480.17 & 367 & 1.65 & 3.321@xmath249 & 2.47 & 1.004 & apmcc782 & 0.0970 & 14 & 3 + rxcj2227.8 - 3034 & 22 27 52.9 & -30 34 11 & 0.522@xmath2480.043 & 0.05@xmath2480.06 & 308 & 1.09 & 3.800@xmath250 & 10.10 & 1.415 & a3880 & 0.0570 & 13 & 11 + rxcj2231.2 - 3802 & 22 31 15.1 & -38 02 39 & 0.088@xmath2480.022 & 0.88@xmath2480.44 & 233 & 1.20 & 5.000@xmath249 & 1.71 & & & & & + rxcj2234.5 - 3744 & 22 34 30.4 & -37 44 05 & 0.573@xmath2480.051 & 0.09@xmath2480.07 & 241 & 1.22 & 7.900@xmath250 & 11.20 & 11.986 & a3888 & 0.1590 & 8 & 70 + rxcj2234.8 - 3033 & 22 34 53.5 & -30 33 56 & 0.119@xmath2480.027 & 0.64@xmath2480.28 & 305 & 1.11 & 6.983@xmath249 & 2.32 & 6.182 & a3889 & 0.2512 & 8 & 19 + rxcj2235.6 + 0128 & 22 35 40.6 & + 01 28 21 & 0.510@xmath2480.080 & 0.76@xmath2480.12 & 167 & 5.81 & 4.113@xmath249 & 11.20 & 1.702 & a2457 & 0.0594 & 3 & 18 + rxcj2239.2 - 1719 & 22 39 16.4 & -17 19 46 & 0.102@xmath2480.033 & 0.76@xmath2480.33 & 110 & 2.99 & 2.559@xmath249 & 2.05 & 0.503 & a2462 & 0.0751 & 17 & 1 + rxcj2241.9 - 4235 & 22 41 56.7 & -42 35 51 & 0.085@xmath2480.022 & 0.49@xmath2480.25 & 259 & 1.66 & 5.000@xmath249 & 1.68 & & a3902 & & & + rxcj2243.0 - 2010 & 22 43 05.1 & -20 10 34 & 0.246@xmath2480.098 & 0.71@xmath2480.20 & 185 & 2.60 & 5.892@xmath249 & 4.99 & 4.085 & a2474 & 0.1385 & 1 & 1 + rxcj2244.5 - 1744 & 22 44 34.9 & -17 44 21 & 0.128@xmath2480.035 & 0.38@xmath2480.29 & 151 & 2.79 & 4.361@xmath249 & 2.60 & 1.963 & a2478 & 0.1325 & 1 & 2 + rxcj2246.1 - 3600 & 22 46 11.8 & -36 00 22 & 0.100@xmath2480.022 & 0.40@xmath2480.25 & 260 & 1.14 & 2.290@xmath249 & 1.90 & 0.374 & & 0.0673 & 1 & 2 + rxcj2247.2 + 0204 & 22 47 15.2 & + 02 04 35 & 0.168@xmath2480.064 & 0.93@xmath2480.18 & 151 & 5.84 & 5.000@xmath249 & 3.70 & & & & & + rxcj2248.7 - 4431 & 22 48 43.4 & -44 31 41 & 0.431@xmath2480.047 & 0.47@xmath2480.10 & 208 & 1.79 & 11.946@xmath249 & 8.59 & 22.736 & s1063 & 0.2520 & 7 & ? + rxcj2249.2 - 3727 & 22 49 17.6 & -37 27 26 & 0.154@xmath2480.028 & 0.23@xmath2480.20 & 253 & 1.13 & 1.388@xmath249 & 2.66 & 0.096 & s1065 & 0.0289 & 23 & 2 + rxcj2250.8 - 4521 & 22 50 48.9 & -45 21 33 & 0.104@xmath2480.026 & 0.44@xmath2480.30 & 203 & 1.88 & 1.898@xmath249 & 1.99 & 0.226 & s1067 & 0.0511 & 21,22 & 6 + rxcj2251.0 - 1624 & 22 51 00.4 & -16 24 21 & 0.255@xmath2480.052 & 0.83@xmath2480.16 & 103 & 3.16 & 5.426@xmath249 & 5.24 & 3.342 & a2496 & 0.1221 & 1 & 3 + rxcj2253.5 - 3343 & 22 53 34.2 & -33 43 21 & 0.155@xmath2480.029 & 0.29@xmath2480.16 & 195 & 1.24 & 7.133@xmath249 & 3.04 & 6.504 & a3934 & 0.2250 & 8 & 6 + rxcj2305.5 - 4513 & 23 05 34.7 & -45 13 11 & 0.220@xmath2480.094 & 0.11@xmath2480.20 & 159 & 1.69 & 5.132@xmath249 & 4.34 & 2.919 & a3970 & 0.1253 & 1 & 3 + rxcj2306.8 - 1324 & 23 06 51.7 & -13 24 59 & 0.200@xmath2480.043 & 0.51@xmath2480.24 & 139 & 3.07 & 5.000@xmath249 & 4.10 & 0.765 & & 0.0659 & 2 & 2 + rxcj2307.2 - 1513 & 23 07 15.5 & -15 13 41 & 0.242@xmath2480.063 & 1.12@xmath2480.12 & 90 & 2.78 & 4.894@xmath249 & 4.92 & 2.599 & a2533 & 0.1110 & 5 & 1 + rxcj2308.3 - 0211 & 23 08 23.1 & -02 11 32 & 0.200@xmath2480.029 & 1.11@xmath2480.22 & 354 & 4.29 & 10.221@xmath249 & 4.26 & 15.544 & a2537 & 0.2966 & 1 & 2 + rxcj2312.2 - 2129 & 23 12 16.3 & -21 29 35 & 0.335@xmath2480.051 & 0.63@xmath2480.20 & 86 & 2.03 & 4.100@xmath250 & 6.65 & 3.532 & a2554 & 0.1111 & 3 & 28 + rxcj2313.0 - 2138 & 23 13 02.2 & -21 38 07 & 0.273@xmath2480.070 & 0.41@xmath2480.15 & 88 & 2.03 & 4.700@xmath250 & 5.43 & 1.746 & a2556 & 0.0865 & 3 & 2 + rxcj2313.9 - 4243 & 23 13 58.2 & -42 43 59 & 1.185@xmath2480.085 & 0.20@xmath2480.06 & 183 & 1.85 & 5.337@xmath249 & 23.50 & 3.211 & s1101 & 0.0564 & 1 & 4 + rxcj2315.7 - 3746 & 23 15 44.2 & -37 46 53 & 0.295@xmath2480.060 & 0.54@xmath2480.17 & 128 & 1.50 & 7.709@xmath249 & 5.82 & 7.852 & am 2312 - 380/a3984 & 0.1786 & 1 & 3 + rxcj2315.7 - 0222 & 23 15 46.0 & -02 22 38 & 0.425@xmath2480.034 & 0.80@xmath2480.18 & 337 & 4.18 & 2.024@xmath249 & 8.73 & 0.268 & ngc 7556 & 0.0267 & 1 & 2 + rxcj2316.1 - 2027 & 23 16 08.0 & -20 27 19 & 0.373@xmath2480.068 & 0.48@xmath2480.15 & 95 & 2.07 & 4.589@xmath249 & 7.43 & 2.222 & a2566 & 0.0834 & 1 & 2 + rxcj2321.4 - 2312 & 23 21 25.9 & -23 12 31 & 0.164@xmath2480.042 & 0.42@xmath2480.21 & 98 & 1.96 & 5.100@xmath250 & 3.26 & 4.883 & a2580 & 0.1870 & 16 & 1 + rxcj2321.5 - 4154 & 23 21 33.0 & -41 54 04 & 0.440@xmath2480.059 & 0.40@xmath2480.12 & 143 & 2.15 & 5.178@xmath249 & 8.80 & 2.983 & a3998 & 0.0889 & 1 & 2 + rxcj2325.3 - 1207 & 23 25 20.1 & -12 07 40 & 1.016@xmath2480.059 & 0.44@xmath2480.05 & 322 & 2.50 & 9.100@xmath250 & 20.60 & 6.365 & a2597 & 0.0852 & 8 & 3 + rxcj2326.2 - 2406 & 23 26 16.0 & -24 06 13 & 0.105@xmath2480.023 & 0.73@xmath2480.26 & 185 & 1.82 & 2.888@xmath249 & 2.06 & 0.692 & apmcc895 & 0.0880 & 14 & ? + rxcj2331.1 - 3630 & 23 31 11.3 & -36 30 22 & 0.562@xmath2480.124 & 0.54@xmath2480.17 & 51 & 1.44 & 6.005@xmath249 & 11.00 & 4.279 & a4010 & 0.0954 & 8 & 8 + rxcj2336.2 - 3135 & 23 36 14.9 & -31 35 55 & 0.448@xmath2480.057 & 0.16@xmath2480.19 & 125 & 1.18 & 3.712@xmath249 & 8.67 & 1.320 & s1136 & 0.0594 & 26,27 & 2 + rxcj2337.6 + 0016 & 23 37 40.9 & + 00 16 34 & 0.169@xmath2480.033 & 0.55@xmath2480.15 & 378 & 3.85 & 8.948@xmath249 & 3.56 & 11.264 & a2631 & 0.2753 & 6 & 1 + rxcj2338.2 - 4944 & 23 38 12.7 & -49 44 43 & 0.094@xmath2480.024 & 0.75@xmath2480.28 & 175 & 1.89 & 5.000@xmath249 & 1.87 & & & & & + rxcj2341.2 - 0901 & 23 41 17.3 & -09 01 34 & 0.168@xmath2480.044 & 0.95@xmath2480.16 & 346 & 2.55 & 8.182@xmath249 & 3.41 & 9.067 & a2645 & 0.2517 & 8 & 5 + rxcj2344.2 - 0422 & 23 44 16.2 & -04 22 07 & 0.425@xmath2480.034 & 0.69@xmath2480.07 & 335 & 3.54 & 4.691@xmath249 & 8.81 & 2.340 & & 0.0786 & 1 & 2 + rxcj2344.4 - 2153 & 23 44 28.9 & -21 53 34 & 0.115@xmath2480.028 & 0.61@xmath2480.23 & 198 & 1.88 & 3.603@xmath249 & 2.27 & 1.234 & a2655 & 0.1122 & 1 & 3 + rxcj2346.7 - 1028 & 23 46 42.4 & -10 28 11 & 0.129@xmath2480.024 & 0.38@xmath2480.20 & 340 & 2.78 & 5.872@xmath249 & 2.63 & 4.093 & a2661 & 0.1911 & 3 & 2 + rxcj2347.4 - 0218 & 23 47 24.5 & -02 18 49 & 0.180@xmath2480.070 & 1.22@xmath2480.23 & 343 & 3.62 & 1.197@xmath249 & 3.08 & 0.065 & hcg 97 & 0.0221 & 1 & 2 + rxcj2347.7 - 2808 & 23 47 43.2 & -28 08 32 & 2.546@xmath2480.097 & 0.35@xmath2480.03 & 310 & 1.55 & 4.250@xmath249 & 49.90 & 1.843 & a4038 & 0.0293 & 8 & 44 + rxcj2348.2 - 2849 & 23 48 13.0 & -28 49 01 & 0.103@xmath2480.021 & 0.09@xmath2480.16 & 309 & 1.45 & 6.077@xmath249 & 2.03 & 4.408 & a4041/s1151 dc & 0.2260 & 18 & ? + rxcj2350.1 - 0339 & 23 50 09.1 & -03 39 10 & 0.125@xmath2480.026 & 1.00@xmath2480.23 & 340 & 3.65 & 4.733@xmath249 & 2.60 & 2.399 & a2664 & 0.1466 & 6 & 2 + rxcj2351.6 - 2605 & 23 51 40.5 & -26 05 09 & 0.468@xmath2480.047 & 0.48@xmath2480.07 & 308 & 1.66 & 11.318@xmath249 & 9.29 & 19.901 & a2667 & 0.2264 & 1 & 1 + rxcj2354.2 - 1024 & 23 54 13.5 & -10 24 53 & 0.473@xmath2480.057 & 0.68@xmath2480.08 & 310 & 2.92 & 3.900@xmath250 & 9.62 & 2.403 & a2670 & 0.0761 & 8 & 220 + rxcj2355.1 - 2834 & 23 55 08.6 & -28 34 27 & 0.082@xmath2480.020 & -.09@xmath2480.16 & 300 & 1.49 & 5.000@xmath249 & 1.61 & & a4054 & & & + rxcj2356.0 - 0129 & 23 56 02.4 & -01 29 43 & 0.160@xmath2480.028 & -.01@xmath2480.13 & 367 & 3.58 & 5.000@xmath249 & 3.32 & & & & & + rxcj2357.0 - 3445 & 23 57 02.4 & -34 45 45 & 1.688@xmath2480.093 & 0.30@xmath2480.05 & 255 & 1.10 & 3.500@xmath250 & 32.60 & 3.404 & a4059 & 0.0492 & 1 & 2 + rxcj2359.9 - 3928 & 23 59 56.8 & -39 28 53 & 0.258@xmath2480.047 & 0.56@xmath2480.20 & 152 & 1.33 & 4.595@xmath249 & 5.03 & 2.229 & a4068 & 0.1016 & 8 & 2 + rxcj0000.4 - 0237 & 00 00 24.7 & -02 37 30 & 0.092@xmath2480.019 & 1.00@xmath2480.37 & 372 & 3.56 & 1.460@xmath249 & 1.77 & 0.110 & & 0.0379 & 24 & 1 + rxcj0001.6 - 1540 & 00 01 39.0 & -15 40 52 & 0.116@xmath2480.021 & 1.00@xmath2480.27 & 337 & 2.24 & 3.960@xmath249 & 2.32 & 1.552 & & 0.1246 & 1 & 3 + rxcj0003.1 - 3555 & 00 03 11.1 & -35 55 21 & 0.400@xmath2480.071 & 0.29@xmath2480.14 & 142 & 1.11 & 3.046@xmath249 & 7.70 & 0.799 & a2717 & 0.0490 & 3 & 56 + rxcj0003.2 - 0605 & 00 03 12.0 & -06 05 12 & 0.218@xmath2480.027 & 0.71@xmath2480.14 & 350 & 3.16 & 8.577@xmath249 & 4.50 & 10.167 & a2697 & 0.2320 & 2 & ? + rxcj0003.8 + 0203 & 00 03 50.9 & + 02 03 45 & 0.205@xmath2480.039 & 0.83@xmath2480.12 & 362 & 3.00 & 4.185@xmath249 & 4.18 & 1.776 & a2700 & 0.0994 & 1 & 2 + rxcj0006.0 - 3443 & 00 06 04.8 & -34 43 26 & 0.301@xmath2480.041 & 0.23@xmath2480.13 & 226 & 1.16 & 5.381@xmath249 & 5.86 & 3.276 & a2721 & 0.1143 & 8 & 72 + rxcj0007.4 - 2809 & 00 07 25.3 & -28 09 21 & 0.103@xmath2480.022 & 0.43@xmath2480.27 & 313 & 1.77 & 5.000@xmath249 & 2.04 & & a2726 & & & + rxcj0011.3 - 2851 & 00 11 19.5 & -28 51 41 & 0.612@xmath2480.055 & 0.39@xmath2480.09 & 254 & 1.84 & 4.426@xmath249 & 12.10 & 2.034 & a2734 & 0.0625 & 3 & 80 + rxcj0012.9 - 0853 & 00 12 54.7 & -08 53 03 & 0.090@xmath2480.020 & 0.32@xmath2480.17 & 336 & 3.32 & 5.000@xmath249 & 1.86 & & & & & + rxcj0013.6 - 1930 & 00 13 38.2 & -19 30 08 & 0.306@xmath2480.035 & 0.48@xmath2480.11 & 313 & 2.04 & 4.699@xmath249 & 6.09 & 2.355 & a13 & 0.0949 & 1 & 4 + rxcj0014.3 - 3022 & 00 14 20.1 & -30 22 33 & 0.250@xmath2480.047 & 0.23@xmath2480.15 & 122 & 1.65 & 11.184@xmath249 & 4.96 & 19.331 & a2744 & 0.3067 & 8 & 34 + rxcj0015.5 - 2346 & 00 15 31.8 & -23 46 13 & 0.175@xmath2480.032 & 0.36@xmath2480.19 & 277 & 2.45 & 2.814@xmath249 & 3.48 & 0.647 & a14 & 0.0655 & 3 & 9 + rxcj0016.3 - 3121 & 00 16 19.8 & -31 21 55 & 0.178@xmath2480.032 & 0.00@xmath2480.14 & 296 & 1.88 & 3.292@xmath249 & 3.50 & 0.981 & a2751 & 0.0805 & 20 & 1 + rxcj0017.5 - 3511 & 00 17 34.9 & -35 11 09 & 0.166@xmath2480.033 & 0.28@xmath2480.17 & 354 & 1.31 & 3.686@xmath249 & 3.22 & 1.304 & a2755 & 0.0969 & 8 & 18 + rxcj0019.0 - 2026 & 00 19 03.9 & -20 26 17 & 0.147@xmath2480.027 & 0.42@xmath2480.17 & 322 & 1.78 & 8.314@xmath249 & 2.93 & 9.432 & s26 & 0.2773 & 1 & 1 + rxcj0020.5 - 4913 & 00 20 34.1 & -49 13 40 & 0.124@xmath2480.030 & 0.77@xmath2480.25 & 151 & 2.12 & 2.620@xmath249 & 2.44 & 0.536 & a2764 & 0.0711 & 3 & 19 + rxcj0020.7 - 2542 & 00 20 43.3 & -25 42 53 & 0.289@xmath2480.035 & 0.48@xmath2480.12 & 281 & 2.26 & 6.360@xmath249 & 5.81 & 4.921 & a22 & 0.1410 & 3 & 3 + rxcj0024.0 - 1704 & 00 24 03.6 & -17 04 32 & 0.088@xmath2480.019 & 0.63@xmath2480.24 & 314 & 2.00 & 4.956@xmath249 & 1.75 & 2.680 & a2768 & 0.1890 & 1 & 2 + rxcj0025.5 - 3302 & 00 25 32.2 & -33 02 53 & 0.460@xmath2480.045 & 0.37@xmath2480.09 & 328 & 1.69 & 3.215@xmath249 & 9.00 & 0.921 & s41 & 0.0487 & 8 & 2 + rxcj0028.6 - 2338 & 00 28 39.5 & -23 38 27 & 0.245@xmath2480.033 & 0.55@xmath2480.13 & 294 & 1.82 & 4.933@xmath249 & 4.85 & 2.650 & a42 & 0.1129 & 8 & 5 + rxcj0030.6 - 2410 & 00 30 37.6 & -24 10 53 & 0.094@xmath2480.019 & 0.60@xmath2480.22 & 314 & 1.51 & 3.927@xmath249 & 1.84 & 1.516 & a47 & 0.1382 & 3 & 1 + rxcj0033.8 - 0750 & 00 33 51.6 & -07 50 29 & 0.087@xmath2480.021 & 0.56@xmath2480.24 & 270 & 3.50 & 5.000@xmath249 & 1.80 & & a56 & & & + rxcj0034.2 - 0204 & 00 34 13.9 & -02 04 31 & 0.412@xmath2480.033 & 0.67@xmath2480.11 & 597 & 2.82 & 4.763@xmath249 & 8.38 & 2.433 & sh518 & 0.0822 & 11 & 1 + rxcj0035.6 + 0138 & 00 35 36.3 & + 01 38 20 & 0.105@xmath2480.022 & 0.75@xmath2480.23 & 291 & 2.66 & 5.000@xmath249 & 2.13 & & & & & + rxcj0037.4 - 2831 & 00 37 27.0 & -28 31 52 & 0.113@xmath2480.023 & 0.39@xmath2480.21 & 326 & 1.73 & 3.387@xmath249 & 2.22 & 1.057 & a2798 & 0.1050 & 3 & 20 + rxcj0041.8 - 0918 & 00 41 50.0 & -09 18 12 & 3.554@xmath2480.114 & 0.63@xmath2480.03 & 328 & 3.58 & 6.200@xmath250 & 74.00 & 9.810 & a85 & 0.0556 & 12 & 116 + rxcj0042.1 - 2832 & 00 42 08.8 & -28 32 12 & 0.498@xmath2480.049 & 0.28@xmath2480.07 & 325 & 1.49 & 6.768@xmath249 & 9.82 & 5.726 & a2811 & 0.1170 & 8 & 32 + rxcj0043.5 - 0443 & 00 43 34.6 & -04 43 06 & 0.119@xmath2480.018 & 0.64@xmath2480.13 & 446 & 3.63 & 5.000@xmath249 & 2.48 & & & & & + rxcj0048.6 - 2114 & 00 48 38.0 & -21 14 50 & 0.217@xmath2480.032 & 0.09@xmath2480.13 & 339 & 1.52 & 2.759@xmath249 & 4.20 & 0.614 & a2824 & 0.0581 & 17 & 7 + rxcj0049.3 - 2931 & 00 49 23.8 & -29 31 27 & 0.265@xmath2480.042 & 0.40@xmath2480.13 & 323 & 1.80 & 5.172@xmath249 & 5.25 & 2.974 & s84 & 0.1150 & 8 & 20 + rxcj0051.1 - 4833 & 00 51 11.2 & -48 33 35 & 0.123@xmath2480.027 & 0.63@xmath2480.24 & 204 & 2.12 & 5.646@xmath249 & 2.46 & 3.684 & a2830 & 0.1873 & 1 & 2 + rxcj0051.3 - 2831 & 00 51 19.2 & -28 31 10 & 0.127@xmath2480.023 & 0.40@xmath2480.16 & 324 & 1.63 & 3.750@xmath249 & 2.49 & 1.359 & a2829 & 0.1125 & 13 & 6 + rxcj0056.0 - 3732 & 00 56 01.0 & -37 32 45 & 0.425@xmath2480.034 & 0.55@xmath2480.21 & 351 & 2.59 & 8.538@xmath249 & 8.65 & 10.102 & & 0.1663 & 2 & 11 + rxcj0056.3 - 0112 & 00 56 18.3 & -01 12 53 & 1.644@xmath2480.087 & 0.74@xmath2480.06 & 321 & 3.10 & 5.100@xmath250 & 33.70 & 2.855 & a119 & 0.0444 & 17 & 23 + rxcj0058.4 - 1425 & 00 58 28.2 & -14 25 39 & 0.108@xmath2480.018 & 0.17@xmath2480.15 & 492 & 1.87 & 3.105@xmath249 & 2.12 & 0.840 & a123 & 0.0957 & 5 & 2 + rxcj0102.7 - 2152 & 01 02 42.3 & -21 52 46 & 0.970@xmath2480.065 & 0.33@xmath2480.06 & 268 & 1.48 & 3.800@xmath250 & 18.90 & 2.574 & a133 & 0.0562 & 8 & 9 + rxcj0104.5 - 2400 & 01 04 35.0 & -24 00 00 & 0.119@xmath2480.021 & 0.64@xmath2480.19 & 335 & 1.61 & 4.672@xmath249 & 2.34 & 2.322 & a140 & 0.1520 & 8 & 12 + rxcj0105.5 - 2439 & 01 05 35.4 & -24 39 33 & 0.196@xmath2480.031 & 0.32@xmath2480.12 & 326 & 1.61 & 8.018@xmath249 & 3.88 & 8.635 & a141 & 0.2300 & 8 & 1 + rxcj0106.8 - 0229 & 01 06 51.5 & -02 29 18 & 0.152@xmath2480.032 & 0.91@xmath2480.23 & 423 & 4.10 & 6.386@xmath249 & 3.21 & 4.972 & a145 & 0.1909 & 5 & 1 + rxcj0107.8 - 3643 & 01 07 49.4 & -36 43 44 & 0.180@xmath2480.024 & 0.82@xmath2480.16 & 378 & 1.94 & 4.642@xmath249 & 3.57 & 2.285 & a2871 & 0.1221 & 2 & 11 + rxcj0108.2 + 0210 & 01 08 12.8 & + 02 10 48 & 0.230@xmath2480.037 & 0.59@xmath2480.15 & 415 & 3.01 & 2.342@xmath249 & 4.60 & 0.398 & a147 & 0.0447 & 3 & 11 + rxcj0108.5 - 4021 & 01 08 33.2 & -40 21 00 & 0.150@xmath2480.035 & 0.40@xmath2480.23 & 377 & 3.04 & 4.898@xmath249 & 3.07 & 2.610 & a2874 & 0.1408 & 13 & 6 + rxcj0108.8 - 1524 & 01 08 49.3 & -15 24 30 & 0.358@xmath2480.035 & 0.29@xmath2480.08 & 446 & 1.69 & 3.149@xmath249 & 7.00 & 0.872 & a151 & 0.0537 & 8 & 37 + rxcj0108.9 - 1537 & 01 08 55.5 & -15 37 42 & 0.189@xmath2480.029 & 0.46@xmath2480.09 & 437 & 1.69 & 3.954@xmath249 & 3.72 & 1.541 & & 0.0981 & 4,19 & 9 + rxcj0110.0 - 4555 & 01 10 00.7 & -45 55 24 & 0.756@xmath2480.060 & 0.26@xmath2480.08 & 301 & 2.10 & 3.500@xmath250 & 15.00 & 0.404 & a2877 & 0.0250 & 8 & 109 + rxcj0113.9 - 3145 & 01 13 54.2 & -31 45 07 & 0.080@xmath2480.017 & 0.58@xmath2480.24 & 333 & 2.40 & 1.000@xmath249 & 1.21 & 0.018 & s141 & 0.0184 & 8 & 18 + rxcj0114.7 - 2118 & 01 14 47.3 & -21 18 02 & 0.082@xmath2480.018 & 0.56@xmath2480.24 & 346 & 1.49 & 5.000@xmath249 & 1.61 & & & & & + rxcj0114.9 + 0024 & 01 14 57.5 & + 00 24 20 & 0.520@xmath2480.048 & 0.47@xmath2480.09 & 406 & 3.32 & 2.600@xmath250 & 10.60 & 1.043 & a168 & 0.0477 & 1 & 2 + rxcj0116.1 - 1555 & 01 16 11.8 & -15 55 24 & 0.130@xmath2480.023 & 0.23@xmath2480.17 & 394 & 1.69 & 1.862@xmath249 & 2.46 & 0.214 & scg 16 & 0.0448 & 1 & 2 + rxcj0118.1 - 2658 & 01 18 10.2 & -26 58 18 & 0.193@xmath2480.027 & 0.54@xmath2480.14 & 366 & 1.47 & 7.987@xmath249 & 3.81 & 8.554 & a2895 & 0.2310 & 1 & 1 + rxcj0120.9 - 1351 & 01 20 57.9 & -13 51 19 & 0.425@xmath2480.034 & 0.30@xmath2480.07 & 344 & 1.73 & 3.240@xmath249 & 8.33 & 0.939 & cid 10 & 0.0511 & 1 & 3 + rxcj0121.8 - 2022 & 01 21 48.1 & -20 22 34 & 0.110@xmath2480.020 & 0.70@xmath2480.28 & 465 & 1.45 & 5.000@xmath249 & 2.16 & & a2902 & & & + rxcj0122.2 - 2131 & 01 22 12.4 & -21 31 04 & 0.097@xmath2480.020 & 0.05@xmath2480.17 & 494 & 1.39 & 5.000@xmath249 & 1.90 & & a185 & & & + rxcj0122.9 - 1245 & 01 22 55.1 & -12 45 50 & 0.096@xmath2480.018 & 0.87@xmath2480.25 & 391 & 1.68 & 3.598@xmath249 & 1.88 & 1.229 & a188 & 0.1230 & 3 & 2 + rxcj0125.4 + 0145 & 01 25 29.9 & + 01 45 46 & 0.425@xmath2480.034 & 0.77@xmath2480.15 & 413 & 3.08 & 1.446@xmath249 & 8.03 & 0.107 & ngc 533 & 0.0176 & 2 & 4 + rxcj0125.6 - 0125 & 01 25 38.0 & -01 25 03 & 0.610@xmath2480.086 & 0.84@xmath2480.14 & 431 & 3.56 & 1.900@xmath250 & 12.20 & 0.176 & a194 & 0.0183 & 8 & 81 + rxcj0126.0 - 3758 & 01 26 04.6 & -37 58 36 & 0.105@xmath2480.019 & 0.76@xmath2480.23 & 351 & 1.68 & 2.703@xmath249 & 2.04 & 0.581 & a2911 & 0.0810 & 4 & 31 + rxcj0126.7 - 1810 & 01 26 44.4 & -18 10 20 & 0.103@xmath2480.025 & 0.32@xmath2480.21 & 328 & 1.62 & 5.000@xmath249 & 2.03 & & a197 & & & + rxcj0127.2 - 1746 & 01 27 13.9 & -17 46 17 & 0.109@xmath2480.021 & 0.28@xmath2480.19 & 351 & 1.52 & 4.358@xmath249 & 2.14 & 1.960 & a199 & 0.1459 & 1 & 2 + rxcj0131.7 + 0033 & 01 31 42.4 & + 00 33 43 & 0.088@xmath2480.017 & 0.26@xmath2480.18 & 432 & 2.92 & 2.535@xmath249 & 1.77 & 0.490 & a208 & 0.0798 & 3 & ? + rxcj0131.8 - 1336 & 01 31 53.5 & -13 36 27 & 0.277@xmath2480.029 & 0.33@xmath2480.10 & 453 & 2.27 & 8.513@xmath249 & 5.59 & 9.982 & a209 & 0.2060 & 3 & 2 + rxcj0137.2 - 0912 & 01 37 15.3 & -09 12 10 & 0.425@xmath2480.034 & 0.68@xmath2480.08 & 444 & 2.75 & 2.672@xmath249 & 8.50 & 0.564 & & 0.0392 & 1 & 3 + rxcj0137.4 - 1259 & 01 37 29.2 & -12 59 10 & 0.088@xmath2480.016 & 0.55@xmath2480.18 & 458 & 2.45 & 5.503@xmath249 & 1.78 & 3.462 & a222 & 0.2134 & 8 & 6 + rxcj0137.9 - 1248 & 01 37 56.4 & -12 48 01 & 0.115@xmath2480.018 & 0.60@xmath2480.20 & 454 & 1.84 & 5.935@xmath249 & 2.28 & 4.160 & a223 & 0.2070 & 8 & 4 + rxcj0139.1 - 1915 & 01 39 10.1 & -19 15 34 & 0.083@xmath2480.015 & 0.28@xmath2480.20 & 479 & 1.45 & 5.000@xmath249 & 1.63 & & & & & + rxcj0139.9 - 0555 & 01 39 59.7 & -05 55 23 & 0.084@xmath2480.018 & 0.43@xmath2480.20 & 441 & 2.85 & 5.000@xmath249 & 1.71 & & & & & + rxcj0143.4 - 4614 & 01 43 29.8 & -46 14 20 & 0.082@xmath2480.017 & 0.41@xmath2480.25 & 455 & 2.30 & 5.000@xmath249 & 1.65 & & a2937 & & & + rxcj0144.6 - 2213 & 01 44 41.8 & -22 13 43 & 0.097@xmath2480.015 & 0.25@xmath2480.13 & 494 & 1.18 & 6.991@xmath249 & 1.90 & 6.196 & a2938 & 0.2781 & 1 & 1 + rxcj0148.2 - 3155 & 01 48 14.8 & -31 55 08 & 0.105@xmath2480.016 & 0.37@xmath2480.17 & 521 & 1.53 & 3.400@xmath250 & 2.05 & & a2943 & & & + rxcj0148.3 - 0406 & 01 48 22.2 & -04 06 13 & 0.104@xmath2480.024 & 0.25@xmath2480.19 & 255 & 2.59 & 5.000@xmath249 & 2.10 & & & & & + rxcj0152.5 - 2853 & 01 52 32.8 & -28 53 32 & 0.084@xmath2480.016 & 0.16@xmath2480.16 & 391 & 1.51 & 5.000@xmath249 & 1.65 & & & & & + rxcj0152.7 + 0100 & 01 52 43.6 & + 01 00 58 & 0.151@xmath2480.022 & 0.42@xmath2480.13 & 407 & 2.84 & 7.309@xmath249 & 3.09 & 6.900 & a267 dc & 0.2300 & 9 & 1 + rxcj0152.9 - 1345 & 01 52 59.8 & -13 45 07 & 0.425@xmath2480.034 & 0.25@xmath2480.13 & 462 & 1.69 & 1.000@xmath249 & 6.30 & 0.009 & ngc 720 & 0.0057 & 15 & 1 + rxcj0153.5 - 0118 & 01 53 32.0 & -01 18 44 & 0.123@xmath2480.030 & 0.54@xmath2480.17 & 433 & 2.66 & 7.025@xmath249 & 2.50 & 6.277 & & 0.2438 & 1 & 2 + rxcj0157.5 - 0549 & 01 57 30.5 & -05 49 32 & 0.164@xmath2480.024 & 0.55@xmath2480.19 & 414 & 2.31 & 4.681@xmath249 & 3.29 & 2.332 & a281 dc & 0.1285 & 2 & 2 + rxcj0158.4 - 0146 & 01 58 28.4 & -01 46 51 & 0.116@xmath2480.020 & 0.48@xmath2480.17 & 406 & 2.55 & 4.948@xmath249 & 2.34 & 2.672 & a286 & 0.1632 & 1 & 1 + rxcj0201.7 - 0211 & 02 01 44.3 & -02 11 58 & 0.209@xmath2480.024 & 0.52@xmath2480.11 & 406 & 2.57 & 7.327@xmath249 & 4.25 & 6.941 & a291 & 0.1965 & 9 & 1 + rxcj0202.3 - 4447 & 02 02 18.8 & -44 47 37 & 0.124@xmath2480.020 & 0.31@xmath2480.18 & 465 & 2.70 & 2.359@xmath249 & 2.46 & 0.405 & & 0.0616 & 1 & 2 + rxcj0202.3 - 0107 & 02 02 20.6 & -01 07 06 & 0.191@xmath2480.026 & 0.22@xmath2480.13 & 405 & 2.58 & 2.103@xmath249 & 3.76 & 0.298 & a295 & 0.0428 & 5 & 3 + rxcj0206.4 - 1453 & 02 06 29.9 & -14 53 37 & 0.149@xmath2480.027 & 0.52@xmath2480.16 & 274 & 2.45 & 5.192@xmath249 & 3.00 & 3.003 & a305 & 0.1529 & 1 & 2 + rxcj0208.0 - 1537 & 02 08 00.2 & -15 37 05 & 0.085@xmath2480.019 & 0.92@xmath2480.24 & 286 & 2.38 & 5.000@xmath249 & 1.71 & & & & & + rxcj0211.4 - 4017 & 02 11 25.2 & -40 17 09 & 0.166@xmath2480.019 & 0.54@xmath2480.14 & 583 & 1.43 & 3.859@xmath249 & 3.24 & 1.457 & a2984 & 0.1022 & 8 & 6 + rxcj0212.8 - 4707 & 02 12 53.9 & -47 07 58 & 0.197@xmath2480.026 & 0.35@xmath2480.18 & 517 & 2.12 & 4.598@xmath249 & 3.93 & 2.232 & a2988 & 0.1150 & 25 & ? + rxcj0213.9 - 0253 & 02 13 56.6 & -02 53 53 & 0.101@xmath2480.022 & 0.46@xmath2480.20 & 350 & 2.16 & 5.000@xmath249 & 2.02 & & & & & + rxcj0214.6 - 0433 & 02 14 41.1 & -04 33 48 & 0.106@xmath2480.021 & 0.68@xmath2480.30 & 357 & 2.21 & 4.181@xmath249 & 2.12 & 1.772 & a329 & 0.1393 & 1 & 2 + rxcj0216.3 - 4816 & 02 16 19.1 & -48 16 23 & 0.140@xmath2480.019 & 0.85@xmath2480.23 & 542 & 2.95 & 5.573@xmath249 & 2.86 & 3.568 & a2998 & 0.1709 & 1 & 1 + rxcj0216.7 - 4749 & 02 16 43.3 & -47 49 31 & 0.180@xmath2480.030 & 0.89@xmath2480.28 & 523 & 2.95 & 2.809@xmath249 & 3.63 & 0.644 & s239 & 0.0640 & 23 & 1 + rxcj0218.3 - 3141 & 02 18 19.6 & -31 41 34 & 0.099@xmath2480.016 & 0.28@xmath2480.14 & 486 & 1.87 & 5.000@xmath249 & 1.96 & & & & & + rxcj0219.1 - 1724 & 02 19 08.3 & -17 24 40 & 0.085@xmath2480.016 & 0.82@xmath2480.23 & 390 & 2.99 & 5.000@xmath249 & 1.74 & & & & & + rxcj0220.9 - 3829 & 02 20 56.8 & -38 29 02 & 0.425@xmath2480.034 & 0.74@xmath2480.17 & 558 & 1.85 & 10.965@xmath249 & 8.48 & 18.532 & & 0.2287 & 2 & 6 + rxcj0225.1 - 2928 & 02 25 10.2 & -29 28 23 & 0.425@xmath2480.034 & 0.63@xmath2480.15 & 445 & 1.70 & 3.710@xmath249 & 8.35 & 1.327 & & 0.0607 & 1 & 2 + rxcj0225.8 - 4154 & 02 25 53.7 & -41 54 19 & 0.258@xmath2480.029 & 0.29@xmath2480.10 & 460 & 2.14 & 8.693@xmath249 & 5.19 & 10.503 & a3017 & 0.2195 & 1 & 1 + rxcj0227.2 - 2851 & 02 27 12.6 & -28 51 18 & 0.071@xmath2480.018 & 0.59@xmath2480.21 & 343 & 1.55 & 5.003@xmath249 & 1.40 & 2.743 & edcc651 & 0.2138 & 2 & ? + rxcj0229.3 - 3332 & 02 29 22.3 & -33 32 13 & 0.198@xmath2480.026 & 0.34@xmath2480.12 & 471 & 2.07 & 3.352@xmath249 & 3.92 & 1.029 & apmcc269 & 0.0779 & 1 & 2 + rxcj0229.9 - 1316 & 02 29 56.4 & -13 16 12 & 0.127@xmath2480.026 & 0.25@xmath2480.20 & 312 & 2.04 & 5.000@xmath249 & 2.53 & & & & & + rxcj0230.8 - 3305 & 02 30 51.2 & -33 05 49 & 0.166@xmath2480.026 & 0.71@xmath2480.24 & 456 & 1.96 & 3.116@xmath249 & 3.27 & 0.848 & a3027 & 0.0774 & 13 & 4 + rxcj0231.7 - 0451 & 02 31 47.3 & -04 51 21 & 0.124@xmath2480.033 & 0.55@xmath2480.24 & 125 & 2.55 & 5.619@xmath249 & 2.51 & 3.640 & a362 & 0.1843 & 1 & 3 + rxcj0231.9 + 0114 & 02 31 57.5 & + 01 14 38 & 0.425@xmath2480.034 & 0.97@xmath2480.26 & 227 & 2.56 & 1.714@xmath249 & 8.12 & 0.171 & & 0.0221 & 1 & 2 + rxcj0232.2 - 4420 & 02 32 17.2 & -44 20 43 & 0.425@xmath2480.034 & 0.14@xmath2480.09 & 369 & 2.61 & 13.183@xmath249 & 8.66 & 28.933 & & 0.2836 & 1 & 2 + rxcj0236.6 - 1923 & 02 36 39.6 & -19 23 05 & 0.171@xmath2480.025 & 0.57@xmath2480.19 & 349 & 2.65 & 4.100@xmath250 & 3.45 & 1.179 & a367 & 0.0891 & 17 & 3 + rxcj0237.4 - 2630 & 02 37 29.0 & -26 30 13 & 0.156@xmath2480.024 & 0.43@xmath2480.16 & 325 & 1.56 & 7.042@xmath249 & 3.08 & 6.305 & a368 & 0.2200 & 2 & 1 + rxcj0241.3 - 2839 & 02 41 20.9 & -28 39 27 & 0.130@xmath2480.021 & 0.10@xmath2480.13 & 423 & 1.56 & 4.300@xmath250 & 2.55 & 5.942 & a3041 & 0.2323 & 1 & 3 + rxcj0244.1 - 2611 & 02 44 06.3 & -26 11 07 & 0.152@xmath2480.027 & 0.79@xmath2480.25 & 286 & 1.74 & 4.726@xmath249 & 3.00 & 2.389 & & 0.1362 & 4,19 & 6 + rxcj0245.2 - 4627 & 02 45 12.6 & -46 27 14 & 0.114@xmath2480.017 & 0.66@xmath2480.22 & 539 & 3.15 & 2.986@xmath249 & 2.32 & 0.757 & a3047 & 0.0868 & 1 & 1 + rxcj0248.0 - 0332 & 02 48 02.6 & -03 32 09 & 0.229@xmath2480.042 & 0.49@xmath2480.16 & 224 & 4.12 & 7.521@xmath249 & 4.85 & 7.396 & a383 & 0.1899 & 1 & 1 + rxcj0248.2 - 0216 & 02 48 16.3 & -02 16 44 & 0.220@xmath2480.033 & 0.66@xmath2480.22 & 117 & 3.65 & 8.748@xmath249 & 4.60 & 10.743 & a384 & 0.2360 & 2 & ? + rxcj0249.6 - 3111 & 02 49 36.9 & -31 11 19 & 0.469@xmath2480.040 & 1.31@xmath2480.27 & 419 & 1.80 & 1.837@xmath249 & 8.89 & 0.206 & s301 & 0.0232 & 8 & 34 + rxcj0251.4 - 2456 & 02 51 28.8 & -24 56 58 & 0.144@xmath2480.023 & 0.04@xmath2480.12 & 365 & 1.90 & 4.500@xmath250 & 2.85 & 1.511 & a389 & 0.1111 & 8 & 6 + rxcj0251.7 - 4109 & 02 51 42.3 & -41 09 00 & 0.102@xmath2480.018 & 0.00@xmath2480.17 & 496 & 1.86 & 5.000@xmath249 & 2.02 & & & & & + rxcj0252.8 - 0116 & 02 52 50.1 & -01 16 28 & 0.425@xmath2480.034 & 1.34@xmath2480.23 & 135 & 5.25 & 1.833@xmath249 & 8.84 & 0.205 & ngc 1132 & 0.0232 & 10 & 1 + rxcj0258.2 - 2105 & 02 58 15.0 & -21 05 29 & 0.114@xmath2480.023 & 0.57@xmath2480.21 & 287 & 1.92 & 3.859@xmath249 & 2.25 & 1.458 & a3073 & 0.1226 & 1 & 1 + rxcj0303.2 - 2735 & 03 03 14.8 & -27 35 59 & 0.122@xmath2480.028 & 0.25@xmath2480.20 & 266 & 1.55 & 5.000@xmath249 & 2.40 & & a3082 & & & + rxcj0303.3 + 0155 & 03 03 21.3 & + 01 55 35 & 0.219@xmath2480.033 & 0.83@xmath2480.15 & 222 & 7.82 & 6.421@xmath249 & 5.08 & 5.062 & a409 & 0.1530 & 9 & 1 + rxcj0304.0 - 3656 & 03 04 06.0 & -36 56 32 & 0.151@xmath2480.023 & 0.76@xmath2480.23 & 384 & 1.95 & 6.958@xmath249 & 3.01 & 6.120 & a3084 & 0.2192 & 2 & ? + rxcj0307.0 - 2840 & 03 07 04.0 & -28 40 15 & 0.197@xmath2480.034 & 0.38@xmath2480.15 & 237 & 1.36 & 8.674@xmath249 & 3.88 & 10.434 & a3088 & 0.2534 & 7 & ? + rxcj0311.4 - 2654 & 03 11 25.3 & -26 54 00 & 0.154@xmath2480.078 & 0.52@xmath2480.22 & 210 & 1.56 & 2.722@xmath249 & 2.98 & 0.593 & a3094/edcc736 dc & 0.0677 & 3 & 67 + rxcj0313.6 - 3817 & 03 13 38.0 & -38 17 55 & 0.119@xmath2480.023 & 0.20@xmath2480.22 & 321 & 2.03 & 2.905@xmath249 & 2.34 & 0.704 & a3098 & 0.0833 & 3 & 6 + rxcj0314.3 - 4525 & 03 14 19.6 & -45 25 24 & 0.401@xmath2480.038 & 0.45@xmath2480.10 & 313 & 3.57 & 4.252@xmath249 & 8.31 & 1.845 & a3104 & 0.0718 & 2 & 4 + rxcj0317.7 - 4848 & 03 17 46.4 & -48 48 42 & 0.097@xmath2480.016 & 0.56@xmath2480.23 & 487 & 2.06 & 5.000@xmath249 & 1.93 & & a3113 & & & + rxcj0317.9 - 4414 & 03 17 58.2 & -44 14 16 & 1.447@xmath2480.056 & 0.30@xmath2480.04 & 481 & 2.53 & 4.100@xmath250 & 29.10 & 7.130 & a3112 & 0.0754 & 1 & 2 +

the tmt project conducted an extensive site monitoring campaign to identify the most suitable site to host the observatory . knowing the vertical distribution of the optical turbulence strength ( ts ) , as expressed by the structure constant of the refractive index @xmath0 , in earth s atmosphere the turbulence profile ( tp or @xmath1 ) is of great importance for modern astronomical observing techniques . especially adaptive optics ( ao ) instrumentation in its various forms relies in its inital design on a good assumption of the typical altitude distribution of the ts above the telescope . ao will become an important observing technique of the future extremely large telescopes ( elts ) and it is therefore essential to obtain measurements of the tps above the elt candidate sites . such measurements can in principle be obtained in - situ by means of balloon borne turbulence sensors ( @xcite ) . but a continuous monitoring of the tp using balloons would be very costly . therefore , site testing programs use remote sensing techniques to obtain measures of the total ts . the instruments used during such campaigns have to be rugged and should only require small size telescopes . these requirements limited early site testing programes to measurements of the total seeing only and were first based on visual observations ( @xcite ) , photoelectric ( @xcite ) or photographic techniques ( @xcite ) . a radical improvement of measuring the total seeing was obtained by the development of the differential image motion monitor ( dimm , @xcite ) . remote sensing of the actual tp based on measurements of stellar scintillations has been done since some time by means of the scintillation detection and ranging ( scidar ) technique ( @xcite ) . this technique , however , requires telescopes with apertures of at least 1 m for being effective . therefore , in recent years several other techniques , like slope detection and ranging ( slodar , @xcite ) or multi aperture scintillation sensor ( mass , @xcite ) have been developed to overcome the requirement for such large apertures and still obtain tps with reasonable altitude coverage and resolution . the tmt site testing employed combined mass - dimm instruments ( @xcite ) to assess the low resolution tp through the entire atmosphere . in order to measure the tp within the lowest few hundred meters above the sites , acoustic turbulence profilers sodars ( sound detection and ranging , @xcite ) were also employed . here we report on the results obtained with these turbulence profilers during the tmt site testing . in section [ sitesdata ] the candidate sites , the deployed instrumentation and data are described . sections [ massbasedprofiles ] and [ sodarbasedprofiles ] show and discuss the observed `` typical '' tps . sections [ seasonalprofiles ] , [ nightprofiles ] and [ meteorelations ] discuss seasonal and nighttime changes and the dependence on the ground wind speed of the ts at the different altitudes . after a preselection based on satellite data and general topographic conditions ( @xcite ) , tmt deployed site testing stations on five candidate sites : 13n on mauna kea ( hawaii ) , san pedro mrtir or spm ( mexico ) , cerro tolar , cerro armazones and cerro tolonchar ( all in northern chile ) . the general properties of these mountains are summarized in table [ sites ] . more detailed information about these sites is given in @xcite and @xcite . these stations have been operated for at least two years on each mountain , with a considerable time span being simultaneous between all stations ( see table [ sites ] ) . all tmt site testing stations were equipped with a custom built 35 cm cassegrain telescope , manufactured by teleskoptechnik halfmann . these telescopes were mounted on 6.5 m tall towers , resulting in an elevation of the telescopes of approximately 7 m above ground . all components of the tmt site testing equipment are summarized in @xcite and @xcite . on each telescope a combined mass - dimm device ( @xcite ) was deployed . the dimm channel of these instruments measures the differential image motion of the star , which then results in the total seeing from the telescope level to the top of the atmosphere ( @xcite ) . the used dimm system and its precision is described in detail by @xcite , who reported the seeing precision the comparability between the tmt dimms to be better than 002 . the mass ( multi aperture scintillation sensor ) reconstructs , by measuring the spatial structure of the flying shadows , a tp at six altitudes @xmath2 km above the telescope ( @xcite ) . the mass tps consist of the integrals of @xmath0 over each altitude bin , weighted with functions which peak at the the layer height @xmath3 and go to zero at the neighboring mass layer altitudes ( thus causing some overlap between neighboring layers ) . therefore , the mass tps are given in @xmath4 , which is equivalent to the seeing of the layer to the power of @xmath5 . from such a profile the seeing which would be seen by an observer 500 m above the telescope to the top of the atmosphere is computed . also the isoplanatic angle @xmath6 can directly be obtained from these profiles . the precision of the strength of the individual mass layers was found by @xcite to be better than @xmath7 m@xmath8 and of the mass seeing to be better than 005 . all turbulence values in this paper are given at a wavelength of @xmath9 nm and corrected for zenith distance . from the difference between mass and dimm seeing data , the seeing originating between the site testing telescope and the first mass layer can be computed by @xmath10 we refer to this lowest layer seeing as ground layer seeing ( gl ) . in order to obtain the gl , only mass data were used for which dimm data exist within 60 seconds . this is usually the case as the dimm channel is used to acquire the target star ( see below ) . when the gl is so weak that its strength is comparable to the measurement accuracy of mass and dimm , the mass seeing can become larger than the dimm seeing . we treat these cases by applying the exponent ( @xmath11 ) in eq . [ glformula ] on the absolute of the difference and reverting the sign of the result in case @xmath12 , same as in @xcite . even though negative gl values are unphysical , not taking them into account would bias the resulting statistics ; in particular the low seeing percentiles . the atmospheric turbulence time constant @xmath13 requires the additional knowledge of the vertical wind speed profile @xmath14 at the time of observation , which is usually unknown however , @xcite found that mass data can in principle be used to infer a @xmath15 value based on the differential exposure time scintillation index . but this @xmath15 has to be properly calibrated to obtain the real @xmath16 and is generally unreliable for @xmath17ms in the case of the tmt mass systems ( @xcite ) . a thorough investigation of the @xmath16 measurements obtained by the tmt site testing is given by @xcite . here , we make use of the uncalibrated @xmath15 data but only for data selection purposes for specific analyses within section [ massbasedprofiles ] . mass and dimm measurements are triggered simultaneously and the dimm integrates the stellar light for 36 sec , whereas the mass samples 60 sec . including processing time , this results in a mass and dimm measurement every 7090 sec . as the dimm ccd is used to acquire the target star , typically for more than 95% of the collected mass data a dimm measurement was taken within 60 sec . however , sometimes the dimm or mass computations failed to provide valid results . this is in part due to data taken under cloudy conditions . as the robotic system operated the telescope whenever the environmental conditions did allow a safe operation , a high data collection efficiency was assured ( @xcite ) . nightly operations commenced approximately one hour after sunset and ceased approximately one hour before sunrise . the tmt site testing project also operated acoustic turbulence profilers sodars ( sound detection and ranging , i.e. , @xcite ) at the candidate sites . these systems emit sound pulses into the atmosphere and obtain from the backscatter signal , the turbulence and wind speed profiles above the site . two sodar systems , sfas and xfas ( both manufactured by scintec ) , were operated . the sfas profiler , measures from 10 m to 200 m above the sites with an altitude resolution of 5 m. xfas profiler range from 40 m to 800 m with a resolution of 20 m. the tmt site testing project operated three sfas and xfas pairs which were rotated among most sites ( no sodar was operated on tolar ) . these systems obtained turbulence and wind speed profiles every 30 min . the details on the operations , calibrations and accuracy are given in @xcite . table [ sites ] indicates the time spans covered by the data shown in this work . figure [ datapts ] shows the amount of simultaneous mass - dimm data collected at each site during each month of a year . even though data collection was ongoing on several sites until mid 2008 , only the data obtained during the indicated periods were considered in the tmt site testing process and are presented in this work . the site testing stations at cerro tolar , mauna kea 13n , spm and cerro tolonchar were dismantled in august 2007 , june 2008 , august 2008 and november 2008 , respectively . at some of these sites , mass - dimm data collection had already been terminated before these dates . the collected mass observations allow the computation of the characteristic turbulence profile above each of the candidate sites . there is , however , no unique definition of a _ characteristic _ vertical turbulence profile that is ideally suited for all applications . in the following we describe two different definitions of representative turbulence profiles . in both methods we include the gl layer strength in these profiles and therefore present only data when mass and dimm data were available simultaneously . as described before , this does not significantly bias or limit the data sample . the first way is to compute the median and mean ts for each mass layer over the entire data set . the statistics 25%ile , median and 75%ile of the individual layer tss are shown in tab . [ profilestable ] . for completeness that table also shows the mean and rms values for each layer . these data represent the statistical properties of each individual layer , but these profiles can not be considered as realistic or representative . in particular because the layers will not be entirely independent from each other . for example , for the lowest mass layers , which are spaced close to each other , one might expect that shearing forces could cause interaction between these layers . the other way to define a tp is to compute the ts of the layers under particular seeing , isoplanatic angle and coherence time conditions . for example , we looked for the typical tp for the @xmath18%ile of the dimm seeing , which is inverse proportional to the fried parameter @xmath19 . to this end , we first identified @xmath20% of all dimm data with values closest to @xmath19 . then , we compute the statistics of the mass data which were taken simultaneous with the selected dimm measurements . in this study we set @xmath21 % , e.g. , @xmath22 % of all data around the 25 , 50 and 75 dimm seeing percentiles were used to obtain the turbulence profile . the so defined profiles for the median @xmath23 and @xmath16 conditions are shown in the panels b , c and d of figure [ massprofiles ] . for reference , tables [ profiledimmseltable ] , [ profiletheta0seltable ] and [ profiletau0seltable ] in the appendix , also show the mean and medians of the profiles computed around the 25 , 50 and 75 percentiles of respective integrated turbulence parameter . it must be noted that even though these profiles represent the typical vertical distribution of turbulence during the times when a certain value of an integrated turbulence parameter is observed , computing that parameter from these profiles does not result in the value of the respective percentile of that parameter computed from the overall data sample . for example in the case of tolar , computing the total seeing from the median profile from mass data selected around the median dimm seeing results in 054 . but the overall dimm seeing statistics results in a median seeing of 063 ( @xcite ) . this is because in the first case the statistics is done before the exponent of @xmath11 is applied to the layer ts , whereas the median seeing is obtained after that step . these turbulence profiles now allow to understand the differences in the integrated turbulence parameters between the sites ( @xcite ) . the three sites in chile show a very similar vertical distribution of the ts . compared to the northern hemisphere sites , 13n on mauna kea and spm , the chilean sites show a weaker ground layer ts by approximately 50% . however , the ts above the chilean sites is stronger than the northern sites at altitudes above 4 km . at these altitudes 13n shows the weakest ts , even less than spm whose 16 km layer is approximately as strong as that of the chilean sites . this explains why spm shows a mass seeing similarly low as 13n but an isoplanatic angle of the same level as the chilean sites . because strong weight is put on the highest layers in the isoplanatic angle calculation ( @xmath24 ) , this makes the 16 km mass layer dominant for this specific parameter and results in 13n showing the largest isoplanatic angle . for the seeing calculation all altitudes are weighted equally , explaining the low mass seeing of spm . the different altitude weighting is also visible in the selected turbulence profiles . the spread between the profiles selected from the various @xmath25 percentiles is larger than the spread between the @xmath26 selected profiles . as example , figure [ t6profiles ] shows the profiles from 13n ( from tables [ profiledimmseltable ] and [ profiletheta0seltable ] ) . it can be seen that the difference between the 25%ile and 75% gl ts median values selected from @xmath25 is approximately a factor of 30 larger than in the case of the @xmath26 selected profiles . again , the strong weight put on the heighest layers ( @xmath24 ) in the @xmath26 calculation , causes that mainly variations of the highest layers will drive the spread of the distribution of the @xmath26 selected tp . the seeing selected profiles reflect the variability of all layers equally , thus explaining the much wider statistical distribution of the seeing selected tps . as the chilean sites show a weaker gl than the northern sites , but the northern sites show a weaker mass seeing than the chilean sites , a crossover between the tps above the northern and southern hemisphere sites has to occur within the lowest 500 m. we show now that the sodar based profiles allow to detect this crossover . in order to compare the sodar results directly to the mass dimm measurements , we selected only mass - dimm observations obtained during the 30 minute sodar data acquisition times . for a thorough discussion on the data acquistion , analysis and calibration of the sodar data we refer to @xcite . figure [ sodarprofiles ] shows the seeing which an observer would see at various altitudes above the ground up to 200 m. to obtain this profile , the xfas profile was integrated from 200 m to 500 m and added to the median mass seeing value during the time of observation . this mass seeing is indicated by the markers to the right of the sodar profiles . this 200 m xfas+mass seeing was then added to the sfas layers , resulting in the profiles shown in figure [ sodarprofiles ] . the profiles were first added and the median was calculated afterwards . the observed dimm seeing taken simultaneously with the sodar observations is indicated by the markers to the left of the profiles , at 7.5 m above ground . even though the sodar profiles start at 10 m above ground , the 10 m datapoint is an integral from 7.5 m to 12.5 m. and , as the sodars were always mounted approximately 1 m to 3 m below the bases of the 6.5 m tall telescope towers , it can be assumed that the sodars do not miss turbulence compared to the dimm . the sodars were typically operated during much less time than the mass - dimm on the sites . these profiles can therefore not be taken as representative on absolute scales for the sites . nevertheless , these profiles allow to draw some interesting conclusions . as can be seen from figure [ sodarprofiles ] , the sodar profiles provide a good match between the mass and dimm observations . this demonstrates the quality of the data analysis and calibration ( @xcite ) . the agreement between sfas , xfas , mass and dimm is better than 10% , showing that these three methods , each based on a different physical concept , measures the ts with good accuracy . the only site where the agreement is worse is spm . this can be attributed to the increased acoustic noise due to the trees surrounding the site , which affect the sodar measurements ( see @xcite for details ) . it can also be seen in figure [ sodarprofiles ] that the profiles obtained at the chilean sites are similar to each other . their tps show that the bulk of the gl is located below 40 m and 60 m at tolonchar and armazones , respectively . in comparison , the northern sites start from a weaker ts at higher altitudes , but below 100 m they increase much stronger than the chilean sites . between approximately 60 m and 90 m , their profiles are crossing the ts of the chilean sites . from the presented turbulence profiles we conclude , that the topography and surface properties of the sites dominate the optical ts over the lowest hundred meters above ground , which are the main drivers for the overall seeing . the site at spm is surrounded by trees and we believe that the therefore increased surface roughness , heat capacity , and release of humidity , cause the enhanced ts ; this raises the effective height of the ground . as 13n is located approximately 150 m below the summit ridge of mauna kea , it falls below the wake region when winds are coming from eastern directions , as reported in @xcite . in comparison , the chilean sites are located on the actual summit of the candidate mountains . the terrain shows low surface roughness and is basically free of vegetation . these sites are are very dry and all of similar soil properties . these arguments hold specifically for armazones and tolar . cerro tolonchar is different in the sense that it represents a summit plateau with an extension of approximately 500 m in north south and 250 m in east west direction and that some sparse vegetation is present . but as the location of the tolonchar site testing equipment is on the northern most edge of this plateau and winds are predominantely coming from the north - west ( @xcite ) , the air above that location resembles closely the free air flow . in recent years , indications for seasonal variability of ts above other observatories were found ( @xcite , @xcite ) . clearly revealing the presence of seasonal variability in the ts at various altitudes will help to shed light on the mechanisms driving optical turbulence and will be essential for validating numerical models which link meteorological modeling to optical ts . to look for seasonal variations of the vertical turbulence distribution throughout the atmosphere in the data we present here , we constructed what we call a standard year . for each month of the year , the median ts of each layer was computed from all measurements taken during the corresponding months . the resulting standard years for the turbulence distribution above the tmt candidate sites are shown in figure [ txcnyear ] . due to technical problems , bad weather and the daylight seasons , not all months contain the same amount of data . to judge the significance of features seen in the median ts data , we again refer to figure [ datapts ] which shows how many mass - dimm profiles were used for the computation of ts in each month of the standard year . several features in figure [ txcnyear ] are coherent at several sites and in various layers to a level of a few @xmath27 m@xmath8 . this indicates that the resolution of the mass is likely better than a few @xmath27 m@xmath8 . as can be seen , the seasonal behaviour of the mass layers below approximately 4 km is anticorrelated with the mass layers at 8 km . the 4 km and 16 km layer exhibit less pronounced seasonal variations at some sites , if any . the 8 km mass layer ( @xmath28 10 km to 12 km a.s.l . ) at the sites falls closest to the altitude domain of the tropopause . the lower mass layers thus trace the troposphere , whereas the 16 km mass layer ( @xmath28 18 km to 20 km a.s.l . ) rather traces the lower stratosphere . that the parts of the atmosphere below and above this altitude behave differently is therefore not unexpected . we will discuss these two altitude regimes separately . first we discuss the chilean sites . above tolar all layers up to 4 km show the lowest ts between may and august , raising to their highest values during the summer months . above armazones this behaviour is less pronounced and only clearly detectable in the 0.5 km and 1 km layer , maybe some weak variation is visible in the 2 km layer as well . tolonchar shows seasonal trends also only in its 0.5 km and 1 km layer . the ts of the 2 km and 4 km mass layers at tolonchar show very low ts during april and may . as at tolonchar only few data are available during the months of may and june , the feature in may represents a single year only and it is questionable whether it can be taken as typical for this site . still , there appears to be a maximum altitude of approximately 6 km asl below which seasonal variations weak ts in winter and strong ts during summer can be detected in the atmosphere above northern chile . at these low altitudes , local terrain effects will play an imporant role . like the proximity of tolar to the pacific ocean which acts as heat reservoir ; and tolonchar rising steeply out of its surrounding arid terrain , ranging already into the westerly winds which dominate at higher elevations ( @xcite ) . but also the large scale atmospheric conditions might be the driver of the ts . the `` altiplano winter '' phenomenon is a result of the shift of the inner tropical convergence zone southwards over the amazonia region . over south america this results in increased low altitude winds and convective , monsoon - like , activity ( @xcite , @xcite ) . during summer the chilean sites are located within the region where the parts of these northerly winds which cross the andes and the anti - cyclonic southerlies meet and turn westwards off the pacific coast . even though wind speeds in the region of the sites are low , one might expect that , as these flows merge , will result in increased large scale wind shear ; in addition to the convectivity of the notherlies . the situation at the sites of spm and 13n is somewhat different to what is observed in chile . the layers up to 2 km , in particular the 0.5 km one , are basically turbulence free until may . the 0.5 km layer then becomes strong ( up to @xmath7 m@xmath8 ) from june onwards . this behaviour does not show a smooth seasonal variation , like temperature or humidity measured at these sites . only the 1 km and 2 km layers above 13n show a rather smooth seasonal variation on a very low level ( few @xmath27 m@xmath29 ) . these variations are opposite to what is observed in chile : strong ts in winter and weak ts during summer . at spm the seasonal variation is not reaching into the 4 km mass layer , whereas 13n might show some faint seasonal trend of this layer . interestingly , the behaviour of the 0.5 km layer does not correspond to the gl . for example , during jja the 0.5 km layer becomes strongest at spm whereas its gl is weakest during these months . both sites are not far to the pacific ocean and we expect it to play an important role in the ts . but their local terrain is quite complex ( vegetation at spm and 13n being not on the actual summit ) , not allowing an easy qualitative assessment of the ts behaviour . for the large scale flows it can be said that both sites are located in the flow of the hawaiian high which extends further south during the summer months . during the winter it shifts towards north and the north - westerly flow in the area of hawaii interacts with the south - easterly wind which is driven by the aleutian low ( @xcite ) . this might explain the seasonal variation in the 1 km and 2 km layers at 13n . of the higher altitudes , the 8 km layer shows the stronger seasonal variation , whereas the 16 km layer shows essentialy no seasonal trend . the very strong seasonal variation of the 8 km layer , of up to @xmath30 m@xmath8 , is the main driver for the seasonal variability of the isoplanatic angle at some of the sites ( @xcite ) . with altitudes of the tmt candidate sites ranging from 2.3 to 4.5 km asl , the 8 km mass layer traces the domain of the jetstream , whose typical altiude is around 200 mb ( corresponding to @xmath3111 km asl ) . an influence of the jetstream activity , or more general of the wind speed , on the ts has been discussed since some time in the literature ( @xcite , @xcite , @xcite ) . to see whether the wind speed is indeed one of the main drivers for the ts at these altitudes , we make use of the vertical wind speed profiles provided by the the ncep / ncar reanalysis project ( @xcite ) . the used locations of the ncep models are given in table [ ncepcoord ] . the ncep wind speed profiles are provided every six hours and at 16 altitudes , ranging from 761 m to 25 km asl . of these , only 13 or 14 levels are above the candidate sites , depending on the sites elevation . to correlate these data with the mass observations , the wind speed profiles have to be convolved with the mass weighting function of the corresponding layer . this was done after interpolating the wind speed profile linearly onto a grid with 100 m vertical resolution . again , we computed a standard year for these wind speed profiles , but only from those data for which mass - dimm observations were taken within @xmath323 hours . figure [ ncepcn2 ] shows the resulting scatter plots for the standard year data of the @xmath33 km layer above the sites . in particular armazones shows an excellent linear correlation . tolar and tolonchar do show a similar increase of the ts with increasing wind speed , however , with more scatter than armazones . in the case of tolonchar this enhanced scatter is in part due to low statistics during certain months ( see above ) . nevertheless , the slope ( increase of ts with wind speed , marked as parameter @xmath34 in fig . [ ncepcn2 ] ) is at the same level at all three sites . on the other hand , above spm and 13n the ts remains at very low levels up to @xmath35 m/s and only the higher wind speeds show a trend towards stronger @xmath36m@xmath8 values . in other words , for the same level of wind speed , the turbulence strength at the northern sites is only about half of what is observed above chile . also the rate of increase is only half of what is observed at the southern sites . we tested similar correlations for the other mass layers but these do not show significant correlations . this confirms the findings by @xcite who showed a direct correlation between the 200 mbar wind speed and the total seeing observed at telescopes on mauna kea and la silla . @xcite deduced a threshold of 20 m / s at the 200 mbar level for good total seeing conditions . this is remarkably similar to our observations taking into account that @xcite had to employ correction factors for the various seeing values and could only use the total seeing . @xcite later found from balloon data a correlation of @xmath16 with the 200 mbar wind speed above pachn . this led to a number of studies of high altitude winds above other existing and potential observatory sites ( e.g. , @xcite , @xcite ) . it is in principle understandable that high wind speeds are more likely to develop stronger ts , but our observations of the weaker increase of ts above the sites of 13n and spm are at first puzzling and are in contrast to the common threshold found in the @xcite data . this could indicate that the mass misses to detect the jetstream above spm and 13n . but as the mass calculates the total seeing from the integral of the reconstructed @xmath4 values and also from the total scintillation , it is possible to test the overall integration scheme of the mass system . we found that both methods deliver values of the total seeing which agree to within 10% : the armazones median mass seeing from the scintillation calulation resulted in 040 , as compared to 043 from the @xmath0 reconstruction ( for 13n mauna kea the numbers read 030 and 034 , respectively ) . we believe that this , in combination with the previously shown agreement with other instruments , demonstrates that no turbulence is acutally missed by the mass . on the other hand , the interpolated wind speed will certainly not be very accurate as the original ncep altitude grid is not dense . we experimented with different interpolation schemes , as well as using the raw wind speed from the closest level to 8 km above site altitude and all resulted in very similar figures . the observed differences between the correlations would then show that the 200 mb wind speed alone is not a good indicator on a global scale for the median ts at these altitudes and that the jetstream above the northern sites during the times of our observations was somewhat less turbulent than above the chilean mountains . this might be caused by the lack of a large continental land mass and a high mountain range , as in the case of the andes mountains . in northern chile , the airflow over the south american continent , which rises steeply out of the sea , can be expected to show enhanced formation of atmospheric waves . such waves can also be forced by differential heating patterns like the proximity of the atacama desert to the pacific ocean . atmospheric waves are commonly thought to result in an increased ts . unfortunately , such waves and their turbulence production can not yet be modeled properly , even though comparisons between models and observations show promising results ( e.g. , @xcite ) . our results might indicate that it is the actual interaction between wave and jet stream which is driving the ts . to investigate the typical behaviour of the ts during the night we computed , similar to the standard year , a standard night . it consists of the median ts of all data collected during each hour after sunset . figure [ txcnnight ] shows the results for the sites . to avoid artefacts introduced by the seasonal variations of the times of sunrise and sunset , the graphs in figure [ txcnnight ] represent the merged behaviour for the hourly medians computed with respect to sunset and sunrise . we found that the curves computed with respect to sunrise and sunset match best when midnight is set to 5 hours and the standard night for which data are available , is therefore 11 hours long . the jumps of the curves at 5 hours in figure [ txcnnight ] are the offset between the curves computed with respect to sunset and sunrise . again , it can be noted that coherent structures are apparent in these graphs on a scale of a few @xmath27 m@xmath8 , consistent with what was found in section [ seasonalprofiles ] . also , the median night time evolution of the ts indicates that the atmosphere above approximately 4 km above the sites is basically decoupled from air below . the high altitude layers remain close to constant throughout the night , whereas the lower layers show a decreasing ts with time . only 13n and tolar exhibit almost no change during the course of the night in the lower layers . at spm the 0.5 km layers at this site shows a decrease of its ts until one two hours before sunrise . the ts in the layers up to 2 km above armazones decrease during the first half of the night and then remain at almost constant values . finally , the air above tolonchar shows a continuous drop of ts throught the night , which maybe only levels off in the 0.5 km layer . the observed behaviour appears similar to a cooling process and the differences between the sites on how the ts evolves below the tropopause , are related to the evolution of wind shear and temperature gradients . we do not have measurements in order to compute the richardson number or any other stability parameter in this altitude range . nevertheless , we make the assumption that the temperature and wind speed measured at the sites are related to their respective gradients throughout the air above the site . this assumption is only strictly true below the height of the boundary layer which , based on the sodar profiles presented earlier , we expect to be in the order of 100 m. figure [ mednightwst ] shows the standard night for the wind speed and temperature , recorded 2 m above ground simultaneous to the tp measurements . it can be seen that the temperature descreases , similarly at all sites . in contrast , the nightly evolution of the wind speed shows remarkable differences between the sites . in particular armazones shows an increase during the second half of the night , whereas tolonchar shows a continuous decrease of the measured wind speed . similar to tolonchar , but at a lower level , is the behaviour at 13n and tolar . note that our wind speed measurements are unreliable at spm due to the site being surrounded by trees . therefore , we will not try to interpret the behaviour at spm . from these observations we suspect that cooling of the terrain during the night is important for the decrease of the ts . this is in particular the case for continental sites such as tolonchar and armazones . we attribute the constant ts at armazones during the second half of the night to increased wind shear . due to its much higher heat capacity and its better efficiency to carry thermal energy below its surface ( @xcite ) , the sea surface temperature of the pacific ocean remains almost constant during night time . this certainly affects the turbulent fluxes within the air above and close to the ocean . we therefore suspect the proximity of the sites of tolar and 13n to the pacific ocean as reason for the almost constant ts observed at these sites . spm being located in the center of the peninsula baja california appears to be rather an inland case . the wind speed plays a crucial role as turbulence driver by introducing shear in the flow . here the turbulence measurements of each mass layer are correlated to the wind speed which was measured 2 m above ground by the automated weather stations . these weather stations were equipped with cup anemometers of type an3 by monitor sensors . in principle we could have also used the wind measurements taken by the csat-3 sonic anemometers which were installed close to the telescopes at 7 m above ground ( @xcite ) . the advantage of sonic anemometer measurements is that mechanical anemometers are prone to problems induced by dust or ice in the mechanisms . however , we found the two systems to agree very well during most of the time . a particular case is tolonchar , where a strong gradient exists between the 2 m and 7 m level , due to the steep terrain drop towards the prevailing wind direction . still , we show here only the aws anemometer data , as the sonic anemometers were installed only during the second half of the site testing program and thus do not cover the full time span of mass - dimm measurements . the panels of figure [ cn2vsws ] show the dependence of the tss of the individual mass layers on the wind speed . the median @xmath4 value was computed in 1 m / s wind speed bins for wind speeds up to the operational limit of the site testing telescope . this limit was typically 15 m / s , only the tolonchar setup was limited to 12 m / s due to the strong gradient mentioned before . we expect that the measured wind speed at 2 m above ground is indicative of the wind speed within the lowest mass layers . for the higher layers this assumption can not be expected to hold , as the wind adjusts here to the geostrophic conditions which differ from the situation close to the ground . nevertheless , some correlation might still be present . to visualize this behaviour figure [ awsncepcorr ] shows the correlation coefficient of each ncep model wind speed with the measured wind speed 2 m above the sites . again , measured wind speeds were simultaneous to mass - dimm observations and the median was calculated for data taken within @xmath323 hours around the ncep model times . the strongest correlations , which is always positive , are found at altitudes slightly above the sites . the correlations break down at altitudes below the sites . as was seen in the overall profiles , as well as in @xcite , the ground layer at all sites shows the strongest ts by up to a factor of a few . this is not unexpected , as the mass - dimm defined ground layer covers ( most of ) the atmospheric boundary layer , i.e. , the layer in which turbulence production is dominated by the interaction between air and ground . shearing forces are strong in this layer as the air speed at surface has to reach zero . also the ground might heat or cool the air above it by means of the sensible heat flux , introducing temperature gradients . as was already shown in @xcite , the gl shows the lowest @xmath37 in the range of wind speeds between approximately 2 m / s and 8 m / s for the chilean sites and approximately 2 m / s and 3 m / s for 13n mauna kea . spm shows a continuous increase of @xmath37 of the gl with increasing wind speed . as that site is surrounded by trees the wind speed measurements , especially at low values , are doubtful . nevertheless , the observed increase of the gl with wind speed appears reasonable also at that site . the behaviour of the 0.5 , 1 and 2 km layers above the sites seems to be dominated by the orography of the surrounding terrain . these layers above tolar remain at the same level of ts up to 10 m / s . at armazones the 0.5 km layer rises strongly , whereas its 1 and 2 km layers rise less steep . these trends are similar to what happens above 13n . above tolonchar the 0.5 km and 1 km layers remain at very low ts , but the 2 km layer rises up to 5 m / s and then shows a quite strong increase by @xmath38 m@xmath8 towards 7 m / s . tolonchar is about 2000 m above the level of the salar de atacama ( a flat dry salt lake ) , which is in the prevailing wind direction and within 20 km from the mountain base . these 20 km are smoothly sloped terrain ranging up 500m-800 m below the tolonchar summit . therefore , one could assume that the wind reaching the summit effectively resembles an altitude of 1000 - 2000 m above the surface and the 0.5 and 1 km mass layers trace the free air flow between 2 and 3 km above the effective terrain . tolar and armazones reach out similarly but not as high and steep as tolonchar . in the case of armazones a mountain range which is slightly higher ( 3400 m asl ) is located in the prevailing wind direction . we speculate that these orographic features cause the onset of waves , driving the ts at these altitudes . this picture seems reasonable also for the 13n site , as it is below the actual summit . as our spm wind speed data are not reliable due to the presence of trees , we rather do not want to fit it into this picture . nevertheless , the increase in ts of the 0.5 km layer appears reasonable . the 8 and 16 km layers show as well an increase in their ts with wind speed at the sites , in particular tolonchar . from figure [ awsncepcorr ] it turns out that tolonchar shows the highest correlation of all sites between the measured and modeled wind speeds . this indicates that its altitude brings it into the regime of the geostrophic westerly flows which are probably better modelled by ncep and are less dependent on the local orography and therefore the correlation is higher . the mass and sodar based turbulence profiles collected at the tmt candidate sites have been presented . the average profiles explain the origin of the observed behaviour of the integrated turbulence parameters . at all sites the turbulence strength of the lowest 500 m the ground layer is found to be the dominating contributor to the overall seeing . this layer is stronger at 13n and spm as compared to the chilean sites . on the other hand , 13n and spm show lower turbulence strengths at higher altitudes , resulting in a weaker mass seeing than at the chilean sites . the sodar measurements , even though these cover much less time than the mass observations , support this finding . they indicate that the altitude dependence of the overall seeing at 13n and spm is steeper than at the chilean sites between approximately 60 m and 100 m above ground . this results in a stronger total seeing below this altitude at the northern sites as compared to the chilean sites . the agreement found between the measurements of mass , dimm and sodar of 10% demonstrates that these techniques do all have a good accuracy to measure @xmath0 . however , we expect the sodar measurements to be less accurate ( @xcite ) than the mass - dimm measurements . the isoplanatic angle differences between the sites depends on the distribution of the high altitude turbulence . in this respect , spm shows similar high altitude turbulence strengths as the chilean sites and therefore a similar isoplanatic angle , despite its mass seeing being comparable to 13n . seasonal and nocturnal variations of the median turbulence strength recorded by the mass have also been shown . these data indicate that the resolution of the mass is in the order of a few @xmath27 m@xmath8 . clear seasonal variations are observed up to 2 km and at 8 km above the sites . the 4 km and 16 km layers show only little or no seasonal trends . the 8 km mass layer turbulence strength was found to correlate well with the wind speeds at these altitudes above the chilean sites . as seasonal turbulence strengths of the 8 km and the lower mass layers are basically anit correlated , it can be expected that ground layer adaptive optics will be most efficient during the summer months at all sites . nocturnal variations are typically confined to regions below the 4 km mass layer . these trends seem to indicate that orographic features and surface energy properties on scales of tens of km are important drivers for the turbulence strength at altitudes below the tropopause . in particular the close proximity of the pacific ocean to some of the sites is dominating the behaviour of the turbulence strengths at altitudes above approximately 200 m. astronomical turbulence profilers provide a powerful tool to sense clear sky atmopsheric turbulence . the collected data are thus not only providing the necessary information for the decision at which site to build tmt and for the design of future astronomical instruments . these data have also to be analysed in view of their relation to meteorological mechanisms . the present study could only provide a brief overview of some aspects of what can be extracted from such data . more detailed studies using climatological and meteorological data , computational models in combination with new observations from turbulence profilers should result in a better understanding of the actual drivers of optical turbulence in earths atmosphere . the tmt project gratefully acknowledges the support of the tmt partner institutions . they are the association of canadian universities for research in astronomy ( acura ) , the california institute of technology and the university of california . this work was supported as well by the gordon and betty moore foundation , the canada foundation for innovation , the ontario ministry of research and innovation , the national research council of canada , the natural sciences and engineering research council of canada , the british columbia knowledge development fund , the association of universities for research in astronomy ( aura ) and the u.s . national science foundation . barletti r. , ceppatelli g. , paterno l. , righini a. , speroni n. , 1977 , , _ 54 _ , 649 birkle k. , elssser h. , neckel th . , schnur g. , schwarze b. , 1976 , , _ 46 _ , 397 butterley t. , wilson r.w . , sarazin m. , 2006 , , _ 369 _ , 835 carrasco e. , avila r. , carramina a. , 2005 , , _ 117 _ , 104 crescenti g.h . , 1997 , bull . 78 _ , 651 cuevas o. , et al . , 2009 , `` meteorological study of an event of occurence of optical turbulence at the tmt candidate sites '' , in optical turbulence meteorology meets astronomy , masciadri e. & sarazin m. ( eds . ) , imperial college press , _ in press _ els s.g . , schck m. , seguel j. , tokovinin a. , kornilov v. , riddle r. , skidmore w. , travouillon t. , vogiatzis k. , blum r. , bustos e. , gregory b. , vasquez j. , walker d. , gillett p. , 2008 , , _ 47 _ , 14 , 2610 garcia - lorenzo b. , fuensalida j.j . , muoz - tuon c. , mendizabal e. , 2005 , , _ 356 _ , 849 irwin j.b . , 1966 , , _ 71 _ , 28 jumper g.y . , murphy e.a . , ruggiero f.h . , roadcap j.r . , ratkokwski a.j . , vernin j. , trinquet h. , 2007 , environ . fluid mech . , _ 7 _ , 351 kalnay e. , kanamitsu m. , kistler r. , et al . , 1996 , bull . soc . , _ 77 _ , 437 kornilov v. , tokovinin a. , shatsky n. , voziakova o. , potanin s. , & safonov b. , 2007 , , _ 382 _ , 1268 masciadri e. & garfias t. , 2001 , , _ 366 _ , 708 masciadri e. & egner s. , 2006 , , _ 118 _ , 1604 riddle r. , schck m. , skidmore w. , 2006 , spie , 6267 riddle r. , schck m. , skidmore w. , els s. , travouillon t. , bustos e. , seguel j. , walker d. , gillett p. , 2009 , _ in prep . _ sarazin m. & roddier f. , 1990 , , _ 227 _ , 294 sarazin m. & tokovinin a. , 2002 , `` the statistics of isoplanatic angle and adaptive optics time constant derived from dimm data '' , in beyond conventional adaptive optics , raggazoni r. , hubin n. , esposito s. ( eds . ) , eso conference and workshop proceedings , vol . 58 , 321 satyamurty p. , nobre c.a . , silva dias p.l . , 1998 , `` meteorology of the tropics south america '' , in ` meteorology of the southern hemisphere ' , koroly d.j & vincent d.g . ( eds . ) , meteorological monographs , vol . 49 , 119 schck m. , els s. , riddle r. , skidmore w. , travouillon t. , 2007 , , _ 31 _ , 10 schck m. , els s. , riddle r. , skidmore w. , travouillon t. , blum r. , bustos e. , chanan g. , djorgovski g. , gillett p. , gregory b. , nelson j. , otrola a. , seguel j. , vasquez j. , walker a. , walker d. , wang l. , 2009 , pasp , _ submitted _ skidmore w. , els s. , schck m. , riddle r. , travouillon t. , seguel j. , bustos e. , walker d. , 2009 , pasp , _ in prep . _ skidmore w. , els s. , riddle r. , schck m. , travouillon t. , 2009a , pasp , _ in prep . _ smith t.m . , & reynolds r.w . , 1998 , j. climate , _ 11 _ , 3320 stock j. , 1964 , `` procedures for location of astronomical observatory sites '' , in ` le choix des sites dobservatoires astronomiques ' , rosch j. ( ed . ) , iau symp . 19 stull r.b . , 1988 , `` an introduction to boundary layer meteorology '' , kluwer academic publishers tokovinin a. , 2002 , , _ 41 _ , 957 tokovinin a. & kornilov v. , 2007 , , _ 381 _ , 1179 travouillon t. , els s. , riddle r. , schck m. , skidmore w. , blum r. , bustos e. , gillett p. , gregory b. , seguel j. , vasquez j. , walker d. , 2009a , , _ in prep . _ travouillon t. , els s. , riddle r. , schck m. , skidmore w. , 2009b , pasp , _ submitted _ walker a. , djorgovski g. , blum r. , schck m. , els s. , riddle r. , skidmore w. , travouillon t. , bustos e. , deyoung d. , nelson j. , seguel j. , vasquez , vogiatzis k. , walker d. , pasp , _ in prep . _ vernin j. & roddier f. , 1973 , josa , _ 63 _ , 270 vincent d.g . , 1998 , `` meteorology of the tropics pacific ocean '' , in ` meteorology of the southern hemisphere ' , koroly d.j & vincent d.g . ( eds . ) , meteorological monographs , vol . 49 , 101 vernin j. , 1986 , spie , _ 628 _ , 142 wang l. , schck m. , chanan g. , et al . 2007 , , _ 46 _ , 25 , 6460 zou j. & lau k .- , 1998 , journal of climate , _ 11 _ , 1020 crrrrrrrrrrr site name & latitude & longitude & elevation & + and telescope # & [ deg n ] & [ deg w ] & [ m a.s.l . ] & mass - dimm & sodar + cerro tolar t1 & -21.9639 & 70.0997 & 2290 & oct . 2003 apr . 2007 & + cerro armazones t2 & -24.5800 & 70.1833 & 3064 & nov . 2004 feb . 2008 & sep . 2005 dec . 2005 + cerro tolonchar t3 & -23.9333 & 67.9750 & 4480 & jan . 2006 feb . 2008 & mar . 2007 feb . 2008 + spm t4 & 31.0456 & 115.4691 & 2830 & oct . 2004 feb . 2008 & may 2006 jun . 2006 + 13n mauna kea t6 & 19.8330 & 155.4810 & 4050 & jun . 2005 feb . 2008 & oct . 2005 dec . 2007 + crrrrrrrrrrr site name & latitude & longitude + and telescope # & [ deg n ] & [ deg w ] + cerro tolar t1 & -21.25 & 71.25 + cerro armazones t2 & -23.75 & 71.25 + cerro tolonchar t3 & -23.75 & 68.75 + spm t4 & 31.25 & 116.25 + 13n mauna kea t6 & 18.75 & 156.25 + c c c c c c c height & + @xmath39km@xmath40 $ ] & 25%ile & median & 75%ile & & mean & rms + + 0.0 & 5.26e-14 & 1.11e-13 & 2.02e-13 & & 1.62e-13 & 2.88e-13 + 0.5 & 4.21e-18 & 3.84e-15 & 3.09e-14 & & 4.26e-14 & 1.37e-13 + 1.0 & 3.52e-16 & 3.21e-15 & 1.37e-14 & & 2.74e-14 & 1.02e-13 + 2.0 & 1.78e-16 & 3.88e-15 & 3.49e-14 & & 3.77e-14 & 8.69e-14 + 4.0 & 9.80e-15 & 2.77e-14 & 5.52e-14 & & 4.42e-14 & 5.79e-14 + 8.0 & 4.09e-15 & 2.54e-14 & 7.72e-14 & & 5.05e-14 & 6.21e-14 + 16.0 & 2.29e-14 & 3.51e-14 & 5.49e-14 & & 4.57e-14 & 3.96e-14 + + 0.0 & 6.04e-14 & 1.19e-13 & 2.15e-13 & & 1.71e-13 & 2.77e-13 + 0.5 & 1.23e-17 & 9.78e-15 & 5.58e-14 & & 6.93e-14 & 1.85e-13 + 1.0 & 6.59e-17 & 1.99e-15 & 1.10e-14 & & 2.38e-14 & 9.17e-14 + 2.0 & 5.27e-16 & 7.16e-15 & 3.23e-14 & & 3.28e-14 & 7.44e-14 + 4.0 & 6.34e-15 & 1.93e-14 & 4.09e-14 & & 3.33e-14 & 4.66e-14 + 8.0 & 8.66e-15 & 3.03e-14 & 7.96e-14 & & 5.63e-14 & 6.87e-14 + 16.0 & 1.96e-14 & 3.04e-14 & 4.84e-14 & & 4.05e-14 & 3.68e-14 + + 0.0 & 4.72e-14 & 1.04e-13 & 1.69e-13 & & 1.14e-13 & 1.93e-13 + 0.5 & 2.11e-21 & 1.76e-15 & 2.60e-14 & & 3.71e-14 & 1.21e-13 + 1.0 & 2.87e-17 & 1.29e-15 & 8.27e-15 & & 2.38e-14 & 9.18e-14 + 2.0 & 9.02e-16 & 1.29e-14 & 6.21e-14 & & 6.18e-14 & 1.30e-13 + 4.0 & 1.65e-14 & 3.87e-14 & 7.59e-14 & & 6.06e-14 & 7.11e-14 + 8.0 & 6.75e-15 & 3.47e-14 & 1.03e-13 & & 6.49e-14 & 7.41e-14 + 16.0 & 2.24e-14 & 3.62e-14 & 6.12e-14 & & 5.77e-14 & 6.96e-14 + + 0.0 & 1.75e-13 & 2.76e-13 & 5.09e-13 & & 4.88e-13 & 7.05e-13 + 0.5 & 1.67e-21 & 3.10e-15 & 3.62e-14 & & 6.39e-14 & 2.66e-13 + 1.0 & 1.56e-21 & 5.39e-16 & 5.93e-15 & & 2.22e-14 & 1.09e-13 + 2.0 & 9.82e-19 & 4.30e-16 & 9.41e-15 & & 2.55e-14 & 8.61e-14 + 4.0 & 8.00e-16 & 1.37e-14 & 4.27e-14 & & 3.55e-14 & 5.86e-14 + 8.0 & 2.70e-17 & 1.06e-14 & 5.57e-14 & & 4.15e-14 & 6.57e-14 + 16.0 & 2.25e-14 & 3.60e-14 & 6.10e-14 & & 5.43e-14 & 6.00e-14 + + 0.0 & 1.38e-13 & 2.41e-13 & 4.29e-13 & & 3.63e-13 & 5.07e-13 + 0.5 & 1.22e-20 & 5.11e-15 & 4.21e-14 & & 7.71e-14 & 3.11e-13 + 1.0 & 1.88e-16 & 2.46e-15 & 1.13e-14 & & 4.03e-14 & 1.83e-13 + 2.0 & 3.26e-17 & 8.98e-16 & 1.18e-14 & & 2.55e-14 & 8.15e-14 + 4.0 & 4.81e-15 & 1.72e-14 & 4.12e-14 & & 3.48e-14 & 5.56e-14 + 8.0 & 4.34e-16 & 6.21e-15 & 3.35e-14 & & 3.07e-14 & 5.58e-14 + 16.0 & 1.27e-14 & 2.03e-14 & 3.48e-14 & & 3.02e-14 & 3.53e-14 + c c c c c c c c c height & & & + @xmath39km@xmath40 $ ] & 25% @xmath25 & 50% @xmath25 & 75% @xmath25 & & 25% @xmath25 & 50% @xmath25 & 75% @xmath25 + + 0.0 & 9.41e-14 & 1.31e-13 & 1.81e-13 & & 8.72e-14 & 1.20e-13 & 1.68e-13 + 0.5 & 1.17e-15 & 4.29e-15 & 9.85e-15 & & 1.24e-14 & 2.35e-14 & 4.62e-14 + 1.0 & 2.14e-15 & 3.53e-15 & 6.03e-15 & & 6.20e-15 & 1.24e-14 & 2.80e-14 + 2.0 & 9.79e-16 & 5.26e-15 & 2.08e-14 & & 8.15e-15 & 2.14e-14 & 4.98e-14 + 4.0 & 2.04e-14 & 2.99e-14 & 4.51e-14 & & 2.54e-14 & 3.69e-14 & 6.19e-14 + 8.0 & 2.09e-14 & 3.15e-14 & 4.24e-14 & & 3.76e-14 & 5.27e-14 & 6.73e-14 + 16.0 & 3.32e-14 & 3.66e-14 & 3.96e-14 & & 3.90e-14 & 4.55e-14 & 5.33e-14 + + 0.0 & 1.02e-13 & 1.37e-13 & 2.05e-13 & & 9.64e-14 & 1.29e-13 & 1.91e-13 + 0.5 & 2.03e-15 & 1.02e-14 & 3.71e-14 & & 1.20e-14 & 2.94e-14 & 7.86e-14 + 1.0 & 9.61e-16 & 2.17e-15 & 5.49e-15 & & 4.29e-15 & 9.47e-15 & 2.62e-14 + 2.0 & 2.90e-15 & 1.04e-14 & 1.99e-14 & & 1.01e-14 & 2.42e-14 & 4.54e-14 + 4.0 & 1.41e-14 & 2.14e-14 & 3.24e-14 & & 1.84e-14 & 2.75e-14 & 4.69e-14 + 8.0 & 2.49e-14 & 3.69e-14 & 4.52e-14 & & 3.89e-14 & 5.76e-14 & 7.48e-14 + 16.0 & 2.75e-14 & 3.09e-14 & 3.53e-14 & & 3.23e-14 & 3.85e-14 & 4.72e-14 + + 0.0 & 1.03e-13 & 1.24e-13 & 1.41e-13 & & 9.59e-14 & 1.17e-13 & 1.33e-13 + 0.5 & 8.53e-16 & 2.46e-15 & 4.17e-15 & & 1.15e-14 & 2.23e-14 & 4.61e-14 + 1.0 & 1.02e-15 & 1.33e-15 & 2.81e-15 & & 4.19e-15 & 7.76e-15 & 2.24e-14 + 2.0 & 3.17e-15 & 1.30e-14 & 4.32e-14 & & 1.23e-14 & 2.95e-14 & 6.76e-14 + 4.0 & 2.91e-14 & 3.96e-14 & 5.98e-14 & & 3.39e-14 & 4.68e-14 & 7.55e-14 + 8.0 & 2.14e-14 & 3.86e-14 & 7.46e-14 & & 3.65e-14 & 6.01e-14 & 8.67e-14 + 16.0 & 3.07e-14 & 3.66e-14 & 4.61e-14 & & 3.38e-14 & 4.28e-14 & 6.01e-14 + + 0.0 & 2.06e-13 & 3.09e-13 & 5.55e-13 & & 1.94e-13 & 2.90e-13 & 5.07e-13 + 0.5 & 1.07e-15 & 2.35e-15 & 6.73e-15 & & 1.56e-14 & 2.75e-14 & 6.10e-14 + 1.0 & 2.52e-16 & 6.51e-16 & 1.33e-15 & & 4.04e-15 & 9.02e-15 & 2.56e-14 + 2.0 & 1.89e-16 & 3.80e-16 & 1.67e-15 & & 4.55e-15 & 1.03e-14 & 3.71e-14 + 4.0 & 9.77e-15 & 1.62e-14 & 2.80e-14 & & 1.71e-14 & 3.00e-14 & 5.49e-14 + 8.0 & 3.91e-15 & 1.38e-14 & 2.86e-14 & & 1.90e-14 & 3.75e-14 & 6.12e-14 + 16.0 & 3.08e-14 & 3.62e-14 & 4.76e-14 & & 3.55e-14 & 4.47e-14 & 6.60e-14 + + 0.0 & 1.83e-13 & 2.85e-13 & 4.89e-13 & & 1.67e-13 & 2.56e-13 & 4.28e-13 + 0.5 & 1.45e-15 & 4.89e-15 & 1.22e-14 & & 1.78e-14 & 3.43e-14 & 6.49e-14 + 1.0 & 1.33e-15 & 2.36e-15 & 5.03e-15 & & 5.06e-15 & 1.02e-14 & 3.24e-14 + 2.0 & 2.93e-16 & 9.77e-16 & 3.62e-15 & & 4.75e-15 & 1.33e-14 & 3.21e-14 + 4.0 & 1.25e-14 & 1.95e-14 & 2.58e-14 & & 1.95e-14 & 3.03e-14 & 4.48e-14 + 8.0 & 4.30e-15 & 7.30e-15 & 1.29e-14 & & 1.67e-14 & 2.64e-14 & 4.46e-14 + 16.0 & 1.89e-14 & 2.06e-14 & 2.34e-14 & & 2.40e-14 & 2.72e-14 & 3.48e-14 + c c c c c c c c c height & & & + @xmath39km@xmath40 $ ] & 25% @xmath26 & 50% @xmath26 & 75% @xmath26 & & 25% @xmath26 & 50% @xmath26 & 75% @xmath26 + + 0.0 & 1.32e-13 & 1.20e-13 & 9.28e-14 & & 2.00e-13 & 1.82e-13 & 1.47e-13 + 0.5 & 7.43e-15 & 5.92e-15 & 9.65e-16 & & 4.45e-14 & 5.10e-14 & 4.33e-14 + 1.0 & 3.08e-15 & 3.86e-15 & 3.67e-15 & & 1.82e-14 & 2.52e-14 & 3.28e-14 + 2.0 & 2.17e-15 & 2.93e-15 & 8.35e-15 & & 2.39e-14 & 3.42e-14 & 5.18e-14 + 4.0 & 2.67e-14 & 3.42e-14 & 2.87e-14 & & 3.24e-14 & 4.39e-14 & 4.84e-14 + 8.0 & 7.49e-15 & 2.18e-14 & 6.76e-14 & & 1.53e-14 & 3.14e-14 & 7.11e-14 + 16.0 & 2.68e-14 & 3.83e-14 & 5.47e-14 & & 2.52e-14 & 3.60e-14 & 5.28e-14 + + 0.0 & 1.32e-13 & 1.23e-13 & 1.06e-13 & & 1.90e-13 & 1.76e-13 & 1.68e-13 + 0.5 & 1.53e-14 & 1.37e-14 & 5.12e-15 & & 6.45e-14 & 8.82e-14 & 7.45e-14 + 1.0 & 2.35e-15 & 2.60e-15 & 2.56e-15 & & 1.73e-14 & 2.25e-14 & 3.08e-14 + 2.0 & 4.86e-15 & 7.31e-15 & 1.17e-14 & & 2.05e-14 & 2.95e-14 & 3.97e-14 + 4.0 & 1.83e-14 & 2.26e-14 & 2.27e-14 & & 2.29e-14 & 3.00e-14 & 3.75e-14 + 8.0 & 1.24e-14 & 2.85e-14 & 7.16e-14 & & 1.75e-14 & 3.46e-14 & 7.57e-14 + 16.0 & 2.25e-14 & 3.30e-14 & 4.86e-14 & & 2.15e-14 & 3.15e-14 & 4.69e-14 + + 0.0 & 1.17e-13 & 1.12e-13 & 8.79e-14 & & 1.40e-13 & 1.32e-13 & 1.05e-13 + 0.5 & 7.59e-15 & 4.53e-15 & 1.17e-18 & & 4.20e-14 & 4.16e-14 & 3.11e-14 + 1.0 & 1.87e-15 & 2.42e-15 & 5.66e-16 & & 1.25e-14 & 2.15e-14 & 1.96e-14 + 2.0 & 7.00e-15 & 8.38e-15 & 2.16e-14 & & 3.18e-14 & 3.87e-14 & 6.41e-14 + 4.0 & 3.34e-14 & 4.58e-14 & 3.60e-14 & & 3.85e-14 & 5.44e-14 & 5.51e-14 + 8.0 & 1.09e-14 & 2.55e-14 & 9.83e-14 & & 1.79e-14 & 3.55e-14 & 1.02e-13 + 16.0 & 2.56e-14 & 3.98e-14 & 5.88e-14 & & 2.42e-14 & 3.76e-14 & 5.66e-14 + + 0.0 & 2.39e-13 & 2.61e-13 & 3.22e-13 & & 3.44e-13 & 3.90e-13 & 5.36e-13 + 0.5 & 5.03e-15 & 4.08e-15 & 1.73e-15 & & 3.91e-14 & 5.42e-14 & 6.44e-14 + 1.0 & 7.48e-16 & 9.78e-16 & 5.68e-16 & & 8.99e-15 & 1.51e-14 & 3.12e-14 + 2.0 & 3.09e-16 & 3.93e-16 & 7.34e-16 & & 9.43e-15 & 1.85e-14 & 3.65e-14 + 4.0 & 1.27e-14 & 1.86e-14 & 2.07e-14 & & 1.95e-14 & 3.29e-14 & 4.58e-14 + 8.0 & 9.21e-16 & 1.01e-14 & 4.77e-14 & & 6.73e-15 & 2.00e-14 & 5.60e-14 + 16.0 & 2.51e-14 & 3.89e-14 & 6.20e-14 & & 2.41e-14 & 3.72e-14 & 6.02e-14 + + 0.0 & 2.32e-13 & 2.44e-13 & 2.42e-13 & & 3.33e-13 & 3.69e-13 & 3.59e-13 + 0.5 & 7.29e-15 & 1.14e-14 & 4.59e-15 & & 4.10e-14 & 7.89e-14 & 1.12e-13 + 1.0 & 1.79e-15 & 2.68e-15 & 3.46e-15 & & 1.62e-14 & 2.76e-14 & 5.44e-14 + 2.0 & 5.80e-16 & 8.99e-16 & 1.10e-15 & & 1.02e-14 & 2.06e-14 & 3.38e-14 + 4.0 & 1.30e-14 & 2.12e-14 & 2.90e-14 & & 1.79e-14 & 2.95e-14 & 4.25e-14 + 8.0 & 1.48e-15 & 5.93e-15 & 2.39e-14 & & 4.89e-15 & 1.32e-14 & 3.51e-14 + 16.0 & 1.50e-14 & 2.29e-14 & 3.70e-14 & & 1.43e-14 & 2.15e-14 & 3.47e-14 + c c c c c c c c c height & & & + @xmath39km@xmath40 $ ] & 25% @xmath16 & 50% @xmath16 & 75% @xmath16 & & 25% @xmath16 & 50% @xmath16 & 75% @xmath16 + + 0.0 & 1.41e-13 & 1.11e-13 & 7.97e-14 & & 2.02e-13 & 1.76e-13 & 1.29e-13 + 0.5 & 9.42e-15 & 4.13e-15 & 7.54e-16 & & 5.24e-14 & 4.72e-14 & 3.42e-14 + 1.0 & 4.12e-15 & 3.02e-15 & 3.07e-15 & & 3.26e-14 & 2.32e-14 & 2.42e-14 + 2.0 & 5.87e-15 & 2.69e-15 & 4.23e-15 & & 3.86e-14 & 3.50e-14 & 4.38e-14 + 4.0 & 2.94e-14 & 2.81e-14 & 2.86e-14 & & 3.87e-14 & 4.00e-14 & 4.88e-14 + 8.0 & 8.51e-15 & 2.39e-14 & 6.83e-14 & & 1.98e-14 & 3.63e-14 & 7.49e-14 + 16.0 & 2.80e-14 & 3.64e-14 & 4.83e-14 & & 3.17e-14 & 4.01e-14 & 5.46e-14 + + 0.0 & 1.31e-13 & 1.12e-13 & 1.01e-13 & & 1.87e-13 & 1.58e-13 & 1.58e-13 + 0.5 & 1.36e-14 & 1.25e-14 & 6.33e-15 & & 6.86e-14 & 7.97e-14 & 7.56e-14 + 1.0 & 1.99e-15 & 2.35e-15 & 2.35e-15 & & 1.76e-14 & 2.35e-14 & 2.69e-14 + 2.0 & 4.75e-15 & 7.67e-15 & 1.12e-14 & & 1.98e-14 & 2.89e-14 & 3.93e-14 + 4.0 & 1.59e-14 & 2.12e-14 & 2.49e-14 & & 2.21e-14 & 2.91e-14 & 3.64e-14 + 8.0 & 1.13e-14 & 2.84e-14 & 6.51e-14 & & 1.83e-14 & 3.73e-14 & 7.31e-14 + 16.0 & 2.36e-14 & 3.06e-14 & 3.95e-14 & & 2.67e-14 & 3.50e-14 & 4.60e-14 + + 0.0 & 1.20e-13 & 1.01e-13 & 6.26e-14 & & 1.43e-13 & 1.23e-13 & 7.65e-14 + 0.5 & 4.47e-15 & 1.49e-15 & 7.96e-18 & & 4.05e-14 & 3.79e-14 & 3.52e-14 + 1.0 & 1.53e-15 & 1.60e-15 & 8.31e-16 & & 1.57e-14 & 1.91e-14 & 2.66e-14 + 2.0 & 6.94e-15 & 8.55e-15 & 2.46e-14 & & 3.28e-14 & 4.06e-14 & 6.38e-14 + 4.0 & 2.98e-14 & 3.63e-14 & 5.92e-14 & & 3.70e-14 & 4.54e-14 & 7.11e-14 + 8.0 & 1.57e-14 & 3.91e-14 & 1.10e-13 & & 2.82e-14 & 5.07e-14 & 1.12e-13 + 16.0 & 3.14e-14 & 3.71e-14 & 5.35e-14 & & 3.55e-14 & 4.29e-14 & 6.85e-14 + + 0.0 & 2.29e-13 & 2.95e-13 & 3.87e-13 & & 3.24e-13 & 4.26e-13 & 6.38e-13 + 0.5 & 5.56e-15 & 3.70e-15 & 1.40e-15 & & 4.61e-14 & 5.30e-14 & 7.65e-14 + 1.0 & 8.20e-16 & 8.35e-16 & 2.40e-16 & & 1.30e-14 & 1.82e-14 & 3.75e-14 + 2.0 & 3.71e-16 & 3.99e-16 & 4.32e-16 & & 1.47e-14 & 2.52e-14 & 3.23e-14 + 4.0 & 1.15e-14 & 1.67e-14 & 1.91e-14 & & 2.46e-14 & 3.46e-14 & 4.22e-14 + 8.0 & 1.32e-15 & 8.79e-15 & 5.51e-14 & & 1.14e-14 & 2.45e-14 & 6.61e-14 + 16.0 & 2.77e-14 & 3.63e-14 & 5.53e-14 & & 3.23e-14 & 4.23e-14 & 6.58e-14 + + 0.0 & 2.46e-13 & 2.50e-13 & 2.21e-13 & & 3.70e-13 & 3.95e-13 & 3.74e-13 + 0.5 & 1.04e-14 & 1.06e-14 & 2.51e-15 & & 5.74e-14 & 8.16e-14 & 1.55e-13 + 1.0 & 2.50e-15 & 2.88e-15 & 3.99e-15 & & 2.68e-14 & 4.07e-14 & 7.53e-14 + 2.0 & 8.59e-16 & 8.83e-16 & 2.31e-15 & & 1.97e-14 & 2.31e-14 & 4.04e-14 + 4.0 & 1.52e-14 & 2.20e-14 & 3.55e-14 & & 2.30e-14 & 3.18e-14 & 5.17e-14 + 8.0 & 2.84e-15 & 9.31e-15 & 4.18e-14 & & 9.47e-15 & 2.17e-14 & 5.48e-14 + 16.0 & 1.75e-14 & 2.35e-14 & 3.68e-14 & & 2.03e-14 & 2.77e-14 & 4.22e-14 +

the phase transition driven by quantum fluctuations is one of the fundamental issues in quantum many body systems . a number of novel phenomena associated with the transition such as the quantum critical behavior have been observed in a variety of condensed matter materials @xcite . one of the prototype model systems exhibiting the quantum phase transition is the one - dimensional spin-1/2 ising lattice with a transverse field , namely the transverse ising model ( tim)@xcite , defined by @xmath1 where @xmath2 is the spin operator at the site @xmath3 on a one - dimensional lattice of length @xmath4 . the transverse field @xmath5 introduces quantum fluctuation to the system , leading to a quantum phase transition from the ferromagnetic / antiferromagnetic ordered to the paramagnetic disordered states above a critical value @xmath6 . actually the model is equivalent to a free spinless fermion system and can be exactly solved by applying the jordan - wigner transformation@xcite . based on the exact solution , all physical quantities , including the ground state energy , the low - energy excitations , the specific heat and other thermodynamic functions can be evaluated . this provides a thorough understanding of the quantum critical behavior of this system . however , in real materials , the moments of atoms may be larger than @xmath7 . despite of immense efforts in the past three decades it is still very difficult to find exact solutions of the @xmath8 or other higher spin quantum ising systems . this is partly due to the existence of the spin neutral states @xmath9 ( called as holes hereafter ) in addition to the two spin polarized states @xmath10 at each site in the @xmath8 spin chain . in these models , a hole can decay into a pair of polarized spin states , and vice verse , thus rendering the jordan - winger approach invalid in exactly solving the @xmath8 tim . in this paper , we study a class of one - dimensional @xmath8 quantum ising model , defined by the following hamiltonian @xmath11 where , @xmath12 is the single - ion anisotropy term with site - dependent @xmath13(@xmath14 ) . the above model has a classical limit where @xmath15 . this corresponds to the blume - capel model@xcite . so the present model can be regarded as the quantum generalization of the blume - capel model . the simplest quantum case is the uniform chain defined by@xcite @xmath16 . \label{eq : uqim}\ ] ] in two or higher dimensions , this kind of quantum ising model with single - ion anisotropy was studied by a number of authors , based mainly on the mean - field approximations@xcite . in particular , the ground state of the model ( [ eq : uqim ] ) was shown to be equivalent to the @xmath0 tim defined by eq . ( [ htim ] ) @xcite . such equivalence is also valid for quantum ising models with bond- and site - alternations@xcite or geometrical frustrations , such as a fully frustrated spin-1 ising delta - chain@xcite . the purpose of the present paper is to study the physical properties of the model described by eq . ( [ eq : hqim ] ) based on the exact solution . the key idea in solving the proposed @xmath8 model is to divide the total hilbert space of the @xmath8 system into a number of subspaces labeled by the number of holes . this is what we call the hole decomposition scheme ( hds ) . this hds was developed in our recent work @xcite where a recursion approach based on the hds is suggested for the uniform chain . in the present paper , we shall give more comprehensive investigations for various properties of the @xmath8 model , including the case with dimerization @xcite . in particular , we show that for a given hole configuration , each sub - lattice system with purely polarized spins can be exactly solved in that case . based on the exact solution we study how the quantum phase transitions and thermodynamic properties are affected by the interplay between the dimerization and the single - ion anisotropy . depending on the strength of dimerization , we find that the system undergoes a quantum phase transition where the criticality is the same as that of the uniform @xmath0 tim . we note that the spin-1/2 tim can be realized in certain low - dimensional magnetic materials @xcite . for the systems with local moments larger than @xmath7 , the single - ion anisotropy generated by crystal fields and the dimerization may also become important@xcite . the @xmath8 tim with the crystal field splitting was used to describe the ferroeletric transition in srtio@xmath17@xcite . in a class of quasi - one dimensional spin chains , such as [ ni@xmath18(medpt)@xmath18(@xmath19-ox)(h@xmath18o)@xmath18](clo@xmath21)@xmath22h@xmath18o , the magnetic ni@xmath23 ion shows not only a single - ion anisotropy@xcite , but also a @xmath8 bond - alternating pattern , where medpt is the bis(3-aminopropyl ) methylamine@xcite . recently , cold atoms or polar molecules in optic lattices were shown to be ideal systems to realize various quantum spin models@xcite . in particular , the spin-1 models can be implemented by trapping polar molecules where the spin degrees of freedom can be described by the hyperfine vibrational states@xcite . this paper is organized as follows . in sec.[sec : model ] and sec.[sec : mapping ] , we discuss some general properties of the model and introduce the hds . in sec.[sec : diag ] , we solve exactly the hamiltonian in the presence of dimerization . in sec.[sec : excitation ] , we study the low energy excitation spectra and quantum phase transitions based on the exact solutions . in sec.[sec : thermal ] , we discuss in detail the recursion method introduced in ref . @xcite for evaluating thermodynamic quantities . finally , we give a summary in sec . [ sec : summary ] . let us consider a s=1 ising lattice with single - ion anisotropy , defined by eq . ( [ eq : hqim ] ) . @xmath24 are the @xmath8 spin operators at lattice site @xmath25 , with the lattice length @xmath4 . the uniform classical blume - capel model@xcite corresponds to the symmetric case with @xmath26 . because only two of these @xmath27 terms are independent , we shall mainly consider the quantum case with @xmath28 and @xmath29 , without losing generality . the case with @xmath30 will be discussed later . we shall mainly focus on the dimerization case where @xmath31 by definition , one has @xmath32 so it is straightforward to show that @xmath33 does not couple @xmath34 state to @xmath35 states . thus @xmath33 commutes with @xmath36 . this leads to the following theorem : _ theorem 1 . when @xmath8 , the total hole number operator @xmath37 commutes with the hamiltonian for any site - dependent @xmath38 and @xmath39 _ @xmath40=0.\end{aligned}\ ] ] it means that the total number of holes is a conserved quantity if @xmath8 . consequently , the eigenstates of @xmath41 can be classified by the eigenvalue @xmath42 of @xmath43 . in the discussion below , we will call the subsystem with @xmath42 holes as the @xmath42-th sector . let @xmath44 be the hilbert space of the @xmath42-th sector , the total hilbert space is then given by the sum of all the subspaces : @xmath45 a complete set of the eigenstates in the @xmath42-th sector forms a sub - band of the whole spectrum . generally the lowest eigen - energy of the @xmath42-sector satisfies the following theorem : _ theorem 2 . let @xmath46 be the eigen - energy , corresponding to the eigenstate @xmath47 of @xmath41 , and @xmath48 be the lowest eigen - energy in the @xmath42-th sector , then the following inequality holds _ @xmath49 this relationship is an extension of the lieb - mattis theorem derived initially for the uniform systems with ferromagnetic ising couplings@xcite . it holds still no matter whether the ising couplings are ferromagnetic or antiferromagnetic . it indicates that in the absence of the @xmath50-term the ground state always lies in the @xmath51 sector and the energy spectrum of the system has a hierarchy structure . there are two kinds of excitations in the system . one is the fermionic excitation within a given sector . the corresponding excitation energy is defined by @xmath52 . the other is the hole excitation and the excitation energy with respect to the ground state is given by @xmath53 . there are two kinds of minimal excitation gaps corresponding to these excitations @xmath54 fig . [ fig : band ] shows schematically the hierarchy band structure of the system . within each bands ( or each boxes shown in fig . [ fig : band ] ) there are fermionic excitations , with the minimal gaps @xmath55 . while the minimal gaps of the hole excitations @xmath56 increase with @xmath42 when the longitudinal anisotropy @xmath57 . -th sectors . @xmath55 and @xmath56 are the minimal gaps for the fermoinic and hole excitations respectively in the @xmath42-sector . several explicit results of these gaps will be given in section iv . ] the eigenstates of the hamiltonian eq . ( [ eq : hqim ] ) can be generally expressed as @xmath58 , with @xmath59 being the wave function . as the total spin and its @xmath60-component are not conserved , the summation runs over all @xmath61 states . however , by theorem 1 , the local holes are good quantum numbers and can be regarded as non - magnetic local impurities embedded in the @xmath0 ising system . assuming in the @xmath42-th sector the holes are located at @xmath62 ( @xmath63 ) , then the corresponding eigenstates can be expressed as @xmath64 where @xmath65 if @xmath66 , and @xmath67 . in the @xmath51 sector , there are only two spin states at each site , corresponding to @xmath65 respectively . they are in one - to - one correspondence with the states of the spin-1/2 ( pauli ) operators @xmath68 : @xmath69 . one has then the mapping relationship : @xmath70 . thus , @xmath33 acts like @xmath71 . the original hamiltonian , when acting on the @xmath51 subspace , has the following reduced form @xmath72 this is just the spin-1/2 ising model with bond ( site)-dependent ising couplings and transverse fields . now let us turn to the @xmath73 sector . if the hole is located at the site @xmath74 , the corresponding state can be written as @xmath75 . now because the bonds connecting the hole are broken , the reduced hamiltonian , obtained by acting the original one on the @xmath73 sector , is given by @xmath76 where @xmath77 is the hamiltonian of the spin-1/2 tim segment of lengths @xmath78 imposed by the open boundary condition . if @xmath79 and @xmath80 are the eigenstates of the segments @xmath81 and @xmath82 , respectively , then @xmath83 is the eigenstate of @xmath84 . this hds can be easily generalized to the multi - hole sectors . for instance , the reduced hamiltonian in a @xmath42-hole sector is given by @xmath85 where @xmath86 s are the positions of holes . similarly , the eigenstate of @xmath87 can then be expressed in terms of those of the individual segments . therefore , all the eigenstates of the @xmath8 qim ( [ eq : hqim ] ) can be obtained by solving a set of the spin-1/2 tims . for the @xmath42-th sector , the number of the s=1/2 tim segments are @xmath88 or @xmath42 , depending on whether the original chain is periodic or open . note that the hole positions may vary along the chain , so there are many different hole configurations in a given sector . the decomposition of the total hilbert space into the sum of subspaces can be formally represented by @xmath89^{\otimes l}= \textbf{2}^{\otimes l}\oplus \textbf{2}^{\otimes ( l-1)}\otimes \textbf{1}\oplus \textbf{2}^{\otimes ( l-2)}\otimes \textbf{1}^{\otimes 2}\oplus \cdots \oplus \textbf{1}^{\otimes l}$ ] , where the dimension of the @xmath8 system is given by @xmath90 in this section , we follow the standard approach introduced in refs . @xcite to diagonalize the @xmath51 sector of the s=1 model , eq.([eq : h0 ] ) . we introduce the fermion operators @xmath91 and @xmath92 and use the jordan - wigner transformation to rewrite eq.([eq : h0 ] ) as follows @xmath93\nonumber\\ & ~~ & + h_{pb},\end{aligned}\ ] ] where @xmath94 , @xmath95 , and @xmath96 . notice that @xmath97 is the boundary term ( which disappears for the open chains ) and can be neglected in the thermodynamics limit . next , we introduce the bogoliubov transformation as in the following : @xmath98 where , @xmath99 and @xmath100 are the fermionic quasi - particle operators with quasi - momentum @xmath101 . they satisfy the usual anti - commutation relations . while , @xmath102 and @xmath103 are coefficient matrices , which should be complex in general . then , @xmath104 becomes @xmath105 if we define @xmath106 , @xmath107 , the eigenvalue @xmath108 can be solved by @xmath109 with @xmath110 being a symmetric matrix defined by @xmath111 , or @xmath112 . for the dimerized ( or alternating ) chain , the @xmath110-matrix depends on the boundary conditions and the parity of the chain length ( even or odd ) . for simplicity , we here consider the case of periodic chain with even @xmath4 , where @xmath113 with @xmath114 , @xmath115 , @xmath116 , and @xmath117 . in order to solve eq.([eq : psi ] ) with this @xmath110-matrix , we employ the following ansatz @xmath118 , \\ ( \phi_k)_{2j+1}&=a_{o}\exp[ik(2j+1 ) ] . \end{split}\end{aligned}\ ] ] the ratio @xmath119 is then a measure of dimerization , determined by @xmath120},\end{aligned}\ ] ] where @xmath121 note that when @xmath122 and @xmath123 , we have @xmath124 , as it should be for the uniform chain . the eigenvalue @xmath125 is then obtained by @xmath126 , so that one has four branches of quasi - particle excitations : @xmath127 with @xmath128 , @xmath129 , @xmath130 , @xmath131 , and @xmath132 , @xmath133 . qim with @xmath134 , @xmath135 , @xmath136 , @xmath137 , @xmath138 . this kind of excitations for a sector with fixed hole number is similar though the quasi - momentum is determined by the corresponding secular equation . ] it should be noticed that @xmath139 ( for @xmath128 ) are invariant under the shift @xmath140 . thus the quasi - momentum @xmath101 is constrained in one of the following regimes : ( 1 ) @xmath141 or ( 2 ) @xmath142 . here , we choose @xmath101 to be in the first regime for @xmath143 and in the second regime for @xmath144 . [ fig : spectra12 ] shows the typical energy dispersions for a dimerized system with @xmath134 . the case with @xmath145 can be obtained by a reflection under the shift . in the @xmath73 sector , the hole breaks two bonds connecting to it . for convenience , we assume that the hole is located at the site @xmath4 . the reduced hamiltonian can be rewritten as @xmath146 where @xmath147 is defined by eq . ( [ eq : h03 ] ) but with the open boundary conditions . the procedure for diagonalizing @xmath147 is the same as for @xmath148 . the eigen wavefunction is determined by eq.([eq : psi ] ) in a similar way . the only difference is that now the @xmath110-matrix is a @xmath149 matrix , with @xmath150 , and @xmath151 . the previous approach is extended to the subsystems or sectors with more holes . the holes break the ising couplings , leading to a series of segments of the spin-1/2 tim s . for the periodic chain , the number of these segments is equal to the number of holes @xmath42 ( including the segment of zero length ) . if the holes are located at @xmath163 , then the corresponding reduced hamiltonian is given by eq.([eq : h0 ] ) , i.e. , @xmath164 , with @xmath165(here , @xmath166 ) being the length of the @xmath167-th segment and @xmath168 each segment hamiltonian @xmath169 can be diagonalized as @xmath170 , where @xmath101 is the quasi - momentum satisfying the secular equation eq . ( [ k - dimer ] ) ( by replacing @xmath4 by @xmath171 ) . notice that associated with the fixed length of the segment there are four different kinds of configurations , depending on whether the two edge holes are located at odd or even sites . in the previous section , we show how to diagonalize the reduced hamiltonians of different sectors . in a given hole sector , there are four branches of quasi - particle excitations , given by eq . ( [ lamb1 ] ) . obviously , the two negative branches @xmath172 ( @xmath128 ) will be filled in the ground state of that sector . in the @xmath51 sector , the ground state is given by @xmath173 , where @xmath174 ( @xmath128 ) are the fermionic quasi - particle operators in the branches @xmath172 , @xmath101 and @xmath175 are the allowed momenta for @xmath128 , respectively . the ground state energy is given by @xmath176 . the low energy excitations can be obtained by applying operators @xmath177 or @xmath178 on the ground state . the excitation energy is given by @xmath179(@xmath180 ) . when @xmath134 , the energy gap between the lowest excitation and the ground state is given by @xmath181 where @xmath182 . if @xmath183 and @xmath184 , one has @xmath185 with @xmath186 , reproducing the result of the uniform chain . in this case , the branches 1 and 2 connect smoothly . in the presence of dimerization , however , the two branches will split and produce a dimerization gap @xmath187 the ground state of the @xmath73 sector has @xmath4-fold degeneracy in the uniform chain , because the energy does not depend on the position of the hole . but in the presence of dimerization the degeneracy will be @xmath188-fold . while the momentum - dependence of the spectra can be determined as in the @xmath51 sector , the quasi - momentum @xmath189 must satisfy the secular equation associated with the open chain of length @xmath190 . in the lowest state of the @xmath73 sector , the bands with negative energies are fully filled . the corresponding energy is then given by @xmath191 -\frac{1}{2}\max{(d_1,d_2)}$ ] . the last term is contributed from the hole which may locate at either even or odd sites . according to theorem 2 , the ground state of the original hamiltonian should be the one in the @xmath51 sector . so in addition to the fermionic excitations , inclusion of the @xmath73 sector ( adding a hole to the system ) will induce the hole excitations . the minimal hole excitation gap is given by @xmath192 . the typical behavior of @xmath193 ( as a function of @xmath194 ) is shown in fig . [ dimerholegap ] for fixed @xmath195 , @xmath196 , and @xmath197 , with either periodic or open boundary conditions . in order to understand the difference in @xmath193 for the periodic and open chains , we consider a special case where the chain is uniform and at the critical point , i.e. , @xmath198 , @xmath199 , and @xmath200 . in this simple case , the energy spectrum is given by @xmath201 , and the allowed quasi - particle momenta are @xmath202 , @xmath203 for the period boundary condition , and @xmath204 , @xmath205 for the open boundary condition . thus for sufficiently large @xmath4 we obtain @xmath206 and @xmath207 for the open and periodic chains , respectively . their difference , @xmath208 , does not change for larger @xmath4 . so it is the surface energy cost to turn into an open chain . as a function of @xmath194 for the system with periodic ( left panel ) or open ( right panel ) boundary conditions . three curves corresponding to @xmath209 and @xmath210 are plotted and @xmath195 and @xmath196 are fixed . ] it is straightforward to extend the above discussion to a multi - hole system . let us start with @xmath211 . the eigenfunction in the @xmath211 sector is a direct product of the wavefunctions for the two @xmath73 segments , but with smaller lattice lengths @xmath212 and @xmath213 , respectively , here @xmath214 is the distance between the two holes sited at @xmath74 , @xmath215 . for the uniform chain , the lowest energy state of the @xmath211 sector should correspond to the configuration where two holes are close together , as in this case the total surface energy is minimized . so the lowest hole excitation gap in the @xmath211 sector is given by @xmath216 . for @xmath217 , one can further show that the lowest state in the @xmath42-th sector is the configuration in which all holes are close to each other . this is the hole condensation phase in one dimension . the remain spin chain has a length @xmath218 . when @xmath219 and @xmath220 , the finite size effect is negligible so one has @xmath221 . we now discuss how dimerization influence the phase diagram . we have shown that the ground state still lies in the @xmath51 sector in the presence of dimerization . however , dimerization splits the bands and induce a gap at higher energies . the quantum critical points extend to lines , which are determined by the gapless condition , i.e. @xmath222 , or according to eq . ( [ eq : delta0 ] ) , @xmath223 qim . the @xmath224-axis and @xmath225-axis describe the dimerization strength and the competition parameter , see eq.([ro ] ) in the main text . the dashed area is the magnetic ordered phase which extends slightly for small @xmath224 but split into two parts for @xmath226 . ] the influence of dimerization can be seen more clearly by introducing ( assuming @xmath227 ) @xmath228 where @xmath225 describes the competition between @xmath39 and @xmath38 , and @xmath224 is a measure of the dimerization strength . in the uniform chain limit , @xmath229 , @xmath230 . with increasing @xmath224 , we have the following three situations : \(1 ) in the weak dimerization regime , @xmath231 , there is a pair of symmetric critical points , @xmath232 , with @xmath233 . the ground state is magnetically ordered when @xmath234 . \(2 ) in the strong dimerization regime , @xmath226 , there are two pairs of symmetric critical points , @xmath235 , with @xmath236 . the magnetically ordered phase appears when @xmath237 . \(3 ) when @xmath238 ( @xmath239 ) , there are three critical points , which take values @xmath240 respectively . the ground state is magnetically disordered when @xmath241 , but ordered ( either ferromagnetic or anti - ferromagnetic , depending on the signs of @xmath242 ) when @xmath243 . note that the @xmath244 point corresponds to an alternating ( or staggered ) array of single - ion anisotropy @xmath245 . it becomes critical when @xmath246 . the @xmath247 ground state phase diagram is plotted in fig . [ fig : phase ] . the magnetic ordered phase and paramegnetic disordered phase are separated by the critical lines ( red ) . in this section , we study the thermodynamic properties of the @xmath8 qim . as was shown previously , this model is exactly solvable not only for the ground state , but also for all excited states . however , exactly evaluating the thermodynamic quantities is still a very hard task , particularly for large system size and dimerization . here , we shall develop the recursion method proposed in ref . @xcite in the evaluation of the partition function as well as other thermodynamic quantities for either uniform and dimerized chains . based on the hds , the partition function @xmath248 of the system with lattice length @xmath4 can be expressed as the sum of all the partition functions of the subsystems , i.e. , @xmath249 where , @xmath250 is the partition function of the @xmath42-th sector . because for fixed @xmath42 , there are many different hole configurations . so @xmath250 can be further rewritten as a sum over all possible hole configurations @xmath251 for each hole configuration @xmath252 , the corresponding partition function of the open chain is given by ( @xmath253 ) @xmath254}.\end{aligned}\ ] ] where , @xmath255 is the hamiltonian of the n - th segment . the partition functions of each segments can be regarded as the building blocks of total partition of the original system . these building blocks are denoted by @xmath256 , the partition functions of the spin-1/2 tim segments with length @xmath171 . then , in the uniform case , @xmath248 can be expressed as @xmath257 , where @xmath258 is the partition function of a hole , @xmath259 , and @xmath260 . as the length of the allowed segment may vary , one has the summation constraint @xmath261 . if we denote @xmath262 , then the total partition function can be rewritten as @xmath263 where , @xmath264 is contributed from the hole s . to numerically evaluate the partition function , it is practically convenient to use the following recursion formula : @xmath265 where , @xmath266 and @xmath267 . by this way , we firstly calculate the building block , @xmath256 , and then by iterative use of above relation , evaluate the partition function of the @xmath8 system . we find that this method is particular efficient for the uniform chain , where the system size @xmath4 could be as large as @xmath268 . it is non - trivial to extend the above recursion method to the spin chain in the presence of dimerization . here , there are four kinds of blocks , associated with the parity of the two ends . thus we can denote them by @xmath269 , with @xmath270 indicating the left / right ends respectively . the @xmath0 tim segments with odd or even end sites can be solved exactly . the analytical expressions for @xmath269 will be provided in a separated supplementary material@xcite . in the following , we present some numerical results obtained by the recursion method while the system size is kept at @xmath197 . for simplicity , we use the open boundary condition for the original @xmath8 chain . the extension to the periodic boundary condition is straightforward . thermodynamics quantities can be calculated from the partition function . in our model , an important physical quantity is the thermal average of the hole number , defined by @xmath271 ( a ) and the entropy @xmath272 ( b ) for a uniform spin chain with @xmath273 , respectively . ] [ fig : holeuniform](a ) shows the temperature dependence of @xmath274 for several @xmath275 . at low temperatures , @xmath274 increases rapidly with increasing temperature for small and positive @xmath275 . the proliferation of the hole number at low temperatures is obviously due to the smallness of the hole excitation gap . however , for the negative @xmath275 , say , @xmath276 , the hole excitation gap is relatively larger , so @xmath274 increases much slowly with temperatures . another interesting physical quantity is the entropy @xmath272 , which we plot as a function of temperature for several different @xmath275 in fig . [ fig : holeuniform](b ) . we find that the suppression of the entropy @xmath272 is stronger for larger @xmath275 . the suppression is even more pronounced for the negative @xmath275 . these behavior are similar to the temperature dependence of the hole number and are also due to the hole excitation gap . we also find that the @xmath274 approaches about @xmath277 in the high temperature limit for all @xmath275 . correspondingly , the entropy saturates at the value @xmath278 in the high temperature limit(not fully shown in fig . [ fig : holeuniform ] ) . and ( b ) the entropy @xmath272 for a dimerized spin chain in three representative cases : @xmath279 ( dashed line ) , @xmath280 ( real line ) , @xmath281 ( dash - dotted line ) , respectively . where , @xmath195 , @xmath282 , and @xmath283 are fixed . ] , @xmath284 ( blue - dashed line ) , @xmath285 , @xmath286 ( red - real line ) , @xmath287 , @xmath288 ( dash - dotted line ) with @xmath289 , @xmath282 , and @xmath283 , respectively ; ( b ) the specific heat c ( dash - dotted line ) and the corresponding specific heat c(p=0 ) for the @xmath51 subsystem ( green open circle ) for @xmath287 and @xmath195 , @xmath282 , @xmath283 , @xmath288 , respectively . ( c ) the specific heat @xmath290 ( blue - real line ) and c(p=0 ) ( red - doted line ) with @xmath291 , @xmath283 , and @xmath292 . ] the recursion method , after some extensions discussed previously , is also used to evaluate thermodynamic quantities for a dimerized system . [ fig : holedimer ] and [ fig : heatdimer](a ) show the temperature dependence of the entropy @xmath272 , the average hole concentration @xmath274 , and the specific heat @xmath290 in the ordered ( @xmath279 ) , critical ( @xmath285 ) , and disordered phases(@xmath287 ) , respectively . here we fix @xmath195 , @xmath282 , @xmath283 , and choose several @xmath194 (= @xmath293 , @xmath294 , and @xmath295 ) in figs . [ fig : holedimer ] and [ fig : heatdimer](a ) . note that both the dimerization strength and competition parameter are tuned by varying @xmath194 . the general features of the hole number and the entropy are similar to that in the uniform case as the hole excitation gap plays the role at low temperatures ( note that when @xmath195 , @xmath282 , and @xmath283 , @xmath296 , @xmath297 , and @xmath298 for @xmath284 , @xmath294 , and @xmath295 , respectively ) . in figs . [ fig : heatdimer ] , the specific heat is plotted as a function of temperature in several cases . we find that at low - temperatures the specific heat is peaked at the temperature scale near the hole excitation gap . the peak behavior changes depending on the competition of the fermionic and hole excitations . the influence of the dimerization on the specific heat can be clearly seen at relatively higher temperatures , as the dimerization induced gap is much larger than the hole or fermion excitation gaps . [ fig : heatdimer](b ) shows the specific heats of the @xmath8 qim system and its @xmath51 sector for the case @xmath299 . note that the @xmath51 sector is identical to the corresponding @xmath0 tim . in this case @xmath300 is very small , leading to a sharp peak in the @xmath8 qim system . we also find a well separated and relatively round peak in the higher temperature regime where the energy scale is close to the dimerization gap @xmath301 . the competitions among the dimerization effect , the hole excitations and the fermion excitations can be seen more clearly from fig . [ fig : heatdimer](c ) where @xmath291 , @xmath283 , and @xmath292 . in this case , the three kinds of gaps are well separated , @xmath302 , so that the specific heat exhibits three peaks . the first one is a sharp peak around @xmath303 @xcite , other two peaks are around @xmath304 and @xmath305 , respectively . by contrast , the low temperature sharp peak disappears in the @xmath51 sector and other two peaks still persist [ see the red dotted line in fig . [ fig : heatdimer](c ) ] . this is in agreement with the fact that no hole excitation exists in the @xmath51 sector . . two typical situations are plotted : ( @xmath306 , @xmath307 ) and ( @xmath308 , @xmath286 ) , with small and large hole excitation gaps , respectively . other parameters are @xmath282 and @xmath283 . as a comparison , the corresponding @xmath0 tim case is also plotted . ] we also calculated the scaled free energy @xmath309 at the critical point @xmath285 for the dimerized spin chain . @xmath309 is defined by @xmath310}{t^2}.\ ] ] this quantity , which is identical to the specific heat coefficient at low temperatures , was introduced in ref.@xcite in order to show the temperature persistence of the quantum critical scaling behavior in the @xmath0 tim . in the critical region where quantum critical fluctuations dominate this quantity should be a constant . it was shown that in the @xmath0 uniform tim it deviates from the constant only when @xmath311 , indicating a rather higher temperature scale below which the quantum critical scaling behavior persists @xcite . in our recent work ref.@xcite , we found that in the uniform @xmath8 qim this behavior is strongly suppressed by the hole excitations . here , we find that the similar conclusion can be inferred in the presence of dimerization . in fig . [ fig : phidimer ] , we plotted the results for two cases ( i ) @xmath312 , @xmath282 , @xmath283 , @xmath307 , and ( ii ) @xmath195 , @xmath282 , @xmath283 , and @xmath286 . as a comparison , the corresponding result for the dimerized @xmath0 tim is also plotted . we find that the quantum critical scaling behavior at @xmath313 ( or @xmath314 in the figure [ fig : phidimer ] ) persists approximately at finite @xmath315 for the dimerized @xmath0 tim and the case ( ii ) , and at @xmath316 for the case ( i ) , respectively . this is because that the hole gap in the case ( i ) is much smaller than the case ( ii ) . consequently , the hole excitations play more significant role in suppressing the quantum critical scaling behavior in the former case . this result is consistent with the conclusion in ref . @xcite . now we turn to the case with non - zero @xmath317 . it is straightforward to show that the @xmath317-term plays the role of chemical potential for holes in the @xmath8 qim @xcite . more specifically , the energy of a @xmath42-hole sector with finite @xmath317 is related to that with @xmath318 by the following relationship [ @xmath319 : @xmath320 thus , with finite @xmath317 , the fermion excitation spectra remain unchanged , but the hole excitation gap becomes@xcite @xmath321 for @xmath318 or @xmath322 . other parameters are @xmath195 , @xmath282 , and @xmath283 . ] therefore , the ground state depends strongly on the value of @xmath317 . when @xmath323 , the hole excitation is positive , and the ground state is still in the @xmath51-sector . but when @xmath324 , the hole excitation gap is negative . this indicates that the @xmath51 sector is no longer the lowest energy state and there are holes in the ground state . as a result , theorem 2 is no longer valid in the present case . when @xmath4 is sufficiently large and @xmath42 is relatively small , the lowest energy of the @xmath42-sector can be approximated by @xmath325 as discussed in the previous section . then , @xmath326 $ ] , so that we have @xmath327 thus , the order of the band structure hierarchy is completely overturned . in this case , the ground state is in the @xmath328 sector and all sites are occupied by holes@xcite . this can be also seen clearly from fig . [ fig : dimerhole2 ] , where the temperature dependence of the hole concentration @xmath274 is shown . we find that in the zero temperature limit , @xmath274 is equal to 1 when @xmath329 or 0 when @xmath330 . in this paper , we have studied a class of exactly solvable @xmath8 qims with single - ion anisotropy . they exhibit a hierarchy of the band structure with both fermionic and hole excitations . the hole excitation gap can be tuned by the longitudinal crystal field @xmath317 . it becomes zero when @xmath331 is equal to @xmath332 . the ground state exhibits three distinct phases : the magnetically ordered or disordered phases when @xmath333 , or the hole condensation phase when @xmath334 . the transition to the hole condensation phase is of the first order . we have shown that dimerization does not destroy the exact solvability of this model . to our knowledge , this is the first example in dimerized @xmath8 quantum spin systems where all the eigen states as well as the wavefunctions and the eigen energies can be solved exactly . the hole excitations enhance the thermodynamic fluctuations as evidenced in the specific heat which shows a sharp peak in the low - temperature region where the hole excitations proliferate . this strongly reduces the characteristic temperature below which the quantum criticality persists . all these results are robust against the dimerization . however , dimerization deforms the phase diagram and affects the high energy behavior . we have developed a recursion method to sum over all hole configurations efficiently . this provides a powerful approach for evaluating rigourously all thermodynamic quantities as well as static and dynamic correlation functions of the qims in the thermodynamic limit . the detailed derivations for these quantities in the presence of dimerization will be provided as a supplementary material@xcite . the recursion method holds not just for the model studied here . it can be easily extended and applied in other physical systems whose hamiltonian can be written as a sum of independent spin segments , separated by nonmagnetic impurities , such as pd- or zn - doped quasi - one - dimensional antiferromagnets sr@xmath18(cu@xmath335pd@xmath336)o@xmath17 or cu@xmath335zn@xmath336geo@xmath17@xcite . z.h.y . would like to thank z.x . xu for helpful discussions . this work was supported in part by the national natural science foundation of china , the national program for basic research of china , the pcsirt ( irt-0754 ) and srfdp ( no.j20050335118 ) of education ministry of china . 99 m. greiner , o. mandel , t. esslinger , t. w. hnsch , and i. bloch , nature * 415 * , 39 ( 2002 ) . p. gegenwart , q. si , f. steglich , nat . phys . * 4 * , 186 ( 2008 ) . s. l. sondhi , s. m. girvin , j. p. carini , and shahar , rev . * 69 * , 315 ( 1997 ) . s. sachdev , _ quantum phase transitions _ , cambridge university press , new york , 1999 . b. k. chakrabarti , a. dutta , and p. sen , _ quantum ising phase and transitions in transverse ising models _ , springer , berlin , 1996 . e. lieb , t. schultz , and d. mattis , ann . phys . n.y . * 16 * , 407 ( 1961 ) . p. pfeuty , ann . phys . n.y . * 57 * , 79 ( 1970 ) . m. blume , phys . * 141 * , 517 ( 1966 ) ; h. w. capel , physica * 32 * , 966 ( 1966 ) . z. h. yang , l. p. yang , j. dai , t. xiang , phys . * 100 * , 067203 ( 2008 ) . x. f. jiang , j. l. li , j. l. zhong , c. z. yang , phys . b * 47 * , 827 ( 1993 ) . n. c. eddeqaqi , m. saber , a. el - atri , and m. kerouad , physica a * 272 * , 144 ( 1999 ) . j. oitmaa and a. m. a. von brasch , phys . b * 67 * , 172402 ( 2003 ) . h. n. wu , z. h. yang , j. dai , h. p. ying , mod . a * 22 * , 727 ( 2007 ) . y. fukumoto and a. oguchi , prog . phys . * 115 * , 847 ( 2006 ) ; y. fukumoto and a. oguchi , aip conf . proc . * 850 * , 1081 ( 2006 ) . while `` dimerization '' usually refers to the alternating antiferromagnetic coupling , here we use it to indicate that both the ising coupling and the single - ion anisotropy can be of alternating pattern with period 2 . j. richter , s. e. kruger , d. j. j. farnell , and r. f. bishop , in _ series on advances in quantum many - body theory _ , world scientific , singapore , vol.5 , p.239 , 2001 . d. bitko , t. f. rosenbaum , and g. aeppli , phys . lett . * 77 * , 940 ( 1996 ) . a. abragam and b. bleaney , _ electron paramagnetic resonace of transition ions _ , oxford university , oxford , 1970 . y. yamada , n. todoroki , and s. miyashita , phys . b * 69 * , 024103 ( 2004 ) . a. esucer , r. vieente , x. solans and m. font - baida , inorg . chem . * 33 * , 6007 ( 1993 ) . j. j. borrs - almenar , e. coronado , j. curely and r. georges , inorg . chem . * 34 * , 2699 ( 1995 ) . s. kimura , s. hiraia , y. narumib , k. kindoa , and m. hagiwarac , physica b * 294 * , 47 ( 2001 ) . y. narumi , m. hagiwara , r. sato , k. kindo , h. nakano and m. takahashi , physica b * 246 * , 509 ( 1998 ) ; y. narumi , r. sato , k. kindo and m. hagiwara , j. magn . magn . mater . * 177 * , 685 ( 1998 ) . a. micheli , g. k. brennen , and p. zoller , nat . * 2 * , 341 ( 2006 ) . g. v. brennen , a. micheli , and p. zoller , new j. phy . * 9 * , 138 ( 2007 ) . e. lieb and d. mattis , phys . rev . * 125 * , 164 ( 1962 ) . t. d. shultz , d. c. mattis and e. lieb , rev . phys . * 36 * , 856 ( 1964 ) . see epaps document no.e - prbmdo-79 - 081921 about some details of the exact solutions , together with some additional discussions on the green functions , the spin - spin correlation functions , as well as the spin susceptibility . for more information on epaps , see http://www.aip.org/pubservs/epaps.html . in the presence of dimerization the hole excitation gap is also dependent on the odd / even sites where the hole located . here we define the smaller one by @xmath193 . a. kopp and s. chakravarty , nature phys . * 1 * , 53(2005 ) . in the following , we use @xmath337 to explicity denote the hole excitation gap in the case with @xmath318 , in order to distinguish the one with @xmath338 . j. sirker , n. laflorencie , s. fujimoto , s. eggert and i. affleck , phys . * 98 * , 137205 ( 2007 ) . m. hase , i. terasaki , y. sasago , k. uchinokura and h. obara phys . lett . * 71 * , 4059 ( 1993 ) .

the high pressure and temperature equation of state ( eos ) and phase diagram of carbon have received considerable attention of late . recent laser - shock and ramp - compression studies @xcite , together with shock measurements performed with magnetically - driven flyer plates @xcite , have produced data in the range from @xmath3 mbar . in addition , theoretical work on the eos and phase diagram in this same range have yielded predictions which are largely ( though not completely ) in accord with these experimental data @xcite . these studies were conducted with density functional theory ( dft ) molecular dynamics ( md ) . much of this recent focus on the carbon eos , specifically in states of compression reached when starting in the diamond phase , has arisen from the interest of using high density carbon as an ablator material for capsules designed to achieve fusion at the national ignition facility @xcite . while these experimental and theoretical studies have been useful in constraining the eos of carbon for this and related applications , it is crucial to note that the states reached by the ablator in inertial confinement fusion ( icf ) are expected to include temperatures in excess of tens of ev @xcite . such conditions have not yet been sufficiently characterized . at such high temperatures , dft - md is extremely challenging to perform , due to the large number of high - lying , partially - occupied single - electron states that must be included for an accurate rendering of the electronic thermal excitations . indeed , an orbital - free variant of dft - md has been developed and used to handle such high temperature applications @xcite , but as yet , a satisfactory treatment of atomic shell structure in this approach is lacking . the approach as implemented thus far @xcite is ( somewhat ) reminiscent of a thomas - fermi treatment for the electrons @xcite . another altogether different approach to describing the electronic structure of the high-@xmath4 plasma state is path integral monte carlo ( pimc ) @xcite . in this approach , unlike in typical implementations of dft , there is no mean - field assumption made for the many - electron problem , and the imaginary time treatment makes high-@xmath4 simulations more efficient to perform than low-@xmath4 simulations . although assumptions regarding the nodal surface of the many electron density matrix necessarily introduce approximations for the treatment of atomic shells , it is very encouraging that recent work on the c plasma has demonstrated that eos predictions can be made with this method which smoothly interpolate between dft - md results at the lower temperatures and the high-@xmath4 debye - hckel limit @xcite . the ultra - high pressure regime near @xmath5 has been studied recently as well , using ab initio electronic structure techniques of the dft variety together with a random search method to find the thermodynamically stable crystalline phases at pressures between 10 - 1000 mbar @xcite . this work extended earlier quantitative studies of the c phase diagram that treated the diamond and bc8 phases up to @xmath6 mbar @xcite . the random search established the sequence of stable phases to be diamond @xmath7 bc8 @xmath7 ( slight modification of ) simple cubic ( sc ) @xmath7 simple hexagonal ( sh ) @xmath7 face - centered cubic ( fcc ) @xcite . though these authors did not attempt to predict the details of the c melt curve in these extreme conditions , their analysis using a quasiharmonic description of the phonons for these solid phases suggested the existence of a bc8-sc - sh triple point , as seen in fig . 3 of their manuscript @xcite . a few years ago , correa et al . used the available theoretical understanding of the c phase diagram at the time @xcite to produce a three - phase eos model for c focusing specifically on the regime of @xmath8 mbar and @xmath5 - 20,000 k @xcite . the phases considered were : diamond , bc8 , and the liquid . the individual phase free energy models were fit to the results of dft - md calculations . since the final multiphase eos model was designed to be used in hydrocode simulations spanning a wider range of conditions , these researchers embedded their detailed three - phase eos model into a more coarse - grained model @xcite that respected the ideal gas limits at high-@xmath4 and low-@xmath2 and the thomas - fermi limit @xcite at high-@xmath2 . while satisfactory in a broad sense , this older eos model suffers from three main drawbacks : 1 . the embedding of the dft - based 3-phase model into the coarse - grained model created kinks in the thermodynamic functions , even though an attempt was made to smooth out such features by interpolation . there are no solid phases beyond bc8 ; sc and sh phases , for instance , are not included . the electronic excitations of the high-@xmath4 liquid are treated with an average - atom thomas - fermi model , so atomic shell structure is not included in sufficient detail . in this work , we remedy these deficiencies by constructing a 5-phase eos for c , again based on ab initio calculations ( for densities between @xmath9 1 - 25 g / cc ) , in which the free energy models for each phase are defined over wide ranges of @xmath2 and @xmath4 and exhibit sensible limiting behavior . no embedding into a simpler model @xcite is performed . the two extra phases included are the sc and sh phases . while our phase diagram differs slightly from that of ref.@xcite , the inclusion of these extra phases allows our eos to be in accord with ab initio predictions for pressures up to well over 100 mbar . the electronic excitations in the liquid are treated with an average - atom model called purgatorio which , unlike thomas - fermi , includes the treatment of atomic shell structure within a kohn - sham dft framework @xcite . to constrain the ion - thermal model for the liquid , we perform dft - md in c up to 100,000 k and pimc caculations @xcite from @xmath9 200,000 k ( depending on the density ) to over @xmath10 k. we establish that the best fit to these eos predictions requires a model in which the decay of the ion - thermal specific heat is much faster than previously expected . our newly - developed `` cell model '' @xcite , in conjunction with the purgatorio electron - thermal term , provides an excellent fit throughout the range of both the dft - md and the pimc data . in what follows , we describe the details of our dft and pimc calculations , present the models we use for the free energies of the individual phases , and compare various thermodynamic tracks through our eos with the results of recent experiments performed on high - energy laser platforms @xcite . our computational results fall into two categories : 1 . calculations of the eos , by which we mean internal energy , @xmath11 , and pressure , @xmath12 , as functions of density , @xmath2 , and temperature , @xmath4 , and 2 . calculations of intermediate quantities we which use to build the eos . these include cold curves ( @xmath11 and @xmath12 as functions of @xmath2 , for ions fixed in position @xcite ) , phonon densities of states ( pdos ) , and electronic excitation contributions to the free energy . the first category of quantities we extract from dft - md ( for solid and liquid phases ) and pimc ( liquid phase ) . the second group of quantities we extract solely from dft calculations for the solid phases . it may seem at first as if our multitude of results overdetermines the eos , particularly in the liquid phase where both dft - md and pimc results reside . however we find it crucial to make use of this full suite of results ; the approximations inherent in our dft and pimc calculations are altogether different , and it is important to check whether certain features of our results are artifacts of the simulation method or are robust indicators of the true eos of c. this is especially the case for the liquid at high temperatures , where we find ( with both methods ) results which contradict some of our previous assumptions ( see below ) . fig.[pd_data ] shows the locations of our dft - md and pimc simulation points in the @xmath13-plane . also shown for reference are our predicted c phase lines , presented and discussed in detail below , as well as portions of five separate isentropes computed from the eos model described in this work . we stress that while the dft - md results are concentrated entirely in the diamond and liquid phases , we use the aforementioned dft - derived intermediate quantities ( cold curves , pdoss , etc . ) to constrain the eos model for each of the four solid phases considered here . all of our dft - md and pimc data , consisting of internal energy and pressure at different densities and temperatures , are reproduced in tables contained in the supplementary material . in the following two subsections , we describe in detail our approaches to obtaining these data . and pimc points for higher-@xmath4 . also shown are portions of five isentropes ( green lines ) computed from our eos model with entropy values increasing from the bottom : 3.78 @xmath14/atom , 5.49 @xmath14/atom , 6.61 @xmath14/atom , 12.56 @xmath14/atom , and 12.72 @xmath14/atom . note that the 0.1 g / cc isochore ( upper left - most red curve ) exists throughout a range in which the eos model of this work is _ not _ valid . ] with dft methods , we perform calculations of cold curves , phonon densities of states , electronic excitation contributions to the free energy , and the eos itself ( @xmath11 and @xmath12 as functions of @xmath2 and @xmath4 ) . we also present a limited number of calculations of ionic diffusivity , from which we can infer melt behavior . for all of these , we use the vasp code @xcite , together with projector augmented wave ( paw ) pseudopotentials @xcite . we use a `` hard '' paw with a core radius of 1.1 bohr and 4 valence electrons and the generalized gradient approximation ( gga ) of dft with the perdew - burke - ernzerhof ( pbe ) exchange correlation functional @xcite . the plane - wave cutoff is set to 1300 ev . for the phonon calculations , we use the primitive cell with a mp k - point grid of 40x40x40 for the simple cubic ( sc ) , the simple hexagonal ( sh ) phases and for fcc . the sh is done with a @xmath15 ratio of 0.986 evaluated by performing cell optimization for several compressions . no appreciable variation in the @xmath15 ratio with volume was observed . for the diamond and bc8 phases we used a 8 atom unit cell with a 20x20x20 k - points grid . for the sc phase , we compared with a 2 atom unit cell with a pmma spacegroup ( tetragonal distorsion ) @xcite which was calculated with a 20x40x40 k - point grid . for the molecular dynamics simulations , we use 64 atoms and perform born - oppenheimer md ( bomd ) within the nvt - ensemble with a nos - hoover thermostat @xcite . we use a time step of 0.75 fs in order to converge the internal energy and pressure to the desired accuracy . the electronic density is constructed from single - particle wave functions by sampling at the ( @xmath16,@xmath16,@xmath16 ) point of the brillouin zone . the electron occupation numbers are taken to be a fermi - dirac distribution set at the average temperature of the ions . for the different densities and temperatures , we use a sufficient number of bands such that we have at least 40 bands with occupation numbers smaller than 0.00001 . the md is run for 10000 steps with the last 5000 steps used for averaging the internal energy and the pressure . the use of pseudopotentials under extreme conditions of pressure and temperature can be problematic . for high compressions or high temperatures the inter - atomic distances can become smaller than the diameter of the paw sphere . in figure [ hpgofr ] , we plot the pair distribution functions for the densities 26.59 g / cc , 16.48 g / cc and 6.93 g / cc and temperatures of 50000 k and 100000 k , where clearly the paw spheres overlap significantly ( the black and blue dashed vertical lines are the diameter of the 1.1 and 0.8 bohr paw potentials respectively ; see below ) . at higher temperatures still , the thermal excitations of the 1s electrons can no longer be neglected . this can be addressed by having all 6 electrons in the valence ( for the current dft - md calculations , however , we find this not to be necessary for temperatures of 100,000 k or lower ) . at the highest compressions , the perturbation of the 1s orbital may be significant and can again be addressed by having all 6 electrons in the valence . in order to assess the sensitivity of our results to the paw potential chosen , we construct paw potentials with different cutoff radii , which range from 1.1 to 0.8 bohr radii . the vanderbilt projectors generation scheme is utilized @xcite . as the cutoff radii decrease , in order to maintain and optimize the paw performance , we add additional partial waves for both @xmath17 and @xmath18 angular momenta . different sets of reference energies are chosen , which are determined not to affect our computational results . moreover , when the carbon - carbon distance is sufficiently small , the 1@xmath17 core state is also included in the paw . for all cases tested , we find that a 5442 ev plane - wave cutoff energy is enough to reach convergence . for the paw test , we use fcc cells with lattice constants chosen so that the carbon - carbon nearest neighbor distances are representative of the distances observed during the highest-@xmath12 cold curve and md runs . these tests are performed with the quantum - espresso ( qe ) package @xcite . figure [ pawtest ] shows a comparison of the pressure obtained for this fcc cold curve with the different paws ( for both vasp and qe ) . t 50,000 k and 100,000 k from pbe - dft - md . the black and blue dashed vertical lines are the diameter of the 1.1 and 0.8 bohr paw potentials respectively . ] and for fixed ionic positions ) versus inter - ionic distance for fcc carbon with various paw potentials using pbe - dft . the black and blue dashed vertical lines are the diameter of the 1.1 and 0.8 bohr paw potentials respectively . ] we only observe appreciable deviations between the different paws for densities above 60 g / cc , corresponding to carbon - carbon distances of @xmath9 1.3 bohr ( 0.7 @xmath19 ) . the difference in the calculated pressure between the vasp @xmath20=1.1 bohr paw potential and the 6 electron paw is only 1% up to densities of 100 g / cc . this corresponds to carbon - carbon distances of 1.1 bohr which is essentially the smallest carbon - carbon distances sampled during the highest temperature and highest density md runs . however , this is no guarantee that the dynamics are not affected by the overlapping paw spheres . in this sense , we submit that our dft - md eos data for the very highest densities and temperatures reported here may be of lower accuracy than those for the more moderate conditions . nevertheless , we stress that the bulk of our conclusions below are unaffected by this fact . pimc is the most accurate and efficient first - principles simulation technique to study the equilibrium properties of quantum systems in high temperature plasma states . fermionic pimc simulations have been applied to study hydrogen @xcite , helium @xcite , hydrogen - helium mixtures @xcite , one component plasmas @xcite and most recently to simulate carbon and water plasmas @xcite . in pimc simulations , electrons and nuclei are treated equally as feynman paths in a stochastic framework for solving the full , finite - temperature , quantum many - body problem . the natural operator to work with in this context is the thermal density matrix represented as a path integral in real space , @xmath21 @xmath22 @xmath23\big\ } \label{pimceq1}\ ] ] where @xmath24 is a many - body hamiltonian , @xmath25 assumes the role of imaginary time , @xmath26 , @xmath27 is the position of the @xmath28th particle , and @xmath29 is the step size in imaginary time intervals . $ ] is the action , which determines the weight of every path . the kinetic energy operator controls the diffusion of the paths in imaginary time and keeps bead @xmath31 at adjacent time slices close together . a thermodynamic function corresponding to an operator , @xmath32 , can be derived from @xmath33 , where @xmath34 is the canonical partition function . pimc explicitly addresses all the physics of high-@xmath4 plasmas including effects of bonding , ionization , exchange - correlation , and quantum degeneracy on an equal footing , and thereby circumvents both the need to occupy single - particle states , and the need to employ exchange - correlation approximations inherent in dft . the coulomb interaction between electrons and nuclei is introduced using pair density matrices derived using the eigenstates of the two - body coulomb problem @xcite . the periodic images are treated using an optimized ewald break - up @xcite applied to the pair action @xcite . challenges with the pimc formalism arise in fermionic simulations because only antisymmetric eigenstates contribute to the partition function . those can be projected out by introducing an additional sum over all permutations , @xmath35 . straightforward integration methods all lead to unstable algorithms due to the _ fermion sign problem _ , which is the result of the near complete cancellation of positive and negative contributions to the many - body density matrix . a fixed - node approximation @xcite was initially introduced to solve the sign problem for ground - state quantum monte carlo calculations . for the finite - temperature pimc simulation , a restricted - path method was developed @xcite , where one only integrates over all paths that do not cross the nodes of a _ trial _ density matrix , @xmath36 , which must be available in analytic form . the fermionic version of eq . [ pimceq1 ] then reads @size9@mathfonts @xmath37^{\rho_t({\bf r}(\tau),{\bf r'};\tau))>0 } \end{array } } \!\!\!\!\!\!\!\!\!\!\!\!\ ! \mathcal{d}{\bf r}(\tau ) \exp\big\{-s[{\bf r}(\tau)]\big\}.\ ] ] in addition to weighting all paths according to their action and summing over permutations , one must also check whether the sign of the trial density matrix has remained positive during every monte carlo move . if a sign change occurs anywhere along the new path , the proposed move is rejected . the @xmath38 argument in the nodal restriction plays a special role because the sign of the trial density matrix at all other time slices depends on it . the restricted - path method provides the exact answer if the nodes from the exact many - body density matrix are used . since exact nodes are not known for interacting systems , one must work with approximate nodes . however , within any given set of nodes , the restricted path method will obtain the best possible solution that includes all interaction and correlation effects . the pimc method used here employs a free - particle nodal structure , which has been shown to be sufficient for carbon at temperatures where atoms still have occupied 1s states , but partially occupied 2s states ( @xmath39 k ) @xcite . using the restricted - path , free - particle nodal framework , we use all - electron pimc to compute equations of state for carbon at densities of 0.1 , 3.18 , 8.5 , and 11.18 g@xmath40@xmath41 over a temperature range of 10@xmath4210@xmath43 k ( though we present data pertaining to all four isochores in the supplementary material , we stress that 0.1 g / cc is below the low - density limit of applicability of the eos model described below ) @xcite . a sufficiently small time step associated with pimc path slices is determined by converging total energy as a function of time step until it changes by less than 0.2% . we use a time step of 0.0078125 ha@xmath44 for temperatures below 5@xmath45 k and , for higher temperatures , the time step decreases as @xmath46 while keeping at least four time slices in the path integral . in order to minimize finite size errors , the total energy was converged to better than 0.2% for a 24 atom cubic cell . we make the fundamental assumption that the helmholtz free enegy , @xmath47 , of each phase can be decomposed into the following terms : @xmath48 where @xmath49 and @xmath50 represent the contributions from ionic and electronic excitations , respectively . while the identification of a temperature - independent piece ( @xmath51 ) is always possible by virtue of the fact that it is merely definitional , the separation of the thermal part into decoupled ionic and electronic pieces is a major assumption . in a practical sense , this assumption is at least a reasonable starting point since for solids at lower-@xmath4 , it has been shown to be accurate @xcite , and for the liquid at very high temperature ( ideal gas limit ) it is also trivially satisfied because interactions between particles are of no importance . it is not necessarily justified for intermediate temperatures , but we shall invoke the approximation at any rate , as has been done in the past @xciteand take comfort in the fact that the ab initio eos data to which we fit in the liquid ( obtained from dft - md and pimc ) does _ not _ make this assumption . for the solid phases we consider diamond , bc8 , simple cubic ( sc ) , and simple hexagonal ( sh ) , we take @xmath52 to be the internal energy of the perfect crystalline lattice ( with motionless ions ) at @xmath5 @xcite with the electrons in their ground state . we fit this term to the dft data for the internal energy of the perfect lattice using the vinet equation of state @xcite , @xmath53,\ ] ] where @xmath54.\ ] ] = 0 ( and for fixed ions ) in ev / atom versus @xmath55/atom in @xmath56 for various phases of c. points indicate the results of pbe - dft calculations , while the thin curves are fits to these data used in our eos model . the phases represented are : diamond ( red ) , bc8 ( blue ) , sc ( cyan ) , and sh ( black ) . ( a ) shows the whole range over which the eos model is constructed , while ( b ) shows a closer view of the range between @xmath57 2 and 5 @xmath56/atom . , title="fig : " ] = 0 ( and for fixed ions ) in ev / atom versus @xmath55/atom in @xmath56 for various phases of c. points indicate the results of pbe - dft calculations , while the thin curves are fits to these data used in our eos model . the phases represented are : diamond ( red ) , bc8 ( blue ) , sc ( cyan ) , and sh ( black ) . ( a ) shows the whole range over which the eos model is constructed , while ( b ) shows a closer view of the range between @xmath57 2 and 5 @xmath56/atom . , title="fig : " ] here , @xmath58 is the bulk modulus at the volume @xmath59 , @xmath60 is the pressure - derivative of the bulk modulus at @xmath59 , and @xmath61 is the internal energy at @xmath59 . fig . [ colds ] shows the computed cold curves of the various phases , together with our fits to them , using eq.[vinet ] . the cold curve parameters for each phase ( @xmath61 , @xmath58 , @xmath60 , and @xmath59 ) are given in table 1 , along with the other phase - dependent parameters we discuss below . our ( pbe ) gga - dft results for the solid cold curves are very similar to those of ref.@xcite , but there are differences which result presumably from the use of different pseudopotentials and associated choices for the plane wave energy cutoff in the determination of the internal energy . though small , these differences give rise to rather pronounced changes in the predicted phase transition pressures ( bc8 @xmath7 sc @xmath7 sh ) when compared to the phase diagram presented in ref.@xcite . this will be discussed more below , after the thermal components of the free energies are considered and our phase diagram is presented . at this point , we note that we have checked that our inferred _ cold _ transition pressures , determined from the two - phase maxwell construction @xcite , should be inaccurate by no more than a few percent due to our use of the vinet fitting form , eq.[vinet ] ( though we in no way imply that pbe gga - dft is necessarily this accurate for predicting these transition pressures at such high compressions ) . for @xmath62 , we use the ion excitation free energies of the various lattices computed for cases in which the electrons are in their instantaneous ground state ( for each ionic configuration ) . in other words , we invoke the born - oppenheimer approximation . in addition , the quasiharmonic approximation is assumed for each phase . we investigated anharmonicity for diamond and bc8 phases in ref.@xcite and showed its effects to be small ; the effects of anharmonicity in sc and sh phases await further study . given this quasiharmonic description , we use double - debye models @xcite for the diamond , bc8 , and sc phases , and a regular single - debye model for the sh phase . all are fit to the @xmath55-dependent phonon densities of states ( pdoss ) for these various phases computed with dft using a linear response approach . in the double - debye prescription , the pdos for a given volume is presumed to consist of two overlapping features , each of which is of the form : @xmath63 for @xmath64 and 0 for @xmath65 , where the two unequal @xmath66 functions are related to various moments of the pdos @xcite . since the double - debye model allows for two different values of @xmath67 ( say , @xmath68 and @xmath69 ) for each @xmath55 , it is useful in cases where the high-@xmath4 phonon moment , @xmath70 $ ] , is rather different from the low-@xmath4 zero - point moment , @xmath71 . this is certainly the case for the diamond and bc8 phases , especially at high compressions , as discussed in detail in ref.@xcite . our calculations for the sh phase shows that @xmath72 and @xmath73 are very nearly equal throughout the predicted range of its stability . thus , we use a single - debye free energy for the sh ion - thermal term to reduce the total number of fitting parameters . [ thetas ] shows the three phonon moments , @xmath74 , and @xmath75 ( @xmath76 , as functions of @xmath55 for bc8 , sc , and sh phases as determined from our computed pdoss . note that the @xmath77 for the different phases nearly coincide in the neighborhood of the volumes where their cold curves cross . for this reason , the @xmath4 vs. @xmath12 phase lines are rather vertical in nature ( see below ) . it is from these moments that we derive ( for the diamond , bc8 , and sc phases where the double - debye model is used ) the two debye temperatures , @xmath68 and @xmath69 , by solving the system of equations given as eqs.9 - 12 of ref . @xcite . for all @xmath78 functions for the various phases , we assume the @xmath55-dependence : @xmath79 , \label{gamma}\ ] ] where @xmath80 is the value at a reference volume @xmath81 . this arises from the assumption that the ion - thermal grneisen parameter for @xmath66 is equal to @xmath82 @xcite . the relevant ion - thermal parameters ( @xmath80 , @xmath83 , @xmath84 , and @xmath81 ) which determine the various debye temperatures for each phase are presented in table 1 . in terms of @xmath68 , @xmath69 , and @xmath73 , the double - debye free energy is written as @xmath85f_{\rm a}(v , t ) + \left[\frac{\theta_{1}(v ) - \theta_{\rm a}(v)}{\theta_{\rm b}(v ) - \theta_{\rm a}(v)}\right]f_{\rm b}(v , t),\ ] ] where @xmath86 and @xmath87 are the single - debye free energies for the individual a and b peaks associated with a given pdos . ( circles ) , @xmath73 ( squares ) , @xmath75 ( diamonds ) in kelvins versus @xmath55/atom in @xmath56 for bc8 ( blue ) , sc ( cyan ) , and sh ( black ) phases as computed with linear response methods using pbe - dft . ] the electronic excitation term for each solid phase , @xmath88 , is taken be of the form , @xmath89 which is motivated by a performing a sommerfeld expansion of the electronic excitation free energy , assuming @xmath90 to be a small parameter @xcite . this contribution is extracted from pbe gga - dft calculations in which electronic occupancies are constrained by a fermi - dirac distribution with temperature @xmath4 , but for ions _ fixed _ in their lattice positions . as discussed in ref.@xcite , this term is essentially zero for the diamond phase throughout its stability field , and is exceedingly small ( though non - zero ) for the bc8 phase as well @xcite . for this reason , electronic excitation contributions were intentionally neglected by those authors in their solid - phase eos models . here we choose to include these terms ; their values for the sc and sh phases are somewhat larger than those of bc8 , because the electronic density of states near @xmath91 generally increases with compression . still , their inclusion has a rather small effect on eos and the resulting phase lines . we fit these small dft - based terms by assuming @xmath92 @xcite ; our choices for @xmath93 , @xmath94 , and @xmath95 are presented in table 1 for each phase . note that there are nonzero values for @xmath96 which appear for the diamond and bc8 phases in this table ; these result from earlier extractions of anharmonic ion - thermal free energy contributions @xcite ( which happen also to have a @xmath97 dependence ) , and are _ not _ due to electronic excitations @xcite . ( k ) for the diamond phase . ( b ) pressure ( gpa ) isochores versus @xmath4 ( k ) for the diamond phase . red points are the results of dft - md , blue lines are the results of our eos model . in both plots , @xmath55 ranges from 3.0683 @xmath56/atom to 5.3903 @xmath56/atom . , title="fig : " ] ( k ) for the diamond phase . ( b ) pressure ( gpa ) isochores versus @xmath4 ( k ) for the diamond phase . red points are the results of dft - md , blue lines are the results of our eos model . in both plots , @xmath55 ranges from 3.0683 @xmath56/atom to 5.3903 @xmath56/atom . , title="fig : " ] in order to check the validity of our solid - phase eos model for c , we compare to our dft - md results for pressure and internal energy in the diamond phase . these results exist throughout a range where diamond is predicted to be the stable phase ( @xmath98 , and @xmath99 ) . [ diamond ] shows that the model agrees with the diamond - phase dft - md quite well , even though these data , per se , were not used in the fitting ( rather , only cold curves , pdoss , and electronic doss were used ) . various factors contribute to the slight disagreement between the model and the dft - md data : 1 . the model is built from the diamond cold curve as calculated for an effectively infinite system ( affected by a large number of @xmath100-points ) , while the dft - md was performed with a limited number of atoms ( 64 ) in a supercell and a single @xmath100-point . the ion - thermal term of the diamond free energy model is taken to be that of ref.@xcite ; the phonon calculations therein were performed with different pseudopotentials and different plane wave cutoff energies . this is likely to be a very small contributor to the eos differences , however , for the relevant moments of the pdoss at various volumes are strikingly similar to those resulting from the computational framework adopted in this work . .phase - dependent eos model parameters for the solid phases of our multiphase c eos . the upper segment of the table concerns cold curve parameters , the middle segment quasiharmonic ion - thermal parameters , and the lower segment electron - thermal / anharmonic - ion - thermal parameters . all volumes ( @xmath55 ) are in @xmath56/atom , @xmath58 is in gpa , @xmath60 is unitless , @xmath61 is in ev / atom , all characteristic temperatures ( @xmath67 ) are in kelvins , @xmath83-parameters are in @xmath101 , @xmath84-parameters are unitless , @xmath93 is in kelvins@xmath44 , and @xmath94 is unitless . [ cols="^,^,^,^,^",options="header " , ] given the free energy models for the individual solid ( diamond , bc8 , sc , sh ) and liquid phases , we construct the phase diagram and multiphase eos by invoking the maxwell construction for each pair of phases ( here denoted 1 and 2 ) , and at each temperature , @xmath4 , of interest @xcite : @xmath102 where @xmath103 are total free energies for each phase , @xmath104 are the transition volumes , and @xmath12 is the transition pressure . since the individual solid phases are only meaningful ( as in metastable ) in the neighborhood of their fields of thermodynamic stability , we only consider each solid phase free energy function throughout a restricted range of @xmath55 , so as to minimize the spurious effects of improper extrapolation . in particular , we consider diamond from 2.5 @xmath56/atom @xmath7 7.0 @xmath56/atom , bc8 from 1.3 @xmath56/atom @xmath7 3.5 @xmath56/atom , sc from 1.0 @xmath56/atom @xmath7 7.0 @xmath56/atom , and sh from 0.01 @xmath56/atom ( undoubtedly well beyond its actually range of stability ) @xmath7 7.0 @xmath56/atom . the liquid is considered throughout the wide range : 0.01 @xmath56/atom @xmath7 7.0 @xmath56/atom , though the actual intended range of validity of the eos model reported here is limited to the regime covered by our _ ab initio _ calculations ( see supplementary material ) : 0.75 @xmath56/atom @xmath105 @xmath55 @xmath106 5.4 @xmath56/atom ( as we stated above , we choose not to require agreement with the 0.1 g / cc pimc isochore in this work ) . [ pd]a shows the resulting phase diagram in the regime of pressure and temperature addressed in the work of ref.@xcite . our phase lines computed in this work differ only slightly from those in that earlier study . in particular , the diamond - bc8 transition pressure is slightly lower ; this is due to our diamond and bc8 cold curves being slightly different from those of ref.@xcite ( higher plane wave cutoffs and different pseudopotentials were used here ) . also , the bc8-phase melt temperature is slightly lower , as evidenced by our comparison to the kechin fits to the carbon melt curves of ref.@xcite to which the eos model of ref.@xcite was fit . nevertheless , our phase diagram is quite close to that of the older model in the range of its applicability . also shown in fig . [ pd ] are various isentropes as computed with our multiphase eos model ; values of the entropy for each are listed in the figure caption . the principal hugoniot of carbon computed with the model is shown as well in fig . this is defined to be the locus of final states accessible via a planar one - dimensional shock , given a particular assumed initial state , here taken to be @xmath107 3.52 g / cc and @xmath108 300 k. this curve has two branches : one for which the final state is the diamond phase , and one in which the liquid is the final state . they are separated by a flat region in @xmath4 straddling the pressure at which the diamond @xmath7 bc8 transition is predicted to occur . in this portion of the hugoniot curve the final state is in a mixed - phase region , the size of which is directly related to the latent heat of melting , as discussed in ref.@xcite . we predict a principal hugoniot which is similar to that predicted by the earlier carbon eos model @xcite ( indeed , it is nearly identical in the diamond - phase portion ) . however , there is a pronounced increase ( @xmath9 100 gpa ) in the hugoniot final state pressure of the liquid portion , which is explained by the added latent heat of melting resulting from the addition of the purgatorio electron - thermal contribution . our change in entropy from solid to liquid at constant @xmath55 at the shock melt conditions is @xmath109/atom , while in the model of ref.@xcite it is @xmath110/atom . this in turn moves our liquid branch of the hugoniot a bit closer to the measured @xmath4 vs. @xmath12 hugoniot data of eggert et al.@xcite ( the upper and lower bounds of their experimental error bars are displayed as well in fig . [ pd]a ) . our hugoniot still fails to fall within the bounds of their measurements , however . the phase diagram of our multiphase c eos model over a larger range of pressure and temperature is shown in fig . [ pd]b . the higher-@xmath12 solid phases , sc and sh , appear as well , in addition to diamond , bc8 , and liquid phases . several important features are worth noting : 1 . the bc8 @xmath7 sc transition pressure is nearly independent of temperature , as is the case for the diamond @xmath7 bc8 transition pressure . this is because the relevant moments , @xmath66 , of the pdoss of the phases on either side of each transition are very similar at the densities at which the transitions occur , as discussed above . the sc @xmath7 sh phase line is less vertical , and curves to lower-@xmath12 for higher @xmath4 , but we do _ not _ predict that this phase line intersects the bc8-sc phase line before intersecting the melt curve . this is in stark contrast to the prediction of ref.@xcite , which shows a sc - sh phase line which intersects the bc8-sc line at @xmath111 k. though we use similar ab initio and eos modeling methods to theirs , we suspect that their treatment of the @xmath4-dependence of the transition pressure in their gibbs free energy matching prescription is sufficiently different from our approach to produce this discrepancy . we have checked that alternate ways of fitting our eos data ( cold curves , pdoss and resulting ion - thermal free energy terms ) still give the general phase diagram topology we present here , though we do not discount the possibility that the extreme sensitivity of the phase lines to small changes in the individual phase free energies can render our predictions somewhat inaccurate , especially at these higher pressures where the @xmath112 curves are nearly parallel . the value of the sc - sh transition pressure at low-@xmath4 is @xmath113 higher than that reported in ref.@xcite , though the position of the bc8-sc transition line is in reasonable agreement between the two studies . this is likely a result of the use of different pseudopotentials and plane wave convergence criteria . again , the nearly coincident free energy functions for different phases at high pressures makes the phase lines extremely sensitive to any otherwise subtle changes . 4 . the melt line of the sh phase possesses a maximum at @xmath114 gpa . this is striking , but is one of the least certain predictions we make in this work ; we have seen that slight changes to our eos model parameters which produce small changes to our liquid free energy ( which are within both our assumed model and dft - md uncertainties ) yield sc melt temperatures which exhibit markedly different behavior . our reason for favoring this particular eos model parameterization over other nearly equivalent ones can be found in our computation of the ionic diffusivities using the dft - md . [ diff ] shows two diffusivity isotherms , the lower one for @xmath108 10,000 k and the upper one for @xmath108 20,000 k. a very small ( approaching zero ) diffusivity is indicative of a solid phase . the ionic diffusivity at 20,000 k shows the expected decrease as @xmath55 decreases , but is otherwise large as expected for a liquid . the diffusivity at 10,000 k shows a pronounced dip towards zero at a @xmath115 1.2 @xmath56/atom , corresponding to a pressure of roughly 10,000 gpa ( see fig . [ highpliqp ] ) . this is a clear indication of solidification , and indeed , the phase diagram of fig . [ pd]b shows a melt curve which attains its maximum ( of just over 10,000 k ) at @xmath116 gpa . though this provides a nice consistency check on our dft - md - derived c eos model , we caution that the diffusivity calculations were performed in a cell of 64 atoms ( as were the dft - md calculations of eos ) , and solidification can be biased in simulations with small cells of fixed shape @xcite . the overarching points to be made about our prediction of the phase diagram of c are two - fold : first , our predictions of phase lines are necessarily less accurate the higher the pressure is , for the reasons mentioned directly above . these inaccuracies are the combined result of the inherent uncertainties in the underlying ab initio electronic structure data , and the deficiencies of eos models which have a finite number of adjustable parameters . second , the predicted volume changes across such high-@xmath12 phase lines are often exceedingly small ( which is what makes the accurate determination of these transition lines difficult ) . thus , the least accurate phase lines have a minimal impact on the eos itself ( @xmath12 and @xmath11 at a given @xmath2 and @xmath4 ) . /atom , 5.49 @xmath14/atom , 6.61 @xmath14/atom , 12.56 @xmath14/atom , and 12.72 @xmath14/atom . , title="fig : " ] /atom , 5.49 @xmath14/atom , 6.61 @xmath14/atom , 12.56 @xmath14/atom , and 12.72 @xmath14/atom . , title="fig : " ] /s vs. @xmath55 in @xmath56/atom for two isotherms : @xmath108 10,000 k ( lower ) and @xmath108 20,000 k ( upper ) , as computed by dft - md . ] we have already discussed the comparison of the results of our c eos model to principal hugoniot data in the @xmath13-plane . these recent measurements made use of optical pyrometry techniques @xcite which lie outside the typical purview of classic shock - compression experiments . now we compare to a larger set of principal hugoniot data for elemental carbon transformed more typically to the @xmath117-plane , as well as to room - temperature isotherm data from static compression experiments . [ rh ] shows the 300 k isotherm ( red ) and the principal hugoniot ( blue ) in @xmath117-space as calculated from our multiphase c eos model . the isotherm ( which is very similar to the principal isentrope ) shows flat regions where the phase transitions occur ( bc8 @xmath7 sc and sc @xmath7 sh in this plot ) . the @xmath12 vs. @xmath2 hugoniot is nearly indistinguishable from the isotherm below roughly 1000 gpa , but deviates dramatically from it at higher stresses . the point at which they diverge corresponds to the point where the hugoniot final states begin to reside in the liquid phase ( see fig . [ pd ] ) . also shown on this figure are data from static compression measurements of the room - temperature isotherm @xcite ( green circles at the lowest @xmath12 ) , magnetically - driven flyer plate studies @xcite ( dense set of magenta + symbols with @xmath118 6 - 7 g / cc ) and multiple sets of laser - shock data on the principal hugoniot of diamond ( symbols with error bars ) @xcite . note that the much of the highest-@xmath12 hugoniot data @xcite seems to straddle the 300 k isotherm , even well above @xmath119 gpa ; indeed , the data of brygoo et al . @xcite even falls below our prediction of the room-@xmath4 isotherm . this is very puzzling in light of our theoretical results which suggest that the hugoniot should be much stiffer than the isotherm at these larger compressions . our previous ab initio - based c eos @xcite was only fit to dft - md data up to @xmath108 20,000 k , which translates to @xmath120 1600 gpa on the principal hugoniot . at this pressure , the hugoniot and the 300 k isotherm are still rather close . the eos model of this work , however , is validated by comparing to calculations ( dft , pimc ) that span the full range of temperature up to ideal gas conditions ; note in particular the comparison to dft - md data on a dense grid of @xmath121 as pictured in figs . [ liqe ] and [ liqp ] which covers the entire range of conditions relevant for the comparison of fig . [ rh ] . in order to check the robustness of our conclusions that the principal hugoniot should indeed be far stiffer than the 300 k isotherm for @xmath122 1000 gpa , we have verified that the different variants of liquid c eos we discussed in section iii.b.1 ( cowan + purgatorio , cell + thomas - fermi , etc . ) all show this same basic relationship between the low-@xmath4 isotherm and the hugoniot . looking back again to the experimental results pictured in fig . [ rh ] , we note that there are data with @xmath122 1000 gpa that seem to be closer to our hugoniot predictions , in particular those of nagao et al . @xcite and hicks et al . it must be mentioned that the hicks et al . data shown here is _ not _ that as presented in the original hicks et al . reference , but is instead a data set which results from a reanalysis @xcite of the quartz eos @xcite which was used as an impedance - matched standard in that work . the original uncorrected data set @xcite lies much closer to our low-@xmath4 isotherm than to our principal hugoniot prediction in this higher-@xmath12 region . these experimental improvements notwithstanding , we still submit that further work must be done to resolve the theory - experiment discrepancy at the very highest compressions shown here . we have constructed a 5-phase eos for elemental carbon based entirely on ab initio electronic structure calculations of the density functional theory and path integral monte carlo varieties . the pimc and high - temperature dft - md data were particularly useful in helping us differentiate between different choices of ion - thermal ( cowan , cell ) and electron - thermal ( thomas - fermi , purgatorio ) models . we found that these data strongly favor the cell + purgatorio combination . the multiphase eos is constructed entirely without patching different models together in an ad hoc fashion ; rather , full thermodynamic consistency is maintained and the appropriate high-@xmath4 limit is reached in a seamless manner . our eos model includes , in addition to diamond , bc8 , and liquid , two of the ultra - high pressure solid phases recently predicted by martinez - canales et al . @xcite : simple - cubic , and simple - hexagonal . though our prediction of the phase diagram of carbon in this high-@xmath12 region is broadly similar to theirs , the detailed positions of phase lines and resulting triple - points are somewhat different , owing to the sensitivity of the phase lines to the detailed prescriptions for obtaining the individual phase free energies and computing the transition pressures . further work should be done to resolve these details , though we maintain that the eos itself should be reasonably accurate at high compressions in spite of this . comparison to recent laser - shock compression data @xcite shows notable disagreement with the subset of results @xcite which suggest that the principal hugoniot is nearly coincident with the low-@xmath4 isotherm even at pressures in excess of 1000 gpa . it is not at all clear at the moment as to what is causing such a discrepancy , particularly since the general features of our predictions ( seen in fig . [ rh ] ) are quite robust and seemingly independent of many of the choices we have made in modeling the eos of liquid c. we are encouraged , however , that a more recent reanalysis @xcite of the data of hicks et al . @xcite seems to bring their principal hugoniot into better agreement with our prediction . one of the central results of this work is that our high-@xmath4 dft - md and pimc data on carbon are found to be in excellent agreement with a new ion - thermal free energy model . this so - called cell model exhibits an evolution from dulong - petit to ideal gas limits as @xmath4 is increased which is more rapid than that of the older cowan model . although the results presented here are specific to carbon , existing eos models for other materials that are based on the widely - used cowan model may need to be re - examined in this light , as those materials may also possess a somewehat rapid decay of @xmath123 with @xmath4 . recent work with simpler classical inter - ionic potentials suggests that such rapid decays may be common @xcite . we thank c.j . pickard , m. martinez - canales , j. dubois , m.a . morales , g.w . collins , p.a . sterne , h.d . whitley , d.m . sanchez , j.i . castor , d. ho , d. braun , and r.f . smith for helpful discussions . this work was performed under the auspices of the u.s . department of energy by lawrence livermore national laboratory under contract no . de - ac52 - 07na27344 . correa , s.a . bonev , and g. galli , proc . sci . u.s.a . * 103 * , 1204 ( 2006 ) . an important earlier work first suggesting the existence of a maximum in the diamond melt curve is : m.p . grumbach and r.m . martin , phys . b * 54 * , 15730 ( 1996 ) . strictly speaking , our cold curves describe the material at @xmath5 , assuming _ infinite ionic mass _ ; quantum zero - point motion is therefore not included in this contribution . although the zero - point free energy ( and hence , @xmath11 and @xmath12 as well ) is independent of @xmath4 , we choose to lump the zero - point energy into our ion - thermal term , since it arises from the consideration of ionic excitations . g. kresse and j. hafner , phys . b , * 47 * , 558 ( 1993 ) ; g. kresse and j. hafner , phys . rev . b , * 49 * , 14251 ( 1994 ) ; g. kresse and j. furthmller , comput . mat . , * 6 * , 15 ( 1996 ) ; g. kresse and j. furthmller , phys . rev . b , * 54 * , 11169 ( 1996 ) . it is important to note that gga - dft generally underestimates the electronic excitation gaps of materials , resulting in metalization pressures that are too low , and electronic excitation free energies ( for systems on the border between insulating and metallic ) which are too large . in this sense , our predictions of the effects of electronic excitation on the eos of solid c are likely to be overestimates . for all solid phases whose models possess a @xmath97 ( electron - thermal or anharmonic ion - thermal ) free energy term , we apply corrections above @xmath4= 20,000 k , well beyond their ranges of stability , which smoothly force @xmath49 to a constant independent of @xmath4 . this is done to prevent these solid phases from becoming more stable than the liquid at high-@xmath4 , which would eventually occur if the @xmath97 term , meant strictly as a low-@xmath4 correction , is allowed to persist at high temperatures . for a given break - point , the cold curve function is changed according to : @xmath124^{n}}{b + [ ( v_{b}/v ) - 1]^{n}}$ ] for @xmath125 , and no correction for @xmath126 . when there are multiple break - points applied for a given phase ( there are four in our liquid model ) , it is understood that these additions are applied _ in succession _ ( i.e. , first `` 1 '' , then `` 2 '' , etc . ) . the requirement that @xmath127 satisfies the lindemann relation with the liquid ion - thermal grneisen parameter ensures that @xmath128 ( ideal gas pressure ) as @xmath129 @xcite . see this reference for the original discussion of the lindemann relation : f. lindemann , z. phys . * 11 * , 609 ( 1910 ) . note that @xmath127 is _ not _ the melting temperature of c ( as in the original cowan prescription used in ref.@xcite ) , but is rather the temperature above which the ion - thermal specific heat begins to deviate from the dulong - petit limit of @xmath130 and head toward the ideal gas value of @xmath131 @xcite . note also that the conclusions we draw in this work regarding the lack of suitability of the cowan model also apply if we lower the value of @xmath132 to 10,000 k , which is just above the melt temperature of c over a wide range of compressions . our definitions for the fractional differences in @xmath11 and @xmath12 would be problematic if @xmath133 or @xmath134 were close to zero for some @xmath4 ; this is not the case for the @xmath4 and @xmath2 considered here ( and given the definition of the zero of internal energy employed ) . we have also computed the internal energy of an isochore with @xmath135/atom using dft - md , and have shown that there is a pronounced jump in @xmath11 as @xmath4 is raised above 10,000 k , which again is right in the neighborhood of the maximum of the melt curve of the sh phase . for this set of calculations , a periodic cell with 392 c atoms was used . m. hafland , h. beister , and k. syassen , phys . rev . b * 39 * , 12598 ( 1989 ) . note that this work described static compression of graphite , however for densities above @xmath9 3.5 g / cc , a clear transition to diamond - like compressibility is seen .

chiral symmetry plays an essential role in particle physics . in qcd it is a global symmetry which forbids an additive quark mass renormalization and whose spontaneous breakdown provides us with ( nearly ) goldstone bosons with their specific interactions . on the other hand , in electroweak interactions chiral symmetry plays an even more fundamental role : it is a local gauged symmetry . its presence is necessary to provide a renormalizable theory . accordingly , when trying to regularize these theories on the lattice ( which is the only non - perturbative regularization known for gauge theories ) one meets two basic problems . as it is ( or rather _ was _ ) well known one can not realize chiral symmetry exactly on the lattice without sacrificing some even more important principles , like locality or absence of extra physical particles . this is the content of the famous no - go theorem by nielsen and ninomiya @xcite . in wilson s formulation of lattice qcd chiral symmetry is explicitly broken by an irrelevant term which does not affect the continuum limit . it causes , however , o@xmath0 lattice artifacts and other unwanted effects at finite lattice spacing . for the electroweak theory , on the other side , the problem of chiral symmetry on the lattice is a principal one : is it possible to define a chiral gauge theory non - perturbatively ? this second problem is much more difficult , and assumes that the first one a solution for global chirality has been successfully solved . several ways to solve these problems have been proposed . these were mostly covered on previous lattice conferences . here i mention only the domain wall fermions @xcite and the related overlap formalism @xcite . on can state that these approaches provide a solution to the problem of global chirality . the numerical results using domain wall fermions @xcite are encouraging . in the overlap formalism also important steps were made towards the solution of the gauged chiral symmetry ( for references see ref . @xcite ) . it was unclear , however , to what extent are the theoretical constructions used in these approaches necessary , or what is the basic ingredient behind the chiral properties observed . as it became clear recently , the basic relation which has to be satisfied to realize chiral symmetry at finite lattice spacing , is the ginsparg - wilson ( gw ) relation @xcite suggested a long time ago ! in this talk i shall concentrate on the role of the gw relation and its consequences . some of the results were obtained earlier in the approaches mentioned above , but even these are more transparent within the new framework . consider the free fermionic action @xmath1 describing a massless fermion on the lattice . ( colour , dirac and flavour indices are suppressed . ) the desirable properties of the dirac operator @xmath2 should be * @xmath3 is local ( bounded by @xmath4 ) * @xmath5 for @xmath6 * @xmath7 is invertible for @xmath8 ( no massless doublers ) * @xmath9 here @xmath7 is the fourier transform of @xmath3 . the theorem by nielsen and ninomiya @xcite states that properties ( a)-(d ) can not hold simultaneously . following wilson , to avoid the massless doublers present in the naive discretization of the dirac operator , one introduces a chiral symmetry breaking , irrelevant operator . the corresponding wilson - dirac operator is given by @xmath10 where @xmath11 ( @xmath12 ) is the forward ( backward ) lattice derivative . due to the explicit breaking of chiral invariance , however , some extra problems appear : \(1 ) chiral symmetry is broken at finite lattice spacing by o(@xmath13 ) lattice artifacts and it is recovered only in the continuum limit . \(2 ) one has to fine tune the bare quark mass since there is an additive quark mass renormalization . \(3 ) operators of different chiral representations get mixed , etc . one of the conditions of the no - go theorem was locality . the natural definition ( used here ) is that the dirac operator in an arbitrary gauge field background falls off exponentially i.e. satisfies the bound @xmath14 where @xmath15 and @xmath16 are independent of the gauge field @xmath17 . in addition , it is also assumed that @xmath18 depends negligibly on @xmath17 s far apart from @xmath19 and @xmath20 , i.e. that @xmath21 also falls exponentially in @xmath22 , @xmath23 and @xmath24 . it is generally assumed that such interactions satisfy universality . note that it is physically too restrictive to consider only nearest neighbour interactions or even those with strictly finite support . in a real system the interaction coefficients die away exponentially rather then having a finite number of strictly nonzero terms . on the other hand , interactions decreasing more slowly then exponentially should be considered as non - local and non - acceptable , since universality is not expected to hold for these . some non - local interaction can give the same correlations as a local one ( e.g. when a bad rg transformation is used , or when some physical fields are integrated out ) . this is , however , no sufficient excuse to use them a slight change in the interaction could modify the picture drastically . theoretically any decay rate @xmath25 in eq . ( [ bd ] ) is allowed . in practice , however , one works at finite lattice spacing and @xmath26 must hold . for that reason it is important to have a sufficiently small interaction range . in principle it could happen that the decay rate of @xmath18 is not uniformly bounded by an exponential , but there exists a bound @xmath27 ( with @xmath28 ) . in general , such interactions should be considered non - local . however , it may happen that those configurations when @xmath29 are so strongly suppressed that they practically ` never ' occur in a mc simulation . this is a potentially dangerous possibility , but could perhaps be accepted with extra caution . back in 1982 ginsparg and wilson @xcite suggested a way to avoid the no - go theorem and preserve consequences of the chiral symmetry . they suggested to require instead of relation @xmath9 the following milder condition @xmath30 here @xmath13 is the lattice spacing and @xmath31 is a _ local _ operator . for simplicity it will be assumed also that @xmath31 is trivial in dirac indices . that @xmath31 commutes with @xmath32 follows from eq . ( [ 1 ] ) . the locality of @xmath31 expresses the requirement that the propagating states are effectively chiral , so at distances larger than the range of @xmath31 its presence is not felt . obviously , this is a highly nontrivial condition since the propagator @xmath33 on the lhs . is non - local . ( the coefficient 2 on the rhs . is for historical reasons only . ) accordingly , the dirac operator @xmath2 should satisfy the ginsparg - wilson relation @xmath34 in ref . @xcite this relation originates from rg considerations applied to the free fermionic action and has been generalized to fermions in a gauge field . the local operator @xmath31 comes from a chiral symmetry breaking term in the _ block transformation _ only , not from the original action of the rg procedure . for that reason the physics described by @xmath2 is expected to be chirally invariant . ( for a chirally symmetric block transformation @xmath2 becomes chirally invariant but non - local , in accordance with the no - go theorem . ) two comments are in order here : \(1 ) the naive commutation property for @xmath2 is recovered in the continuum limit @xmath35 . \(2 ) the rhs . of eq . ( [ 2 ] ) is zero on solutions , i.e. for @xmath36 . ginsparg and wilson suggested that a solution to eq . ( [ 2 ] ) is the _ mildest _ way to break chiral invariance : it should maintain the physical consequences of chiral invariance ( soft pion theorem , etc . ) . however , no solution was found for the interacting case ( qcd ) , and the paper has been practically forgotten for 15 years ! last year peter hasenfratz realized @xcite that the fixed point ( fp ) action ( or classically perfect action ) for qcd @xcite satisfies the gw relation . this observation revived the interest in this relation . let me recall the definition of the fp action . for the pure gauge part of the action one has the saddle point ( i.e. classical ) equation @xmath37 here the gauge field @xmath38 lives on the coarse lattice , @xmath17 on the fine one , and @xmath39 is the blocking kernel . this is a recursive definition , but its recursive evaluation converges rapidly since the minimizing field @xmath40 is much smoother than the original @xmath38 . in the classical approximation the fermionic part remains quadratic , and the corresponding propagator satisfies the recursion relation @xmath41 here @xmath42 label the sites on the coarse lattice , @xmath43 on the fine lattice , while on the rhs . @xmath40 , the minimizing configuration of the pure gauge problem is taken . the term proportional to @xmath44 comes from the chiral breaking term in the blocking kernel for fermions ; @xmath45 characterizes the averaging procedure for the fermions in rg transformation it is assumed to be trivial in dirac indices and usually it only extends over the hypercube . applying eq . ( [ 4 ] ) recursively and using the fact that the original @xmath2 ( in the continuum ) anti - commutes with @xmath32 , it is easy to see that @xmath46 satisfies eq . ( [ 1 ] ) with @xmath47 ( in fact , for the rg transformation used , @xmath48 or it lives on the hypercube , i.e. is even ultra - local . ) in ref . @xcite hasenfratz has demonstrated that , indeed , the gw relation implies the chiral ward identities . the argument is the following : when the symmetry breaking part appears in some expressions ( integrated out in fermionic fields ) @xmath49 then the @xmath33 factors from contractions of the fermionic fields are cancelled by the @xmath2 s on both sides in the breaking term , leaving only the local operator @xmath31 . as a consequence , the symmetry relation @xmath50 is satisfied up to a contact term i.e. one obtains the ward identities . in this paper it has also been shown that the proper order parameter for spontaneous chiral symmetry breaking is given by @xmath51 it has been proven as well that the current is not renormalized ( @xmath52 ) and operators in different chiral representations do not mix . i shall discuss these issues later , in the light of newer developments . in ref . @xcite it has been shown that there is an exact atiyah - singer index theorem on the lattice , and the fermionic spectrum ( in arbitrary gauge background ) has nice chiral properties . the discussion was based on the fp action , but it used mostly only the gw relation to establish these facts . as mentioned in the introduction , the domain wall and the related overlap formalism also produced dirac operators with nice chiral properties . neuberger @xcite has pointed out that the dirac operator obtained earlier @xcite also satisfies the gw relation . to be on more general ground , i discuss first the overlap dirac operator . this is an explicit construction . introduce @xcite @xmath53 where @xmath54 is the standard massless wilson - dirac operator given by eq . ( [ dw ] ) . note that @xmath55 then define @xmath56 it satisfies the gw relation @xmath57 ( i.e. eq . ( [ 2 ] ) with @xmath58 ; sometimes this special case will be referred to as the gw relation ) . introducing @xmath38 through @xmath59 together with the relation @xmath60 one concludes that eq . ( [ 12 ] ) is equivalent to @xmath61 i.e. @xmath38 is unitary . in other words , @xmath62 with a unitary @xmath38 defines a general solution to eq . ( [ 12 ] ) . of course , eq . ( [ 14 ] ) in itself is not enough to define an acceptable dirac operator . for the free case the overlap dirac operator has the properties ( a)-(c ) listed in the no - go theorem . it will be argued later that @xmath2 remains local in a gauge field background as well . as a consequence , @xmath63 defines an _ acceptable _ lattice dirac operator ( properties ( a)-(c ) of no - go theorem ) . let me make a few simple comments . ( some of these straightforward observations in this or modified form have been made independently by several people and i shall not always give a reference . see , however , refs . @xcite . ) \(1 ) instead of @xmath54 in eq . ( [ 8 ] ) one could start with any acceptable dirac operator @xmath64 ( massless , no doublers , local ) . \(2 ) if the original @xmath64 satisfies the gw relation eq . ( [ 12 ] ) then @xmath65 hence @xmath66 , i.e. the operator reproduces itself . \(3 ) @xmath67 is hermitian and @xmath68 . as a consequence @xmath69 where @xmath70 is the sign function \(4 ) for @xmath71 ( where @xmath72 ) one has @xmath73 which satisfies a slightly modified relation @xmath74 the choice of @xmath75 will be useful as discussed later . note that for @xmath76 there are no massless fermions at all , while for @xmath77 there are too many , since the modes around @xmath78 ( the massive doublers ) also become massless modes of @xmath2 . therefore only @xmath72 produces an acceptable dirac operator . \(5 ) it is also easy to generalize to the case of arbitrary local @xmath31 : @xmath79 and @xmath80 this satisfies eq . ( [ 2 ] ) . for simplicity , however , we shall consider mostly the case @xmath58 . for a massless dirac operator in the continuum theory the atiyah - singer index theorem @xcite holds : @xmath81 here @xmath82 is the topological charge of the gauge field , and @xmath83 where @xmath84 , @xmath85 is the number of solutions to @xmath86 with @xmath87 chiralities , i.e. @xmath88 . the spectrum of @xmath2 in the continuum consists of * @xmath89 ; the corresponding modes have definite chirality * @xmath90 ( @xmath91 real , non - zero ) ; for the corresponding modes @xmath92 . obviously , @xmath93 is topological invariant : an unpaired eigenvalue at @xmath89 can not be continuously produced or destroyed . in the lattice formulation one has problems with both sides of eq . ( [ 17 ] ) . on the lhs . there are several ways to assign @xmath94 to a given gauge configuration ( field theoretical definition , geometrical definition ) . to avoid ambiguities for strongly fluctuating gauge fields , cooling is often applied . an additional problem is that , in general , dislocations could be present , i.e. configurations which violate the inequality @xmath95 on the rhs . in an arbitrary gauge field @xmath54 has no exactly zero eigenvalues they are distorted by lattice artifacts . the spectrum of @xmath54 consists of * real eigenvalues , with @xmath96 but not exactly @xmath97 * complex eigenvalues , in cc . pairs , @xmath98 and @xmath99 with @xmath92 . for smooth instanton background ( @xmath100 ) @xmath54 has an isolated real , nearly zero eigenvalue with @xmath101 or @xmath102 . however , for a small instanton , or a typical mc configuration ( say at @xmath103 ) there are a lot of small real eigenvalues . therefore the definition of topological charge based on investigating the spectrum of @xmath54 is quite ambiguous @xcite . let me remind you that the rg approach has a theoretically appealing ( although difficult in practice ) way to define the topological charge of a given configuration , in spin models @xcite and in gauge theories @xcite . the solution is the following : take the minimizing configuration on the fine lattice @xmath104 and define the fp topological charge on @xmath105 through @xmath106 it is easy to show that with this definition there are no dislocations , i.e. @xmath107 holds for any configuration . as has been shown in ref . @xcite the fp fermion operator has nice chiral properties which allow one to establish the exact index theorem on the lattice . let me discuss the main points . the gw relation eq . ( [ 12 ] ) ( with @xmath108 ) and @xmath109 implies @xmath110 hence @xmath2 and @xmath111 commute , i.e. @xmath2 is normal . ( as a consequence , the eigenvectors corresponding to different eigenvalues are orthogonal to each other . ) the spectrum of @xmath2 ( cf . ( [ 14 ] ) ) is on the circle @xmath112 . there are three types of eigenvalues * @xmath89 , with @xmath113 ( denote their numbers by @xmath84 , @xmath85 ) , * @xmath114 , with @xmath113 ( with multiplicities @xmath115 , @xmath116 ) , * @xmath117 complex , with @xmath118 and @xmath119 note that @xcite since @xmath120 the index of @xmath2 and @xmath121 is opposite : @xmath122 define the topological charge density by @xmath123 for the topological charge @xmath124 we have @xmath125 here we have used the properties of eigenvectors of @xmath2 listed above . obviously , @xmath126 is topological invariant since the integer @xmath93 can not change smoothly . this fact can also be obtained directly from the gw relation . equation ( [ 28 ] ) is the index theorem on the lattice : it states that one can define a ( gauge invariant , real ) topological charge density @xmath127 on a given gauge configuration which leads to an integer topological charge . the topological charge defined this way equals to the index of @xmath2 . both sides are defined through @xmath2 . is this a tautology ? no : the important point here is that one has a local topological charge density not just an integer number defining the charge . it is easy to see that for smooth configurations @xmath128 since @xmath129 is a gauge invariant , pseudoscalar , dimension 4 operator and the prefactor is fixed by the fact that in the continuum limit @xmath35 one should recover the continuum topological charge . i would like to stress that _ any acceptable _ solution to the gw relation defines a topological charge for which the exact index theorem holds . note , however , that different solutions will give , in general , different values for the topological charge they need to agree only for sufficiently smooth configurations . from a practical point of view the operator @xmath2 has to provide a topological charge which is robust enough , giving a reasonable answer also for relatively rough configurations . to illustrate this point , consider neuberger s operator in eq . ( [ 15 ] ) with a small @xmath130 , say @xmath131 . for this choice only very large instantons are counted as such when defined through eq . ( [ 28 ] ) perhaps only those with @xmath132 . smaller instantons `` fall through the lattice '' since the smallest eigenvalue of @xmath54 would shift beyond @xmath130 . obviously , this would be a bad dirac operator and a bad topological charge in spite of the fact that they satisfy the exact index theorem , eq . ( [ 28 ] ) . any _ acceptable _ discretization of the dirac operator ( without requiring to be a solution to the gw relation ) could be used to determine the correct topological charge for _ sufficiently small _ lattice spacing @xmath13 the exactly zero eigenvalue of the continuum dirac operator will move only slightly away from zero , remaining real , and no other nearly zero eigenvalues will appear . the extra requirement of gw relation forbids the eigenvalue to move away from zero for such configurations . hence no extra conditions ( beyond acceptability and gw relation ) are required for the index theorem to hold . however , as pointed out above , these conditions do not guarantee that the topological charge defined this way is the ` correct ' one except for sufficiently large instantons . of course , there is no unique way to define the topological charge on the lattice , but for not - too - small instantons ( with a radius of a few lattice spacing , say ) the geometrical definition is natural . for practical reasons , one would like to have a gw dirac operator @xmath2 which reproduces the natural topological charge of not too large instantons as well . i discuss these points perhaps more than necessary , because there is some controversy in the literature @xcite . in ref . @xcite it has also been shown that for the fp dirac operator one has in addition @xmath133 i.e. the topological charge defined by @xmath134 _ coincides _ with the old definition in pure gauge theory . in this case , for example , there are no dislocations , and one expects small lattice artifacts . an important point is that @xmath18 is discontinuous in the gauge field . the reason is that on the lattice one can connect continuously any two configurations , but @xmath135 has a jump on the path connecting , say , @xmath136 and 1 . with the fp action the origin of this discontinuity is that the minimizing configuration @xmath137 becomes ambiguous when the rhs . of eq . ( [ 3 ] ) has a degenerate minimum ( when the instanton `` falls through the lattice '' ) . for neuberger s dirac operator the discontinuity appears when a real eigenvalue of @xmath54 passes 1 . in this case an eigenvalue of @xmath138 changes sign , and @xmath139 in eq . ( [ 14b ] ) is discontinuous . note that a similar discontinuity appears when the fermions are defined in the continuum , in a continuous gauge background obtained by interpolating the lattice gauge field @xcite . in this case the process of interpolation becomes discontinuous , similarly to the situation with the fp action . an important question : does @xmath2 become non - local when this discontinuity occurs ? this would be a very bad news for the present approach . fortunately , this is not true in general . in the fp action the two solutions @xmath140 and @xmath141 corresponding to the degenerate minima are expected to differ only locally . neuberger s @xmath2 , as shown by lscher @xcite , remains also local when an isolated eigenvalue of @xmath142 passes zero , so a discontinuity of d does not imply a loss of locality . restoring the lattice spacing @xmath13 , the spectrum of @xmath2 is given by a circle @xmath143 with @xmath144 , crossing the real axis at @xmath89 and @xmath145 . correspondingly , the spectrum of @xmath33 is a straight line parallel to the imaginary axis , of the form @xmath146 , @xmath147 for arbitrary gauge field . denote the unnormalized expectation value of @xmath148 in a fixed gauge background by @xmath149 then we have @xmath150 obviously , the @xmath151 part of @xmath152 is independent of the dynamics and has to be subtracted to define a proper chiral order parameter . this yields to @xmath153 the same as suggested in ref . @xcite . adding a mass term @xmath154 to @xmath2 would shift the whole circle of eigenvalues . since the real eigenvalues at @xmath145 also contribute to chiral relations it is better to leave them unchanged . this is simply achieved by the replacement @xmath155 which not only shifts the circle but also rescales its radius appropriately . with this definition of @xmath154 one obtains @xmath156 this expression is identical to eq . ( [ 33 ] ) , i.e. the proper order parameter . this equivalent form on the rhs . has been suggested first by chandrasekharan @xcite , who obtained it from lscher s exact symmetry . at this moment we are in a peculiar situation : all signs of a symmetry are present ( ward identities , index theorem , chiral spectrum , ) but the action is not chirally invariant . lscher made an important observation that in fact the action possesses an exact symmetry . perform an infinitesimal change of variables @xmath157 and @xmath158 with a global flavour singlet transformation @xmath159 from the gw relation eq . ( [ 12 ] ) it follows that the action is invariant @xmath160 the corresponding flavour non - singlet transformation is given by @xmath161 now it seems that we have more symmetry than possible the flavour singlet transformation should be anomalous . where is the anomaly hidden ? as pointed out by lscher , the fermionic integration measure is _ not _ invariant under the singlet transformation in eq . ( [ 35 ] ) @xmath162 & = & { \rm tr}(a\gamma_5 d ) [ d\overline{\psi}d\psi ] \label{37 } \\ & = & 2n_{\rm f}(n_- - n_+)[d\overline{\psi}d\psi ] \nonumber\,.\end{aligned}\ ] ] that is , the measure breaks the flavour singlet chiral symmetry in a topologically non - trivial gauge field where @xmath163 . the corresponding ( global ) ward identity for the flavour singlet transformation is @xmath164 where @xmath165 . note that the non - singlet symmetry is not anomalous , @xmath166 since eq . ( [ 37 ] ) for this case contains @xmath167 . using these exact symmetries of the action , the ward identities follow , of course , directly . by taking @xmath168 and a non - singlet transformation one has @xmath166 i.e. @xcite @xmath169 this is the order parameter which is expected to be broken spontaneously in the thermodynamic limit @xcite : @xmath170 in ref . @xcite there is also a relation connected with the u(1 ) problem ( massive @xmath171 meson ) this will be discussed later . whether the chiral symmetry is indeed broken spontaneously in qcd is a delicate dynamical question . the chiral invariant dirac operator provides a firm basis to investigate this question by avoiding an explicit breaking . note that the nielsen - ninomiya no - go theorem is avoided here by the non - standard form of the lattice chiral symmetry : @xmath2 is not invariant under the naive @xmath32 symmetry , nevertheless it is invariant under the generalized transformation . in the continuum , chiral symmetry allows one to decompose the fermion field into left and right chiral components , transforming independently . besides yielding a convenient formalism for vector theories like qcd , this decomposition is a necessary first step towards the chiral gauge theories . the matrix @xmath32 anti - commutes with the continuum dirac operator and @xmath172 . this allows to introduce the chiral projectors @xmath173 and the action decomposes as @xmath174 . on the lattice it is useful to introduce @xcite the operator @xmath175 where @xmath2 satisfies the gw relation eq . ( [ 12 ] ) . remember that @xmath38 is unitary and @xmath176 . then @xmath177 one can introduce the projectors @xmath178 they satisfy the relations @xmath179 and allow the decomposition @xmath180 to conclude the definitions let us introduce the chiral components of @xmath181 : @xmath182 in terms of these components we have @xmath183 analogously to the continuum case . note that eqs . ( [ 50],[51 ] ) are not symmetric in @xmath181 and @xmath184 . as discussed later , this asymmetric definition seems to play an essential role . analogously to the continuum case , the mass term should couple the l and r components ( but not the l , l or r , r components ) : @xmath185 the corresponding pseudoscalar is given by @xmath186 the scalar ( pseudoscalar ) densities @xmath187 and @xmath188 will often appear in the formulae below . performing a left chiral transformation means @xmath189 using this transformation one has @xmath190 the combinations in eqs . ( [ 53],[54 ] ) are the natural generalizations of the corresponding continuum expressions . note that the @xmath191 factor cancels the contribution of the unphysical real eigenmode at @xmath145 . using these combinations of fields , one obtains very simple , compact relations , for example : @xmath192 note that only the @xmath89 eigenmodes contribute when evaluating the trace . it is instructive to consider the flavour singlet ward identity which leads to a relation for the @xmath171 propagator : @xmath193 if the @xmath171 stays massive in the chiral limit @xmath194 then the lhs . ( after averaging over the gauge fields ) should go to zero . this leads to the interesting relation @xcite between the topological susceptibility and chiral condensate in full qcd @xmath195 this relation has been obtained previously in the continuum @xcite but it is important to have its lattice version as well . it shows , for example that the topological susceptibility must be a well behaved physical quantity on the lattice . for the flavour non - singlet variant of eq . ( [ 58 ] ) the term proportional to @xmath196 is absent since this symmetry is not anomalous . therefore a non - vanishing chiral condensate , eq . ( [ 42 ] ) , yields to ( quasi ) goldstone bosons with mass @xmath197 . we conclude this section with a few remarks . note that @xmath198 coincides with @xmath139 introduced in the overlap formalism . another point worth mentioning is that for the modes with @xmath89 one has @xmath199 while for the modes at @xmath145 they differ in sign , @xmath200 , as seen from the definition eq . . this observation will be important when we consider chiral fermions . we can now introduce the @xmath201 parameter , as in the continuum . here i shall restrict the discussion to some basic points only . define the lattice lagrangean with a complex mass @xmath154 and the @xmath201-parameter : @xmath202 where @xmath203 is the gauge action and @xmath129 is defined through @xmath2 by eq . ( [ 27 ] ) . by changing the variables @xmath204 one can restore the invariance of the action by @xmath205 and the non - invariance of the measure can be compensated by @xmath206 . therefore the free energy should depend only on the combination @xmath207 , as in the continuum . one can proceed now e.g. by repeating ( now on the lattice ) the steps of ref . @xcite where e.g. useful sum rules have been obtained for the spectrum of the dirac operator . to illustrate this , consider the partition function @xmath208 under some natural assumptions @xcite one has for the free energy @xmath209 taking then the second derivative with respect to @xmath201 ( for real @xmath154 ) one obtains @xmath210 which is equivalent to eq . ( [ 59 ] ) . as another example , it is easy to show that the fermion determinant is given by @xmath211 for @xmath212 , and the product is over the complex eigenvalues only . for @xmath213 the prefactor is replaced by @xmath214 . ( [ 67 ] ) differs from its continuum counterpart by the presence of the factor @xmath215 . an important consequence of the exact chiral symmetry on the lattice is that there are no o(@xmath13 ) lattice artifacts , to any order of @xmath216 . the reason is the following . the o(@xmath13 ) lattice artifacts of any action could be compensated by adding an extra term which is a lattice version of @xmath217 with an appropriately chosen coefficient @xcite . the original action in our case has the exact ( generalized ) chiral symmetry of eq . ( [ 35 ] ) and its artifacts could be compensated only by a similarly invariant extra term . any lattice regularization of @xmath217 will violate , however , this symmetry , hence the o(@xmath13 ) artifacts should be absent for the original lattice action as well . note also that the clover term needed for o(@xmath13 ) improvement has been determined by requiring that some ward identity holds independently of the gauge field . this is , however , automatically satisfied here for @xmath218 hence there is no need for extra clover term . in addition , in the presence of a mass @xmath154 there are also no o(@xmath219 ) artifacts , if the mass is introduced properly , as in eq . . then the symmetry @xmath220 forbids such term . indeed , under a finite chiral transformation @xmath221 @xmath222 is invariant while the mass term @xmath223 changes sign . although due to the anomaly the partition function is not invariant for @xmath224 , this fact is not relevant for our considerations . indeed , for @xmath225 and sufficiently large volume one can restrict the system to the topologically trivial sector @xmath226 @xcite . note , however , that the o(@xmath227 ) lattice artifacts could still be large at the relevant lattice spacings , and could be quite different for different solutions of the gw relation . finally , it is interesting to see ( although it is not necessary after the previous discussion ) that the absence of o(@xmath13 ) artifacts on the tree level follows directly from the gw relation : expanding @xmath2 in powers of @xmath13 it implies that @xmath228 ( with some constant @xmath229 ) i.e. a term with @xmath230 is automatically generated . here we shall discuss some basic properties of neuberger s dirac operator and its straightforward generalizations . consider first the free case . the spectrum @xmath231 of @xmath2 defined in eq . ( [ 11 ] ) is shown in fig . [ fig : spect ] . together with the spectrum for the standard wilson - dirac operator for @xmath232 . the latter is denoted by w and given by @xmath233 while n denotes the two branches for neuberger s dirac operator where @xmath234 is given by @xmath235 ( note that the lower curve extends only up to the point where the two branches join each other this peculiar behaviour is due to the non - analyticity in the square root . ) the spectrum for @xmath236 is worse than that for @xmath54 . this illustrates that a solution of gw relation can have large o(@xmath227 ) artifacts the fact that @xmath2 is spread over some distance does not necessarily make the artifacts small ! the locality of @xmath2 for the free case is guaranteed by the inequality @xmath237 where @xmath238 and @xmath239 . this asymptotic behaviour is shown on fig . [ fig : loc ] by dashed line . the decay of @xmath240 for neuberger s @xmath2 is shown by circles . two more solutions to gw relation are added . crosses show the case with @xmath241 where @xmath242 is a fp action truncated to the hypercube . since there is a non - trivial @xmath31 in this case the more natural choice , @xmath243 , ( cf . ( [ 16],[16a ] ) ) is also shown ( + signs ) . finally , the true fp action is also indicated ( triangles ) . it is obvious that there are rather large differences in the range of interactions for different gw dirac operators , and this is an important factor in making a choice for numerical simulations ( see also ref . @xcite ) . a less trivial point is the question of locality in a gauge field background . in ref . @xcite it is proven that for small gauge fields @xmath244 , the corresponding @xmath18 is local with exponentially decaying tail , and the localization length is uniformly bounded . although eq . ( [ 74 ] ) should hold sufficiently close to the continuum limit , it is violated at practical values of @xmath13 . theoretically one can restrict the gauge fields to satisfy eq . ( [ 74 ] ) . this is sufficient to show that there exists an exactly chiral invariant , acceptable lattice regularization of qcd . however , the restriction leads to an astronomical correlation length , hence it is only of theoretical interest . as mentioned before , in ref . @xcite it is also shown that an isolated zero of @xmath245 does not spoil locality . an important question is the locality in gauge fields generated by mc simulations . first results on this are presented in refs . @xcite . in fig . [ fig : f2r ] results from ref . @xcite are shown , for su(3 ) gauge fields generated at @xmath246 , @xmath247 and @xmath248 ( triangles , squares , circles ) . the full symbols are for @xmath249 while the open ones for ( roughly ) optimized values , @xmath250 for @xmath246 and @xmath251 for @xmath252 . it is also important to have a sufficiently effective way of evaluating the inverse square root . this issue is discussed in refs . @xcite . i only mention here that it seems to be a feasible task but still much has to be done in this direction . a few comments are in order : \(1 ) unlike evaluating @xmath33 for massless quarks , taking the inverse square root is not a critical problem appearance of zero modes of @xmath245 is accidental . \(2 ) starting with a good approximation of the fp action instead of @xmath54 ( as in eq . ( [ 16 ] ) ) the probability to have a zero mode of @xmath245 is strongly suppressed , since with the exact fp action @xmath245 would be unitary . note , however , that for any continuous parametrization of the fp action , @xmath245 must have a zero mode for specific configurations , when an instanton `` falls through the lattice '' . this happens on the boundary separating regions of gauge field configurations with different topological charges . \(3 ) using a gw dirac operator there will be no `` exceptional configurations '' in the usual sense @xcite these would correspond to eigenvalues @xmath253 . ( one should distinguish these from configurations where @xmath254 appearing in this approach . ) \(4 ) one can construct , in principle , configurations when many `` instantons fall through '' simultaneously at different sites of a large lattice . this would produce a situation when @xmath245 has many nearly zero eigenvalues . will @xmath2 be local on such configurations ? one expects that such configurations have negligible weight in the partition function . will their effect be also negligible ? i think , these questions have to be investigated . there are several equivalent ways to measure the topological charge defined by @xmath2 through eqs . ( [ 27],[28 ] ) : \(1 ) find the zero eigenmodes of @xmath2 ( or of hermitian @xmath255 ) , \(2 ) count the real eigenvalues of @xmath54 in the region @xmath256 , with chirality signs , \(3 ) trace the `` level crossings '' of @xmath257 while @xmath258 runs from @xmath259 to @xmath130 . note that although we rely here on the exact index theorem on the lattice , the value of the topological charge determined through @xmath2 with @xmath260 is the same as the method using @xmath54 suggested a long time ago @xcite . the last method has also been used in ref . @xcite in the framework of the overlap formalism . this definition of the topological charge works well for small instantons , as expected . but how does it work for typical mc configurations ? does @xmath2 ( generated using @xmath54 ) give a ` better ' definition than the geometrical one ? hernandez @xcite argues that the answer is no . gattringer and hipp @xcite simulated su(2 ) on a @xmath261 lattice at @xmath262 and found that the value of the topological charge , @xmath263 on individual configurations depends strongly on the parameter @xmath130 since there are usually many real eigenvalues in the region of interest , and only very few configurations have a gap in the spectrum where @xmath263 does not depend on @xmath130 chosen in a reasonable range . they also note that a naive inclusion of the clover term makes the situation worse . the same conclusion is found by degrand , a. hasenfratz and kovcs @xcite . this last fact is not too surprising since the corresponding artifact is not an o(@xmath13 ) effect . note , however , that in the average this definition of charge works better : edwards , heller and narayanan @xcite found that the topological susceptibility stays roughly constant ( independent of @xmath130 ) for @xmath264 . here i would like to point out that the @xmath130-independence of @xmath265 does not necessarily mean the ` true ' value : if @xmath266 then @xmath267 . a good definition of the topological charge ( at a given lattice spacing ) should give relatively unambiguous definition of @xmath263 on individual configurations . in other words , the value of the lattice spacing @xmath13 when a gw dirac operator gives a ` good ' definition of the topological charge could strongly depend on the choice of @xmath2 . with the presently used lattice spacings the original choice of neuberger ( using @xmath54 ) seems to produce a topological charge which is too ambiguous . on the other hand , measuring the topological charge with the fp action in the schwinger model @xcite gives quite convincing results . for 4d gauge theories , however , the situation is less satisfactory @xcite mostly because of technical problems to perform the required multigrid minimization on the gauge configurations . for qcd a good strategy would be a compromise : take a dirac operator @xmath64 with good properties ( small cut - off effects , nearly satisfying the gw relation , ) and build @xmath2 from it , which in turn could be used in the definition of the topological charge . it would have smaller o(@xmath227 ) artifacts , better locality , fewer ` dangerous ' configurations , larger gap in the real eigenvalues hence less ambiguity of @xmath126 , but will also need more programming effort and extra computational overhead . measuring the topological charge density using @xmath268 is perhaps too time consuming to be feasible . the question whether it is possible to define chiral gauge theories ( like the standard model ) on the lattice is a purely theoretical problem , but of great importance . much has been done in this direction and has been reported earlier @xcite . nevertheless , there remain several open questions . a gw dirac operator seems to be a good starting point : it has exact gauge invariance and one can define weyl fermions ( cf . ( [ 50],[51 ] ) ) . here i shall report briefly about new developments ( partially yet unpublished ) by lscher @xcite based on the properties of gw dirac operators . the discussion will touch only a few interesting points which could be easily understood on the basis of previous sections . consider fermion fields which are purely left handed ( see eq . ( [ 50 ] ) ) @xmath269 these are local , gauge invariant conditions . introduce a basis for @xmath270 and @xmath271 through relations @xmath272 where @xmath273 and @xmath274 are grassmann variables . in this basis the fermion action reads @xmath275 with @xmath276 . the fermion propagator ( when there are no zero modes of @xmath2 ) is given by @xmath277 it is easy to show that this is chiral in the generalized sense . note that @xmath278 the first term on the rhs . is chiral in the usual sense ( with @xmath32 ) while @xmath279 is a contact term . consider a chiral gauge theory with fermion multiplet @xmath280 , with integer charges @xmath281 . for the number of zero modes @xmath282 , @xmath283 the index theorem implies @xmath284 since @xmath285 is replaced by @xmath286 for smooth gauge fields when @xmath287 . a simple but important observation is that the number of @xmath181 and @xmath184 fields are not necessarily equal . they are given by @xmath288 and @xmath289 , respectively . for their difference we have @xmath290 to illustrate this , consider a background field with @xmath291 , where @xmath292 , @xmath293 and @xmath294 , @xmath295 . then for the @xmath89 mode @xmath296 while for the eigenvector @xmath297 at @xmath114 @xmath298 according to eqs . ( [ 76]-[78 ] ) , at @xmath89 we have only @xmath181 ( but not @xmath184 ) while at @xmath114 both @xmath181 and @xmath184 are present since @xmath32 and @xmath198 differ by sign at this point . obviously for complex values of @xmath98 both @xmath181 and @xmath184 are present again . altogether there is one more @xmath181 than @xmath184 . of course , this statement depends on the topological charge of the gauge field . as a consequence the fermionic average @xmath299 of any operator in a given background gauge field vanishes unless the fermionic number of type @xmath91 ( the number of @xmath300 s minus the number of @xmath301 s ) in @xmath302 equals to @xmath303 . perform now ( for the u(1 ) case ) a global gauge transformation @xmath304 . this modifies @xmath302 by a phase factor @xmath305 where @xmath306 . from the other side a global u(1 ) gauge transformation does not change the gauge fields at all , hence invariance under this transformation requires @xmath307 i.e. the condition of anomaly cancellation of the continuum theory . these observations , however , constitute only the starting point of a difficult mathematical problem to find a satisfactory choice of the fermionic integration measure ( `` chiral determinant '' ) , i.e. of the basis @xmath308 . lscher has proven that it is possible to construct a basis for the u(1 ) case that satisfies the following requirements : \(1 ) the fermionic expectation values @xmath309 are smooth functions of @xmath17 \(2 ) the field equations ( obtained by some change of variables ) are _ local _ : @xmath310 if @xmath20 is far from @xmath311 . \(3 ) @xmath309 is gauge invariant if @xmath302 is . note that the last property may not be necessary to require . in principle , it is possible that after the leading violation of gauge invariance is cancelled by the condition @xmath312 , the rest is small and goes away in the continuum limit , without the need of extra counter terms . this possibility is based on a simple but surprising result of foerster , nielsen and ninomiya @xcite , who show that gauge invariance on the lattice is robust in the sense that the effect of a sufficiently small violating term vanishes in the continuum limit . the constructions of the overlap and gauge fixing @xcite formalisms assume this scenario . it is clear , however , that it is much safer if possible to preserve chiral gauge invariance at finite lattice spacing . it has been illustrated that dirac operators satisfying the gw relation provide an elegant solution to the chirality problem in lattice qcd . in particular , there is no additive quark mass renormalization , and ward identities , soft pion theorem , etc . are valid , there are no o(@xmath13 ) artifacts . for practical implementations of these ideas new algorithms are needed . that the task is feasible is illustrated by mc calculations in the domain wall approach @xcite which could be considered as a particular approximation to a gw dirac operator . not restricting itself to the given ideology could open ways to more efficient algorithms . t. degrand , a. hasenfratz and d. zhu , nucl . b475 ( 1996 ) 321 ; nucl . b478 ( 1996 ) 349 ; nucl . phys . b(proc.suppl . ) 53 ( 1997 ) 945 . + t. degrand , a. hasenfratz and t. kovcs , nucl . phys . b(proc.suppl . ) 63 ( 1998 ) 528 .

geomagnetism is still one of the unsolved fundamental problems of the earth since william gilbert discovered the magnetized earth ( gilbert 1600 ) . not only the very existence of the dipole magnetism , motonori matsuyama discovered the polarity reversal events ( matsuyama 1927 ) and vivid dynamics of geomagnetism . the qualitative nature of this polarity flip was explained by a simple disk model by tsuneji rikitake ( rikitake 1957 ) . recent numerical simulations of magnetohydrodynamics ( mhd ) for geomagnetism ( roberts & glatzmaier 2000 ) , ( kono & roberts 2002 ) seem to describe such dynamics . inspired by these simulations , we recently studied to clarify the physics behind this remarkable polarity reversals proposing a simple model ( domino model ) ( mori , schmitt , ferriz - mas et al . 2011 ) composed from many macro - spins which are interacting with each other . this _ coupled spin model _ is based on the idea that the whole dynamo mechanism is described by global interactions of many small dynamo elements ( macro - spins ) . this model naturally yields many of the observed features of geomagnetism : its time evolution , the power spectrum , the frequency distribution of stable polarity periods , etc . ( mori et al . 2011 ) . in the case of the earth , the dynamo element , that a macro - spin describes , is considered to be the taylor cell in the iron fluid core produced and supported by the coriolis force . in this paper , we would like to develop the above idea by elaborating the domino model in a form that is _ minimal and general_. in order to construct the _ minimal _ model , we introduce a _ long - range coupling _ of macro - spins , i.e. all the spins interact with each other with the same amplitude of interactions . this is essentially different from the previous domino model in which only neighboring spins interact with each other . it turns out that this long - range coupling model is also effective to describe the observed features of geomagnetism as the domino model . in order to show the _ generality _ of the model , we consider it in wider contexts . the spin model , so far , is at most a phenomenological model to describe geomagnetism . therefore it needs some _ supplement from the basic mhd equations _ for better descriptions . on the other hand , magnetic fields associated with celestial objects are quite common in the universe ; many planets , satellites , and the stars including our sun . although the magnetic fields of these objects have variety that is quite different from the geomagnetism , all the magnetic fields are thought to be generated and supported by the robust dynamo mechanism . therefore we try to fit our model , supplemented with mhd equations , in _ other celestial objects _ and see how they are consistent with each other . the long - range coupled spin model can also describe the _ synchronization _ physics as well if we simply change the parameter value in the model . this property is particularly useful to describe the quasi - periodic variation of solar magnetic field . furthermore the long - range coupling in our model itself can produce sufficient chaoticity and randomness without introducing explicit random force as in the domino model . thus the long - range coupling spin model has rich physics and generality . we start our study by the basic description for geomagnetism . the _ fluid iron _ in the center of the earth is now considered to be the place where geomagnetism is created ( inglis 1981 ) . the conductive fluid motion and its interaction with electromagnetic fields should control the dynamics of geomagnetism through dynamo mechanism . they are described by the _ magneto - hydrodynamics _ ( mhd ) for incompressible ( @xmath0 ) fluid , @xmath1@xmath2@xmath3@xmath4with appropriate boundary conditions ( roberts & glatzmaier 2000 ) . the dynamics of fluid velocity @xmath5 in eq.([eqforv ] ) is governed by the nonlinearity @xmath6 the pressure @xmath7 the centrifugal force potential @xmath8 , gravity @xmath9 , and the lorentz force @xmath10 where @xmath11 is the electromagnetic current and @xmath12 is the mass density . the coriolis force @xmath13 yields the vorticity and convective pattern of the fluid . the magnetic field @xmath14 is amplified or reduced by the flow @xmath15 as described by eq.([eqforb ] ) . the dissipations ( @xmath16 ) and the energy injections ( from the heat generation @xmath17 and possibly the rotational driving force ) should balance with each other in the stationary state . main non - dimensional quantities which characterize the above set of equations , and their typical values in the iron fluid core in the earth , are as follows:@xmath18@xmath19@xmath20@xmath21these values are important to estimate the scale of dynamo elements in later sections . straightforward approach based on these highly complicated equations requires sophisticated numerical calculations . such analysis have recently been developed and the generation , maintenance and even reversals of geomagnetism have been obtained ( kida & kitauchi 1998),(roberts & glatzmaier 2000),(kono & roberts 2002 ) . these works have provided us with valuable knowledge to develop fundamental understanding of geomagnetism . however the parameter range in numerical calculations is still far from the above realistic values . therefore it would be very favorable if our coupled spin model is complementary to the mhd calculations and if it describes some basic features of dynamo mechanism . these are the aim of this paper . the plan of this paper is as follows . in _ section 2 _ , we describe our idea of the coupled dynamo elements which plays the central role in the present paper . we emphasize the necessary condition for the elements in order to clarify what the macro - spin represents in our model . readers who immediately want to see the model can skip this section . in _ section 3 _ , we briefly review our previous study on the domino model ( short - range coupled spin model ( scs ) ) , which is a first realization of the idea of the coupled dynamo elements to describe geomagnetism . in _ section 4 _ , we introduce the generalized long - range coupled - spin model ( lcs ) in which macro - spins have long range interactions . then we apply this model to geomagnetism and successfully describe many of the characteristic features of geomagnetism . in _ section 5 _ , we study the physical relevance of lcs model showing the similarity of it to the mean field model ( hmf ) and kuramoto model . readers who immediately want to see further applications of the model can skip this section . in _ section 6 _ , we apply our model , supplemented with mhd equations , to other planets and satellites in order to estimate the scale of macro - spins . in _ section 7 _ , we apply the same model , with different parameter values , to the solar magnetism and successfully describe the quasi - periodic dynamics of the solar magnetism . the last _ section 8 _ is devoted to summarize our work , and speculate possible generalizations of our idea of coupled dynamo elements . we now describe the idea of coupled dynamo elements ( mori et al . 2011 ) which plays the fundamental role in this paper . the central equation to describe the dynamo mechanism in mhd is eq.([eqforb ] ) , in which the flow @xmath22 is determined by eq.([eqforv ] ) . this is practically a very complicated set of equations to describe geomagnetism and the dynamo mechanism though we can extract some basic insight from it . geomagnetism is characterized by the coexistence of many time scales , from million years to thousand years , with clear power - law power spectrum . the dynamo mechanism is ubiquitous in the sense that most of the celestial objects have magnetism generated from this mechanism . the ubiquitousness implies the basic dynamo mechanism is simple and the many time scales with power - law power spectrum imply the dynamo system is composed from many components . therefore we expect that _ the whole dynamo system may be composed of many elements , each of which has minimal dynamo function , that interact with other elements in a simple form_. it is often happen that a simple interaction yields very complicated structures . then what is the element which has the minimal dynamo function in the case for geomagnetism ? we notice that the above eq.([eqforb ] ) is similar to the general equation for vorticity @xmath23 @xmath24which is derived by taking the rotation of eq.([eqforv ] ) , neglecting the lorentz force term . this fact reveals the apparent duality of the magnetic field @xmath14 and the vorticity @xmath25 . the iron fluid core is estimated to be highly turbulent from eq.([re ] ) and is full of vorticities . the coherent vorticity @xmath25 is generated from the convection and coriolis force . these effects seem to be significant from the values in eqs.([ta]-[ra ] ) . therefore the iron fluid is full of vorticity , and possibly of magnetic fields . it is clear that the coriolis force dominates in the iron fluid as in eq.([ta ] ) . in general , if the coriolis force dominates in almost the stationary flow , then eq.([eqforv ] ) becomes @xmath26 by taking the rotation of this form , we have @xmath27 . therefore the flow has a tendency to be two - dimensional perpendicular to the rotational axes @xmath28 , _ i.e. _ the taylor - proudman theorem . therefore this theorem suggests that a convection of iron fluid forms convective columns ( usually called taylor cell ) parallel to the rotation axes . the scale of the taylor cell will be determined by geometry and various parameters in eqs.([re]-[ra ] ) . thus coherent vorticity @xmath29 is naturally expected in the earth core . however the vorticity @xmath25 is not sufficient to produce magnetism @xmath14 ( tsinober 2007 ) . this is because the circular vorticity itself is not endowed with energy production . any deviation from the complete circular motion would be necessary for the production of magnetic energy . we would like to emphasize that the inward - winding vorticity is essential for dynamo mechanism . suppose that we prepare a typical inward - winding solution for the navier - stokes equation , the burgers vortex , whose velocity field is expressed in the cylindrical coordinate @xmath30 as @xmath31 with some constants @xmath32 , and @xmath33 for inward - winding and @xmath34 for outward - winding . then putting the above eq.([bergers ] ) into eq.([eqforb ] ) yields , in the cartesian coordinate , @xmath35 for the magnetic field @xmath36 parallel to the symmetry axes of the vorticity . only the first term in rhs can definitely enhance @xmath37 if @xmath38 the inward - winding flow squeezes the magnetic force lines against the positive pressure from @xmath37 . summarizing all the above , the iron fluid in the earth core may allow _ multiple vortex columns parallel to the earth rotation axes and their interaction may cooperatively yield the whole geomagnetism , if the columns are inward - winding_. on the other hand in the super computer simulations , such column structure is commonly formed ( kida & kitauchi 1998 ) and robustly persists . these columns are called taylor cell(tc ) . a close look into tc in the numerical data , reveals that tc are either inward - winding ( _ i.e. _ high - pressure due to the compression ) or outward - winding ( _ i.e. _ low - pressure due to the spread ) . they are called , respectively , anti - cyclone and cyclone . that is the anti - cyclone counter rotates to the earth and the cyclone co - rotates . the coriolis force associated with the earth rotation makes these difference and the smooth flow may favor the alternating configuration of cyclone and anti - cyclone . according to the above argument , only the anti - cyclone , which has inward - winding flow , can produce magnetic fields ( kageyama & sato 1997 ) . in the real situations in the earth , the flow is highly turbulent and non - linear on top of such tc structures reflecting the large reynolds number eq.([re ] ) . complicated magnetic fields in the outer core are captured by inward - winding tc and aligned with the rotation axes , and then compressed to enhance the poloidal magnetic fields . these non - linear flow , as well as huge electric current , yields the interaction among tc , either short - range or long - range . each dynamo element of inward - winding vorticity , anti - cyclone , accompanies generated magnetic fields which can be characterized by a direction and a strength in the first approximation . therefore we characterize each dynamo element as a _ vector _ quantity @xmath39 . this vector represents the direction of the magnetic fields associated with tc and not the direction of tc itself . since each element is supposed to have the minimum dynamo function , this vector can be identified as a macroscopic spin or magnetic moment . then the whole system should be a set of such macro - spins @xmath40 , which is located on each site @xmath41 in the iron fluid core . although the interactions between the macro - spins should be very complicated , the potential energy of it must be a _ scaler _ or the vector inner product of the spins @xmath42 . this simplicity may guarantee the robustness and universality of the geomagnetism . however we can not determine the amplitude of the coupling parameter @xmath43 by a simple argument . it depends on at least the iron flow pattern , electric currents , and configuration of cyclone and anti - cyclone . on the other hand we can guess the signature of @xmath44 . electric currents are smoothly winding each tcs , and cyclone and anti - cyclone align alternately . therefore the smooth electric current yields the winding direction is the same for all anti - cyclones ( and opposite direction for all cyclones ) . this suggests the magnetic fields of anti - cyclones themselves ( and also cyclones themselves ) have a tendency to align to the same direction . this suggests the _ _ negative value for the coupling strength : _ _ @xmath45 . ] , like in the ferromagnetism in ferromagnetism has the quantum origin , which has nothing to do with our macro - spin model . ] . then the effective curie point of these macro - spin system can be high enough and therefore , the whole system may actually show ferromagnetism . this makes a quite contrast to the case of the ( micro ) spins of the iron in the earth ; no ferromagnetism is produced in the core of temperature @xmath46k which well exceeds the curie point about @xmath47k . we have emphasized the necessary condition for the macro - spin @xmath48to be relevant and their interaction although the construction of the macro - spin @xmath48 is not our aim in this paper . we would like to study the inter - relation between many of such elements @xmath48 which yields various non - trivial dynamics . before our study , there have been some works ( mazaud & laj 1989 ) , ( seki & ito 1993 ) , ( dias , franco , & papa 2008 ) in which the spin models with the interaction @xmath42 were studied to reveal the polarity flip dynamics of geomagnetism . they are based on the two - dimensional ising model and have found the polarity flip near the critical temperature . our aim to study geomagnetism in this paper is not simply to extend the similarity with phase transitions but to study wider point of view such as _ phase coexistence _ and _ synchronization_. there are two types of coupling between the spins according to the range of the interaction . if the fluid motion is responsible to the coupling , then only the neighboring spins interact with each other and yield short - range coupled - spin ( _ scs _ ) system . on the other hand , if the electric current or magnetic fields are responsible to the coupling , then all the spins interact with the same strength and yields long - range coupled - spin system ( _ lcs _ ) . these are the two extreme ends of the series of models with finite interaction rage . we have already studied scs model in our paper ( mori 2011 ) , which successfully describes many characteristic features of geomagnetism as we now review in the next section before we study lcs model is the subsequent sections . we now briefly review our previous model for geomagnetic dynamics ( mori et al . 2011 ) , in which the concept of coupled dynamo elements is realized as short - range interacting macro - spins ( scs model ) . our intention was to establish a simple model which realizes various basic properties of geomagnetic observations based on the idea that the coupled dynamics of dynamo elements . we assume that each element has dynamo function to produce element magnetism putting the very origin of the dynamo mechanism aside . we first introduce the scs model , which is described by the lagrangian @xmath49 where@xmath50@xmath51and by appropriate dissipation and fluctuation terms . the spins are assumed to be located on a circle with equal separation ( therefore @xmath52 ) . this circle is set on the plane which includes the equator and fully immersed inside the fluid core region of the earth . the spin @xmath53 at the cite @xmath41 may be represented by a single angle parameter @xmath54 as @xmath55 . we measure the angle @xmath56 from the angular velocity vector @xmath57 . ) . the angular velocity @xmath58 points to the top . ] this plane the spin @xmath48 moves is set perpendicular to the radius toward the cite @xmath41 . the full evolution equation then becomes the stochastic differential equation ( langevin equation),@xmath59for @xmath60 . the parameter @xmath61 is the dissipation coefficient . the fluctuation force @xmath62 is assumed to be gaussian white noise as usual , characterized by the correlation:@xmath63then we have , so far , four parameters in our model @xmath64 . the parameter @xmath65 represents the tendency for each spin to be parallel to the rotational axes . larger value of @xmath65 means the dominance of the coriolis force over the other effects and the taylor - proudman theorem better holds . the parameter @xmath44 measures the strength of interaction between the element spins . it reflects the nature of the fluid flow and electric current across the neighboring spins . the parameter @xmath61 represents the whole energy dissipation @xmath66 and @xmath67 in eqs.([eqforv],[eqforb ] , [ eqfort ] ) . the parameter @xmath68 represents the inhomogeneous heat generation from the inner iron core and the random perturbation from the other element spins . we also have to specify the number @xmath69 of spins . it may depend on the geometry of the fluid core region of the earth in which cyclones and anti - cyclones are formed . we can save the number of parameters if we focus on the dynamics in long time scale and neglect the inertial term @xmath70 . then the reduced equation of motion becomes ( @xmath60),@xmath71which now has three parameters @xmath72 . the full equation eq.([eomlcs ] ) is solved in ( mori et al . the neglect of the inertial term is justified by the fact that both the methods yield almost the same results . each dynamo element represented by the spin @xmath48 naturally corresponds to a magnetic dipole located at the site @xmath41 of the dynamo element . we are interested in the order parameter @xmath73 of the set of equations eq.([redlcs ] ) defined by @xmath74this quantity @xmath73(@xmath75 $ ] ) is a good indicator of the whole magnetic field when we analyze the history of the geomagnetism though it is not the magnetic field itself . it is also possible to describe the realistic configuration of the magnetic field on the earth by using the set of @xmath76 , ( @xmath60 ) . as we have explained before , the magnetic moments @xmath48 ( @xmath60 ) are located on the ring with equal separation , on the surface which includes the equator . then the whole magnetic field @xmath14 becomes @xmath77where each spin @xmath48 is located at the position @xmath78 , and @xmath79 is the unit vector @xmath80 . this simple model naturally describes many of the observed properties of geomagnetism including the intermittent polarity reversal , power - law power spectrum , chron interval distribution , etc . the results are summarized in the paper ( mori et al . the short - range coupled spin ( scs ) model reviewed in the last section was simple and effective to describe the statistical properties of geomagnetism . we now study a slightly modified model introducing the long - range interaction coupled spin ( lcs ) model . the simplest one would be described by the same lagrangian @xmath81 with eq.([k ] , [ v ] ) but the interaction of macro spins is long ranged , in the denominator in the coupling term is simply from a technical reason so that all the terms in @xmath82 formally becomes additive . ] @xmath83 according to the naive extension of scs , the full evolution equation seems to become ( @xmath60),@xmath84as in the previous section , introducing the fluctuation force @xmath62 and the dissipation term @xmath85 . however , thanks to the long - range coupling , these _ fluctuation - dissipation terms are not necessary in lcs _ model . the long - range coupling by itself can yield sufficient chaoticity which is necessary to describe the intermittent polarity flip . it is still possible to reserve the fluctuation - dissipation terms , which does not alter the characteristic feature of the system . therefore we would like to choose the simpler description which has no fluctuation - dissipation terms . then the system becomes conservative system in which the total energy is conserved . the evolution equation becomes simple ( @xmath60),@xmath86which has only two parameters @xmath65 and @xmath44 through the potential @xmath82 as eq.([vingcs ] ) . the order parameter and the magnetic fields are described by the same expressions eqs.([m ] , [ b ] ) as in the scs model . we next report the numerical results of lcs model with comparison with observations . we have solved the set of equations for @xmath87 spins which are located on a circle . we have set parameters of the model as @xmath88 , all of which is chosen to be of order one and we did not perform systematic fine tuning of them . further we set the initial condition so that all the spins are rest and their direction angles are equally separated within the width of angle @xmath89 . the reason that we did not distribute the initial spin angles in the full range @xmath90 $ ] is because we had to choose low energy to make the coexistence possible of the cluster and expanded spins , as we will explain in the next section . time evolution of the total magnetic moment projected on the rotation axes @xmath91 is shown in fig.(fig4 - 1 ) top . the end time @xmath92 is chosen so that the total number of polarity - reversal becomes several hundreds as is observed number @xmath93 in the history of geomagnetism within @xmath94 year in the past . the total number of the polarity reversal is @xmath95 in this calculation . therefore the unit calculation time ( @xmath96 ) corresponds to about @xmath97 year . we have obtained the following results . 1 . unit calculation time : @xmath98 @xmath97 year 2 . average flipping time : @xmath99 yeas . ( observation @xmath100year ) 3 . average chron time : @xmath101 @xmath102 year . ( observation @xmath103 year . ) 4 . superchron : @xmath104 year ( observation @xmath105 year ) + the power spectrum of the time series @xmath106 is shown in fig.([fig4 - 1 ] ) bottom . 5 . the power index : @xmath107 and @xmath108 , respectively , for right and left of the knee . ( observation : index @xmath109 in high frequency regime @xmath110 , and is @xmath111 in the low frequency regime @xmath112 . ) 6 . the location of the knee : @xmath113 year ( observation : @xmath114my ) + these values are not much changed for the projected time series @xmath115 because the original time series @xmath116 is already almost the same as the projected shape , relatively long steady chrons and sharp reversal . the frequency of chron intervals is calculated . the power index : @xmath117 ( observation : @xmath118 ) next we pick up a typical chron and calculated the power spectrum within this period . + the power index : @xmath119 for @xmath120year ( odp : @xmath121 for @xmath122 ) in the numerical calculation of lcs model . the positive and negative values in the vertical axis correspond , respectively , to the normal and the reversed polarity . ( botom ) the power spectrum of this time series . the data is sucessfully fitted by the the two power laws with the indeces -0.5 for low frequency and -1.7 for high frequency regions . , width=453 ] . the left - right symmetry , or the symmetry in normal - reversed polarites , is not respected . this indicates the model has strong fluctuation in the long time variation . this accord with the ever increasing power spectrum toward low frequency region fig.([fig4 - 1 ] ) . , width=453 ] the apparent asymmetry in fig.([fig4 - 3 ] ) is not caused by the initial condition but suggests long range temporal fluctuations in our model . this seems to be consistent with the ever increasing power spectrum in the low - frequency direction in fig.([fig4 - 1 ] ) . in some chron . the valuea are mostly negative but perpetually fluctuating strongly . ( botom ) the power spectrum of this time series . the data is sucessfully fitted by the the two power laws with the indeces -0.46 for low grequency and -2.0 for high frequency regions . , width=453 ] for each spin with shifted zero points for visibility . the first spin from top was rapidly rotating before the flip but the nineth spin took over this role after the flip . these perpetually rotating spins form the gas phase . see the next section for detail . ( bottom ) the countour / density graph for horizontal component of the magnetic field eq.([b ] ) on the earth surface around this polarity flipping era . each ellipse represents the entire earth surface drown in mollweide mapping , before , right in the middle , and after the flip from left to right . the radius of the spin ring is assumed to be one - third of the earth radius . , width=453 ] the results of lcs model for geomagnetism is almost the same as our previous study on scs model and are not so far from the observational values . thus the range of interaction is shown to be irrelevant for geomagnetic characteristics . in the scs model , even the inertial term was not necessary to yield the similar results . it is interesting to notice that this lcs system , we explored in the last section , is a slight extension of the _ hamiltonian mean field model _ ( hmf ( antoni & ruffo 1995 ) , ( campa , dauxois , & ruffo 2009 ) by adding the potential term @xmath123 . hmf model a good tool to describes phase transitions and statistical fluctuations of a deterministic system composed from many elements . slightly modified version of hmf model was applied to describe the self gravitating systems(sota , iguchi , morikawa , tatekawa , & maeda 2001 ) . however this hmf model itself can not describe the polarity flip of geomagnetism because of the lack of the preferred angle . the polarity flip is made possible simply by introducing the above potential term . in this sense our lcs model is the _ minimal model _ to describe geomagnetism . acute readers might have some concern that the geodynamics is a dissipative system rather than the conservative system like hmf model . however the difference is not relevant in our context . hmf and lcs systems have strong chaoticity and the mean field @xmath124 has apparent statistical fluctuations . in short , these systems have the relevant central degrees of freedom ( _ i.e. _ mean field ) and the rest ( _ i.e. _ environmental degrees of freedom ) together . we focus on the mean field . the mean field immersed in the rest behaves as the dissipative system . this point will be clarified in the rest of this section . dynamics of both hmf and lcs models can be visualized by the interacting @xmath69-particles which are moving on a circle of unit radius as shown in fig.(fig5 - 1 ) . this is a typical snapshot of the spin angles @xmath125 represented by the points on a circle . the whole ring represents all the directions of spins @xmath126 and the top position of the ring is the direction of the earth rotation axes @xmath127 @xmath128 . . the whole ring represents @xmath129 and the top position of the ring is @xmath130 . the earth rotation axes @xmath28 has the fixed angle @xmath130 . as is shown the most spins are condenced to the almost fixed angle ( i.e. fixed polarity ) , and few rapidly rotating spins exit . the latter trigges the flip of the condensed spins intermittently . , width=453 ] furthermore it is remarkable that both hmf and lcs models have two classes ( phases ) of elements : the spins in _ expanded phase _ which are almost freely moving all the angle and the spins in _ condensed phase _ which are bounded in a localized angle . there exists a critical energy in hmf model in the limit @xmath131 ; the condensed phase appears only when the system energy is below this critical energy and the portion of the condensed phase increases for lower energy . similarly in our lcs model , expanded and condensed phases coexist in wide rage of parameters is not so large . ] . we have chosen our parameter so that the system energy is low enough to guarantee that the both phases coexist in one system . the condensed phase yields finite order parameter defined by eq.([m ] ) . free spins in the expanded phase perpetually disturb the condensed cluster spins and is considered to trigger the intermittent signature change of the order parameter eq.([m ] ) . it is interesting that the expanded and condensed phases often exchange their constituent spins when the angles of the spins coincide with each other , _ i.e. _ collide in the above visualization . in terms of geomagnetism , coexistence of the expanded and condensed phases are essential ; the latter yields the well defined polarity in magnetism eq.([m ] ) and the former triggers the intermittent flip of the polarity . this phase coexistence is naturally realized in lcs model when we choose mildly low energy values . thus , this model naturally deduces that the system energy required for intermittent polarity flip necessitates _ the coexistence of the relatively steady component and the rapidly changing component in geomagnetism_. the former is of course from the condensed phase and the latter from the expanded phase . since the macro - spin has magnetic dipole , these freely moving spins in the expanded phase yields _ rapidly moving pairs of local magnetic spots of reverse polarity _ superposed on the average steady dipole magnetism caused by the spins in condensed phase . these pairs also have a tendency to align on north - south direction according to the potential form in eq.([vingcs ] ) . these pairs have some similarity with the south atlantic magnetic anomaly ( sama ) observed on the core - mantle boundary ( korte 2010 ) . the above interpretation that the spins in the expanded phase trigger the flip of the global magnetic polarity suggests that a possible _ correlation between the activity of the spins in expanded phase and the duration of the fixed polarity_. actually in our numerical calculations , the expanded phase temporally disappears in the superchron , i.e. the longest period of fixed polarity . this correlation should be further studied . anomalous acceleration in the movement of the local magnetic spots of reverse polarity may predict a polarity flip in the near future . it is also interesting to notice that this set of equations eq.([eomgcs ] ) is a slight extension of the _ kuramoto model for synchronization_(kuramoto 2003 , acebron , bonilla , perez , ritort , & spigler 2005 ) by the change of potential : @xmath132 , and the reduction of the order of time derivative : @xmath133 . the potential @xmath134 forces each angle @xmath56 to rotate with the fixed frequency @xmath135 despite the lack of the inertial term . therefore the latter simplification @xmath136 is possible in kuramoto model . kuramoto model is a set of long - range coupled oscillators with different frequencies . if there were no interaction between the oscillators , then each phase of the oscillator behaves independently . however if the long - range interaction is sufficiently strong , then the phases of oscillators _ synchronize _ with each other despite the difference in each frequencies @xmath137 . kuramoto model very generally describes the synchronization processes of many type of oscillators and limit cycles including chemical and biological systems . it is remarkable that lcs model can also describe the synchronization of element spins and yield _ quasi - periodic motion of the mean field _ , by simply changing the parameters . the first term of rhs in eq.([vingcs ] ) gives independent non - linear oscillations of each spin angle . since the oscillation is nonlinear and the amplitudes are different from spin to spin , the phases of the spins behave very chaotically . the second term of rhs in eq.([vingcs ] ) gives the cosine potential between all the spin pairs ( @xmath138 ) and thus naturally yields the tendency for all the pairs to synchronize with each other . whether the whole synchronization actually takes place or not depends on the balance between the two terms . if the pairwise interaction @xmath139 is sufficiently large in lcs model , then the synchronization takes place as in the case of kuramoto model . however , reflecting the strong non - linearity of the cosine potential , the synchronized oscillation becomes quasi - periodic in general . this potential ability of lcs model to describe synchronization physics is especially important when we apply lcs to the solar magnetism in later sections . we would like to argue , in this section , to what extent the lcs model can be _ general _ in wider contexts . the concept of the coupled dynamo elements itself , as we have discussed in section 2 , is general and is not in principle restricted to geomagnetism . we expect the same for lcs model , which is the minimal realization of this concept . the magnetic fields associated with celestial objects are quite common in the universe ; many planets , satellites , and the stars including our sun . although the magnetic fields of these objects have variety that is quite different from the geomagnetism , all the magnetic fields are thought to be generated and supported by the robust dynamo mechanism . however the lcs model is at most a phenomenological model to describe dynamo systems . the lcs model is simply a macroscopic description and can not be directly deduced from the microscopic mhd equations eqs.([eqforv]-jb ) , which are especially important for quantitative argument in wider contexts . therefore we _ need to supplement lcs model with the mhd equations_. first we examine the _ energy flow _ for geomagnetism . geomagnetic fields would be amplified by the inward - winding motion of the convection vortex , anti - cyclone . this inward motion is caused by the coriolis force induced by the earth rotation although the rotation itself can not directly transfer its energy to geomagnetism . on the other hand , the convective motion of the iron fluid is supported by the thermal flow from the central region of the earth . this flow is mainly supported by the release of the gravitational potential by the core sinking . thus the thermal flow accelerates the convective motion and supports the inward - winding force to amplify the magnetic fields through the coriolis force . one essential fact for this thermal energy flow is that the main dissipation is due to the joule heat . this is because the magnetic prandtl number @xmath140 , i.e. the ratio of the time scale of magnetic diffusion ( @xmath141 ) and that of convective diffusion ( @xmath142 ) , is actually very small : @xmath143 . we now closely look at the basic equations eqs.([eqforv ] , [ eqforb ] ) in the above picture of energy flow . from eq.([eqforb ] ) , we first notice that the steady state for the magnetic field requires the _ energy balance on average _ : @xmath144 , which reduces to the relation which is independent of @xmath14 , @xmath145where @xmath146 is the fe - core radius and @xmath147 is the radius of the taylor cell . the existence of this special scale is essential to our model to characterize the dynamo elements ; most of the relevant process of dynamo takes place on this scale . this small scale compared with the core size is foreseen from the order in eq.([rm ] ) . the magnetic field is amplified by the inward winding motion of the convection vortex through the term @xmath148 until this steady state eq.([v - steady ] ) is realized . on the other hand , the inward - winding motion of flow is generated by the term @xmath149 in eq.(eqforv ) , and this term should eventually _ balances _ with the backreaction from the generated magnetic pressure through the term @xmath150 in the same equation . therefore in the steady state , we have the estimate of the typical strength of the magnetic field @xmath151 at the scale @xmath147 , the scale of the taylor cell , as @xmath152 the observed magnetic field @xmath153 at the earth surface is related with this estimated @xmath151 through the magnetic flux conservation as , @xmath154where @xmath69 is the number of taylor cells in the core and @xmath155 is the earth radius . we simply assumed that all the taylor cell is the same scale and the magnetic fields associated with tc are parallel to the rotation axes of the earth for simplicity . combining the above three equations , we have @xmath156and the corresponding ( virtual ) magnetic moment @xmath157 becomes@xmath158 in the geomagnetic case , we have chosen @xmath87 and putting reasonable values , @xmath159 km , @xmath160 km , @xmath161emens / m , @xmath162kg / m@xmath163 , and assuming @xmath164 , we have @xmath165tesla from eq.([bout ] ) , and @xmath166m/@xmath167 from eq.([v - steady ] ) . although the value of @xmath153is consistent with observations the value of @xmath168 is only @xmath169 of the estimated speed of the convection flow simply from the west - ward drift motion . if this discrepancy is real , the time scale of the magnetic field variation or of the electric current change is about ten times larger than the convective flow speed . this point should be further studied . the estimate of the effective magnetic moment in eq.([d ] ) is general since it is deduced simply from the basic equations ( [ eqforv ] , eqforb ) based on the lcs model . the expression eq.([d ] ) is consistent with the general claim in the literature ( christensen2010 ) , ( stevenson 2010 ) that the planetary magnetic fields are essentially determined from the factor @xmath170 . we can step further and estimate the unknown factor @xmath171 in eq.([d ] ) . by using various data for celestial objects , but assuming @xmath172 is a constant , we can draw a graph for @xmath157 against @xmath173 . ] . this turns out to show a power law relation with the index about 3/4 and the angular moment @xmath174 for celestial objects @xmath175 where @xmath176meter@xmath177ampere)@xmath178 @xmath179 . ] : @xmath180 , though this is not what we needed . simply motivated by this scaling relation , we can draw a similar graph for @xmath181 against @xmath124 in fig.([fig6 - 1 ] ) . this relation turns out to be almost a constant within a factor @xmath182 for the whole mass range of 8-digits as shown in fig.([fig6 - 1 ] ) . is shown in the vertical axes against the mass @xmath124 in the horizontal axes for various celestial objects . these quantities are normalized by the values of earth . the data are from stevenson ( 2010 ) . eulopa and callisto are excluded from the figure because their magnetic fields are probably induced from the external fields ( stevenson 2010 ) . the value @xmath183 t is used for the solar dipole magnetic field . the quantity @xmath184 turns out to be an almost constant within a factor @xmath182 for the mass range over 8 digits . this fact suggests the scaling @xmath185 , which determins the number and the radius of the dynamo element in general . , width=453 ] the factor @xmath186 is important for this constancy yields the smallest variation of the data as far as we have checked . when we choose @xmath124 for the extensive variable , the best index was @xmath187 which is @xmath188 within an error . ] . we can deduce the scaling formula , from this phenomenological relation , as @xmath189 in other words , the scaling eq.([d ] ) holds if we choose eq.(taylorcellscaling ) . this phenomenological scaling is helpful to reveal the number of dynamo elements . for example , if we suppose the parameter @xmath190 is almost a constant for various objects , and @xmath87 for earth , we have @xmath191 for the sun . however , we must keep in mind that the planets , satellites and stars are very different with each other and the structure of dynamo mechanism is different ( stanley 2010 ) despite the above generality . keeping the above caution in mind , it is interesting to try the unified view of planetary magnetism within our model . in the context of the lcs model , observed large dipole tilt more than @xmath192 and the quadrupole moment of magnetic fields in _ uranus and neptune _ are not exceptional . the magnetic field described by lcs model naturally becomes irregular and asymmetric when occasional large excursion takes place . we may observing these irregular and asymmetric fields in the present uranus and neptune . moreover in the context of lcs model , almost axis symmetric magnetic field of _ saturn _ does not conflict with the anti - dynamo theorem , which inhibits dynamo function in the axisymmetic steady configuration . this is because individual element dynamo yields magnetic field independently from others and easily violates axis symmetry of the whole system . the magnetic field in lcs model is rapidly changing and steady condition is also violated . as we have seen in the previous section , a simple argument on the mhd equations based on the lcs model can yield a general scaling law for planets and satellites . this was useful to estimate the scale of dynamo elements . in this section , we would like to explore this generality of the lcs model by applying it to the solar magnetic dynamics . lcs model does not specify the macro - spin or dynamo element . the macro - spin can express any subsystem which has minimal dynamo property . typically the element is the inward - winding vorticity . in case of our sun , the macro - spin may be the convecting vortex deep inside the solar surface . it may happen that this region is turbulent and the hierarchy of vortices yield a network of many dynamo elements . the very end of this hierarchy may appear on the entire surface of the sun , possibly as the supergranulations . recently the pattern of the supergranulation is observed by using the local helioseismology ( nagashima 2010 ) . on top of these supergranulation , strong horizontal magnetic field is observed everywhere on the surface ( ishikawa 2008 ) . we speculate that these vortices associated with dynamo function may fill the entire convecting region inside the sun . the solar magnetism changes its polarity quasi - regularly with the period about 22 years , which synchronizes with the solar activity of period 11 years . the solar activity can be roughly measured simply by counting the sunspots ( sidc 2011 ) . we show this daily count profile in fig . ( [ fig7 - 1 ] ) top , during the past two hundred years . the power spectrum , shown in fig . ( [ fig7 - 1 ] ) bottom , reveals the clear periodicity of about 11 years , as well as the period of about a month . the latter period is an artifact reflecting the systematic sunspot change on the solar surface which rotates with this period . what is interesting for us is the scaling property with index @xmath193 , near to @xmath194 , on top of the two peaks associated with the above periods . one - over - f noise or the pink noise , _ i.e. _ its power spectrum scales with index @xmath194 , has been observed everywhere in nature . ( for a comprehensive list of references on 1/f noise , see ( li 2011 ) ) . similar 1/f noise in the monthly averages of the sunspot numbers has been reported in the high frequency ( @xmath195 ) regions ( franchiotti , sciutto , garcia , & hojvat 2004 ) . and @xmath196 . the former peak corresponds to the 11 year solar activity cycle and the latter peak fictitiously appears corresponds to a month the solar rotation period . , width=453 ] it is remarkable that the lcs model shows quasi - periodic polarity flip , which resembles the solar magnetism , as a result of synchronization of many spins . for the synchronization process to set in , we need large number of spins ( @xmath197 ) and low potential barrier ( @xmath198 ) . then the spins in lcs model can naturally synchronize to yield quasi - periodic motion like in the kuramoto model . for example in a numerical demonstrations , we set @xmath199 @xmath200 @xmath201 , and obtained the result as shown in fig.([fig7 - 2 ] ) . the top figure shows the time evolution of the order parameter @xmath202 defined in eq.([m ] ) . quasi - periodicity in its evolution is apparent . the amplitude irregularly varies and even becomes almost flat in some periods . the bottom figure shows its power spectrum . beside the apparent peak corresponding to the quasi - periodicity , one - over - f noise , actually the power @xmath203 , is observed in low frequency region . however this result needs to be refined before serious discussions . for example , the number of spins @xmath204 is apparently too small compared with the previous estimate @xmath205 in the last section . if we suppose that the solar magnetic activity is linked with the number of sunspots , then the above results suggest that the lcs model can reproduce the observed solar activity especially _ the quasi - periodicity as well as 1/f like power law in the low frequency region in its power spectrum_. in the numerical calculation with the parameter @xmath199 @xmath206 @xmath207 , and the initial condition is chosed so that all the spin angles are equally separated within the 70% of the whole angle @xmath129 . each spin has the interaction with nearest @xmath182 spins . the quai - periodic oscillation is apparent . this oscillation appears as a result of synchronization of many spins . ( botom ) the power spectrum of this time series . there is a peak at @xmath208 reflecting the above quasi - periodicity on top of the rough power law behavior with index @xmath209 in low frequency region . these behaviors accord with the solar observation in the previous figure . , width=453 ] we have developed a long - range coupled macro - spin lcs model for geomagnetism , solar magnetism , and others . we have seen this model is _ minimal and general_. in _ section 2 _ , we developed the idea of the coupled dynamo elements for general dynamo mechanism in magnetohydrodynamical systems such as geomagnetism . the element should have inward - winding vorticity to amplify the magnetic fields . the we introduced a macro - spin representing such element . in _ section 3 _ , we have reviewed the short - range coupled spin ( scs ) model which successfully describes geomagnetism . in _ section 4 _ , we introduced the long - range coupled - spin ( lcs ) model which also successfully describes geomagnetism . thus the spin models are shown to be useful to describe geomagnetism in the context of statistical fluctuations . in _ section 5 _ , we studied the physical relevance of lcs model . first of all in this model , _ two phases of spins can coexist _ : the expanded phase in which the spins can rapidly move all the directions almost freely and the condensed phase in which the spin directions are bounded into narrow region to form the dominant dipole moment which is almost steady . the spins in expanded phase perpetually `` collide '' ( see the figure [ fig5 - 1 ] ) with the condensed phases and trigger its intermittent polarity reversal . on the other hand lcs model shows quasi - periodicity as well as 1/f noise on top of it . this is the _ synchronization _ of the constituent spins . in both cases , the system shows power law behavior reflecting the long range interactions . the phase coexistence is essential for geomagnetism and the synchronization for solar magnetism . these study taught us how the general models such as hamiltonian mean field ( hmf ) model and kuramoto model , if slightly modified , can describe the universality and variety of general dynamo mechanism . in _ section 6 _ , we applied lcs model , supplemented with mhd equations , to other planets and satellites . finding a useful scaling low , we could estimate the scale and the total number of the dynamo element , or macro - spin , in terms of the mass @xmath124 of the body . in _ section 7 _ , we applied lcs model , with different parameter values , to the solar magnetism and successfully describe the quasi - periodic oscillation of the solar magnetism on top of the 1/f noise property in low frequency regions . thus we have actually demonstrated the universality and variety of lcs model . although the lcs model can describe most of the geomagnetic observations , it can not be deduced from the basic mhd equations . it is also possible to generalize lcs model to endow each spin with dynamo function . this is realized for example by setting the homopolar generator ( faraday disk ) for each dynamo element , @xmath210 , and introduce appropriate interactions between them . this is an extension of the original rikitake model ( rikitake 1957 ) . if we introduce nearest neighbor interaction , we obtain the short - range coupled rikitake model ( scr ) , for the currents @xmath211 and rotation angles @xmath212,@xmath213@xmath214 we can obtain the occasional polarity flip and power law power spectrum also in this model . however even @xmath215 case , the time evolution of scr model inherits chaotic spiky profile from the original rikitake model . the long - range coupled version of rikitake model ( lcr ) is also possible,@xmath216@xmath217 where @xmath218 is the mean field . in this case we obtain a regular and spiky oscillation due to the strong synchronization of the elements . we have not yet performed a full parameter study . further extension of such models would be possible and they will be worth extensive research .

active galactic nuclei pose a fundamental problem how does rotationally dominated gas on kpc scales lose orders of magnitude in angular momentum to fall onto an accretion disk on au scales ( e.g. , * ? ? ? we do not know the mechanism that dissipates angular momentum and allows gas to accrete . there are no shortage of ideas , including major or minor mergers ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , bars and bars within bars ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , and nuclear spirals @xcite as perhaps indicated by dust lanes @xcite . even looking directly at molecular gas kinematics on pc scales in active galaxies , it is very hard to find clear evidence for the true driver of nuclear activity @xcite . thanks to the advent of near - infrared integral - field spectrographs ( nir ifu hereafter ) , there has been significant progress in understanding gas flows in the centers of active galactic nuclei ( agn ) in recent years . the nir observations allow us to penetrate gas and dust in the active nuclei , and , in conjunction with adaptive optics , to probe very near to the nucleus . in some nearby cases , inflows are directly observed along circumnuclear spirals in ionized gas on @xmath2 pc scales ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? furthermore , there are intriguing hints of kinematic differences in the nuclei of local seyfert galaxies . for example , @xcite suggest that the ionized gas in the disks of seyfert galaxies ( i.e. , outside of the narrow - line region or nlr ) is more kinematically disturbed than the gas in a control inactive sample . @xcite also report that seyfert galaxies have more concentrated stellar luminosity profiles , lower stellar velocity dispersions , and elevated h@xmath4 1 - 0 s(1 ) emission within @xmath5 pc relative to an inactive subsample . in this work , we specifically focus on megamaser disk galaxies . in these special systems , luminous water megamasers at 22 ghz trace a geometrically thin molecular disk on sub - pc scales in orbit around the central bh ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? the megamaser spots show near perfect keplerian rotation in many ( e.g. , * ? ? ? * ) , but not all cases ( e.g. , * ? ? ? * ; * ? ? ? thanks to the keplerian rotation in the maser disk , we know the bh mass to a few percent , limited in most cases by the distance to the galaxy ( e.g. , * ? ? ? direct geometric distances have also been derived by measuring the accelerations of the systemic masers for the few best cases @xcite . in addition to distance measurements and bh masses , the megamaser disk reveals the spin axis of the accretion disk on sub - pc scales . to get the long path lengths required for masing , the disk must be virtually edge - on ; we thus typically know the orientation of the disk on sub - pc scales to within a couple of degrees @xcite . we have already found that the spin axis of the megamaser disk aligns with the jet on @xmath2 pc scales , while the spin axis of the disk does not appear to align with flattened disk - like structures on @xmath6 pc scales identified in _ hst _ imaging @xcite . in this paper , we explore the use of nir ifu data to define the gas flows on @xmath7 pc scales , for comparison with the megamaser disk . we currently have nir ifu observations of the centers of two galaxies in the kuo et al . sample : ngc 4388 and ngc 1194 . these two are the only two megamaser disk galaxies with published bh masses where we can come close to resolving the gravitational sphere of influence , in the former case because the galaxy is nearby ( in virgo ) and in the latter because the bh is the most massive known with a maser disk @xcite . we observed ngc 4388 with sinfoni under program 386.b-0244a on the very large telescope @xcite on april 12 & may 5 2011 in natural seeing conditions and in service mode . we have been awarded time to observe this galaxy at two spatial resolutions , the higher one involving assistance of the adaptive optics system ( ao ) , however to date these observations are only partially carried out . thus , in this paper we present the lower spatial resolution data set that covers a wider field of view . our observations cover the central 8@xmath8 8 , with a spatial sampling of 0125 @xmath8 0250 . we used the @xmath9-band grating ( 1.95 2.45 @xmath10 m ) that gives a spectral resolution r@xmath113500 ( 6.2 fwhm , as measured on sky lines ) . for the observations we used standard near - ir nodding techniques . observing blocks consisted of a sequence of object and sky frames ( osoosoosoosoo ) . each individual integration was 300 sec and the sky fields were offset by 240 to the north and east . science frames were dithered by 03 in order to reject bad pixels . there were two observing blocks , with a total on - source integration time of 100 min , and 40 min in total on sky frames . additionally , after each observing block and at a similar airmass , we observed a b dwarf star to act as a telluric star . we used the eso sinfoni pipeline v2.3.3 to perform the basic data reduction on each observing block , consisting of 10 object and 4 sky exposures . in brief , the pipeline extracts the raw data , applies distortion , bad pixel and flat - field corrections and wavelength calibration , and stores the combined sky - subtracted spectra from one observing block in a 3-dimensional data cube . the final pixel size is @xmath12 . we reduced the telluric stars in the same way as the science frames . for each telluric star we extracted a one - dimensional spectrum , removed the hydrogen brackett@xmath3 absorption line at @xmath13 m after fitting it with a lorentzian profile , and divided the star spectrum by a black body spectrum with the same temperature as the star . the last step in preparing the telluric spectrum was to apply small shifts in the spectral direction ( @xmath140.05 pixels or 0.123 ) and scalings to minimize the residuals of the telluric features . to do this , we extracted a central one - dimensional spectrum from each science data cube and cross - correlated and fitted it with the corresponding telluric spectrum . we derived the wavelength - dependent correction required to match the continuum shape of the telluric star ( without an absolute zero - point ) . thus , when we divided each individual spaxel in the six galaxy data cubes by the corresponding best - fitting telluric spectrum , we also obtained a relative flux calibration . based on the seeing monitor , during our observations the seeing varied from @xmath15 for the first observation , and between @xmath16 for the second observation . note that the seeing monitor takes measurements at 5000 , thus our seeing is typically better than this . from the standard stars ( taken at the end ) we measure a fwhm of 08 for the first observation and 05 for the second . we aligned the two obs with integer offsets and summed them ; all analysis was done on this final summed data cube . we observed ngc 1194 using the ifu oh - suppressing infrared imaging spectrograph ( osiris * ? ? ? * ) assisted by the laser - guide star adaptive optics system @xcite on the 10 m keck ii telescope . the data were taken over the latter two - thirds of a single night on oct . we observed in the @xmath0band with the kbb filter , for a spectral coverage of @xmath17 and a spectral resolution of @xmath18 . we used the @xmath19 spatial scale , and oriented the long axis of the ifu along the major axis of the galaxy at a pa of @xmath20 . we dithered the galaxy by @xmath21 perpendicular to the long axis of the ifu , both to facilitate bad pixel removal and to slightly increase the field - of - view to @xmath22 . unfortunately , the nucleus of ngc 1194 was too faint to use as a tip - tilt star . thus , we used a nearby ( 54 ) star ( id=0888 - 0029937 ; @xmath23 ) . natural seeing conditions varied over the night from @xmath24 to @xmath25 . from our final science data cube we find that the psf core had a fwhm of @xmath26 as measured from the galaxy core . we also observed a telluric standard ( a0v ) star approximately once per unit change in airmass . finally , we interleaved sky and object observations with an object - sky - object pattern , spending 600s per sky and object exposure . we acquired a total of 170 min on source . we followed closely the reductions outlined in @xcite , using the osiris data reduction pipeline ( v2.3 ) made available by the instrument team . we repeat salient details here for completeness . the pipeline handles sky subtraction , cosmic - ray removal , glitch identification , spectral extraction into a data cube , wavelength calibration , atmospheric dispersion correction , and telluric correction using an a0v star . we experimented with a scaled sky subtraction method from @xcite , but did not find substantial improvements from the direct method . finally , we determined the centroid of each exposure by hand , applied these sub - pixel offsets , and averaged all cubes into our final data cube . here we describe our analysis procedure for the ngc 4388-sinfoni data cube in detail , highlighting any differences in analysis with the ngc 1194 data cube . we use the direct - pixel fitting code ppxf of @xcite to fit the stellar continuum . to achieve adequate signal - to - noise ( s / n ) we use the voronoi binning scheme of @xcite , which determines contiguous bins with uniform s / n , in this case s / n@xmath27 per pixel . the sinfoni pipeline does not return an error array , and so the noise is determined empirically from line - free regions of the spectrum as the variance in the spectrum once large outliers are clipped . in the case of the osiris data , we use the per - pixel error array generated by the reduction pipeline . we then fit the coadded spectrum corresponding to each bin with a weighted combination of stellar templates . each template is shifted to the systemic velocity of the galaxy and convolved with a gaussian line - broadening function . a polynomial of fourth order is added to account for nonthermal continuum from the agn , and potentially flux calibration errors as well . example fits are shown in figure [ fig : specn4388 ] . the measurement uncertainties in both velocity and velocity dispersion are determined with monte carlo simulations . in the case of the ngc 1194 data , the data reduction software generates an error array . however , for the ngc 4388 data , we use residuals from the best fit ( with the emission lines removed ) to determine the average sigma . we create 100 perturbed input spectra assuming gaussian errors in the spectra . we then refit these 100 artificial spectra . the uncertainty in each parameter is calculated as the values encompassing 68% of the trials . we experimented with a number of fitting regions , including a fit to the bandhead only ( @xmath28 , short ) , a fit to the full region ( @xmath29 , full ) , and an intermediate region ( @xmath30 , best ) . we found that the first fit delivered reliable velocity dispersions , but unstable radial velocities because of the narrow wavelength range , and because the bandhead at @xmath31 is often filled in by [ ] @xmath32 emission . fits to the full spectral region suffered from telluric residuals at @xmath33 . the just - right spectral region was therefore the third , intermediate case . the median errors on the velocities are @xmath34 km s@xmath35 for the best , short , and full regions respectively , while the median errors on the dispersion measurements are @xmath36 km s@xmath35 . in all cases , we masked high - ew emission lines , including the h@xmath4 lines at @xmath37 @xmath38 and also [ ] @xmath39 , @xmath40 , br@xmath41 , and [ ] . to bracket uncertainties caused by template mismatch , we utilize two stellar template libraries in fitting the sinfoni data . one is a library of stars observed with sinfoni using the same observational set - up . in this case , no correction for instrumental resolution is required . the native sinfoni templates range from k4iii to m5iii ( plus a k4.5v star ) @xcite . as a check , we also use the stellar template library of @xcite , observed at higher spectral resolution . we use this second template set only to test our sensitivity to template mismatch . the wallace templates cover a wider range in spectral type , including km supergiants and km giants . we discuss the stellar template fits to the ngc 4388 data in detail in s 4.1 . even with different templates , we find good agreement in the two sets of stellar velocity dispersion measurements . taking @xmath42 as the answer based on the sinfoni templates and @xmath43 as the same for the wallace templates , we find @xmath44 ( where the latter number is simply the standard deviation in this ratio ) . we find that the two measurements agree within 20% ( i.e. , within the observational uncertainties ) and we find no significant systematic offset . in the case of the osiris data , our primary fitting uses a single k5iii star , which provides an acceptable fit to our moderate s / n data . we do not a priori know the location of the black hole , or the precise photometric center of the galaxy . we could use the continuum to determine the photometric center , but then we would be sensitive to obscuration , which is still significant in the nir ( see figure [ fig : image ] ) . instead , we use the equivalent width of the co bandhead to determine where the active nucleus peaks @xcite . we use the index definition from @xcite : the index band is @xmath45 , with a continuum band centered on 2.2900 and a width of 0.0003 @xmath46 , and we also followed their prescription to correct for velocity dispersion although this correction is negligible ( see also * ? ? ? * ) . these authors show that in the absence of an active nucleus , star - forming galaxies have a very uniform co ew . a declining co ew towards the galaxy center can be attributed to infill by the nonthermal continuum from the accreting black hole . we show an example in [ sec : stellarkin4388 ] . we create an agn continuum map by taking the co ew at the edge of the cube ( ew=9.5 in ngc 4388 , consistent with stellar - population models ) and assuming that all dilution further inwards results from the agn , such that continuum ( agn ) = [ total continuum ] @xmath8 [ 1 - co ew / co outer ] . the photometric center of the continuum map falls @xmath47 south of the co ew center . the optical narrow - line region cone is only visible to the south ( [ sec : nlr ] ) , so it is sensible that the continuum centroid should be shifted towards the low - reddening side of the disk . the best - fit center also agrees within less than a pixel with the peak h@xmath4 and br@xmath3 emission , although these latter are not as well defined . we will use this position as the center of the cube throughout . in the case of ngc 1194 , the co - derived and photo - centers agree within @xmath48 . while voronoi binning is very useful to create uniform continuum s / n ( e.g. , * ? ? ? * ) , it is more difficult to apply to patchy emission - line maps , where by using large bins we may smear out interesting features in the line emission maps . we therefore create a second data cube using a @xmath49 pixel smoothing in the outer regions and a @xmath50 pixel binning in the inner @xmath50region . with this level of binning , we increase the s / n by a factor of two but still maintain multiple pixels across the psf . we measure the emission lines from this uniformly gridded cube . in the case of ngc 4388 , because we have such high s / n in the continuum , we again perform continuum - subtraction using ppxf before fitting the emission lines . however , there are no strong stellar absorption features beneath the emission lines of interest , and thus for ngc 1194 , where the s / n is poor in the continuum , we simply fit a local continuum value . we fit each emission line independently . we model the intrinsic line shape as a gaussian , although we allow for up to two gaussian components for emission lines with s / n@xmath51 in ngc 4388 . particularly in the br@xmath3 and [ si vi ] lines , the line shapes are often non - gaussian and warrant a second component . in @xmath52 of cases within the central @xmath53 the s / n is sufficient to warrant a two - component fit . we calculate the non - parametric fwhm from the sum of gaussian components , and from that value a corresponding line dispersion . we also experimented with an empirically determined line - broadening function by fitting a gauss - hermite model to the strong sky emission lines , which caused only very minor changes in our fits . we derive errors on all the fitted parameters by fitting 100 artificial spectra that are created as a sum of the original spectrum and a gaussian random noise array generated from the extracted errors . ngc 4388 is an sbb galaxy at a distance of 19 mpc @xcite , corresponding to a physical resolution of @xmath1 pc given our @xmath54 seeing . the galaxy magnitude is @xmath55 mag , with a b / t@xmath56 based on fits to the sdss data @xcite . we show an hst / wfc3 image combining f336w ( broad @xmath57 ) , f435w ( broad @xmath58 ) , and f814w ( broad @xmath59 ) in figure [ fig : image ] . with a photometric p.a . of 90 , the galaxy is very nearly edge - on , which is not true for the majority of the megamaser disk galaxies @xcite . the inclination determined from the large - scale kinematics is 78@xcite , and the systemic velocity from the same study is @xmath60 km s@xmath35 . the gas kinematics show an isophotal twist on arcmin scales , that @xcite successfully model as a bar within the inner 1.5 kpc , with a p.a . of 100 on the sky . the bar is also apparent in the photometry as a boxy bulge . ngc 4388 is well - known as a galaxy that is falling into the virgo cluster @xcite . gas is being stripped from the galaxy , as evidenced by an and ionized gas tail @xcite . there is even a detection of the photoionized gas in the soft x - rays @xcite . we are interested in the center of ngc 4388 , in particular because of the megamaser disk on sub - pc scales @xcite . based on the megamaser rotation curve , the bh mass is found to be @xmath61 @xmath62 @xcite . combined with the bolometric luminosity estimate from @xcite , the eddington ratio is @xmath63@xmath64 , which is in very rough agreement with our estimate from the [ ] luminosity @xcite . like most megamaser disk galaxies @xcite , the active galactic nucleus ( agn ) in ngc 4388 is compton thick and quite bright in hard x - rays ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? ngc 4388 has also been seen to change state in the x - ray @xcite , switching from being compton thick to compton thin . weak broad h@xmath65 has also been reported at the galaxy center @xcite . we first examine stellar rotation and dispersion . we clearly see rotation within our @xmath66 pc field of view that aligns with the disk on larger scales . following @xcite , we first use ppxf to fit the average spectrum over all data , excluding a 05 radius around the agn ( figure [ fig : specn4388 ] ) . in the fiducial fits , we do not fit higher - order gauss - hermite moments . we will examine @xmath67 and @xmath68 in [ sec : stelrotcurve ] below . all of our default fits employ the sinfoni templates . from this high s / n spectrum we can derive the best - fit spectral mix , which may be degenerate in lower s / n individual spectra . based on our fit to this average spectrum , we find a dispersion of @xmath69 km s@xmath35 , which compares well with our previous measurement of @xmath70 km s@xmath35@xcite . the best - fit template is composed of 50% k4/5iii , 38% k7iii , and 12% m5 ii / iii stars . we then fit spatially resolved spectra across the cube , using voronoi binning to ensure comparable s / n in all bins . we both allow the mix of templates to be a free parameter , and use the best - fit mixture from the high s / n fit to the entire cube . the latter fit has the advantage that by decreasing the number of free parameters , we increase the fitting stability in the agn - dominated region @xcite . however , we find excellent agreement between the two measurements , with the scatter in their fractional difference being only 9% . thus , we will present results with the stellar templates allowed to vary . the fitted rotation and dispersion fields are shown in figure [ fig : sigdispfield ] . our best - fit systemic velocity is @xmath71 km s@xmath35 , consistent with the measurement of @xcite . we do find a systematic difference of @xmath72 km s@xmath35 between the absolute velocity derived when we use the sinfoni stellar templates and that from the wallace et al . we are not sure of the origin of this difference ; it is possible that the sinfoni template stars , while all at rest relative to each other , have not been shifted to a rest velocity . we thus quote an uncertainty of @xmath73 km s@xmath35 in the systemic velocity . to derive the rotation curve we use the _ kinemetry _ formalism of @xcite , which is similar to ellipse fitting of photometry , but operates on both even and odd moments of the kinematic field . the kinematics are modeled using a sixth order expansion along ellipses with the position angle and ellipticity as free parameters . we show the resulting rotation curve in figure [ fig : stelrotcurve ] , along with the best - fit ellipse pa , ellipticity @xmath74 , and the amplitudes of the higher - order terms that encapsulate deviations from simple circular rotation . we clearly measure rotation in the stellar kinematics , with a maximum amplitude of @xmath75 km s@xmath35 . the stellar rotation axis of 90 is aligned with the major axis of the kpc - scale disk . on larger scales , @xcite measure a stellar rotation of 60 - 80 km s@xmath35 out to @xmath76 ( @xmath66 pc ) , consistent with ours , which then appears to fall slightly at larger radius . with @xmath77@xmath78 km s@xmath35 beyond 1(100 pc ) , we find a @xmath79 . overall , the stars are dispersion - dominated , but we see signs of kinematic components apart from just a bulge . firstly , for a @xmath79 , we would expect an ellipticity of @xmath80 for an isotropic oblate rotator @xcite , rather than the @xmath81 that we observe in the isophotes ( figure [ fig : image ] ) . secondly , within a radius of @xmath82 the dispersion field actually drops ( although the decrement is more pronounced on the east side of the disk ) . on the same scale , we see a jump in velocity , pa , and ellipticity likely signaling a transition from an inner disk to more bulge - dominated kinematics . thirdly , we also detect a disk - like structure in the inner 100 pc in the _ hst_/wfc3 f160w image ( figure [ fig : image ] ) . the nuclear disk appears to have pa@xmath83 , misaligned with the large - scale disk , and a size very similar to the @xmath84drop region . we detect this same disk component in more detailed two - dimensional fitting of the _ hst_+ground - based data ( r. lsker et al . in preparation ) . the most likely explanation of these observations is that there is a disk within @xmath85 ( 180 pc ) embedded in the larger - scale bulge / bar . only on unextincted sight lines are the kinematics dominated by the disk component , while on heavily extincted sight lines we are seeing the kinematics in the bulge / bar . the falling rotation curve outside 400 pc may also reflect the increasing dominance of the bulge / bar on larger scales . as additional confirmation of this picture , we refit the stellar kinematics and measure the higher - order gauss - hermite moments @xmath67 and @xmath68 ( e.g. , * ? ? ? although individual @xmath67 measurements have only s / n @xmath86 , we do find the classic anti - correlation between @xmath67 and @xmath87 that points to the superposition of a bulge and disk , as seen both in simulations @xcite and data ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? we also find two additional hints that the @xmath0band extinction is higher to the west . firstly , we find a significantly redder continuum slope to the west side of the disk , corresponding to a differential @xmath88 mag assuming a @xcite reddening law . secondly , we clearly see a dip in the h@xmath4 emission on the same scale on the west side of the disk ( see [ sec : gasflux ] ) . next we will ensure that the @xmath84drop is real and not an artifact of our fitting procedure . the agn is the primary source of uncertainty in our modeling of @xmath77 in the galaxy center . contamination from the non - thermal central source dilutes the stellar signal , and is degenerate with @xmath77 ( e.g. , * ? ? ? * ) . first , we look at the s / n of the dispersion measurements as a function of position . we find that the s / n is always @xmath89 except within @xmath21 of the agn . this first test suggests that our measurements at @xmath90 ( 100 pc ) should not be dominated by the agn . next , we re - run ppxf , but instead of fitting the continuum with the default polynomial of order 4 , we only allow a first or second - order polynomial ( figure [ fig : sigdispfield ] ) . while it is clear that , particularly in the case of the first - order polynomial , we achieve very poor fits to the nuclear region ( @xmath91 ) , we still find significant evidence for an asymmetric @xmath84drop . we then return to the wallace templates ( rather than our default sinfoni templates ) and try restricting the template set to k dwarfs . we find the same result . finally , we examine our fits with the template fixed to the mixture derived from the high s / n fit described in [ sec : stellarkin4388 ] . in all of these cases , although the dispersion profile within 05 is not well determined , we recover the @xmath84drop on 1 ( 100 pc ) scales . we therefore conclude that the observed drop is real and that the asymmetry results from patchy reddening . although in general the north side of the disk is more heavily extincted ( [ sec : nlr ] ) , there are also dust lanes extending to the south , which are apparently preferentially affecting the west side of the 100-pc scale disk . a falling dispersion towards the center of megamaser disk galaxies is not unexpected . these so - called `` @xmath92-drops '' are common both in spiral ( e.g , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , and in early - type galaxies @xcite , in the latter often associated with nuclear star clusters @xcite . the typical scale is @xmath93 pc , and the most likely explanation for the falling dispersion is that the galaxy harbors a gas disk that is currently or was recently forming stars . these bright , dynamically cold stars dominate the central dispersion and cause @xmath77 to fall . in fact , ngc 2273 , one of the nearer megamaser disk galaxies , has a dramatic @xmath84drop that coincides with a known disk in the inner @xmath6 pc @xcite . @xcite suggest that low central dispersions are more common in active galaxies . if the @xmath84drop is caused by a young stellar disk , then we might also expect to see younger stars coinciding with the lower stellar velocity dispersion . we create a median spectrum over the region with @xmath77@xmath94 km s@xmath35 , and rerun ppxf to investigate the mixture of stellar populations using the wallace templates , which cover a wider range of spectral type . while there is a slight preference for more supergiant light in the region of the @xmath84drop , we do not find a significant difference in best - fit spectral type as a function of region . the co ew profile is also quite symmetric and has reached its asymptotic values already in the @xmath84drop region ( figure [ fig : sigdispfield ] ) . with a larger spiral sample , @xcite also find similar stellar population ages inside and outside of the @xmath84drop regions . we propose that the observed stellar kinematics are the superposition of three distinct components . on kpc - scales , the galaxy disk has pa@xmath95 . within our 300 pc aperture , the kinematics are dominated by the hotter bulge / bar @xcite . based on the highest @xmath77 measurements within the cube , presumably occurring where dust obscuration is highest , we measure @xmath77@xmath96 km s@xmath35 in the bulge / bar . then , within the inner 100 pc , we see evidence in both the kinematics and the nir isophotes for a nuclear disk component at a pa@xmath97 , misaligned by @xmath98 from the large - scale disk and with a @xmath99 . finally , on pc scales , there is a masing disk at pa@xmath100 , which is misaligned not only from the kpc - scale disk , but also by @xmath101 from the nuclear stellar disk @xcite . with a best - fit stellar rotation field in hand , we turn to the gas kinematics . we will examine three transitions : h@xmath4@xmath102 1 - 0s(1 ) , br@xmath3 , and the coronal line [ ] @xmath103 . we note that the [ ca viii]@xmath104 line , also coronal , has been used to trace the nlr ( e.g. , * ? ? ? * ) , but is blended with the co bandhead . thus we choose to focus on the [ ] line here . the two - dimensional velocity and line - ratio fields will provide information about the origins and excitation mechanisms of each transition . the fluxes that we present are the integrated flux from our one or two gaussian models as described in 3.3 . we calculate the fwhm non - parametrically from the sum of both components , and the dispersions are simply calculated as fwhm/2.35 . we also measure and present the skewness of the lines , which encapsulates the asymmetry in the two - component fits . in figure [ fig : emissionflux ] , we show the distribution of fluxes in the h@xmath4 , br@xmath3 , and [ ] emission lines , while in figures [ fig : h2rot ] , [ fig : oiiisivi ] , and [ fig:2dgas ] , we summarize their kinematics . we see that the overall morphology of the h@xmath4follows the stellar continuum , and is even more flattened than the stellar distribution , suggesting that the emission arises from the nuclear disk . furthermore , close inspection reveals that the h@xmath4emission is asymmetric , being considerably brighter towards the east . the brighter side of the h@xmath4 disk is coincident with the observed @xmath84drop in the stellar distribution ; we can directly see the extinction of the h@xmath4 in the region where the @xmath84drop is also extincted . in contrast , the high - ionization lines are oriented at pa@xmath105 , aligned with the jet on similar scales , and also with the [ ] emission that traces the narrow - line region ( nlr ; fig . [ fig : oiiisivi ] ) . the br@xmath3 mostly follows the high - ionization lines ( fig . [ fig:2dgas ] ) . their light distribution strongly suggests that their emission arises in front of the disk in projection , since the [ ] is actually brighter towards the west ( similar to the [ ] ) . based partially on these flux distributions , as well as the kinematics and line ratios presented below , we will suggest that the h@xmath4 mainly traces star formation in the nuclear disk , while br@xmath3 and [ ] trace the nlr ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . we focus first on the h@xmath4@xmath106 kinematics ( figure [ fig : h2rot ] ) . as with the stellar kinematics above , we fit the h@xmath4 velocity field using kinemetry . the best - fit rotation curve , pa , and eccentricity are shown in blue in figure [ fig : stelrotcurve ] . with v/@xmath107 , the gas is considerably colder than the stars , but we are not seeing pure rotation . the rotation field shows an @xmath108-shaped kink . the twist is apparent as a discontinuity in the h@xmath4 ellipticity @xmath74 and pa in the rotation profiles at @xmath109 ( 150 pc ) , a similar radius to the changing stellar kinematics . in the skewness map , along the high - dispersion kink , we see a tendency for red asymmetry , which is plausibly attributed to the superposition of two velocity components . interior to the twist , the kinematic pa of the gas on @xmath110 pc scales appears to align with the megamaser disk . thus , it is possible that whatever mechanism causes this kink is responsible for supplying material on pc scales . on the same scales , we find a two - armed spiral in the dispersion map , with an inner pa @xmath105 . we find a corresponding three - arm spiral pattern in the residual map by subtracting the best - fit rotation model from the observed velocity field . to interpret these trends , we recall that any @xmath111 mode perturbation can cause @xmath112 deviations in the even velocity moments , including density and dispersion , and due to geometry @xmath113 ( and @xmath114 ) deviations in the odd velocity moments ( e.g. , * ? ? ? * ) . we propose a few mechanisms that may explain these kinematic properties . first , in the study by @xcite of the inner region of ngc 1097 , similar signatures were explained as the result of a two - arm nuclear spiral in the disk plane ( see also * ? ? ? given that the nuclear disk in ngc4388 is nearly edge - on , such a co - planar spiral density wave seems unlikely . the second possibility is that we are seeing the superposition of the kpc - scale disk and the 100 pc nuclear disk that we saw in the stellar kinematics . however , the gas pa@xmath115 , while approximately aligned with the maser disk on sub - pc scales , does not align with the stellar disk of pa@xmath97 . the final possibility is that the twist we observe here is related to the kink from the bar observed on kpc scales ( e.g. , * ? ? ? * ) . only with our upcoming ao observations can we address the @xmath116 pc gas kinematics definitively , and hopefully help determine what is driving the gas inward towards the agn . the h@xmath4 shows very regular rotation along the orientation of the large - scale galaxy disk outside of the inner 1 ( 100 pc ) . additional information is encoded in the ratios of h@xmath4 to br@xmath3 ( figure [ fig : h2brg ] ) , and ratios of different h@xmath4transitions ( figure [ fig : h2rat ] ) . typically , in the hard ionizing field of the nlr , the ratio of h@xmath4 to br@xmath3 takes a value @xmath117h@xmath4/br@xmath3@xmath118 , while it is @xmath119 in star - forming regions @xcite . finally , in dense spiral arms , even higher ratios of h@xmath4/br@xmath3@xmath120 are observed . @xcite suggest that the increased h@xmath4 emission arises from the higher gas densities within spiral arms ( see also * ? ? ? ngc 4388 obeys these trends in general ( figure [ fig : h2brg ] ) . along the disk , the h@xmath4 is strong , and the ratio has values @xmath121 , as is commonly seen in spiral arms , while in the galaxy center , the line ratio ranges from @xmath122h@xmath4/br@xmath3@xmath123 , values common for seyfert galaxies . the only surprising region lies off the disk along a @xmath124angle east of north above and below the galaxy plane , where the br@xmath3emission is strongest . here , the h@xmath4/br@xmath3 ratio is very low @xmath125 . such low ratios are reportedly associated with star - forming regions ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? however , in this case , the br@xmath3 and [ si vi ] morphologies match both the orientation of the radio jet and the nlr traced by [ ] @xmath126 in the optical . thus , we suggest that the low h@xmath4 relative to br@xmath3 in these regions is instead due to very low densities of the nlr @xcite . we now construct spectra in characteristic regions to examine both the h@xmath4 line ratios and the line kinematics . we select regions of the map based on the h@xmath4/br@xmath3 line ratio and construct high s / n spectra from the continuum - subtracted cubes . region 1 is the center , defined as all pixels with co ew@xmath127 that lie within @xmath109 of the center . we extract two regions in the disk ( h@xmath4/br@xmath3@xmath128 ; regions 2 & 3 ) and two regions in the nlr ( h@xmath4/br@xmath3@xmath129 ; regions 4 & 5 ) . finally , we sample gas that falls beyond the main galaxy and overlaps with the [ ] ionization cone ( region 8) , as well as two regions north of the disk , towards the northern radio lobe ( regions 6 & 7 ) . the extracted spectra for the strong transitions ( h@xmath4@xmath130 , br@xmath3 , and [ si vi ] ) from each region are shown in figure [ fig : h2brg ] , the spectra from different h@xmath4 transitions are shown in figure [ fig : h2rat ] , and measured properties from these spectra are presented in table 1 . because we fit the lines using an instrumental profile , the linewidths are naturally corrected for instrumental resolution . each line is fitted independently . in fact , as shown in figure [ fig : h2rat ] , the centroids and line shapes for the different transitions can differ , particularly for regions in the nlr ionization cone . for each region , we can ask about the excitation mechanisms of the h@xmath4 . there are two ways to excite the warm h@xmath4 : thermal processes or fluorescence via uv radiation . the thermal processes that can in principle heat the gas include shocks , x - ray emission , or heating by uv emission . the predicted line ratios of various h@xmath4 transitions provide diagnostics of the primary excitation mechanism . specifically , if fluorescence dominates the excitation , then the ratio of h@xmath4 2 - 1s(1 ) 2.25@xmath131/1 - 0s(1 ) 2.12@xmath131 is expected to be @xmath132 while the 1 - 0s(2 ) 2.03@xmath131/1 - 0s(0 ) 2.22@xmath131 ratio is expected to be @xmath90 @xcite . we show this pair of ratios and the expectation for pure thermal emission as the solid line , while the region dominated by fluorescence is shown schematically with the large x at the lower right in figure [ fig : h2rat ] . there is room for some uv fluorescence , but these line ratios suggest that the bulk of the excitation is thermal . similar conclusions have been drawn for larger samples of seyfert galaxies @xcite . our excitation diagram is constructed for the integrated line profile but , in fact the line ratio is clearly a function of velocity , particularly for the line emission arising from the nlr ionization cones . we do not have sufficient s / n to construct line ratios for the line core and line wings separately , but these different line shapes are pointing to different dominant excitation mechanisms in the line wings . higher s / n is needed to quantitatively examine velocity - dependent line ratios . we are still left with the question of which thermal process dominates the excitation . previous work has drawn various conclusions , with @xcite for instance favoring shocks in outflows , @xcite favoring a mix of processes including emission from photo - dissociation regions , @xcite favoring x - ray heating or shocks , and so on . recent papers have favored x - ray heating in the vicinity of an agn ( e.g. , * ? ? ? however , in our case , the fact that there is little h@xmath4 coming the nlr while the bulk of the h@xmath4emission on these scales follows the morphology of the starlight ( figure [ fig : emissionflux ] ) strongly suggests that the h@xmath4 is predominantly excited by stellar processes ( e.g. , uv radiation and/or shocks from supernovas ) rather than the agn . bridging the gap between the accretion flow on @xmath14pc scales and the large - scale ionized gas outflow , we can examine the ionized gas emission on 100 pc scales . in this region , the radio emission is complex , extends predominantly north - south ( perpendicular to the disk ) and contains two unresolved peaks and a more diffuse `` bubble '' to the north @xcite . @xcite show that the northern radio source is coincident with the optical nucleus , while the southern radio knot is coincident with a bright region of ionized gas seen in [ ] , strongly suggesting a jet - ism interaction . high - ionization gas is present to the south of the nucleus in our data cubes as well . previous near - infrared spectra ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) show extended h@xmath4 coincident with the galaxy disk , and high - ionization gas emission from br@xmath3 and [ ] coincident with the nucleus and ionization cone . with our high spatial resolution and three - dimensional coverage , we can robustly disentangle the disk and nlr kinematics . the br@xmath3 and [ si vi ] emission lines show similar spatial distributions that parallel the nlr emission seen in [ ] ( figures [ fig : emissionflux ] and [ fig : oiiisivi ] ) . their emission is distributed at an angle of pa@xmath105 , tracing the general orientation of the radio jet seen on similar scales @xcite . the ionization cone seen in [ ] continues to the south - west beyond the sinfoni field - of - view @xcite , and is filled with radio plasma @xcite . to the north , beyond our field of view , the jet opens into another wide angle fan . it is likely that the southern ionization cone sits in front of the galaxy disk , while the northern one ( which we can see in this map and in the radio but not in the optical ) is extincted by the galaxy disk . [ tab : region ] in figure [ fig:2dgas ] , we show the velocity and fwhm fields for br@xmath3 and [ si vi ] . both the morphology and the kinematics differ from that seen in the h@xmath4 . specifically , we do not see coherent rotation in either transition . the skewness plots are also interesting . in both transitions we see redshifted gas towards the south - west and blue - shifted gas towards the north - east . our interpretation is that we are seeing bulk motion of the gas along the jet , which apparently points towards us to the north and away from us to the south of the galaxy disk . in both cases we are probing the region of the nlr where the jet has `` broken - out '' of the galaxy disk . the higher velocity gas is moving along the walls of the bicone , as accelerated by the jet ( e.g. , * ? ? ? * ; * ? ? ? the coadded spectra from different regions reveal a similar story . in the disk regions ( regions 2 & 3 ) , where there is little emission from the nlr gas , we find strong and narrow h@xmath4 , and weak , but rather broad , br@xmath3 and [ ] . in the nlr regions ( 4 & 5 ) , br@xmath3 and [ si vi ] are stronger and slightly broader than in the disk regions . the line profiles are most complex at positions 6 , 7 , and 8 , regions that are also bright in [ ] emission and where we see the strongest skewness in the lines ( fig . [ fig : oiiisivi ] & fig . [ fig:2dgas ] ) . we clearly see the red wing in the spectra from regions 6 & 8 , and the corresponding blue wing in the spectra from region 7 . again in line with recent work ( e.g. , * ? ? ? * ) , we find that the high - ionization lines and hydrogen recombination lines trace `` feedback '' . that is , the emission in these lines is dominated by extraplanar gas likely excited by the agn , and perhaps entrained by the nuclear jet . this nlr emission is also aligned roughly perpendicular to the megamaser disk on sub - pc scales . finally , we ask whether there is a broad component to the br@xmath3 , since @xcite detect a weak broad base to the h@xmath65 line in ngc 4388 , with a fwhm of 3900 km s@xmath35 . within the central nuclear point source emission , all the species have lines that are broad and symmetric ( table 1 ) . however , we find no evidence for a true broad - line component in br@xmath3 . instead , all the transitions have comparable widths . neither @xcite nor @xcite detect broad emission from br@xmath3 or br@xmath65 in their long - slit nir spectra either . perhaps the broad component detected in the optical is scattered broad emission rather than directly transmitted light ( e.g. , * ? ? ? ngc 1194 is an inclined ( @xmath133 ) s0 galaxy at @xmath134 mpc . the galaxy pa is 145 , and we aligned the ifu along the major axis . given our psf of @xmath135 , we are resolving 40 pc scales with our ao - assisted observations . thus we achieve very comparable spatial resolution to the ngc 4388 observations presented above . we have detected the large - scale disk in ngc 1194 in , and find a circular velocity of @xmath136 km s@xmath35 @xcite . from sdss imaging , we find a bulge - to - total ratio of @xmath137 @xcite . the bh in ngc 1194 has a mass of @xmath138 @xmath62 , one of the most massive in the megamaser disk sample . ngc 1194 is in the _ _ iras-__selected 12@xmath131 sample @xcite , and in the swift / bat 22 month sample @xcite . it has a hard x - ray luminosity of log @xmath139 from 14 - 195 kev , but is not compton thick @xcite . depending on the bolometric correction , the eddington ratio is @xmath140 @xcite , considerably lower than ngc 4388 . the galaxy was also part of the _ hst_/ [ ] snapshot survey of @xcite . they find extended [ ] emission with an extent of 700 pc along the major axis of the galaxy , and 470 pc along the minor axis . the nucleus was detected in first and at 8.4 ghz by @xcite , with a flux density of 0.9 mjy , and an upper limit on the radio core size of 52 pc . there are three existing @xmath0band spectra of ngc 1194 @xcite on @xmath90 ( 900 pc ) scales . in all cases the co bandhead is detected , but the line emission is very weak . davies et al . do not detect br@xmath3 , but marginally detect the h@xmath4@xmath141 line . due to the heavy binning required to achieve adequate s / n , we can not derive much spatially resolved information for this galaxy . in terms of the gas , the br@xmath3 and [ si vi ] flux distributions are very centrally concentrated , with a fwhm that is consistent with that of the continuum . the h@xmath4@xmath142 emission is clearly extended , but the distribution is quite patchy . we therefore construct average spectra in annuli , taking the central spectrum within the fwhm of the psf ( 016 ) . then we construct six spectra with outer radii of 05 , 1 , and 15 ( 140 , 280 , and 420 pc ) respectively , on either side of the galaxy minor axis . three examples of these spectra are shown in figure [ fig : lineregions ] . we can measure the radial velocities and stellar velocity dispersions reliably in these coadded spectra , using ppxf as above . we do find that the stars are rotating , with an amplitude of @xmath2 km s@xmath35(figure [ fig : n1194stelsigcurve ] ) . in contrast , we do not detect a clear rotation curve in the gas . instead , the velocities are very asymmetric . asymmetric velocities in high - ionization lines are often attributed to outflow ( e.g. , * ? ? ? * ; * ? ? ? * ) , and are sometimes seen in h@xmath4 as well ( e.g. , * ? ? ? * ) , which may be the explanation here . another possibility is that we are seeing spiral arms or some other non - axisymmetric gas distribution that is unrelated to the agn . our limited s / n prevents us from distinguishing between these cases . neither the [ si vi ] nor the br@xmath3 is spatially extended enough to measure reliable velocities beyond the central region . all of the gas emission lines fluxes are weak , and the intrinsic luminosities are also low . it is perhaps not surprising that this s0 galaxy is gas poor compared to ngc 4388 above , based simply on their respective morphology . on the other hand , the center of ngc 1194 is quite dusty , so there must be some associated gas , and indeed we detect atomic hydrogen in this galaxy on large scales @xcite . the strongest gas emission comes from the high - ionization [ si vi ] line , but as mentioned above it is only marginally spatially resolved . thus , we suspect that the emission from this transition and br@xmath3 emerge from the inner nlr . in contrast , the h@xmath4 is clearly spatially resolved . in analogy with ngc 4388 above , these different gas morphologies may be suggesting that the h@xmath4 is excited by both the agn and stellar processes . we also measure the spatially resolved ratio of h@xmath4/br@xmath3 , and find that the ratio is lowest in the galaxy center ( having agn - like values ) and then rises outward , perhaps indicating an increased contribution to the h@xmath4 from stellar processes . these line ratios may suggest that the asymmetric velocities we observe in the h@xmath4 are due to non - circular motions in the inner regions of the galaxy ( e.g. , spiral arms ) rather than outflow . we have analyzed integral - field @xmath0band observations of two megamaser disk galaxies . our data probe @xmath1 pc scales in both galaxies , and thus allow us to study the distributions of stars and gas at the centers of these galaxies to investigate agn fueling . in ngc 1194 , an s0 , there is very little gas , and with our s / n ratios , we can not say much about the two - dimensional velocity or dispersion fields . however , in ngc 4388 we have excellent s / n , and uncover a variety of interesting features : 1 . the stellar velocity field demonstrates well - ordered rotation aligned with the kpc - scale disk . although there is net rotation in the stars , still they are kinematically hot , with @xmath79 , likely dominated by the dispersion in the kpc - scale bulge . 2 . in the inner 1 ( 100 pc ) , we also see evidence for a nuclear disk , offet in pa by 15 from the kpc - scale disk . the evidence includes disky isophotes in the _ hst_/wfc3 f160w image and a distinct drop in stellar velocity dispersion on one side of the putative disk ( the other side apparently extincted ) . features like this , so called @xmath84drops , have been seen in the centers of a number of nearby spirals ( e.g. , * ? ? ? * ; * ? ? ? * ) , and there is some hint that they are more common among actively accreting galaxies @xcite . we note that this 100 pc - scale nuclear disk is misaligned by @xmath1135 from the megamaser disk on sub - pc scales . the h@xmath4 gas also shows regular kinematics on @xmath143 pc scales , with a well - defined rotation field that is aligned with the kpc - scale disk . with a higher rotation amplitude , the gas is considerably colder than the stars ( @xmath144 ) . we also see a kink in the rotation curve on @xmath2 pc scales . intriguingly , the h@xmath4 rotation within 100 pc appears to align with the megamaser disk . however , the physical origin of the kinematic twist is not yet clear . 4 . based on h@xmath4 diagnostic line ratios , we conclude that the molecular hydrogen is mainly excited by thermal processes . based on the morphology and kinematics of the h@xmath4 gas , we speculate that these thermal processes are mostly associated with stars rather than radiation from the agn ( e.g. , * ? ? ? * ; * ? ? ? * ) . 5 . in contrast , the br@xmath3 and high - ionization lines ( particularly [ si vi ] ) have a completely different spatial distribution and kinematics . they trace the inner narrow - line region , also seen on 500 pc scales in [ ] , as well as the 100 pc - scale jet . in the future , we hope to analyze a larger sample of megamaser disk galaxies in this manner . combined with direct tracers of the cold molecular gas ( e.g. , with alma ) , as well as high - resolution imaging from _ hst _ , we hope to build a multi - phase map of agn fueling from 100 to 0.1 pc @xcite . the referee provided a very thorough review that improved the quality of this manuscript . we thank k. gebhardt and r. van den bosch for assistance with the early stages of this project . j.l.w . has been supported by an nsf astronomy and astrophysics postdoctoral fellowship under award no . some of the data presented herein were obtained at the w.m . keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration . the observatory was made possible by the generous financial support of the w.m . keck foundation . the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community . we are most fortunate to have the opportunity to conduct observations from this mountain . keck telescope time was granted by noao , through the telescope system instrumentation program ( tsip ) . tsip is funded by nsf .

the standard model ( sm ) has been tested at an unprecedented level of precision by various experiments . however , some excesses have not yet been completely resolved . the first case is the muon anomalous magnetic moment ( muon @xmath3 ) , where the discrepancy between experimental data and the sm prediction currently is @xmath8 @xcite . the second case is the angular observable @xmath1 of @xmath2 @xcite , where @xmath9 deviation , resulting from an integrated luminosity of 3.0 fb@xmath10 at the lhcb @xcite , recently confirmed an earlier result with @xmath11 deviations @xcite . moreover , the same measurement with @xmath12 deviations was reported by belle @xcite . also the other relevant @xmath13 observables are defined in ref . various possible resolutions to this excess have been widely studied @xcite . the third excess is that a resonance with a mass of 750 gev in the diphoton invariant mass spectrum was indicated by the atlas @xcite and cms @xcite experiments . the width and spin of the diphoton resonance are still uncertain . if the resonance is a fundamental particle , its width prefers to be narrow ; however , if it is a composite state , then its width is wide . the production cross sections for the diphoton excess at @xmath14 tev in the atlas and cms are respectively given by : @xmath15 where @xmath16 stands for the diphoton resonance . the challenge of explaining this excess is to obtain the large production cross section shown in eq . ( [ eq : data_di ] ) . various proposals to explain this excess have been broadly studied @xcite . to resolve the excesses in a specific framework , we propose the extension of the sm by including leptoquarks ( lqs ) and one higgs singlet , where the leptoquarks are colored scalars that simultaneously couple to the leptons and quarks , and the higgs singlet is the 750 gev diphoton resonance candidate . therefore , the @xmath17 decays can arise from the tree - level lq - mediated feynman diagrams when the muon @xmath3 is induced from lq loops . although lepton - flavor violating ( lfv ) processes @xmath4 can put strong constraints on some yukawa couplings , it is found that the wilson coefficient @xmath18 from the lq contributions can satisfy the requirements , which result from the global analysis of @xmath2 , and the muon @xmath3 can fit the current data . in addition , the results are also consistent with the measurement of the rare flavor - changing process @xmath19 . it is known that the production cross section of eq . ( [ eq : data_di ] ) can not be achieved if only a 750 gev particle is added to the sm without introducing other new effects . therefore , we need not only some new colored particles to produce the resonance , but also some mechanism by which to enhance the branching ratio ( br ) for the resonance decay to diphoton . intriguingly , lqs meet the requirements for the resonance production and diphoton decay . some leptoquark explanations of the diphoton excess can be found in refs . @xcite the paper is organized as follows : we introduce the model and discuss the relevant couplings in sec . ii . in sec . iii , we study the phenomenon : the sm higgs diphoton decay , lfv processes , wilson coefficients of @xmath20 from lq contributions for @xmath21 decays , the implication of @xmath19 , and the production cross section for @xmath22 in the model . the conclusion is given in sec . in this section , we briefly introduce the model and relevant interactions with the lqs . to reconcile the measurements of @xmath23 and @xmath24 , we extend the sm by adding two different representations of lq , which are @xmath25 and @xmath26 under @xmath27 gauge symmetry . in addition , we also add one real higgs singlet @xmath28 as the candidate for 750 gev diphoton resonance . the gauge - invariant yukawa interactions of the sm fermions and lqs are written as : @xmath29 where the subscripts @xmath30 are the flavor indices ; @xmath31 and @xmath32 are the @xmath33 lepton and quark doublets , @xmath34 , and @xmath35 are the yukawa couplings . since we do not study the cp violating effects , hereafter , we take all new yukawa couplings as real numbers . we use the representations of the lqs as : @xmath36 where the superscripts are the electric charges of the particles . the interactions in eq . ( [ eq : int1 ] ) are then expressed as : @xmath37 + \tilde{k}_{ij } \left [ \bar \ell_{l i } \ , u_{rj } \phi^{-5/3 } - \bar \nu_{l i}\ , u_{r j } \phi^{-2/3 } \right ] \nonumber \\ & + y_{ij } \left [ \bar u^c_{li } \ , \nu_{lj } \delta^{-2/3 } - \frac{1}{\sqrt{2 } } \bar u^c_{l i } \ , \ell_{l j } \delta^{1/3 } - \frac{1}{\sqrt{2 } } \bar d^c_{l i } \ , \nu_{l j } \delta^{1/3 } - \bar d^c_{l i } \ , \ell_{l j } \delta^{4/3 } \right ] + h.c . \end{aligned}\ ] ] since the lqs are colored scalar bosons , they can couple to the sm higgs @xmath38 and higgs singlet @xmath28 via the scalar potential . in order to get the higgs and @xmath28 couplings to the lqs , we write the gauge - invariant scalar potential as : @xmath39 ^ 2 + \lambda'_\delta tr \left ( \left [ \delta_{1/3}^\dagger \delta_{1/3 } \right]^2 \right ) \nonumber \\ & + \lambda_{h \phi } ( h^\dagger h)\left(\phi_{7/6}^\dagger \phi_{7/6 } \right ) + \lambda_{h \delta } ( h^\dagger h ) tr\left ( \delta_{1/3}^\dagger \delta_{1/3 } \right ) + \lambda_{\phi \delta } \left(\phi_{7/6}^\dagger \phi_{7/6 } \right ) tr \left ( \delta_{1/3}^\dagger \delta_{1/3 } \right ) , \nonumber \\ & + \mu_{sh } s ( h^\dagger h ) + \mu_{s \phi } s \left(\phi_{7/6}^\dagger \phi_{7/6 } \right ) + \mu_{s \delta } s \ , tr \left ( \delta_{1/3}^\dagger \delta_{1/3 } \right ) \nonumber \\ & + \lambda_{sh } s^2 ( h^\dagger h ) + \lambda_{s \phi } s^2 \left(\phi_{7/6}^\dagger \phi_{7/6 } \right ) + \lambda_{s \delta } s^2 \ , tr \left ( \delta_{1/3}^\dagger \delta_{1/3 } \right ) \ , . \end{aligned}\ ] ] as usual , we adopt the representations of the higgs doublet @xmath38 and singlet @xmath28 as : @xmath40 where @xmath41 and @xmath42 are the goldstone bosons ; @xmath43 is the sm higgs field , and @xmath44 is the vacuum expectation value ( vev ) of @xmath38 ( @xmath28 ) . it is known that the vevs of scalar fields are dictated by the scalar potential . thus , by the minimal conditions @xmath45 and @xmath46 , the stable vevs are determined by : @xmath47 using the scalar potential of eq . ( [ eq : potential ] ) , the mass - square matrix of @xmath43 and @xmath48 is found by : ( cc , & s ) ( cc m^2 _ & _ sh v/ + _ sh v v_s + _ sh v/ + _ sh v v_s & m^2_s + _ sh v^2 /2 ) ( c + s ) [ eq : m2 ] with @xmath49 . the parameters @xmath50 and @xmath51 lead to the mixture of @xmath43 and @xmath48 . since two neutral scalar bosons are involved in the model , we only need one mixing angle to parametrize the mixing effect . we formulate the mass eigenstates to be ( c h + h_s ) = ( cc & - + & ) ( c + s ) , [ eq : mixing ] where @xmath52 is the sm - like higgs boson , @xmath53 is the new heavy higgs boson and the new resonance , and their masses are obtained as m^2_h(h_s ) = [ eq : hmass ] with @xmath54 and @xmath55 . the relationship of the @xmath56 and mixing angle can be expressed by @xmath57 . since it is known that the higgs diphoton decay measurement has been consistent with the sm prediction , the data from the higgs to diphoton decay can therefore give strict limits on the new physics effects . in order to suppress the contribution from the mixing of @xmath52 and @xmath53 , we set @xmath58 ; thus , the @xmath59 can be simplified as @xmath60 . with @xmath61 gev , @xmath62 gev , and @xmath63 gev , it can be seen that @xmath64 gev and @xmath65 . based on the scheme of a small @xmath66 , the contributions from the mixing of @xmath52 and @xmath53 can be ignored . based on the introduced new interactions , in this section , we study the implications of the higgs diphoton decay , @xmath24 , @xmath4 , the muon @xmath3 , @xmath67 , and the 750 gev diphoton excess . since each of these processes has its own unique characteristics , we discuss these phenomena one by one below . the higgs measurement is usually described by the signal strength parameter , which is defined as the ratio of observation to the sm prediction and is expressed as : ^f_i = _ i _ f , [ eq : muf ] where @xmath68 stands for the possible channels , and @xmath69 denotes the signal strength of production ( decay ) . although vector - boson fusion ( vbf ) can also produce the sm higgs , we only consider the gluon - gluon fusion ( ggf ) process because it is the most dominant . the diphoton higgs decay approached the precision measurement since the 125 gev higgs was observed . therefore , any significant deviation from the sm prediction ( i.e. , @xmath70 ) can imply the new physics effects . as stated earlier , the sm higgs can couple to the lqs via the scalar potential . from eq . ( [ eq : potential ] ) , it can be seen that after spontaneous symmetry breaking ( ssb ) , the quartic terms @xmath71 and @xmath72 can lead to trilinear couplings of higgs to lqs as : @xmath73 where @xmath74 and @xmath75 . with the couplings in eq . ( [ eq : hlq ] ) , the effective lagrangian for @xmath76 by lq - loop can be formulated as : @xmath77 where @xmath78 and the loop function is given by : @xmath79 with @xmath80 ^ 2 $ ] for @xmath81 . accordingly , the signal strength of the higgs production and decay to diphoton can be respectively obtained as : @xmath82 where @xmath83 , @xmath84 is the number of colors ; @xmath85 and @xmath86 , and the functions for vector - boson and fermion loops are given by @xmath87\ , , \nonumber \\ a_1(x ) & = 2 + 3 x + 3(2x -x^2)f(x)\,.\end{aligned}\ ] ] since the effects of the doublet and triplet lqs are similar , for simplicity , we set @xmath88 and @xmath89 . the @xmath90 as a function of @xmath91 is presented in fig . [ fig : muif](a ) , and that of @xmath92 is shown in fig . [ fig : muif](b ) , where the curves in plot ( a ) are @xmath93 tev , and those in plot ( b ) are @xmath94 tev . for comparison , we also show the results of atlas @xcite and cms @xcite with @xmath95 errors in the plots . from the plots , it can be clearly seen that with @xmath92 of @xmath96 gev , the lq contributions can significantly shift the @xmath90 away from the sm prediction and that the results are consistent with the current data . on the contrary , the @xmath90 approaches the sm result when @xmath92 is of the order of gev . as a function of ( a ) @xmath91 and ( b ) @xmath92 , where the curves in plots ( a ) and ( b ) denote @xmath97 tev and @xmath98 tev , respectively . , title="fig:",width=264 ] as a function of ( a ) @xmath91 and ( b ) @xmath92 , where the curves in plots ( a ) and ( b ) denote @xmath97 tev and @xmath98 tev , respectively . , title="fig:",width=264 ] in the following analysis , we study the rare lepton - flavor violating processes , e.g. @xmath4 and @xmath100 , muon @xmath3 @xmath101 , and the fcnc process @xmath24 . we first discuss the radiative lfv processes for @xmath4 . with the couplings in eq . ( [ eq : int2 ] ) , the lq - loop induced decay amplitude for @xmath4 can be written as : @xmath102 \ell_i f^{\mu \nu } \,,\label{eq : lllga } \end{aligned}\ ] ] where the coefficient @xmath103 is expressed as : @xmath104 \left ( \frac{5}{\delta(m_{t } , m_\phi ) } - \frac{2(1-x)}{\delta(m_\phi , m_{t } ) } \right ) \ , , \nonumber \\ \delta(m_1,m_2 ) & = x m_1 ^ 2 + ( y+z ) m_2 ^ 2 \ , , \nonumber \\ \int [ dx ] & = \int dx dy dz \delta(1-x - y - z)\ , , \label{eq : crji}\end{aligned}\ ] ] @xmath105 can be obtained from @xmath106 by exchanging @xmath107 and @xmath108 . in order to balance the chirality of the leptons , it is found that the contributions from @xmath109 , @xmath110 , @xmath111 , and @xmath112 are suppressed by the lepton masses . since the lq @xmath113 can couple to left - handed and right - handed up - type quarks , the chirality flip by the mass insertion in the propagator of the up - type quark can lead to freeing of the lepton masses in the feynman diagrams , which are associated with @xmath114 and @xmath115 . in addition , the top - quark is much heavier than the @xmath116- and @xmath117-quarks ; therefore , we only present the top - quark contribution in @xmath106 . straightforwardly , the br for @xmath4 can be expressed as : @xmath118 where @xmath119 for @xmath120 and the brs for @xmath121 in the sm have been applied . the current experimental upper limits are shown in table [ tab : cif ] . according to eq . ( [ eq : lllga ] ) , muon @xmath3 can be easily obtained by setting @xmath122 and found as : @xmath123 . current upper bounds on the brs for the decays @xmath4 @xcite . [ cols="^,^,^",options="header " , ] if the photon in @xmath4 is replaced by the higgs , similar feynman diagrams can contribute to @xmath124 . since the upper limit of @xmath125 is of @xmath126 and can give strong constraints on the parameters @xmath127 and @xmath128 , if we set @xmath129 and @xmath130 to be small , then it is apparent that @xmath131 and @xmath132 are much smaller than current upper limits . hence , we just study the decay @xmath133 . the one - loop induced effective couplings for @xmath134 are written as : @xmath135 where @xmath136 is expressed as @xcite : @xmath137 \nonumber \\ & + \frac { n_c \mu_{h \phi } } { ( 4\pi)^2 } \sum_{i=1\text{-}3 , q = u , d } \left [ m_\mu ( k^\dagger)_{2 i } k_{i 3 } g(m_{q_i},m_\phi ) + m_\tau \tilde k^\dagger_{2 i } \tilde k_{i 3 } \tilde g(m_{q_i},m_\phi ) \right ] \nonumber \\ & + \frac{n_c m_\tau \mu_{h \delta}}{(4 \pi)^2 } \sum_{i=1\text{-}3,q = u , d } ( y^\dagger)_{2 i } y_{i 3 } \tilde g(m_{q_i},m_\delta ) \ , , \label{eq : cl}\end{aligned}\ ] ] @xmath138 can be obtained from @xmath139 by exchanging @xmath107 and @xmath108 , and the loop functions are given by : @xmath140 \log \left[z+(1-z)r_t - x y r_h - i \epsilon \right ] \nonumber \\ & + \int_0 ^ 1 dx \log \left [ x + ( 1-x ) r_t - i \epsilon \right]\ , , \nonumber \\ b(r_t , r_h ) = & \int [ dx ] \frac{xy r_h - r_t } { z+(1-z)r_t - xy r_h } \ , , \nonumber \\ g(m_1,m_2 ) \approx & \int [ dx ] \frac { z}{- x z m_h^2 + x m_1 ^ 2 + ( y+z ) m_2 ^ 2 } \ , , \nonumber \\ \tilde g(m_1,m_2 ) \approx & \int [ dx ] \frac { y } { - x z m_h^2 + x m_1 ^ 2 + ( y+z ) m_2 ^ 2 } \,.\end{aligned}\ ] ] the @xmath141 in @xmath142 denotes an infinitesimal positive value . it can be seen that the terms associated with @xmath143 , @xmath144 , and @xmath145 in eq . ( [ eq : cl ] ) are proportional to the lepton masses . the situation is similar to the @xmath106 in the decays @xmath4 . although @xmath146 of tev ( @xmath147 can enhance these effects , due to the effects of being related to @xmath148 , their contributions are at least @xmath149 smaller than those from @xmath150 . accordingly , the br for @xmath151 is formulated as : @xmath152 where @xmath153 is the width of the higgs . due to @xmath154 being less than @xmath155 , we use @xmath156 mev in our numerical estimations . next , we discuss the decays for @xmath17 . in this study , we do not directly investigate the angular analysis of @xmath24 ; instead , we refer to the results , which were done by using the global analysis to get the best - fit value of @xmath157 for the new physics contributions @xcite . thus , we just derive the wilson coefficients of @xmath18 and @xmath158 from the lq contributions . with the yukawa couplings in eq . ( [ eq : int2 ] ) , the effective hamiltonian for @xmath17 mediated by @xmath159 and @xmath160 can be respectively found as : @xmath161 we can decompose the eq . ( [ eq : heff ] ) in terms of the effective operators @xmath162 and @xmath163 , defined as @xmath164 . the associated wilson coefficients of @xmath165 from the lqs then are found as : @xmath166 where @xmath167 is a scale factor from the sm effective hamiltonian . it is worth mentioning that the interaction @xmath168 can contribute to @xmath19 , where the brs , measured by lhcb and cms @xcite and predicted by the sm @xcite , are : @xmath169 to consider the constraint from @xmath19 , we adopt the expression for the br as @xcite : @xmath170 with @xmath95 errors , the allowed range for @xmath171 is obtained as @xmath172 . we use this result to constrain the free parameters . since the parameters in the decays @xmath4 , @xmath133 , @xmath101 , and @xmath24 are strongly correlated , in the following analysis , we take the current upper limits of @xmath173 shown in table [ tab : cif ] as the inputs and attempt to find the allowed parameter space , such that the excesses in @xmath101 and @xmath24 can be satisfied , and the @xmath174 can be as large as possible . from @xmath106 in eq . ( [ eq : crji ] ) , the dominant effects on the radiative lfv processes are from the @xmath113 and the top - quark loop ; thus , there is no possible cancellation in any of the decay amplitudes . with the upper bound of @xmath175 , we see that @xmath176 and @xmath177 have to be very small . in order to explain the excesses of muon @xmath3 and @xmath2 , we set @xmath178 . as a result , @xmath179 is negligible in this model . the related parameters for @xmath180 decays are @xmath181 and @xmath182 , respectively . these parameters simultaneously influence @xmath183 , muon @xmath3 , and @xmath184 ; therefore we have to analyze these processes together to get the allowed parameter space . since eqs . ( [ eq : crji ] ) , ( [ eq : cl ] ) , and ( [ eq : c9c10 ] ) involve many free parameters , in order to efficiently perform a numerical analysis , we set the ranges of relevant parameters as : @xmath185 \ ; \text{gev},\quad \mu_{lq } \in [ 1\ , , 100]\ ; \text{gev}\ , , \quad \ { k_{22 } , \tilde k_{22 } , y_{22 } \ } \in [ -0.1\ , , 0.1]\,,\nonumber \\ & \ { k_{33},\tilde k_{33 } , y_{33 } \ } \in [ -0.01,0.01]\ , , \quad \ { k_{23},\tilde k_{23 } , y_{23 } \ } \in [ -0.1\ , , 0.1]\ , , \nonumber \\ & k_{32 } \in \text{sign}(k_{22 } ) [ 0,0.5 ] \ , , \quad \tilde k_{32 } \in [ -0.5\ , , 0.5]\ , , \quad y_{32 } \in -\text{sign}(y_{22 } ) [ 0\ , , 0.5 ] \ , . \label{eq : range_scanning}\end{aligned}\ ] ] in order to avoid the constraints from @xmath186 ( @xmath187 ) and get @xmath188 , we set ( @xmath189 in eq . ( [ eq : range_scanning ] ) . additionally , the negative value of @xmath190 can be achieved when @xmath191 and @xmath192 are opposite in sign . as mentioned earlier , the yukawa couplings in both decays @xmath193 and @xmath194 are the same , we can not remove the constraints from the radiative lfv processes in this model . the brs for @xmath194 thus are of @xmath195 and much smaller than the current upper limits of @xmath196 @xcite . one way to escape the constraint from @xmath193 is to add a new lq @xcite . since we focus on the excesses of muon @xmath3 and @xmath2 , we leave the more complicated model for further study . with the chosen ranges of parameters in eq . ( [ eq : range_scanning ] ) , we first show the values for @xmath197 and @xmath171 in fig . [ fig : scatter](a ) , where the bounds from @xmath193 have been considered ; the horizontal band is from the measurement of @xmath5 ; the vertical band is the range that can explain the excess of @xmath2 , and we used @xmath198 parameter sets and obtained @xmath199 allowed points that satisfy the constraints . it can be seen that the @xmath197 and @xmath171 from the contributions of doublet @xmath200 and triplet @xmath201 lqs can simultaneously satisfy the constraint of @xmath5 and explain the excess in @xmath2 . from eq . ( [ eq : mug2 ] ) , it is known that muon @xmath3 is associated with the yukawa couplings @xmath202 . although only @xmath203 is related to @xmath197 and @xmath171 , since the yukawa couplings @xmath204 , @xmath205 , and @xmath206 are taken to be the same order of magnitude , we present the correlations of @xmath101 and @xmath190 in fig . [ fig : scatter](b ) , where only the allowed range of @xmath190 is shown . by plot ( b ) , it can be seen clearly that the excesses in @xmath101 and @xmath2 can be simultaneously fitted in the model . and @xmath171 using the ranges of parameters in eq . ( [ eq : range_scanning ] ) , where the bound from the @xmath207 and the allowed range of @xmath197 from global analysis for @xmath2 are shown . ( b ) correlation of @xmath101 and @xmath197 , where we only show the values of @xmath197 that can fit the excess in @xmath2 . [ fig : scatter],title="fig:",width=264 ] and @xmath171 using the ranges of parameters in eq . ( [ eq : range_scanning ] ) , where the bound from the @xmath207 and the allowed range of @xmath197 from global analysis for @xmath2 are shown . ( b ) correlation of @xmath101 and @xmath197 , where we only show the values of @xmath197 that can fit the excess in @xmath2 . [ fig : scatter],title="fig:",width=264 ] as discussed before , the 750 gev scalar boson candidate is a singlet , and it can mix with the sm higgs doublet via the scalar potential . since the mixing angle of @xmath52 and @xmath53 is small in our scheme , we use @xmath28 instead of the physical state @xmath53 and ignore the mixing effect @xmath208 . the relevant interactions for the @xmath28 production and decays are from the trilinear and quartic couplings , i.e. , @xmath209 , @xmath210 . it was discussed earlier that the vev of @xmath28 is smaller than 1 gev , so we thus focus on the trilinear couplings in the numerical analysis . similar to eq . ( [ eq : hgg ] ) , which is for @xmath76 effective coupling , using the trilinear interactions of @xmath28 and lqs in eq . ( [ eq : potential ] ) and the gauge couplings of gluons and lqs , the loop induced effective coupling for @xmath211 is found as : @xmath212 where @xmath213 and @xmath214 are the trilinear couplings with mass dimension - one . the partial decay widths for @xmath215 decays via the lq loops are expressed as : @xmath216 due to @xmath217 , it can be seen that @xmath218 is the dominant channel . although @xmath28 can decay to @xmath219 , their partial decay widths are the same as that of @xmath220 in terms of order of magnitude . additionally , @xmath221 can be also induced by lq loops , however the decay becomes unimportant when @xmath92 is small . hence , the br for @xmath222 can be estimated by : @xmath223 it is found that with @xmath224 and @xmath225 , the br for diphoton decay is insensitive to the parameters and becomes @xmath226 . to calculate the production cross section of @xmath227 using the lq loops , we implement the model to the calchep @xcite and use the cteq6l parton distribution function ( pdf ) @xcite thus , the @xmath228 at @xmath14 tev as a function of @xmath91 is presented in fig . [ fig : xsbr ] , where we use the k - factor @xmath229 for the purpose of including next - to - leading order qcd corrections @xcite , and the solid , dashed , and dotted lines are for @xmath230 tev , respectively . from the results , it can be seen that by taking proper values of @xmath231 , @xmath232 fb can be achieved with @xmath7 tev . at @xmath14 tev as a function of @xmath91 , where the @xmath229 for next - to - leading order qcd corrections is used , and the solid , dashed , and dotted lines are for @xmath233 tev , respectively.,width=264 ] in order to resolve the excesses of muon @xmath3 and @xmath2 decay , we investigate the extension of the sm by including leptoquarks , in which the particles are colored scalar bosons and can couple to quarks and leptons . in order to accommodate the measurement of @xmath5 and the excess of @xmath2 , we study a model with both doublet and triplet leptoquarks . moreover , we add a real singlet field to the model as the candidate of the 750 gev diphoton resonance , which is indicated by the atlas and cms experiments . after ssb , the couplings of the sm higgs the lqs are described by @xmath234 . if @xmath146 is of @xmath96 gev , the signal strength parameter @xmath235 can significantly deviate from the sm prediction and is still consistent with the current higgs measurements . in this study , lepton - flavor violating processes @xmath4 give strict constraints on the yukawa couplings @xmath236 and @xmath237 . as a result , the branching ratios for the lepton - flavor violating higgs decays @xmath238 are less than @xmath239 ) . nevertheless , the sizable couplings @xmath240 , @xmath241 , and @xmath242 can still explain the excess of muon @xmath3 and provide the necessary values for the wilson coefficient @xmath18 , such that the excess in @xmath2 can be resolved . the 750 gev scalar boson @xmath28 can couple to leptoquarks through the trilinear couplings @xmath243 and @xmath244 . in light of these couplings , the @xmath28 production and decays in the @xmath245 collisions can occur through the leptoquark loops . with the @xmath246 of tev , the production cross section for @xmath247 can reach 5 fb , which fits the current atlas and cms results . we remark on the production of the leptoquarks at the collider . the leptoquark pairs can be produced via qcd channels . since the masses of leptoquarks used for the purpose of explaining the excesses are @xmath6 tev , their production cross sections can be sizable . thus , our model can be tested by searching for specific signatures , such as @xmath248 , @xmath249 , @xmath250 , and @xmath251 that depend on the electric charges of leptoquarks and their yukawa couplings . the detailed simulation is beyond the scope of this paper and the relevant study will be presented elsewhere . this work was partially supported by the ministry of science and technology of taiwan r.o.c . , under grant most-103 - 2112-m-006 - 004-my3 ( chc ) . h. o. would like to thank the members of kias for their hospitality during his visit . a. abdesselam _ et al . _ [ belle collaboration ] , arxiv:1604.04042 [ hep - ex ] . j. matias , f. mescia , m. ramon and j. virto , jhep * 1204 * , 104 ( 2012 ) [ arxiv:1202.4266 [ hep - ph ] ] . s. descotes - genon , j. matias and j. virto , phys . d * 88 * , 074002 ( 2013 ) [ arxiv:1307.5683 [ hep - ph ] ] . r. gauld , f. goertz and u. haisch , jhep * 1401 * , 069 ( 2014 ) [ arxiv:1310.1082 [ hep - ph ] ] . a. datta , m. duraisamy and d. ghosh , phys . d * 89 * , no . 7 , 071501 ( 2014 ) [ arxiv:1310.1937 [ hep - ph ] ] . t. hurth and f. mahmoudi , jhep * 1404 * , 097 ( 2014 ) [ arxiv:1312.5267 [ hep - ph ] ] . s. descotes - genon , l. hofer , j. matias and j. virto , jhep * 1412 * , 125 ( 2014 ) [ arxiv:1407.8526 [ hep - ph ] ] . w. altmannshofer and d. m. straub , eur . j. c * 75 * , no . 8 , 382 ( 2015 ) [ arxiv:1411.3161 [ hep - ph ] ] . s. descotes - genon , l. hofer , j. matias and j. virto , jhep * 1606 * , 092 ( 2016 ) [ arxiv:1510.04239 [ hep - ph ] ] . b. gripaios , m. nardecchia and s. a. renner , jhep * 1505 * , 006 ( 2015 ) [ arxiv:1412.1791 [ hep - ph ] ] . a. crivellin , g. dambrosio and j. heeck , phys . lett . * 114 * , 151801 ( 2015 ) [ arxiv:1501.00993 [ hep - ph ] ] . s. sahoo and r. mohanta , phys . d * 91 * , no . 9 , 094019 ( 2015 ) [ arxiv:1501.05193 [ hep - ph ] ] . a. bharucha , d. m. straub and r. zwicky , arxiv:1503.05534 [ hep - ph ] . d. becirevic , s. fajfer and n. kosnik , phys . d * 92 * , no . 1 , 014016 ( 2015 ) [ arxiv:1503.09024 [ hep - ph ] ] . a. crivellin , l. hofer , j. matias , u. nierste , s. pokorski and j. rosiek , phys . d * 92 * , no . 5 , 054013 ( 2015 ) [ arxiv:1504.07928 [ hep - ph ] ] . c. j. lee and j. tandean , jhep * 1508 * , 123 ( 2015 ) [ arxiv:1505.04692 [ hep - ph ] ] . r. alonso , b. grinstein and j. martin camalich , jhep * 1510 * , 184 ( 2015 ) [ arxiv:1505.05164 [ hep - ph ] ] . s. sahoo and r. mohanta , phys . d * 93 * , no . 3 , 034018 ( 2016 ) [ arxiv:1507.02070 [ hep - ph ] ] . g. belanger , c. delaunay and s. westhoff , phys . rev . d * 92 * , 055021 ( 2015 ) [ arxiv:1507.06660 [ hep - ph ] ] . s. sahoo and r. mohanta , phys . d * 93 * , no . 11 , 114001 ( 2016 ) [ arxiv:1512.04657 [ hep - ph ] ] . c. w. chiang , x. g. he and g. valencia , phys . d * 93 * , no . 7 , 074003 ( 2016 ) [ arxiv:1601.07328 [ hep - ph ] ] . i. dorsner , s. fajfer , a. greljo , j. f. kamenik and n. kosnik , phys . rept . * 641 * , 1 ( 2016 ) [ arxiv:1603.04993 [ hep - ph ] ] . s. m. boucenna , a. celis , j. fuentes - martin , a. vicente and j. virto , phys . b * 760 * , 214 ( 2016 ) [ arxiv:1604.03088 [ hep - ph ] ] . k. harigaya and y. nomura , phys . b * 754 * , 151 ( 2016 ) [ arxiv:1512.04850 [ hep - ph ] ] . y. mambrini , g. arcadi and a. djouadi , phys . b * 755 * , 426 ( 2016 ) [ arxiv:1512.04913 [ hep - ph ] ] . m. backovic , a. mariotti and d. redigolo , jhep * 1603 * , 157 ( 2016 ) [ arxiv:1512.04917 [ hep - ph ] ] . a. angelescu , a. djouadi and g. moreau , phys . b * 756 * , 126 ( 2016 ) [ arxiv:1512.04921 [ hep - ph ] ] . y. nakai , r. sato and k. tobioka , phys . * 116 * , no . 15 , 151802 ( 2016 ) [ arxiv:1512.04924 [ hep - ph ] ] . d. buttazzo , a. greljo and d. marzocca , eur . j. c * 76 * , no . 3 , 116 ( 2016 ) [ arxiv:1512.04929 [ hep - ph ] ] . s. di chiara , l. marzola and m. raidal , phys . d * 93 * , no . 9 , 095018 ( 2016 ) [ arxiv:1512.04939 [ hep - ph ] ] . s. knapen , t. melia , m. papucci and k. zurek , phys . d * 93 * , no . 7 , 075020 ( 2016 ) [ arxiv:1512.04928 [ hep - ph ] ] . a. pilaftsis , phys . d * 93 * , no . 1 , 015017 ( 2016 ) [ arxiv:1512.04931 [ hep - ph ] ] . r. franceschini _ et al . _ , jhep * 1603 * , 144 ( 2016 ) [ arxiv:1512.04933 [ hep - ph ] ] . j. ellis , s. a. r. ellis , j. quevillon , v. sanz and t. you , jhep * 1603 * , 176 ( 2016 ) [ arxiv:1512.05327 [ hep - ph ] ] . r. s. gupta , s. jager , y. kats , g. perez and e. stamou , arxiv:1512.05332 [ hep - ph ] . a. kobakhidze , f. wang , l. wu , j. m. yang and m. zhang , phys . b * 757 * ( 2016 ) 92 [ arxiv:1512.05585 [ hep - ph ] ] . a. falkowski , o. slone and t. volansky , jhep * 1602 * , 152 ( 2016 ) [ arxiv:1512.05777 [ hep - ph ] ] . r. benbrik , c. h. chen and t. nomura , phys . rev . d * 93 * , no . 5 , 055034 ( 2016 ) [ arxiv:1512.06028 [ hep - ph ] ] . f. wang , l. wu , j. m. yang and m. zhang , phys . b * 759 * ( 2016 ) 191 [ arxiv:1512.06715 [ hep - ph ] ] . p. s. b. dev and d. teresi , arxiv:1512.07243 [ hep - ph ] . b. c. allanach , p. s. b. dev , s. a. renner and k. sakurai , phys . d * 93 * , no . 11 , 115022 ( 2016 ) [ arxiv:1512.07645 [ hep - ph ] ] . k. cheung , p. ko , j. s. lee , j. park and p. y. tseng , arxiv:1512.07853 [ hep - ph ] . f. wang , w. wang , l. wu , j. m. yang and m. zhang , arxiv:1512.08434 [ hep - ph ] . c. w. chiang , m. ibe and t. t. yanagida , jhep * 1605 * , 084 ( 2016 ) [ arxiv:1512.08895 [ hep - ph ] ] . x. j. huang , w. h. zhang and y. f. zhou , phys . d * 93 * , 115006 ( 2016 ) [ arxiv:1512.08992 [ hep - ph ] ] . s. kanemura , k. nishiwaki , h. okada , y. orikasa , s. c. park and r. watanabe , arxiv:1512.09048 [ hep - ph ] . t. nomura and h. okada , phys . b * 755 * , 306 ( 2016 ) [ arxiv:1601.00386 [ hep - ph ] ] . ko , y. omura and c. yu , jhep * 1604 * , 098 ( 2016 ) [ arxiv:1601.00586 [ hep - ph ] ] . ko and t. nomura , phys . b * 758 * , 205 ( 2016 ) [ arxiv:1601.02490 [ hep - ph ] ] . i. dorsner , s. fajfer and n. kosnik , phys . d * 94 * , no . 1 , 015009 ( 2016 ) [ arxiv:1601.03267 [ hep - ph ] ] . t. nomura and h. okada , arxiv:1601.04516 [ hep - ph ] . x. f. han , l. wang and j. m. yang , phys . b * 757 * , 537 ( 2016 ) [ arxiv:1601.04954 [ hep - ph ] ] . g. belanger and c. delaunay , arxiv:1603.03333 [ hep - ph ] . m. bauer and m. neubert , phys . d * 93 * , no . 11 , 115030 ( 2016 ) [ arxiv:1512.06828 [ hep - ph ] ] . c. w. murphy , phys . lett . b * 757 * , 192 ( 2016 ) [ arxiv:1512.06976 [ hep - ph ] ] . w. chao , arxiv:1512.08484 [ hep - ph ] . c. hati , phys . d * 93 * , no . 7 , 075002 ( 2016 ) [ arxiv:1601.02457 [ hep - ph ] ] . f. f. deppisch , s. kulkarni , h. pas and e. schumacher , phys . d * 94 * , no . 1 , 013003 ( 2016 ) [ arxiv:1603.07672 [ hep - ph ] ] . u. k. dey , s. mohanty and g. tomar , arxiv:1606.07903 [ hep - ph ] . a. di iura , j. herrero - garcia and d. meloni , arxiv:1606.08785 [ hep - ph ] . s. baek and k. nishiwaki , phys . d * 93 * , no . 1 , 015002 ( 2016 ) [ arxiv:1509.07410 [ hep - ph ] ] . s. baek , t. nomura and h. okada , arxiv:1604.03738 [ hep - ph ] . v. khachatryan _ et al . _ [ cms and lhcb collaborations ] , nature * 522 * , 68 ( 2015 ) [ arxiv:1411.4413 [ hep - ex ] ] . a. belyaev , n. d. christensen and a. pukhov , comput . commun . * 184 * , 1729 ( 2013 ) [ arxiv:1207.6082 [ hep - ph ] ] . m. nadolsky , h. l. lai , q. h. cao , j. huston , j. pumplin , d. stump , w. k. tung and c .- p . yuan , phys . d * 78 * , 013004 ( 2008 ) [ arxiv:0802.0007 [ hep - ph ] ] . a. djouadi , phys . * 457 * , 1 ( 2008 ) [ hep - ph/0503172 ] .

for ultracold dilute non - relativistic bosons in three dimensions , bogoliubov theory gives a successful description of most quantities of interest @xcite . this approximation breaks down , however , near the critical temperature for the phase transition , as well as for the low temperature phase in lower dimensional systems , due to the importance of fluctuations . one would therefore like to have a systematic extension beyond the bogoliubov theory , which includes the fluctuation effects beyond the lowest order in a perturbative expansion in the scattering length . such extensions have encountered obstacles in the form of infrared divergences in various expansions @xcite . only recently , a satisfactory framework has been found to cure these problems @xcite . functional renormalization @xcite for the average action @xcite systematically copes with the infrared problems by exploring the difficult long range behavior gradually by means of non - perturbative flow equations , which are based on an exact renormalization group equation @xcite . for zero temperature they lead to a consistent description of bose - einstein condensation and the quantum phase transition for nonrelativistic bosons in arbitrary dimension @xcite . a key ingredient is the generation of a term in the inverse propagator that is quadratic in the frequency @xmath4 . it characterizes the long time behavior of the full propagator at long distances , even if no quadratic frequency dependence is present in the microphysical or classical propagator . the coupling @xmath5 is due to quantum fluctuations and appears during the flow from microphysics to macrophysics . the extreme infrared limit is governed by a `` relativistic '' model where @xmath5 dominates , with an enhanced ( approximate ) space - time symmetry @xmath6 corresponding to lorentz symmetry . in one or two dimensions the relativistic term @xmath7 in the inverse propagator is crucial for a correct description of the low temperature behavior . in contrast , in three dimensions the flow towards the extreme infrared behavior is only logarithmic , such that the quantitative influence remains quite moderate for practical purposes . nevertheless , the effective coupling @xmath5 is needed in order to avoid the infrared problems . in this paper , we extend this formalism to a nonvanishing temperature . we present a quantitative rather accurate picture of bose - einstein condensation in three dimensions and find that the bogoliubov approximation is indeed valid for most quantities . the same method can be applied for one or two dimensions , such that the present work can also serve as a test of the method . this is nontrivial , because insufficient truncations of the flow equations can lead to fake dependencies on the microscopic physics . for dilute non - relativistic bosons in three dimensions we find an upper bound on the scattering length @xmath0 . this is similar to the `` triviality bound '' for the higgs scalar in the standard model of elementary particle physics . as a consequence , the scattering length is at most of the order of the inverse effective ultraviolet cutoff @xmath8 , which indicates the breakdown of the pointlike approximation for the interaction at short distances . typically , @xmath8 is of the order of the range of the van der waals interaction . for dilute gases , where the interparticle distance @xmath9 is much larger than @xmath8 , we therefore always find a small concentration @xmath10 . this provides for a small dimensionless parameter , and perturbation theory in @xmath11 becomes rather accurate for most quantities . for typical experiments with ultracold bosonic alkali atoms one has @xmath12 , @xmath13 , such that @xmath14 is really quite small . bosons with pointlike interactions can also be employed for an effective description of many quantum phase transitions at zero temperature , or phase transitions at low temperature @xmath15 . in this case , they correspond to quasi - particles , and their dispersion relation may differ from the one of non - relativistic bosons , @xmath16 . we describe the quantum phase transitions for a general microscopic dispersion relation , where the inverse classical propagator in momentum and frequency space takes the form @xmath17 ( in units where the particle mass @xmath18 is set to @xmath19 ) . we present the quantum phase diagram at @xmath20 in dependence on the scattering length @xmath0 and a dimensionless parameter @xmath21 , which measures the relative strength of the term quadratic in @xmath22 in @xmath23 . in the limit @xmath24 ( @xmath25 ) our model describes relativistic bosons . a nonzero temperature @xmath15 introduces a new scale into the problem . this modifies the effective scaling behavior . in particular , near the critical temperature @xmath2 , where the bose - einstein condensate dissolves , the long distance physics can be described by classical statistics . the non - relativistic bosons belong to the universality class of the three - dimensional @xmath26 model , with the associated universal critical exponents . the flow equations in the classical regime are well studied and lead , for an appropriate level of the truncation , to very accurate results for the critical behavior @xcite . here we use a rather simple truncation , but we obtain nevertheless a reasonable description of the non - analytical critical behavior . however , the temperature range for the applicability of the universal critical behavior is found to be very small for the small values of @xmath11 encountered for the dilute gases . the value of the critical temperature for interacting bosons can not be computed within the bogoliubov theory . we find a temperature shift @xmath3 , compared to the free theory . in contrast to other estimates , which only take the classical fluctuations into account ( only the zero matsubara frequency ) , we include the full quantum statistics ( all matsubara frequencies ) . we finally compute the sound velocity at @xmath20 . this quantity agrees to high precision with the bogoliubov result . for a known scattering length @xmath0 , it can therefore be used as a precise measure of the density . our paper is organized as follows . as a starting point , we specify our microscopic model in sec . [ sectmicroscopicmodel ] . in the following sec . [ sectnonperturbativeflowequations ] , we recall the method of functional renormalization , explain our truncation of the effective average action and also describe the projection of the exact flow equation onto that truncation . the flow equations in the vacuum are investigated in sec . [ sectupperboundforthescatteringlength ] and an upper bound for the scattering length is derived . in sec . [ sectquantumphasediagram ] we explain our method to determine the density and also study the quantum phase diagram at zero temperature . the effects of a non - vanishing temperature on the condensate are taken into account in sec . [ secttemperaturedependenceofcondensate ] , while sec . [ sectcriticaltemperature ] gives our result for the critical temperature in dependence of the interaction strength . finally , we investigate the sound velocity and draw conclusions in secs . [ sectsoundvelocity ] and [ sectconclusions ] . the appendices [ secttruncation ] and [ sectsymmetriesandnoethercurrents ] contain a motivation of our truncation in terms of a systematic derivative expansion and an analysis of symmetries constraining the form of the effective action . our explicit results for the flow equations of the effective potential and the kinetic coefficients are shown in appendices [ secfloweffectivepotential ] and [ sectflowofkineticcoefficients ] . appendix [ sectpropdisp ] analyzes the dispersion relation . our microscopic action describes nonrelativistic bosons , with an effective interaction between two particles given by a contact potential . it is assumed to be valid on length scales where the microscopic details of the interaction are irrelevant and the scattering length is sufficient to characterize the interaction . the microscopic action reads @xmath27=\int_x \,{\big \{}\phi^*\,(s\partial_\tau - v\partial_\tau^2-\delta-\sigma)\,\phi\,+\,\frac{1}{2}\lambda(\phi^*\phi)^2{\big \ } } , \label{microscopicaction}\ ] ] with @xmath28 the integration goes over the whole space as well as over the imaginary time @xmath29 , which at finite temperature is integrated on a circle of circumference @xmath30 according to the matsubara formalism . we use natural units @xmath31 . we also scale time and energy units with appropriate powers of @xmath32 , with @xmath18 the particle mass . in other words , our time units are set such that effectively @xmath33 . in particular , we use the rescaled chemical potential @xmath34 and @xmath15 stands for the temperature multiplied by @xmath32 . in these units time has the dimension of length squared . for standard non - relativistic bosons one has @xmath35 and @xmath36 , but we also consider quasiparticles with a more general dispersion relation described by nonzero @xmath37 . after fourier transformation , the kinetic term reads @xmath38 with @xmath39 at nonzero temperature , the frequency @xmath40 is discrete , with @xmath41 while at zero temperature this becomes @xmath42 the dispersion relation encoded in eq . obtains by analytic continuation @xmath43 in this paper , we consider homogeneous situations , i.e. an infinitely large volume without a trapping potential . many of our results can be translated to the inhomogeneous case in the framework of the local density approximation . one assumes that the length scale relevant for the quantum and statistical fluctuations is much smaller than the characteristic length scale of the trap . in this case , our results can be transferred by taking the chemical potential position dependent in the form @xmath44 , where @xmath45 is the trapping potential . the microscopic action is invariant under the global @xmath46 symmetry which is associated to the conserved particle number , @xmath47 on the classical level , this symmetry is broken spontaneously when the chemical potential @xmath48 is positive . in this case , the minimum of @xmath49 is situated at @xmath50 . the ground state of the system is then characterized by a macroscopic field @xmath51 , with @xmath52 . it singles out a direction in the complex plane and thus breaks the @xmath46 symmetry . nevertheless , the action itself and all modifications due to quantum and statistical fluctuations respect the symmetry . for @xmath35 and @xmath36 , the situation is similar for galilean invariance . at zero temperature , we can perform an analytic continuation to real time and the microscopic action is then invariant under transformations that correspond to a change of the reference frame in the sense of a galilean boost . it is easy to see that in the phase with spontaneous @xmath46 symmetry breaking also the galilean symmetry is broken spontaneously : a condensate wave function , that is homogeneous in space and time , would be represented in momentum space by @xmath53 under a galilean boost transformation with a boost velocity @xmath54 , this would transform according to @xmath55 this shows that the ground state is not invariant under such a change of reference frame . this situation is in contrast to the case of a relativistic bose - einstein condensate , like the higgs boson field after electroweak symmetry breaking . a relativistic scalar transforms under lorentz boost transformations according to @xmath56 such that a condensate wave function @xmath57 transforms into itself . we will investigate the implications of galilean symmetry for the form of the effective action in app . [ secttruncation ] . an analysis of general coordinate invariance in nonrelativistic field theory can be found in @xcite . we start with a functional integral representation of the grand canonical partition function @xmath58}. \label{eqpartfunction}\ ] ] in this paper , we work with the formalism of quantum statistics for many particle problems . in contrast to classical statistics , the fields @xmath59 are parametrized by an euclidean time variable @xmath29 in addition to the space variable @xmath60 . this euclidean time is wrapped up on a circle of circumference @xmath61 , such that the fields @xmath59 live on a ( generalized ) torus . the microscopic action @xmath62 $ ] is of a form similar to . we generalize eq . by introducing a source @xmath63 for the fields @xmath64 and write @xmath65=e^{w[j]}=\int d \chi \,e^{-s[\chi]+\int_x j\chi}. \label{eqpartfunctwithj}\ ] ] the @xmath1-point correlation functions can now be obtained by functional differentiation of @xmath66 $ ] , while @xmath67 $ ] generates the connected @xmath1-point functions . for example , the expectation value of @xmath64 obtains from @xmath68}{\delta j(x)}.\ ] ] the thermodynamic potential @xmath69 associated with the grand canonical partition function is given by @xmath70.\ ] ] the effective action @xmath71 $ ] is defined as @xmath72=\left(-w[j]+\int_x j \phi\right)_{j = j_{\text{ex}}[\phi]}.\ ] ] where @xmath73 is obtained by the inversion of @xmath74}{\delta j}\bigg{|}_{j = j_\text{ex}}=\phi.\ ] ] it is straightforward to show @xmath75}{\delta \phi}=j$ ] . in the absence of a source @xmath63 , we obtain for the thermodynamic potential of the grand canonical partition function @xmath76\ ] ] where @xmath77 is defined by @xmath78}{\delta \phi}\bigg{|}_{\phi=\phi_{\text{eq}}}=0.\ ] ] the effective action @xmath71 $ ] is the generating functional of the one - point irreducible correlation functions and in a sense , its precise knowledge corresponds to the solution of the theory . our method determines @xmath71 $ ] with the help of an exact flow equation . for that purpose , we include an infrared cutoff term @xmath79 $ ] in eq . and define @xmath80}=\int d\chi\ , e^{-s[\chi]-\delta s_k[\chi]+\int j \chi}.\ ] ] in fourier space , the cutoff term reads @xmath81=\int_q r_k(q ) \chi^*(q)\chi(q)\ ] ] and has the properties @xmath82 the effective average action is defined as a modified legendre transform of @xmath83 @xmath84=\left ( -w_k[j]+\int_x j\phi \right)_{j = j_\text{ex}}-\delta s_k[\phi].\ ] ] it has the important property , that it interpolates between the microscopic action @xmath85 $ ] and the full effective action @xmath71 $ ] @xmath86 & \rightarrow s[\phi ] & \quad ( k\rightarrow \infty),\\ \gamma_{k}[\phi ] & \rightarrow \gamma[\phi ] & \quad ( k\rightarrow 0).\end{aligned}\ ] ] in physical terms @xmath87 $ ] is the effective action in the presence of an infrared cutoff at a momentum scale @xmath88 . only fluctuations with momenta larger than @xmath88 are included . for example , a finite volume @xmath89 of the system under consideration would lead to an situation that is described by @xmath87 $ ] . our method to determine @xmath87 $ ] ( and for @xmath90 also @xmath71 $ ] ) relies on the existence of an exact flow equation @xcite @xmath91 here the trace operation includes a momentum integration @xmath92 , as well as a sum over internal indices @xmath93 , according to the two real components in the decomposition @xmath94 . on the r.h.s . of , @xmath95 is the second functional derivative of @xmath87 $ ] @xmath96)_{ij}(q , p)=\frac{\overset{\rightharpoonup}{\delta}}{\delta\phi_i(-q)}\gamma_k[\phi]\frac{\overset{\leftharpoonup}{\delta}}{\delta \phi_j(p)}.\ ] ] it is therefore a matrix in internal and momentum space . correspondingly , @xmath97 in eq . stands for @xmath98 the flow equation describes the evolution of the effective average action with the cutoff scale @xmath88 . the functional differential equation is hard to be solved exactly . in principle , @xmath99 is described by infinitely many couplings . we will use here an approximation with only a finite number of couplings . this is achieved by an ansatz for a specific form of @xmath87 $ ] . in the absence of anomalies the effective action @xmath71 $ ] is invariant under the same symmetries as the microscopic action @xmath85 $ ] . this holds also for the effective average action @xmath87 $ ] , provided that the cutoff term @xmath100 $ ] is also invariant . our ansatz will respect all symmetries of the classical action . approximate solutions of the exact flow equations obtain from a truncation of the general form of the effective action . we use here terms with up to two derivatives and truncate @xmath101 with @xmath102 . this particular form is motivated by a more systematic derivative expansion and an analysis of symmetry constraints ( ward identities ) in appendix [ secttruncation ] . we introduce the renormalized fields @xmath103 , @xmath104 , the renormalized kinetic coefficients @xmath105 , @xmath106 and we express the effective potential in terms of the renormalized invariant @xmath107 , with @xmath108 this yields @xmath109 for the effective potential , we use an expansion around the @xmath88-dependent minimum @xmath110 of the effective potential and the @xmath88-independent value of the chemical potential @xmath111 that corresponds to the physical particle number density @xmath1 . we determine @xmath110 and @xmath111 by the requirements @xmath112 more explicitly we take a truncation for @xmath113 of the form @xmath114 in the symmetric phase we have @xmath115 , while in the phase with spontaneous symmetry breaking , we have @xmath116 . in summary , the flow of @xmath99 for fixed @xmath117 is described by four running renormalized couplings @xmath118 , @xmath119 , @xmath120 and @xmath37 . in addition , we need the anomalous dimension @xmath121 . a computation of @xmath1 requires a flow equation of @xmath122 , which involves the couplings linear in @xmath123 , namely @xmath124 and @xmath125 . the pressure is calculated by following the @xmath88-dependence of the height of the minimum @xmath126 . all couplings @xmath118 , @xmath119 , @xmath120 , @xmath37 , @xmath127 , @xmath122 , @xmath128 , @xmath124 , @xmath125 depend on @xmath88 and @xmath15 . the physical renormalized couplings obtain for @xmath90 . they specify the thermodynamic potential @xmath129 as well as suitable derivatives of the potential and the correlation function . the `` initial values '' at the scale @xmath130 are determined by the requirement @xmath131=s[\phi],\ ] ] using the microscopic action @xmath85 $ ] in eq . . this implies the initial values @xmath132 we remain with the free microscopic couplings @xmath133 , @xmath134 , @xmath135 . the coupling @xmath133 will be replaced by the scattering length @xmath0 in the next section . we further choose units for @xmath29 where @xmath136 then our second free coupling is @xmath137 in consequence , besides the thermodynamic variables @xmath15 and @xmath111 our model is characterized by two free parameters , @xmath0 and @xmath5 . often , we will concentrate on `` standard '' non - relativistic bosons with a linear @xmath29 derivative , such that @xmath138 . the scattering length @xmath0 remains then the only free parameter . in the vacuum , where @xmath139 , this sets the relevant unit of length . finally , we have for the infrared cutoff @xmath140 we choose the optimized cutoff function @xcite @xmath141 where we recall , that @xmath116 in the regime with spontaneous symmetry breaking . it is convenient to work with real fields @xmath142 , @xmath94 , with fourier components @xmath143 the inverse propagator for the fields @xmath144 becomes a @xmath145 matrix in the space of @xmath146 and @xmath147 , given by the second functional derivative of @xmath99 . for a real constant background field @xmath148 , @xmath149 the latter becomes diagonal in momentum space @xmath150 for our truncation one has at @xmath117 @xmath151 here , primes denote derivatives with respect to @xmath107 ( not @xmath152 ) . in the phase with spontaneous symmetry breaking , the infrared cutoff in the flow equation adds to the diagonal term in a piece @xmath153 . this effectively replaces @xmath154 in eq . whenever @xmath155 , thus providing for an efficient infrared regularization . we project the flow equation of the effective average action onto equations for the coupling constants by using appropriate background fields and taking functional derivatives . the flow equation for the effective potential obtains by using a space- and time - independent background field in eq . , with @xmath156 @xmath157 the propagator @xmath158 is here determined from @xmath159 with @xmath160 again primes denote a differentiation with respect to @xmath107 . we switch to renormalized fields by making a change of variables in the differential equation @xmath161 we can now derive the flow equations for the couplings @xmath110 and @xmath162 by appropriate differentiation of with respect to @xmath107 . the flow equation for @xmath163 is given more explicitly in appendix [ secfloweffectivepotential ] . differentiation with respect to @xmath48 yields the flow of @xmath122 , @xmath124 , @xmath125 . we use in detail @xmath164 where we recall , that @xmath165 in our truncation . to determine the flow equation of @xmath118 , we use the condition that @xmath166 for all @xmath88 , and therefore @xmath167 we show the flow of @xmath119 and @xmath118 in figs . [ figflowoflambda ] , [ figflowofrho ] for @xmath168 , @xmath20 and different values of @xmath133 ( with @xmath138 ) . the change in @xmath118 is rather modest . this will be different for nonzero temperature . with the scale parameter @xmath88 for different start values , corresponding to @xmath169 ( dotted ) , @xmath170 ( solid ) and @xmath171 ( dashed ) . the first case is plotted for zero density ( @xmath172 ) only , while the last two cases are plotted also for unit density ( @xmath168 ) . the curves @xmath172 and @xmath168 are identical within the plot resolution . the solid and the dashed curve correspond to @xmath173 and @xmath174 , respectively . ] . the parameters for the solid and the dashed curves are the same as in figure [ figflowoflambda ] . ] the flow of @xmath122 is given by @xmath175 and similar for the flow of @xmath124 and @xmath125 , @xmath176 where the last equation holds since @xmath177 in our truncation . for a derivation of @xmath178 and the flow equations for @xmath120 and @xmath37 , we have to evaluate the flow equation for a background field depending on @xmath179 and @xmath180 . we use an analytic continuation @xmath181 and obtain the flow equation for @xmath120 from @xmath182 with four - volume @xmath183 . the projection prescription for @xmath37 is @xmath184 and similar for @xmath127 @xmath185 after the functional differentiation , we evaluate the expressions , , at homogeneous background fields . these calculations are a little intricate , but standard and straightforward in principle . more explicit flow equations are given in app . [ sectflowofkineticcoefficients ] . eventually , it is always possible to perform the matsubara sums and also the spatial momentum integration analytically . in figure [ figflowkinetic ] we show the flow of @xmath127 , @xmath120 and @xmath37 at zero temperature and for density @xmath168 . ( solid ) , @xmath120 ( dashed ) and @xmath37 for a scattering length @xmath173 , temperature @xmath20 and density @xmath168 . ] the kinetic coefficient @xmath127 starts on the large scale with @xmath186 , increases a little around @xmath187 and saturates to a constant . in contrast , the coefficient @xmath120 starts to decrease after a short period of increase with @xmath127 . for very tiny scales @xmath88 , @xmath120 would finally go to zero . the frequency dependence of the propagator is then governed by the quadratic frequency coefficient @xmath37 . in three spatial dimensions , however , this decrease of @xmath120 is so slow that it is not relevant on the length scales of experiments . this is one of the reasons why bogoliubov theory , which neglects the appearance of @xmath37 , describes experiments with ultracold bosonic quantum gases in three dimensions with so much success . the coefficient @xmath37 is always generated in the phase with spontaneous symmetry breaking @xcite . its @xmath88-dependence is also shown in figure [ figflowkinetic ] . the vacuum is defined to have zero temperature @xmath20 and vanishing density @xmath172 , which also implies @xmath115 . the interaction strength @xmath119 at the scale @xmath188 determines the four point vertex at zero momentum . it is directly related to the scattering length @xmath0 for the scattering of two particles in vacuum , which is experimentally observable . we therefore want to replace the microscopic coupling @xmath133 by the renormalized coupling @xmath0 . in our units ( @xmath33 ) , one has the relation @xmath189 the vacuum properties can be computed by taking for @xmath20 the limit @xmath190 . we may also perform an equivalent and technically simple computation in the symmetric phase by choosing @xmath191 such , that @xmath192 . this guarantees that the boson field @xmath193 is a gap - less propagating degree of freedom . we first investigate the model with a linear @xmath29-derivative , @xmath194 , @xmath195 . projecting the flow equation , we find the following equations : @xmath196 the propagator is not renormalized , @xmath197 , @xmath198 , @xmath199 , and one finds @xmath200 . the coupling @xmath125 is running according to @xmath201 since @xmath125 appears only in its own flow equation , it is of no further relevance in the vacuum . also , no coupling @xmath37 is generated by the flow and we have therefore set @xmath35 on the r.h.s . of eqs . and . inserting in eq . the vacuum values @xmath116 and @xmath36 , we find @xmath202 the solution @xmath203 tends to a constant for @xmath204 , @xmath205 . the dimensionless variable @xmath206 goes to zero , when @xmath88 goes to zero . this shows the infrared freedom of the theory . for fixed ultraviolet cutoff , the scattering length @xmath207 as a function of the initial value @xmath133 , has an asymptotic maximum @xmath208 the relation between @xmath0 and @xmath133 is shown in fig . [ figscatteringbound ] . in dependence on the microscopic interaction strength @xmath133 ( solid ) . the asymptotic maximum @xmath209 is also shown ( dashed ) . ] as a consequence of eq . , the nonrelativistic bosons in @xmath210 are a `` trivial theory '' in the sense that the bosons become noninteracting in the limit @xmath211 , where @xmath212 . the upper bound has important practical consequences . it tells us , that whenever the `` macrophysical length scales '' are substantially larger than the microscopic length @xmath8 , we deal with a weakly interacting theory . as an example , consider a boson gas with a typical inter - particle distance substantially larger than @xmath8 . ( for atom gases @xmath8may be associated with the range of the van der waals force . ) we may set the units in terms of the particle density @xmath1 , @xmath168 . in these units @xmath213 is large , say @xmath214 . this implies a very weak interaction , @xmath215 . in other words , the scattering length can not be much larger than the microscopic length @xmath8 . for such systems , perturbation theory will be valid in many circumstances . we will find that the bogoliubov theory indeed gives a reliable account of many properties . even for an arbitrary large microphysical coupling @xmath216 , the renormalized physical scattering length @xmath0 remains finite . let us mention , however , that the weak interaction strength does not guarantee the validity of perturbation theory in all circumstances . for example , near the critical temperature of the phase transition between the superfluid and the normal state , the running of @xmath162 will be different from the vacuum . as a consequence , the coupling will vanish proportional to the inverse correlation length @xmath217 as @xmath15 approaches @xmath2 , @xmath218 . indeed , the phase transition will be characterized by the non - perturbative critical exponents of the wilson - fisher fixed point . also for lower dimensional systems , the upper bound for @xmath219 is no longer valid - for example the running of @xmath119 is logarithmic for @xmath220 . for our models with @xmath221 , the upper bound becomes dependent on @xmath222 . it increases for @xmath223 . in the limit @xmath224 , it is replaced by the well known `` triviality bound '' of the four dimensional relativistic model , which depends only logarithmically on @xmath213 . finally , for superfluid liquids , as @xmath225 , one has @xmath226 , such that for @xmath227 one finds a large concentration @xmath11 . the situation for dilute bosons seems to contrast with ultracold fermion gases in the unitary limit of a feshbach resonance , where @xmath0 diverges . one may also think about a feshbach resonance for bosonic atoms , where one would expect a large scattering length for a tuning close to resonance . in this case , however , the effective action does not remain local . it is best described by the exchange of molecules . the scale of nonlocality is then given by the gap for the molecules , @xmath228 . only for momenta @xmath229 the effective action becomes approximately local , such that @xmath230 for our approximation . close to resonance , the effective cutoff is low and again in the vicinity of @xmath231 . the density sets a crucial scale for our problem . its precise determination is mandatory for quantitative precision . we will discuss two different methods for its determination and show that the results agree within our precision . for @xmath20 , we also find agreement with the ward identity @xmath232 . the first method is to derive flow equations for the density . this has the advantage that the occupation numbers for a given momentum @xmath233 are mainly sensitive to running couplings with @xmath234 . in the grand canonical formalism , the density is defined by @xmath235{\big |}_{\phi=\phi_0,\sigma=\sigma_0}\ ] ] we can formally define a @xmath88-dependent density @xmath122 by @xmath236{\big |}_{\phi=\phi_0,\sigma=\sigma_0}=-(\partial_\sigma u)(\rho_0,\sigma_0).\ ] ] the flow equation for @xmath122 is given in eq . and the physical density obtains for @xmath188 . the term @xmath237 that enters eq . is the derivative of the flow equation for @xmath163 with respect to @xmath48 . to compute it , we need the @xmath48-dependence of the propagator @xmath158 in the vicinity of @xmath111 . within a systematic derivative expansion , we use the expansion of @xmath113 and the kinetic coefficients @xmath238 and @xmath239 to linear order in @xmath240 , as described in appendix [ secttruncation ] . here , @xmath241 and @xmath242 are the coefficient functions of the terms linear in the @xmath29-derivative and linear in @xmath243 , respectively . no reasonable qualitative behavior is found , if the linear dependence of @xmath238 and @xmath239 on @xmath240 is neglected . also , the scale dependence of @xmath124 and @xmath125 are quite important . the flow equations for @xmath124 and @xmath125 can be obtained directly by differentiating the flow equation of the effective potential with respect to @xmath48 and @xmath107 , cf . the situation is more complicated for the kinetic coefficients @xmath244 and @xmath245 . their flow equations have to be determined by taking the @xmath48-derivative of the flow equation for @xmath246 and @xmath242 . as discussed in app . [ secttruncation ] , we use in this paper the approximation @xmath247 . as a check of both our method and our numerics , we also use another way to determine the particle density . this second method is more robust with respect to shortcomings of the truncation , but less adequate for high precision calculations as needed e.g. to determine the condensate depletion . the second method determines the pressure @xmath248 as a function of the chemical potential @xmath111 . here , the effective potential is normalized by @xmath249 at @xmath20 , @xmath172 . the flow of the pressure can be read of directly from the flow equation of the effective potential and is independent of the couplings @xmath124 and @xmath125 . we calculate the pressure as a function of @xmath48 and determine the density @xmath250 by taking the @xmath48-derivative numerically . it turns out that @xmath251 is in very good approximation given by @xmath252 , where the constant @xmath11 can be determined from a numerical fit . the density is thus linear in @xmath48 . at zero temperature and for @xmath138 , we can additionally use the ward identities connected to galilean symmetry , which yield @xmath232 . we compare our methods in figure [ densitycompared ] and find that they give numerically the same result . we stress again the importance of a reliable method to determine the density , since we often rescale variables by powers of the density to obtain dimensionless variables . . we use three different methods : @xmath253 from the flow equation ( triangles ) , @xmath232 as implied by galilean symmetry ( stars ) and @xmath254 , where the pressure @xmath255 ( boxes ) was obtained from the flow equation and phenomenologically fitted by @xmath256 ( solid lines ) . units are arbitrary and we use @xmath257 , @xmath214 . ] we want to split the density into a condensate part @xmath258 and a density for uncondensed particles or `` depletion density '' @xmath259 . for our model the condensate density is given by the `` bare '' order parameter @xmath260 in order to show this , we introduce occupation numbers @xmath261 for the modes with momentum @xmath233 with normalization @xmath262 one formally introduces a @xmath233 dependent chemical potential @xmath263 in the grand canonical partition function @xmath264}=\text{tr}e^{-\beta(h-\omega_3\int_{\vec{p}}\sigma(\vec{p})n(\vec{p}))},\ ] ] with three dimensional volume @xmath265 . then one can define the occupation numbers by @xmath266{\big |}_{\phi=\phi_0,\sigma(\vec{p})=\sigma_0}.\ ] ] this construction allows us to use @xmath88-dependent occupation numbers by the definition @xmath267{\big |}_{\phi=\phi_0,\sigma(\vec{p})=\sigma_0}.\ ] ] one can derive a flow equation for this occupation number @xmath268 @xcite : @xmath269{\big |}_{\phi_0,\sigma_0}(\partial_k\rho_0 ) . \label{flowofnp}\end{aligned}\ ] ] we split the density occupation number into a @xmath270-distribution like part and a depletion part , which is regular in the limit @xmath271 @xmath272 one can see from the flow equation for @xmath268 that the only contribution to @xmath273 comes from the second term in equation . within a more detailed analysis @xcite one finds @xmath274 we therefore identify the condensate density with the bare order parameter @xmath275 correspondingly , we define the @xmath88-dependent quantities @xmath276 and compute @xmath277 by a solution of its flow equation . even at zero temperature , the repulsive interaction connected with a positive scattering length @xmath0 causes a portion of the particle density to be outside the condensate . from dimensional reasons , it is clear , that @xmath278 should be a function of @xmath279 . the prediction of bogoliubov theory or , equivalently , mean field theory , is @xmath280 . we may determine the condensate depletion from the solution to the flow equation for the particle density , @xmath281 , and @xmath282 . from galilean invariance for @xmath20 and @xmath138 , it follows that @xmath283 with @xmath284 . this gives an independent determination of @xmath258 . in figure [ figdepletion ] we plot the depletion density obtained from the flow of @xmath1 and @xmath285 over several orders of magnitude . apart from some numerical fluctuations for small @xmath279 , we find that our result is in full agreement with the bogoliubov prediction . as a function of the dimensionless scattering length @xmath286 . for the solid curve , we vary @xmath0 with fixed @xmath168 , for the dashed curve we vary the density at fixed @xmath174 . the dotted line is the bogoliubov - result @xmath287 for reference . we find perfect agreement of the three determinations . the fluctuations in the solid curve for small @xmath286 are due to numerical uncertainties . their size demonstrates our numerical precision . ] for @xmath20 a quantum phase transition separates the phases with @xmath115 and @xmath288 . in this section , we investigate the phase diagram at zero temperature in the cube spanned by the dimensionless parameters @xmath289 , @xmath290 and @xmath291 . this goes beyond the usual phase transition for nonrelativistic bosons , since we also include a microscopic second @xmath29-derivative @xmath292 , and therefore models with a generalized microscopic dispersion relation . for non - vanishing @xmath5 ( i.e. for a nonzero initial value of @xmath293 with @xmath294 in app . [ secttruncation ] ) , the galilean invariance at zero temperature is broken explicitly . for large @xmath5 , we expect a crossover to the `` relativistic '' @xmath26 model . if we send the initial value of the coefficient of the linear @xmath29-derivative @xmath295 to zero , we obtain the limiting case @xmath25 . the symmetries of the model are now the same as those of the relativistic @xmath26 model in four dimensions . the space - time - rotations or lorentz symmetry replace galilean symmetry . it is interesting to study the crossover between the two cases . since our cutoff explicitly breaks lorentz symmetry , we investigate in this paper only the regime @xmath296 . detailed investigations of the flow equations for @xmath25 can be found in the literature @xcite . the phase diagram in the @xmath297 plane with @xmath298 is shown in figure [ figqptsigmava1 ] . the critical chemical potential first increases linearly with @xmath5 and then saturates to a constant . the slope in the linear regime as well as the saturation value depend linearly on @xmath299 for @xmath300 . at @xmath20 , the critical exponents are everywhere the mean field ones ( @xmath198 , @xmath301 ) . this is expected : it is the case for @xmath138 @xcite , and for @xmath302 the theory is equivalent to a relativistic @xmath26 model in @xmath303 dimensions . this is just the upper critical dimension of the wilson - fisher fixed point @xcite . -@xmath5 plane for @xmath298 . ] from section 3 we know that for @xmath138 the parameter @xmath299 is limited to @xmath304 . for @xmath138 and a small scattering length @xmath212 , a second order quantum phase transition divides the phases without spontaneous symmetry breaking for @xmath305 from the phase with a finite order parameter @xmath288 for @xmath306 . it is an interesting question , whether this quantum phase transition at @xmath307 also occurs for larger scattering length @xmath0 . we find in our truncation that this is indeed the case for a large range of @xmath0 , but not for @xmath308 . here , the critical chemical potential suddenly increases to large positive values as shown in fig . [ figqptsigmaofa ] . for @xmath309 this increase happens even earlier . ( for a truncation with @xmath310 , the phase transition would always occur at @xmath311 . ) we plot the @xmath312 plane of the phase diagram for different values of @xmath5 in figure [ figqptsigmaofa ] . the form of the critical line can be understood by considering the limits @xmath313 as well as @xmath314 . -@xmath299 plane for @xmath315 ( dotted ) , @xmath316 ( dashed ) and @xmath138 ( solid ) . ] for a fixed chemical potential , the order parameter @xmath118 as a function of @xmath0 goes to zero at a critical value @xmath317 as shown in fig . [ figqptrhoofa ] . this happens in a continuous way and the phase transition is therefore of second order . for @xmath318 , we find @xmath319 . we emphasize , however , that @xmath317 is of the order of the microscopic distance @xmath8 . universality may not be realized for such values , and the true phase transition may depend on the microphysics . for example , beyond a critical value for the repulsive interaction , the system may form a solid . ultracold atom gases correspond to metastable states which may lose their relevance for @xmath320 . for @xmath309 and @xmath321 the phase transition occurs for @xmath322 such that universal behavior is expected . , with @xmath214 . the density @xmath323 as a function of the scattering length @xmath0 goes to zero at a critical @xmath324 , indicating a second order quantum phase transition at that point . ] so far , we have only discussed the vacuum and the dense system at zero temperature . a non vanishing temperature @xmath15 will introduce an additional scale in our problem . for small @xmath325 we expect only small corrections . however , as @xmath15 increases the superfluid order will be destroyed . near the phase transition for @xmath326 and for the disordered phase for @xmath327 , the characteristic behavior of the boson gas will be very different from the @xmath328 limit . for @xmath329 the particle density @xmath1 receives a contribution from a thermal gas of bosonic ( quasi- ) particles . it is no longer uniquely determined by the superfluid density @xmath118 . we may write @xmath330 and observe , that @xmath331 is enforced by galilean symmetry only for @xmath332 . the heat bath singles out a reference frame , such that for @xmath329 galilean symmetry no longer holds . in our formalism , the thermal contribution @xmath333 appears due to modifications of the flow equations for @xmath334 . we start for high @xmath88 with the same initial values as for @xmath20 . as long as @xmath335 the flow equations receive only minor modifications . for @xmath336 or smaller , however , the discreteness of the matsubara sum has important effects . we plot in fig . [ fignoftemperature ] the density as a function of @xmath15 for fixed @xmath337 . ( solid ) and order parameter @xmath338 ( dashed ) as a function of the temperature @xmath339 . the units are arbitray with @xmath340 and @xmath214 . the plot covers only the superfluid phase . for higher temperatures , the density is given by the thermal contribution @xmath341 only . ] in fig . [ fignofsigma ] we show @xmath342 , similar to fig . [ densitycompared ] , but now for different @xmath0 and @xmath15 . for @xmath20 the scattering length sets the only scale besides @xmath1 and @xmath48 , such that by dimensional arguments @xmath343 . bogoliubov theory predicts @xmath344 the first term on the r.h.s . gives the contribution of the ground state , while the second term is added by fluctuation effects . for small scattering length @xmath0 , the ground state contribution dominates . we have then @xmath345 for @xmath168 and @xmath346 can be treated as a small quantity . for @xmath347 and small @xmath0 one expects @xmath348 . the curves in fig . [ fignofsigma ] for @xmath349 show that the density , as a function of @xmath48 , is below the curve obtained at @xmath20 . this is reasonable , since the statistical fluctuations now drive the order parameter @xmath118 to zero . at very small @xmath48 , the flow enters the symmetric phase . the density is always positive , but for simplicity , we show the density as a function of @xmath48 in figure [ fignofsigma ] only in those cases , where the flow remains in the phase with spontaneous @xmath46 symmetry breaking . for different temperatures and scattering length . we plot @xmath342 in arbitrary units , with @xmath214 , and for a scattering length @xmath350 ( solid and dotted ) , @xmath174 ( dashed and dashed - dotted ) . the temperature is @xmath20 ( solid and dashed ) and @xmath349 ( dotted and dashed - dotted ) . ] for temperatures above the critical temperature , the order parameter @xmath118 vanishes at the macroscopic scale and so does the condensate density @xmath351 . the density is now given by a thermal distribution of particles with nonzero momenta . up to small corrections from the interaction @xmath352 , it is described by a free bose gas , @xmath353 with the `` bose function '' @xmath354 in figure [ figrhooftemperature ] we show the dimensionless order parameter @xmath355 as a function of the dimensionless temperature @xmath356 . as a function of the dimensionless temperature @xmath357 for scattering length @xmath174 . here , we varied @xmath15 keeping @xmath48 fixed . numerically , this is equivalent to varying @xmath48 with fixed @xmath15.[figrhooftemperature ] ] this plot shows the second order phase transition from the phase with spontaneous @xmath46 symmetry breaking at small temperatures to the symmetric phase at higher temperatures . the critical temperature @xmath2 is determined as the temperature , where the order parameter just vanishes - it is investigated in the next section . since we find @xmath358 , the condensate fraction @xmath359 as a function of @xmath356 resembles the order parameter @xmath355 . we plot @xmath127 as a function of @xmath356 in fig . [ abaroftemperature ] . except for a narrow region around @xmath2 , the deviations from one remain indeed small . near @xmath2 the gradient coefficient @xmath127 diverges according to the anomalous dimension , @xmath360 , with @xmath361 the anomalous dimension . the correlation length @xmath362 diverges with the critical exponent @xmath363 , @xmath364 , such that @xmath365 here , @xmath361 and @xmath363 are the critical exponents for the wilson fisher fixed point of the classical three - dimensional @xmath26 model , @xmath366 , @xmath367 @xcite . , as a function of the dimensionless temperature @xmath357 , and for scattering length @xmath174 . here , we varied @xmath15 keeping @xmath48 fixed . numerically , this is equivalent to varying @xmath48 with fixed @xmath15 . the plot covers only the phase with spontaneous symmetry breaking . for higher temperatures , the symmetric phase has @xmath368 . the divergence of @xmath127 for @xmath369 reflects the anomalous dimension @xmath361 of the wilson - fisher fixed point . ] in figure [ figrhoofsigma ] we plot @xmath355 as a function of the chemical potential @xmath48 for different temperatures and scattering lengths . we find , that @xmath370 is indeed approached in the limit @xmath328 , as required by galilean invariance . all figures of this section are for @xmath138 . the modifications for @xmath371 are mainly quantitative , not qualitative . , as a function of the chemical potential . we use arbitrary units with @xmath214 . the curves are given for a scattering length @xmath350 ( solid and dotted ) , @xmath174 ( dashed and dashed - dotted ) and temperature @xmath372 ( solid and dashed ) and @xmath349 ( dotted and dashed - dotted ) . at zero temperature , galilean invariance implies @xmath323 , which we find within our numerical resolution . ] the critical temperature @xmath2 for the phase transition between the superfluid phase at low temperature and the disordered or symmetric phase at high temperature depends on the scattering length @xmath0 . by dimensional reasoning , the temperature shift @xmath373 obeys @xmath374 . the proportionality coefficient can not be computed in perturbation theory @xcite . it depends on @xmath5 and we concentrate here on @xmath138 . monte - carlo simulations in the high temperature limit , where only the lowest matsubara frequency is included , yield @xmath375 @xcite . within the same setting , renormalization group studies @xcite yield a similar result , for composite bosons see @xcite . near @xmath2 , the long wavelength modes with momenta @xmath376 dominate the `` long distance quantities '' . then a description in terms of a classical three dimensional system becomes valid . this `` dimensional reduction '' is achieved by `` integrating out '' the nonzero matsubara frequencies . however , both @xmath377 and @xmath1 are dominated by modes with momenta @xmath378 such that corrections to the classical result may be expected . we have computed @xmath2 numerically by monitoring the zero of @xmath118 , as shown in fig . [ figrhooftemperature ] , @xmath379 . our result is plotted in fig . [ tcofa ] . in the limit @xmath212 we find for the dimensionless critical temperature @xmath380 , which is in good agreement with the expected result for the free theory @xmath381 . for the shift in @xmath2 due to the finite interaction strength , we obtain @xmath382 we expect that the result for @xmath383 depends on the truncation and may change somewhat if additional higher order couplings are included . as a function of the dimensionless scattering length @xmath279 ( points ) . we also plot the linear fit @xmath384 ( solid line ) . ] the macroscopic sound velocity @xmath385 is a crucial quantity for the hydrodynamics of the gas or liquid . it is accessible to experiment . as a thermodynamic observable , the adiabatic sound velocity is defined as @xmath386 where @xmath18 is the particle mass ( in our units @xmath387 ) , @xmath251 is the pressure , @xmath1 is the particle density , and @xmath388 is the entropy per particle . it is related to the isothermal sound velocity @xmath389 by @xmath390 where we use our units @xmath33 . one needs the `` equation of state '' @xmath391 and @xmath392 by dimensional analysis , one has @xmath393 with @xmath394 ( in bogoliubov theory ) , and @xmath395 ( free theory ) , such that for small @xmath11 @xmath396 at zero temperature the second term in eq . vanishes , such that @xmath397 . for the isothermal sound velocity one has @xmath398 we can now use @xmath399 and infer @xmath400 one may also define a microscopic sound velocity @xmath401 , which characterizes the propagation of ( quasi- ) particles . at zero temperature , where we can perform the analytic continuation to real time , we can calculate the microscopic sound velocity from the dispersion relation @xmath402 ( with @xmath403 ) . in turn , the dispersion relation is obtained from the effective action by setting @xmath404 , where @xmath405 is the full inverse propagator . we perform the calculation explicitly in app . [ sectpropdisp ] and find @xmath406 the bogoliubov result for the sound velocity is in our units @xmath407 in three dimensions , the decrease of @xmath120 is very slow and the coupling @xmath37 remains comparatively small even on macroscopic scales , cf . [ figflowkinetic ] . we thus do not expect measurable deviations from the bogoliubov result for the sound velocity at @xmath20 . in figure [ figsound ] , we plot our result over several orders of magnitude of the dimensionless scattering length and , indeed , find no deviations from bogoliubovs result . at zero temperature , as a function of the scattering length @xmath279 . within the plot resolution the curves obtained by varying @xmath0 with fixed @xmath1 , by varying @xmath1 with fixed @xmath0 , and the bogoliubov result , @xmath408 , coincide . ] we finally show that for @xmath20 the macroscopic and microscopic sound velocities are equal , @xmath409 . for this purpose , we use @xmath410 from the minimum condition @xmath411 , it follows @xmath412 combining this with the ward identities from app . [ secttruncation ] , namely @xmath413 and @xmath414 , valid at @xmath20 , it follows that the macroscopic sound velocity equals the microscopic sound velocity @xmath415 the use of functional renormalization yields a quantitative description of the whole phase diagram for dilute non - relativistic bosons in three dimensions . this can describe a gas of ultracold bosonic atoms . more generally , our results can also be applied to quantum phase transitions or low temperature phase transitions , whenever the most relevant excitations correspond to bosonic quasi - particles . for this reason we deal with a general dispersion relation , involving in the classical propagator terms linear and quadratic in the frequency . as a function of temperature @xmath15 and effective chemical potential @xmath48 , we have computed the pressure @xmath251 , the density @xmath1 , condensate @xmath285 , superfluid density @xmath118 , and the sound velocity @xmath401 , in dependence on two system parameters , namely the scattering length @xmath0 and a dimensionless coupling @xmath5 parameterizing the classical dispersion relation . for @xmath20 and @xmath138 we find very good agreement with the bogoliubov theory . as @xmath15 increases , the condensate melts at a critical temperature @xmath2 , which exceeds the one for the free theory by @xmath416 . we find a second order phase transition in the universality class of the three - dimensional @xmath26 model , with the associated universal critical exponents . we have found an upper bound for the scattering length , which is of the order of the microscopic length scale @xmath8 . this indicates the breakdown of the pointlike approximation . for atom gases , @xmath8 typically corresponds to the range of the van der waals interaction . the upper bound is a result of strong fluctuation effects in presence of a large microscopic interaction . even for an arbitrary strong microscopic interaction , the quantum fluctuations reduce the renormalized coupling , which corresponds to the physical scattering length in vacuum . for non - relativistic atoms with pointlike interactions , a strong interaction regime can therefore only be realized at high density , @xmath417 . superfluid liquids , like @xmath225 are good examples for such systems . first functional renormalization studies for such systems have already been performed @xcite . our extended treatment overcomes several problems of this early approach and we will apply it to @xmath225 in the future . except for the vicinity of the phase transition , however , the detailed microphysics may play an important role , since all relevant length scales are of the order of the microphysical length scale @xmath8 . other interesting extensions concern lower dimensional systems . we look forward to a unified functional renormalization description of bosons with pointlike interactions , for arbitrary dimensions and arbitrary interaction strength . we use a derivative expansion for the truncation of the effective average action with derivative operators up to four momentum dimensions @xmath418 here , we employ the renormalized fields @xmath419 and coupling functions @xmath163 , @xmath420 , @xmath421 . we fix the wave function renormalization factor @xmath127 such that @xmath422 . terms of the form @xmath423 or @xmath424 are not included here , since they are expected to play a sub - leading role . for a systematic derivative expansion they have to be added - the terms with up to two derivatives can be found in @xcite . in terms of dimensions , the operator @xmath425 counts as two space derivatives for the nonrelativistic model with @xmath35 , while it counts as one space dimension for the relativistic model with @xmath426 . we expand the @xmath88-dependent functions @xmath113 , @xmath241 , @xmath242 , @xmath427,@xmath428 and @xmath429 around the @xmath88-dependent minimum @xmath110 of the effective potential and the @xmath88-independent value of the chemical potential @xmath111 that corresponds to the physical particle number density @xmath1 . for example , with @xmath430 , one has @xmath431 let us concentrate on the non - relativistic model where @xmath36 , @xmath35 in the microscopic action . at zero temperature , we can perform an analytic continuation to real time @xmath432 . the microscopic action is then @xmath433&=&-\int_{-\infty}^{\infty } dt\int d^3x \\ & & { \big \{}\phi^*\,(-i\partial_t-\sigma-\delta)\,\phi\,+\,\frac{1}{2}\lambda(\phi^*\phi){\big \}}. \label{realtimemicroscopicaction}\end{aligned}\ ] ] in addition to space translations , rotations and time translations , two further symmetries constrain the possible forms of the couplings in @xmath434 . in order to derive these constraints , we extend eq . to a @xmath435-dependent source @xmath436 . first , there is a semi - local @xmath46 symmetry of the form @xmath437 this holds since the combination @xmath438 acts as a covariant derivative . in addition , we have the invariance under galilean boost transformations of the fields @xmath439 while the invariance of the interaction term under this symmetry is obvious , its realization for the kinetic term is more involved . performing the transformation explicitly , one finds @xmath440 such that indeed the combination @xmath441 leads to this invariance . on the other hand , the validity of the galilean symmetry for an effective action guarantees that only the combination or powers of this operator act on @xmath193 . an operator of the form @xmath442 with @xmath443 would break the symmetry . ( note , that @xmath444 is also invariant . ) both the semi - local @xmath46 symmetry and the galilean symmetry are helpful only at zero temperature . at nonzero temperature , the analytic continuation to real time is no longer useful . an analog version of the semi - local @xmath46 transformation for euclidean time @xmath29 would involve the imaginary part of the chemical potential @xmath48 , which has no physical meaning . the dependence of physical quantities on @xmath445 is not restricted . in addition , the galilean symmetry is broken explicitly by the thermal heat bath . combining semi - local @xmath46 symmetry and galilean symmetry at @xmath20 , we find that the derivative operators @xmath446 , @xmath243 and the chemical potential term @xmath240 are combined to powers of the operator @xmath447 in addition to powers of that operator acting on @xmath193 , only spatial derivatives of terms , that are invariant under @xmath46 transformations , like @xmath448 , may appear . since the symmetry transformations act linearly on the fields , the full effective action @xmath71 $ ] is also invariant . this also holds for the average action @xmath87 $ ] , provided that the cutoff term @xmath100 $ ] is invariant . we can write the effective action as an expansion in the operator @xmath449 @xmath450&=&\int_x \bigg{\{}u_0(\rho)\\ \nonumber & & + \frac{1}{2}\tilde{z}(\rho)\left(\phi^*(-i\partial_t-(\sigma-\sigma_0)-\delta)\phi+c.c\right)\\ \nonumber & & + \frac{1}{2}\tilde{v}(\rho)\left(\phi^*(-i\partial_t-(\sigma-\sigma_0)-\delta)^2\phi+c.c\right)\\ & & + \dots\bigg{\ } } \label{eqeffectiveactiongalilean}\end{aligned}\ ] ] performing the wick rotation back to euclidean time , we can compare this to eq . , and find for @xmath20 the relations @xmath451 and therefore @xmath452 we next compute the inverse propagator in a constant background field by expanding @xmath99 to second order in the fluctuations around this background . for this purpose , it is convenient to decompose @xmath453 the constant condensate field @xmath51 can be chosen to be real without loss of generality . the fluctuating real fields are the radial mode @xmath454 and the goldstone mode @xmath455 , and @xmath456 . the truncation of the effective average action reads in that basis @xmath457 = \int_x{\bigg \{}u(\rho,\sigma ) % \\ % \nonumber % & & & + & \frac{1}{2}z_1(\rho,\sigma ) \left(i\sqrt{2}\phi_0\partial_\tau\phi_2+i\phi_1\partial_\tau \phi_2-i\phi_2\partial_\tau\phi_1\right)\\ \nonumber & + & \frac{1}{2}z_2(\rho,\sigma ) \left(\sqrt{2}\phi_0(-\delta)\phi_1+\phi_1(-\delta ) \phi_1+\phi_2(-\delta)\phi_2\right)\\ \nonumber & + & \frac{1}{2}v_1(\rho,\sigma ) \left(\sqrt{2}\phi_0(-\partial_\tau^2)\phi_1+\phi_1(-\partial_\tau^2 ) \phi_1+\phi_2(-\partial_\tau^2)\phi_2\right)\\ \nonumber & + & v_2(\rho,\sigma ) \left(i\sqrt{2}\phi_0(\partial_\tau\delta)\phi_2+i\phi_1(\partial_\tau\delta ) \phi_2-i\phi_2(\partial_\tau\delta)\phi_1\right)\\ & + & \frac{1}{2}v_3(\rho,\sigma ) \left(\sqrt{2}\phi_0(-\delta^2)\phi_1+\phi_1(-\delta^2 ) \phi_1+\phi_2(-\delta^2)\phi_2\right){\bigg \}}.\end{aligned}\ ] ] the inverse propagator @xmath95 can be inferred from an expansion in second order in @xmath454 and @xmath455 . we keep the linear order in @xmath123 , which will be needed for the flow equation for the density . this yields @xmath457 & = & \int_x{\bigg \{}u(\rho_0,\sigma_0)+u^{(\sigma)}\,(\sigma-\sigma_0)+\frac{1}{2}(u^\prime+2\rho_0 u^{\prime\prime})\phi_1 ^ 2+\frac{1}{2}u^\prime \phi_2 ^ 2\\ \nonumber & + & \frac{1}{2}\left(z_1+z_1^\prime \rho_0+z_1^{(\sigma)}(\sigma-\sigma_0)\right ) \left(i\phi_1\partial_\tau \phi_2-i\phi_2\partial_\tau\phi_1\right)\\ \nonumber & + & \frac{1}{2}\left(1 + 2 z_2^\prime \rho_0+z_2^{(\sigma)}(\sigma-\sigma_0)\right ) \left(\phi_1(-\delta ) \phi_1\right)\\ \nonumber & + & \frac{1}{2}\left(1+z_2^{(\sigma)}(\sigma-\sigma_0)\right ) \left(\phi_2(-\delta ) \phi_2\right)\\ \nonumber & + & \frac{1}{2}\left(v_1 + 2 v_1^\prime\rho_0+v_1^{(\sigma)}(\sigma-\sigma_0)\right ) \left(\phi_1(-\partial_\tau^2 ) \phi_1\right)\\ \nonumber & + & \frac{1}{2}\left(v_1+v_1^{(\sigma)}(\sigma-\sigma_0)\right ) \left(\phi_2(-\partial_\tau^2 ) \phi_2\right)\\ \nonumber & + & \left(v_2+v_2^\prime\rho_0+v_2^{(\sigma)}(\sigma-\sigma_0)\right ) \left(i\phi_1(\partial_\tau\delta ) \phi_2-i\phi_2(\partial_\tau\delta)\phi_1\right)\\ \nonumber & + & \frac{1}{2}\left(v_3 + 2v_3^\prime \rho_0+v_3^{(\sigma)}(\sigma-\sigma_0)\right ) \left(\phi_1(-\delta^2 ) \phi_1\right)\\ & + & \frac{1}{2}\left(v_3+v_3^{(\sigma)}(\sigma-\sigma_0)\right)\left(\phi_2(-\delta^2 ) \phi_2\right){\bigg \}},\end{aligned}\ ] ] where we dropped the argument @xmath458 at several places and used the implicit rescaling condition @xmath459 . in our simple truncation , we take at @xmath117 only @xmath460 into account . we neglect the contribution of the other couplings , i.e. set @xmath461 . as shown above , it follows from the symmetry requirements at zero temperature , that @xmath462 , @xmath463 and @xmath464 . the truncation @xmath294 therefore violates the galilean symmetry , as does our choice of the cutoff term @xmath465 . within our approximation , it is consistent to set @xmath466 at zero temperature . also the deviations from this relation at finite temperature are neglected for simplicity in this paper . this yields eq . . in the following we discuss the role of continuous symmetries of the microscopic action @xmath85 $ ] . since all these symmetries are linear in the fields , the full effective action @xmath71 $ ] is also symmetric . in the case that the cutoff term @xmath100 $ ] is chosen invariant under the symmetry transformation in question , this also holds for the average action @xmath467 $ ] for finite @xmath88 . from noether s theorem it follows that there exists a conserved current @xmath468 connected with every such symmetry . if the action is formulated as an integral over the imaginary time @xmath29 the conservation equation implies for the current @xmath469 at zero temperature , we can perform a wick rotation to real time , @xmath470 , and eq . takes the usual form @xmath471 the noether charge @xmath472 is conserved in time , i.e. @xmath473 . this holds if @xmath474 falls off sufficiently fast at spatial infinity . at finite temperature however , the situation is different . a simple analytic continuation to real time is no longer possible , since the configuration space is now a torus with periodicity @xmath475 in the @xmath29-direction instead , we can integrate eq . over complex time @xmath29 , giving @xmath476 from the symmetry , it now follows that there exists a solenoidal vector field or three component current @xmath477 . a global symmetry of an action @xmath71 $ ] ( where @xmath434 could be replaced by @xmath120 or @xmath99 if appropriate ) can be formulated in its infinitesimal form as @xmath478=\gamma[\phi],\ ] ] with @xmath479 independent of @xmath480 . here @xmath388 is the infinitesimal generator of the symmetry transformation . for a local transformation , where @xmath479 depends on @xmath480 , @xmath481 , we can expand @xmath482=\gamma[\phi]+\int_x \left\ { ( \partial_\mu\epsilon){\cal j}^\mu+(\partial_\mu\partial_\nu\epsilon){\cal k}^{\mu\nu}+\dots\right\}. \label{noetherexpansion}\ ] ] the global symmetry implies that the expansion on the r.h.s . of eq . starts with @xmath483 . here and in the following it is implied that @xmath479 as well as its derivatives are infinitesimal , i.e. we keep only terms that are linear in @xmath479 . the index @xmath484 goes over @xmath485 , representing @xmath486 in the real time case and @xmath487 for imaginary time . . implies for arbitrary @xmath488 @xmath489}{\delta\phi}\epsilon s\phi-(\partial_\mu\epsilon){\cal j}^\mu-(\partial_\mu\partial_\nu\epsilon){\cal k}^{\mu\nu}+\dots\right\}=0 . \label{eqnoetherintegral}\ ] ] our notation is for real fields and implies a summation over components , if appropriate . in a complex basis one replaces @xmath490 by @xmath491 . is valid for all field configurations @xmath193 and not only for those that fulfill the field equation @xmath492}{\delta \phi}=0 $ ] . in consequence , the integrand is a total derivative @xmath493}{\delta\phi}\epsilon s\phi-(\partial_\mu\epsilon){\cal j}^\mu-(\partial_\mu\partial_\nu\epsilon){\cal k}^{\mu\nu}+\ldots & \\ & = -\partial_\mu\left(j^\mu \epsilon+\kappa^{\mu\nu}\partial_\nu\epsilon+\ldots\right ) . & \end{aligned}\ ] ] we can now specialize to @xmath494 and find @xmath495}{\delta\phi } s\phi(x)=-\partial_\mu j^\mu.\ ] ] this defines the noether current @xmath496 . for solutions of the field equation , @xmath492}{\delta\phi}=0 $ ] , the current @xmath496 is conserved , @xmath497 . for a given @xmath480 we can also specialize to @xmath498 which leads to @xmath499 this process can be continued , leading us to a whole tower of identities for the conserved current @xmath496 . if the action @xmath71 $ ] includes derivatives only up to a finite order @xmath1 , i.e. can be written in the form @xmath72=\int_x{\cal l}(\phi,\partial\phi , \partial\partial\phi , \ldots , \partial^{(n)}\phi),\ ] ] the expansion on the right hand side of only contains terms up to order @xmath500 such that the tower of equations for @xmath496 can be solved . moreover , for homogeneous situations where @xmath492}{\delta\phi}$ ] is solved by a constant @xmath193 , the second term on the r.h.s . vanishes since it includes a derivative . we have then @xmath501 . a convenient way to find the local currents employs parameters @xmath502 that decay sufficiently fast at infinity such that we can partially integrate eq . @xmath503}{\delta\phi } s\phi+\partial_\mu{\cal j}^\mu-\partial_\mu\partial_\nu{\cal k}^{\mu\nu}+\dots\right\}=0.\ ] ] this yields the local identity @xmath78}{\delta \phi}s\phi=-\partial_\mu{\cal j}^\mu+\partial_\mu\partial_\nu { \cal k}^{\mu\nu}-\dots.\ ] ] an expansion of the l.h.s . in derivatives often yields substantial information on @xmath504 etc . by inspection . our construction yields a unique conserved local current @xmath496 for every generator of a continuous symmetry . we note , however , that @xmath505 is also a conserved local current if @xmath124 and @xmath506 are independent of @xmath480 . this remark is important if we want to associate @xmath496 with the current for a physical quantity . a rotation invariant setting implies @xmath507 , but @xmath508 and @xmath124 may differ from zero . after these general considerations we now specialize to nonrelativistic real time actions of the form @xmath72=\int_{-\infty}^{\infty}dt \int d^3x{\cal l}(\phi , ( i\partial_t+\delta)\phi,(i\partial_t+\delta)^2\phi,\dots).\ ] ] we assume , that @xmath434 invariant under the same symmetries as the action . from the symmetry under time translations @xmath509 we find a conserved current @xmath510 . up to a possible additive constant its @xmath435-component is the energy density , while the spatial components describe energy flux density . the multiplicative constant @xmath124 gets fixed if we choose the units to measure energy . the choice @xmath511 corresponds to @xmath512 . similarly , the invariance under spatial translations @xmath513 implies a conserved current @xmath514 for each spatial direction @xmath515 . up to an additive constant @xmath516 the @xmath435-component is the conserved momentum density , @xmath517 , while the spatial components can be interpreted as a momentum flux density , with the diagonal components @xmath518 describing pressure . from the global @xmath46 symmetry @xmath519 we can infer the conservation of the current @xmath520 associated to the conserved particle number . in order to identify the total particle number with the charge of this current , @xmath521 , we need to fix a possible multiplicative constant @xmath124 . for this purpose , we use the galilean boost invariance , described already in eq . . it reads in its infinitesimal form @xmath522 and the conserved charge of @xmath523 is the center of mass , again up to an additive constant . the generator @xmath523 can be decomposed as @xmath524 this implies for the current @xmath525 specializing to the @xmath435-component , identifying the momentum density @xmath526 and reintroducing the particle mass @xmath33 we find @xmath527 from this we can conclude that up to an additive constant @xmath528 is the particle density @xmath529 . for the effective action we find for @xmath117 and constant @xmath530 the current @xmath531 using the normalization condition @xmath532 , this gives @xmath533 . at zero temperature , this is the particle density and the additive constant @xmath534 vanishes . at nonzero temperature we can compare to eq . and find @xmath535 . for completeness we also mention the symmetry under spatial rotations @xmath536 with orthogonal matrix @xmath537 , generators @xmath538 , and @xmath539 the antisymmetric tensor in three dimensions . the infinitesimal transformation reads @xmath540 the time component of the conserved current @xmath541 is , of course , the angular momentum density . we derive the flow equation for the effective potential by evaluating the flow equation for the average action for constant fields . inserting a real constant field @xmath542 one finds for @xmath543 the flow at fixed @xmath107 @xmath544 here , @xmath545 is the number of spatial dimensions and we use the abbreviations @xmath546 we dropped the arguments @xmath547 at several places on the right hand side . primes denote derivatives with respect to @xmath107 . in the phase with spontaneous symmetry breaking , we have @xmath548 . the matsubara sums over @xmath1 can be carried out by virtue of the formulae @xmath549 with @xmath550 . this brings us to @xmath551 where we introduced @xmath552 in our simple truncation with @xmath553 , @xmath554 , @xmath461 , and at @xmath117 , the integrand in eq . becomes mostly independent of the spatial momentum . the integral can than be carried out and we find @xmath555 with @xmath556 that the momentum integral can be performed analytically is a nice feature of the cutoff . the limit @xmath328 is obtained by substituting the @xmath557 functions with unity . the flow of the effective potential contains a subtlety that can be seen in the limit @xmath558 ( @xmath515 ) , where we find @xmath559 the term @xmath560 in the last line is not present if @xmath293 is set to zero from the outset . if @xmath238 is independent of @xmath107 , this term is independent of @xmath107 and gives only an overall shift of the effective potential . we show in this appendix our results for the flow equation of the kinetic coefficients @xmath120 , @xmath127 and @xmath37 . we neglect all contributions from momentum dependent vertices . in other words , we use @xmath107-independent constants @xmath561 , @xmath562 and @xmath554 . in our truncation with @xmath563 , and with the cutoff , we can perform all momentum integrations analytically , leading us to @xmath564 here , @xmath545 is the number of spatial dimensions , @xmath565 and @xmath566 are as in . the matsubara sums over @xmath1 can be performed analytically again by using and derivatives thereof . in the limit @xmath328 , the matsubara frequencies are continuous @xmath567 and the sum becomes an integral @xmath568 . the expressions for @xmath361 and @xmath569 in eq . agree with those derived in @xcite , while our result for @xmath570 corrects an error in a first version of @xcite ( in equation ( c.8 ) ) . the inverse propagator is given by the second functional derivative of the effective action @xmath571 and we find from the truncation @xmath572 here we use the abbreviations @xmath573 at zero temperature , we can analytically continue to real time @xmath574 , and find @xmath575 the dispersion relation is found from the on shell condition @xmath576 which yields @xmath577 the solutions for @xmath22 define the dispersion relation . we find two branches , according to @xmath578 in the phase with spontaneous symmetry breaking , the @xmath579 branch of this solution is an `` optical mode '' , while the @xmath580 branch is a sound mode . the microscopic sound velocity is @xmath581 . using @xmath582 , @xmath583 , @xmath584 and @xmath585 , we find @xmath586 the `` optical mode '' has at vanishing spatial momentum the frequency @xmath587 which diverges @xmath588 in the limit @xmath589 . 12 n. n. bogoliubov , j. phys . ( moscow ) * 11 * , 23 ( 1947 ) . s. t. beliaev , sov . jetp * 7 * , 289 ( 1958 ) ; * 7 * , 299 ( 1958 ) ; j. gavoret , p. nozires , ann . ( n. y. ) * 28 * , 349 ( 1964 ) ; a. a. nepomnyashchii , y. a. nepomnyashchii , jetp lett . * 21 * , 1 ( 1975 ) . c. castellani , c. di castro , f. pistolesi , g. c. strinati , phys . lett . * 78 * , 1612 ( 1997 ) ; f. pistolesi , c. castellani , c. d. castro , g. c. strinati , phys . b * 69 * , 024513 ( 2004 ) . c. wetterich , phys . b * 77 * , 064504 ( 2008 ) . k. g. wilson , phys . b * 4 * , 3174 ( 1971 ) ; k. g. wilson , j. b. kogut , phys . * 12 * , 75 ( 1974 ) ; f. wegner , a. houghton , phys . rev . a * 8 * , 401 ( 1973 ) ; j. polchinski , nucl . b * 231 * , 269 ( 1984 ) . c. wetterich , z. phys . c * 48 * , 693 ( 1990 ) ; nucl . b * 352 * , 529 ( 1991 ) . c. wetterich , phys . b * 301 * , 90 ( 1993 ) . j. berges , n. tetradis and c. wetterich , phys . rept . * 363 * , 223 ( 2002 ) ; c. bagnuls , c. bervillier , phys . rept . * 348 * , 91 ( 2001 ) ; b. delamotte , arxiv : cond - mat/0702365 . n. dupuis , k. sengupta , europhys . lett . * 80 * , 50007 ( 2007 ) . j. o. andersen and m. strickland , phys . rev . a * 60 * , 1442 ( 1999 ) ; l. canet , b. delamotte , d. mouhanna and j. vidal , phys . b * 68 * , 064421 ( 2003 ) ; phys . d * 67 * , 065004 ( 2003 ) ; c. bervillier , a. juttner and d. f. litim , nucl . phys . b * 783 * , 213 ( 2007 ) . d. t. son and m. wingate , annals phys . * 321 * , 197 ( 2006 ) . d. f. litim , phys . b * 486 * , 92 ( 2000 ) ; j. m. pawlowski , annals phys . * 322 * , 2831 ( 2007 ) . c. wetterich , arxiv:0712.1926 [ cond-mat.stat-mech ] . t. papenbrock and c. wetterich , z. phys . c * 65 * , 519 ( 1995 ) . d. i. uzunov , phys . a * 87 * , 11 ( 1981 ) ; s. sachdev , quantum phase transitions , cambridge university press , cambridge ( 1999 ) . k. g. wilson and m. e. fisher , phys . rev . * 28 * ( 1972 ) 240 . a. pelissetto and e. vicari , phys . rept . * 368 * , 549 ( 2002 ) . j. o. andersen , rev . phys . * 76 * , 599 ( 2004 ) . p. arnold , g. moore , phys . * 87 * , 120401 ( 2001 ) ; v. a. kashurnikov , n. v. prokofev , b. v. svistunov , phys . lett . * 87 * , 120402 ( 2001 ) . g. baym , j. -p . blaizot , m. holzmann , f. lalo , d. vautherin , phys . lett . * 83 * , 1703 ( 1999 ) ; j. -p . blaizot , r. mndez - galain , n. wschebor , phys . e * 74 * , 051116 ( 2006 ) ; j. -p . blaizot , r. mndez - galain , n. wschebor , phys . e * 74 * , 051117 ( 2006 ) . s. ledowski , n. hasselmann , p. kopietz , phys . a * 69 * , 061601(r ) ( 2004 ) ; n. hasselmann , s. ledowski , p. kopietz , phys . a * 70 * , 063621 ( 2004 ) . s. diehl , h. gies , j. m. pawlowski and c. wetterich , phys . a * 76 * , 021602(r ) ( 2007 ) . t. gollisch and c. wetterich , phys . b * 65 * , 134506 ( 2002 ) .

the biological and physical properties of proteins are compelling for many reasons . while just a small amount of the nowadays hundreds of thousands known protein sequences are experimentally characterized , the variety of their functions is overwhelming . though the structure has been resolved only for a subset of these sequences , the number of stable folds that are expressed in nature is seemingly small compared to the number of sequences . the relationship between fold and function is far from obvious , and examples such as intrinsically unstructured proteins and multi - functional folds resist simple schemes for classification . the question what really makes a protein functional hence needs to be addressed in the context of its specific biological environment . from a physical point of view , an attempt to find some unifying concepts for the interpretation of dynamics and thermodynamics is the description of proteins in term of energy landscapes@xcite , in which the evolution of the system is related to the dynamics on a high - dimensional rugged energy surface . the existence of local minima , connected by saddles of different barrier height and rank , leads to a distribution of timescales that are reflected in the dynamics of the proteins . although the energy landscape provides a reduced description , the complex interactions in proteins and their interaction with the environment , which involve multi - body interactions and subtle effects of charges , make its complete characterization neither experimentally nor theoretically conceivable . this situation is somewhat reminiscent of other complex systems such as glasses which , though being more homogeneous systems , share the property of displaying a high - dimensional landscape leading to complex dynamics . the energy landscape picture is useful for a qualitative analysis of protein properties , but for quantitative studies , an exhaustive sampling and and the full characterization are practically infeasible for this high - dimensional representation . therefore , in order to obtain comprehensive quantitative predictions on generic protein properties from the information on the landscape , the picture needs to be simplified . a recent work@xcite has shown that a reduced description of the energy landscape , originally devised for the analysis of super - cooled liquids by stillinger and weber@xcite , can successfully capture the essential thermodynamic aspects of folding in the context of a simplified protein model . in particular , it was shown that the density of states constructed from the local minima of the energy landscape , called inherent structures , can be used to compute the most important thermodynamic observables . this finding is important because it provides a general scheme for theoretical studies of protein thermodynamics , showing how the relevant information can be quantitatively accessed from its imprint on the potential energy surface . the approach has consequently been used in the context of studies of the folding properties of a @xmath0-barrel forming protein@xcite , the construction of the free energy landscape by mechanical unfolding@xcite , and the network of native contacts@xcite . other earlier works including inherent structure analysis but not necessarily seeking to characterize the full thermodynamics of folding can be found in @xcite . however the validity of an analysis based on the inherent structure landscape ( isl ) must be critically examined because the method involves a fundamental assumption which could be questioned : the vibrational free energy within the basin of attraction of an inherent structure is assumed to be independent of the basin . a recent study @xcite tried to go beyond this approximation by assuming that the vibrational free energy can depend on the energy of the inherent structures . still , the question is subtle as we show in the present work that , even when the vibrational free energy depends on the inherent structure energy , the derivation of thermodynamic quantities such as the specific heat from the isl can be validly carried without any change in the procedure . therefore an understanding of the limits of the isl approach requires a deeper analysis . this is the aim of the present work . we proceed in two steps . in a first step ( section [ section3 ] ) , after briefly summarizing the isl formalism and introducing the protein model used in section [ section2 ] , we test the validity of the isl approach by comparing its results to the data obtained from equilibrium molecular dynamics for a set of structures . we selected four previously unstudied two - state folding proteins of varying size and secondary structure elements . in a second step ( section [ section4 ] ) , we critically revisit the major hypotheses of the isl approach , as well as its practical limitations , such as the sampling of the phase space , and suggest routes for improvements . finally , we summarize our findings in section [ section5 ] and give an outlook on possible future studies that stem from our results . in this subsection , we briefly review the major results of @xcite on obtaining reduced thermodynamics from an analysis of the inherent structures . the method is general and not bound to a specific protein model , provided the phase space of the protein can be explored by molecular dynamics , and that the energies of the visited states can be calculated . from simulations at fixed temperature close to the folding temperature @xmath1 , which insures that the system evolves in a large part of the configuration space , the local potential minima , labelled by @xmath2 , are determined by conjugate gradient minimizations performed at fixed frequency along a molecular dynamics trajectory . the global minimum @xmath3 is defined as the reference ground state with zero energy . let @xmath4 , denote the @xmath5 cartesian coordinates of the @xmath6-particle system , and @xmath7 its potential energy function . the probability to find a particular minimum @xmath2 with potential energy @xmath8 can be written as @xmath9 where @xmath10 , @xmath11 denotes the configurational part of the partition function and @xmath12 is the basin of attraction of the minimum @xmath2 . with the definition@xmath13 the unknown integral over the complex landscape of the basin of attraction @xmath12 is summed in an free - energy like function @xmath14 which in principle depends both on the nature of the basin and temperature . notice that although we use the index @xmath15 like vibrational , @xmath14 is obtained from the full nonlinear integral over @xmath12 , and not from its harmonic approximation . + the inherent structure landscape approach makes two key assumptions@xcite which enable to considerably reduce the amount of information needed on the landscape while keeping its most important features . * ( a1 ) the function @xmath16 for two minima @xmath17 that are distinct but close in energy , @xmath18 , is the same for both minima : @xmath19 . consequently , @xmath20 . * ( a2 ) the function @xmath21 does not vary significantly for different minima , i.e. @xmath22 . both assumptions were discussed in @xcite . in section [ section4 ] , we show that assumption ( a2 ) can actually be relaxed to the less strong form @xmath23 while most calculations remain feasible and some of the thermodynamic variables unchanged . with these assumptions , the contribution from the function @xmath24 factorizes in the numerator and denominator of ( [ eq1 ] ) so that it can be eliminated to give @xmath25 where the sum in the partition function includes all inherent structures found from the global minimum @xmath3 to the minimum @xmath26 having the highest energy . here , the energy scale is shifted such that the energy of the global minimum @xmath3 is zero . introducing an energy density function for the inherent structures @xmath27 , the probability to find a minimum in the interval @xmath28 $ ] at temperature @xmath29 is @xmath30 for the model used in this work , the low energy minima are in practice sparsely separated in energy . as the ground state is isolated , one obtains @xmath31 with the probability of the ground state @xmath32 , so that the inherent structure density of states can be estimated from the probability to be in the basin of attraction of a minimum in a fixed temperature simulation at temperature @xmath33 as @xmath34 though above we have chosen a continuous notation to simplify the equations , it should be noted that the density built from an estimate of the probability density function of sampled minima for the present model in practice always comprises discrete and continuous parts which can be integrated separately . once the inherent structure density of states is known , one can compute the inherent structure partition function @xmath35 from ( [ eq : piszis ] ) , from which all thermodynamics functions can be derived , including the free energy @xmath36 and the internal energy @xmath37 . given that most states of the system can be sampled close to the folding temperature , it is sufficient to simulate the system at a single temperature @xmath33 to construct the inherent structure landscape , in contrast to the full thermodynamics where one needs to sample different ranges of temperatures . therefore , the isl approach can be computationally very efficient . in the following , we will restrict ourselves to the computation of specific heat @xmath38 being a quantity of fundamental importance in a physical system , as it is sensitive to fluctuations and , for instance , shows a clear signature of phase transitions . it can be deduced from numerical derivatives of the partition function @xmath39 through @xmath40 and hence @xmath41 since our goal is to analyze the validity of the isl approach and not to derive quantitative data for a particular protein , we decided to choose a simplified model , which allows the sampling of phase space at a reasonable computational cost . however the model must be rich enough to properly describe the complex features of its physics , and should be able to distinguish between proteins which differ , for instance , in their secondary structure . we use frustrated off - lattice g - models identical to the ones introduced in @xcite because they provide a good compromise between all - atom simulations and simplified models that do not fully describe the geometry of a protein . these models provide a representation with a single particle per residue centered at the location of each @xmath42-atom . for details on the model and the parameters , we refer to @xcite and a brief review in the appendix [ ap1 ] . although the validity of such models to provide a faithful representation of protein folding is a recurrent subject of debate , off - lattice g - models have been successfully used to study folding kinetics @xcite and the mechanical resistance of proteins @xcite . from a physical point of the view , despite a strong bias towards the ground state , these models have a complex energy landscape with a large number of local minima well suited for the analysis in terms of inherent structures . + as the results of the isl approach depend on the density of states of the inherent structures , for a reliable test of the method it is important to examine examples which could differ in their properties , i.e. to investigate proteins of different size and structure . to test the inherent structure approach beyond the previously employed immunoglobulin ( ig ) binding domain of protein g ( 2gb1 ) , we selected four two - state folding proteins of varying size and folds from the pdb database@xcite : the trp - cage mini - protein construct ( 1l2y , 20 residues , @xmath43-helical ) , the ww domain fpb28 ( 1e0l , 37 residues , @xmath0-sheets ) , the src - sh3 domain ( 1srl , 56 residues , @xmath0-sheets ) and ubiquitin ( 1ubq , 76 residues , @xmath44-fold ) . the motivation for these choices is discussed in section [ section3 ] for each protein ( see insets of figures [ fig1]-[fig4 ] for structural representations of these four proteins drawn with pymol@xcite ) . the positions of the @xmath45-atoms of the pdb files is chosen as a reference for the construction of the g - model . in the case of nmr resolved structures , the first structure is selected as the reference . the native contacts of the model were established according to the distances between atoms belonging to different residues . a native contact is formed if the shortest distance belonging to atoms of two different residues is smaller than @xmath46 . the number of native contacts according to this criterion are : @xmath47 for the ww domain , @xmath48 for ubiquitin , @xmath49 for src - sh3 and @xmath50 for trp - cage . this definition is simple and includes some arbitrariness . there exist other methods for probing contacts between side - chains , e.g. by invoking the van der waals radii of residue atoms and solvent molecules@xcite . though using the latter method preserves the main structure of the contact map , it leads to quantitative differences in the the number and location of contacts along the sequence . consequently , one can expect that the topology of the energy surface and key thermodynamic properties such as the folding temperature are also altered when the definition of the contact map is varied . for the purpose of the present study which does not attempt to give a quantitative description of side chain contacts , and focusses on global properties of the landscape rather than its detailed relation to the network of contacts , the cutoff - based approach is acceptable . + molecular dynamics simulations were performed using the brooks - brnger - karplus algorithm @xcite with a time - step of @xmath51 and a friction constant of @xmath52 ( all units in this section are dimensionless , see @xcite for details ) . to ensure equilibration , the system was thermalized starting from the native state ( pdb coordinates ) for @xmath53 . the simulation time for a single temperature point and a single initial condition was @xmath54 , and the data obtained for both inherent structure sampling ( fixed temperature ) and thermodynamic sampling ( variable temperature ) were averaged over various initial sets of velocities . minimization was performed using the conjugate gradient method with the polak - ribire algorithm . to estimate the vibrational free energy at the minimum , mass weighted normal mode analysis was performed using lapack diagonalization routines . the second - order derivatives of the potential energy function at the minimum were calculated by numerical differentiation of the analytical first - order derivatives . in this section , the validity of the isl approach is tested by comparing the equilibrium thermodynamics deduced from molecular dynamics simulations to the reduced thermodynamics from inherent structure sampling . as discussed in section [ section2 ] , we evaluate the specific heat @xmath55 as a function of temperature as a representative example of the thermodynamic observables . . the inset shows a close up on the low - energy range . the size of the energy bins for the density estimate is @xmath56 . _ right : _ comparison of the specific heat from equilibrium trajectories @xmath57 ( points ) , from which the specific heat of a harmonic oscillator in @xmath5 dimensions has been subtracted , and @xmath58 from inherent structure analysis ( solid red line ) ; see text for further explanations.,title="fig:",width=249 ] . the inset shows a close up on the low - energy range . the size of the energy bins for the density estimate is @xmath56 . _ right : _ comparison of the specific heat from equilibrium trajectories @xmath57 ( points ) , from which the specific heat of a harmonic oscillator in @xmath5 dimensions has been subtracted , and @xmath58 from inherent structure analysis ( solid red line ) ; see text for further explanations.,title="fig:",width=347 ] the src - sh3 domain was chosen since it has the same number of residues as ig binding domain of protein g studied in @xcite , but contrasts to the latter in terms of structure . the src - sh3 domain is mostly composed of @xmath0-sheets and does not contain an @xmath43-helical - secondary structure element ( only five residues form a small right - handed helix segment ) . the inherent structure density of states , shown on the left - hand side of figure [ fig1 ] , was obtained from various simulations close to the folding temperature ( @xmath59 ) and built from @xmath60 minima according to ( [ ihs ] ) . after computing an energy histogram using @xmath61 bins to yield an estimate of the inherent structure probability density , energy bins with only a single count have been discarded from the analysis to avoid a bias that could be introduced by insufficiently sampled isolated minima . the right hand side of figure [ fig1 ] shows a comparison between the temperature dependence of the specific heat calculated from inherent structures , @xmath58 , and the temperature dependence of the specific heat calculated from equilibrium molecular dynamics simulations at variable temperature @xmath57 . the equilibrium thermodynamics has been determined by averaging the results of 10 initial conditions per temperature step except for points close to the transition where the results of 20 initial conditions have been used . despite this averaging , the variance , indicated by error - bars on the y - axis , is large in the vicinity of the folding transition as the waiting time for a transition to occur becomes comparable to the simulation time . for a harmonic system , @xmath62 where @xmath63 is the number of degrees of freedom . at low temperatures @xmath64 , harmonic contributions are dominating , and the difference between @xmath57 and @xmath58 is approximately @xmath65 , which is subtracted from @xmath57 in figure [ fig1 ] . figure [ fig1 ] shows that the isl approach is able to capture the main features of the thermodynamics of the g - model of the src - sh3 domain . the value of the folding temperature is correctly determined by the isl approach , but @xmath38 underestimates the maximum by more than @xmath66 if the highest point of @xmath57 is selected as a reference . this discrepancy at the maximum had previously also been observed for the inherent structure analysis of the ig binding domain of protein g @xcite . on the other hand , towards higher temperatures , @xmath58 decays slower than @xmath67 . for annotation.the size of the energy bins for the density estimate is @xmath68 . , title="fig:",width=249 ] for annotation.the size of the energy bins for the density estimate is @xmath68 . , title="fig:",width=347 ] ubiquitin , with its 76 amino - acids , is a protein with a fairly rich secondary structure since it contains a @xmath43 helix and 5 @xmath0 sheets . similarly to the src - sh3 domain , the ubiquitin g - model presents a sharp folding transition associated with a large peak in the specific heat ( see right - hand side of figure [ fig2 ] ) . the specific heat @xmath57 was estimated from averages on 8 initial conditions per temperature step , and @xmath69 was obtained using @xmath70 minima from several independent trajectories close to the folding temperature using a histogram of @xmath71 bins . the agreement between the full thermodynamics and the isl approach is better than for src - sh3 , though similar trends of discrepancies can be observed . for annotation . the size of the energy bins for the density estimate is @xmath72.,title="fig:",width=241 ] for annotation . the size of the energy bins for the density estimate is @xmath72.,title="fig:",width=355 ] to contrast with sharp two - state transitions of protein g ( 56 residues ) , src - sh3 ( 56 residues ) and ubiquitin ( 76 residues ) , we selected smaller structures , the ww domain ( 37 residues ) and the trp - cage ( 20 residues ) , to examine the performance of the inherent structure approach for less structured proteins , showing a broader transition . for such small protein domains , the validity of the g - model can be questioned as the model is built from the geometrical structure of the folded state . for small molecules the discrimination between folded and unfolded states becomes subtle due to fluctuations covering a large part of the accessible configurational space in a broad range of temperatures . the point in selecting these structures is not to asses the validity of the g - model itself , but to test the isl approach in very stringent cases to highlight possible limitations . the density of states @xmath69 was obtained from @xmath73 minima from several independent trajectories slightly above the folding transition ( see figure [ fig3 ] ) using a histogram of @xmath71 bins . in contrast to the two previous cases , the histogram of minima does not show a clear separation of basins of local minima associated to the folded / unfolded state . we observe a difference in the apparent shape of the density of states ( left - hand side of figure [ fig3 ] ) , which is globally concave in contrast to the convex densities obtained for src - sh3 and ubiquitin . the same shape was also found for protein g in @xcite , for which the relation between the concave shape and the two - hump structure of @xmath74 was discussed in the vicinity of the folding temperature . moreover , by comparing the insets of the left - hand side of figs . [ fig1]-[fig3 ] , one observes that the low energy range of @xmath69 is more discrete , and states tend to lie less densely packed . + the temperature dependence of @xmath57 was obtained by averaging over 12 initial conditions . an interesting feature of the curve is the shoulder in the low temperature range which indicates a partially unfolded structure associated to the breaking of a small number of contacts . comparing the results of @xmath57 and @xmath58 , it is apparent that though the specific heat reconstructed from inherent structure thermodynamics correctly captures the global shape of @xmath57 , including the existence of the shoulder , important deviations can be observed . similarly to the cases analyzed above , @xmath58 underestimates @xmath67 at lower temperatures while giving an overestimation at high temperatures . in contrast to the results for larger proteins , we also observe a significant shift of the transition temperature . for annotation . the size of the energy bins for the density estimate is @xmath75 . , title="fig:",width=241 ] for annotation . the size of the energy bins for the density estimate is @xmath75 . , title="fig:",width=355 ] with only 20 residues , the last protein fragment studied in this series is also the smallest , and mainly consists of a single @xmath43-helix . its inherent structure density of states @xmath69 was estimated from @xmath76 minima sampled from several independent trajectories close to the folding temperature using a histogram of @xmath71 bins . for such a small system , the low lying energy states are largely separated from each other , and the resulting density of states presents large gaps in a relatively broad range of energies . the continuum representation assumed for the inherent structure landscape is certainly questionable in such a case . the equilibrium thermodynamics were constructed from averages on 10 runs over different initial conditions . it is interesting to notice that the value of the specific heat computed with reduced thermodynamics is still fairly close to the actual specific heat , although we notice again that the peak is underestimated and a high temperature tail is observed as in the previous cases . in contrast to the results for the ww domain , the folding temperature of the trp - cage protein domain is correctly found by the analysis of inherent structures . our studies of four proteins , combined with the earlier results on protein g @xcite , allow us to describe some trends in the inherent structure analysis of g - model proteins . + for the density of states given by ( [ ihs ] ) , a general exponential dependence , @xmath77 is observed for all proteins , with slightly different slopes for the low energy states , corresponding to states occupied in the folded configuration , and for the high energy states , occupied in the unfolded configuration . the value of @xmath78 associated to the low energy range is a good estimate of the folding temperature , as previously reported for the case of protein g @xcite . figure [ newfig ] ( left - hand panel ) , which compares the density of states of the inherent structures for the four proteins shows that , when being presented in reduced units as a function of @xmath79 , the functional form of these densities is highly similar . for the large proteins that we studied , src - sh3 and ubiquitin ( as well as protein g ) , the slope is slightly larger in the high energy range than in the low energy range . the converse is true for the small protein domains ww and trp - cage . a formal calculation of the reduced specific heat @xmath38 from a bi - exponential density of inherent structure energies shows that this property is related to the sharpness of the folding transition . a density of states that is curved downwards for the energies associated to unfolded configurations leads to the broad folding transition expected for small protein domains . including the data for the ig binding domain of protein g from @xcite ; the color coding is : protein g ( black ) , src - sh3 ( green ) , ww domain ( red ) , ubiquitin ( blue ) , trp - cage ( magenta ) . _ right : _ deviation between the specific heat obtained from full thermodynamics and inherent structure analysis . , title="fig:",width=294 ] including the data for the ig binding domain of protein g from @xcite ; the color coding is : protein g ( black ) , src - sh3 ( green ) , ww domain ( red ) , ubiquitin ( blue ) , trp - cage ( magenta ) . _ right : _ deviation between the specific heat obtained from full thermodynamics and inherent structure analysis . , title="fig:",width=302 ] the calculation of the specific heat @xmath38 shows that the isl approach is able to determine the specific heat of a protein with reasonable accuracy , including the overall shape of the folding transition . to obtain such an agreement the ground state probability @xmath80 must be sufficiently well sampled to ensure that the density of inherent states is correctly normalized . however , our studies of several proteins shows that limitations exist since systematic deviations from the full thermodynamics are apparent . at low temperatures , and up to the transition temperature , @xmath58 underestimates the specific heat . the peak of @xmath58 is less pronounced than expected from the equilibrium trajectories , and tends to broaden towards higher temperatures ( @xmath81 ) where @xmath38 is larger than @xmath57 . these deviation are shown in a comparative illustration in the right - hand side panel of figure [ newfig ] for the four proteins . our data also reveals that the g model , with its strong bias towards the native state , does not yield qualitatively differences depending on the secondary structure of the protein under consideration . + owing to the results found for various proteins which show systematic deviations from the results of equilibrium thermodynamics , it is important to examine the assumptions made in approximating the inherent structure of states , which we do in the following section . in section [ section2 ] we introduced the major assumptions ( a1 ) and ( a2 ) of the isl approach . the derivation of thermodynamic quantities such as the specific heat is carried out as if the free energy contribution within a basin of attraction , defined by ( [ fdefine ] ) , did not depend on the particular inherent structure @xmath2 . it is difficult to test this assumption as it would in principle require the determination of the complete basin on the energy landscape , including the calculation of all the saddle points that determine the frontier of the basin as well as the shape of the basin within this frontier . still , one can at least compute @xmath82 _ in the harmonic approximation , _ as done also in @xcite . assuming that the contribution to the integral ( [ fdefine ] ) can be approximated by local normal modes in the vicinity of the energy minimum , the effective free energy can be written as @xmath83 where @xmath84 are the normal mode frequencies at the ground state @xmath85 . this expression allows us to calculate a harmonic approximation to @xmath16 by calculating the normal modes for each minimum that we sample , and subsequently summing their different contributions according to ( [ harmapprox ] ) . figure [ fig5 ] shows the @xmath86-dependent part @xmath87 as a function of the energy minimum for the two examples of src - sh3 domain ( left ) and ww domain ( right ) . the minima were obtained along a single trajectory close to the folding temperature . the distribution of the minima along the energy axis reflects the character of the probability distribution function for the two proteins , one being divided into two basins ( src - sh3 domain ) , the other being a single distribution ( ww domain ) . of the vibrational free energy @xmath88 in the harmonic approximation , calculated from a sample of 1000 minima . _ left : _ src - sh3 domain . _ right : _ ww domain . an arbitrary offset was added to shift the origin of the ordinate . , title="fig:",width=302 ] of the vibrational free energy @xmath88 in the harmonic approximation , calculated from a sample of 1000 minima . _ left : _ src - sh3 domain . _ right : _ ww domain . an arbitrary offset was added to shift the origin of the ordinate . , title="fig:",width=294 ] as a first important observation , we note that the variation of @xmath87 with @xmath89 can not be assumed to be negligible as assumed in ( a2 ) , according to which @xmath90 should be approximately constant in @xmath89 . both proteins show the same trend toward a decreasing effective free energy with increasing @xmath89 . for the src - sh3 domain , a nonlinear dependence can be observed in the high energy range in agreement with previously reported results on protein g @xcite . + a second point to be noticed is that , for a given @xmath89 , a distribution of values of @xmath91 can be found ( variance on the y - axis in figure [ fig5 ] ) . while such a variance ( or fluctuation ) could be attributed to the limited numerical accuracy of the normal modes at high energies , this is also true for the low energy part for which the numerical scheme provides accurate results , as can be checked on the vanishing six lowest eigenmodes . consequently , it appears that the first approximation ( a1 ) , i.e. @xmath92 does not hold as assumed , leaving the possibility that the observed deviations in section [ section3 ] could be caused by this simplification . the results of the previous subsection seem to challenge the validity of the isl approach performed under the assumptions ( a1 ) and ( a2 ) . at a first glance , the main problem seems to arise from the approximation ( a2 ) that @xmath93 does not depend on the particular basin considered , which is obviously untrue . on the other hand , although there is clearly a variance associated to @xmath94 for the same energy range @xmath89 in disagreement with approximation ( a1 ) , one can observe a general evolution of @xmath93 with @xmath89 , which suggests that approximating @xmath93 by @xmath95 may be acceptable . however , we show below that , if the free energy within a basin can be written as a sum @xmath96 with @xmath97 at the ground state , the calculation of @xmath38 _ can be carried out without any change , _ so that approximation ( a2 ) appearing to be to particularly bad at a first glance may not be the decisive one . it should be noticed that , as discussed in the previous section , the property ( [ separation ] ) is verified if the motion in each basin of attraction can be described by a combination of harmonic vibrations . + starting again from ( [ eq1 ] ) , and proceeding as in section [ ihstheory ] , we can eliminate the part of @xmath16 that depends on temperature only in the expression for the partition function and the probability distribution function , keeping only the @xmath43-dependent part . one gets @xmath98 the principle of the calculation is to compute @xmath99 form a _ measurement _ of the probability density function @xmath100 estimated from md simulations at a given temperature @xmath33 . this can be achieved through the intermediate calculation of a density of states of inherent structures @xmath69 , which is temperature independent and from which @xmath99 can be obtained at all temperatures . in equation ( [ newzis ] ) , we notice that the inclusion of the term @xmath101 is equivalent in definition an effective density of states @xmath102 from which the classical thermodynamic expression can be derived as shown below . in the calculation of section [ ihstheory ] , the density of states is given by equation ( [ eq : piszis ] ) . this density @xmath103 and all other observable calculated in section [ ihstheory ] will henceforth denote with and index @xmath104 . in the new scheme including the @xmath43-dependent part of the free energy in the harmonic approximation , the probability density @xmath74 becomes @xmath105 yielding the inherent structure density @xmath106 the latter expression shows that if the variation of @xmath16 with @xmath89 can not be ignored , the previously derived density of states @xmath107 is not the correct one . the two are related by @xmath108 a similar result was also reported in @xcite which considered the particular case of a piecewise linear dependence on @xmath89 . though the densities differ , substituting ( [ relatedens ] ) in equations ( [ newzis ] ) or ( [ eq : piszis ] ) , we immediately have @xmath109 , and the inherent structure observables such as @xmath110 and @xmath111 are unchanged , i.e. @xmath112 and @xmath113 even when @xmath114 . + as a consequence , using equation ( [ newz ] ) , the full free energy @xmath115 of the protein can be written as @xmath116 therefore , in the isl formalism , taking into account the variation of @xmath16 as in equation ( [ separation ] ) does not alter the free energy and can not be expected to be at the origin of the quantitative differences between @xmath55 derived from the isl formalism and the full numerical results presented in section [ section3 ] . we conclude that the origin of these discrepancies is likely to be found in the non - separability between the @xmath43 and the @xmath29 dependency within the basins . such an non - separability can be expected as soon as the anharmonicity of the different basins is taken into account . this is certainly relevant for proteins , in particular as the denaturation involves frequent transitions between basins of different shape and volume associated to the semi - rigid folded and the highly flexible unfolded state . + when @xmath16 can not be separated into @xmath43- and @xmath29-dependent contributions to simplify the calculation of the partition function , the remaining possible approximation that can tackle the computational difficulty would be the saddle point approximation of the free energy @xcite . this approximation is acceptable in the thermodynamic limit for large systems , but can not be justified in the present problem as the number of particles involved is still small and the interactions between particles are heterogeneous . while the harmonic approximation seems to be invalid for the present problem which involves large conformational changes due to the denaturation transition , results on super - cooled liquids @xcite indicate that the correction of the heterogeneity of the basins at low temperatures is small and the decoupling approximation of vibrational and inherent structure contributions appears to be possible at least in these temperature regimes . the main practical difficulty of the isl method comes from the need to properly sample all the inherent structures in order to get a meaningful density of states @xmath69 in all energy ranges . in the present scheme of inherent structure sampling , a single temperature is selected to simulate the dynamics of the protein for a finite period of time . the choice of a temperature close to the folding temperature is natural as the protein samples both the folded and the unfolded configurations at this temperature . there are however two questions that have to be answered : _ i ) _ how long should we follow a protein md trajectory to get a sufficient sampling ? _ ii ) _ is it possible to combine data from simulations at a few different temperatures instead of keeping @xmath29 fixed ? + let us first analyze the effect of the sampling time . it should be chosen long enough to cover the slowest intrinsic timescale of the system , and various trajectories with different initial conditions or realizations of the thermostat should be used to ensure that the order of events does not alter the shape of the distribution . an estimate of the time range that sampling must cover is provided by the folding / unfolding time of the protein , to guarantee that the protein explores both configurational subspaces . a possible check of the choice of the simulation time is to compute the specific heat with an increasing number of samples , and stop when the improvement brought by additional samples is negligible . still , as computer time is limited , it can not be ensured that all relevant states are sufficiently well sampled to yield a converged probability density . in particular , the high energy minima are sampled only with low probability , such that the high energy cut - off in the density of states is likely to be underestimated in finite time sampling . in this section , we analyze the impact of this cutoff on a model density to see how the inherent structure specific heat is possibly affected . to analyze the effect of the sampling independently of a particular case , let us assume a model inherent structure density of states taken as a single exponential @xmath117 similar to the shape of the densities that can be found in limited ranges of energies for the numerical results in section [ section3 ] . the partition function than can be readily calculated as @xmath118 likewise , we can calculate the first two moments @xmath119 and @xmath120 to find the specific heat as a function of the temperature , the parameter @xmath121 and cutoff in energy @xmath122 @xmath123 and analyze the result graphically as a function of the energy cutoff @xmath122 in figure [ fig6 ] . as a function of temperature and the cutoff parameter @xmath122 . _ right : _ location of the maxima of @xmath38 for different values of the cutoff @xmath122.,title="fig:",width=302 ] as a function of temperature and the cutoff parameter @xmath122 . _ right : _ location of the maxima of @xmath38 for different values of the cutoff @xmath122.,title="fig:",width=302 ] as can be inferred from left - hand side of figure [ fig6 ] , @xmath124 $ ] increases with higher cut - off @xmath122 . using symbolical computation @xcite , we can further inspect the result to find the local maxima of the the specific heat for fixed cutoff @xmath122 . on the right - hand side of figure [ fig6 ] , we observe that a lower cutoff in the density of states shifts the maximum of the specific heat towards higher temperatures . in addition to the shift , the curve becomes broader and the value at the maximum decreases . this situation is similar to the physical scenario when protein folding is altered by confinement ( see e.g. @xcite ) . the high energy states disappear from the density of states as the system is prohibited to explore these by external forces . for the present purpose of the inherent structure analysis , although derived for a highly idealized model density , the results indicate that insufficient sampling at high energies can significantly alter the global shape of the transition . we have checked that these conclusions remain unchanged for a piecewise constant density of states . for instance , for the ww domain , one finds @xmath125 which is higher than the range of values for which a shift of the maximum can be expected from figure [ fig6 ] . consequently , the origin of this shift can not be attributed to inefficient sampling in the high energy range . + because the high energy minima are less frequently visited , it is tempting to try to sample the minima from a high temperature molecular dynamics trajectory . on the other hand , as the method relies on the probability to occupy the ground state which determines @xmath126 , it is also necessary to properly sample the ground state , i.e. to select a simulation temperature which is below @xmath1 . to reconcile these two exclusive conditions , one solution is to combine results sampled at two different temperatures to calculate @xmath69 , which should be temperature independent . this is possible because , according to ( [ eq : pz ] ) @xmath127 with @xmath128 , so that the ratio of @xmath129 can be calculated from the probabilities to occupy a basin at temperatures @xmath130 and @xmath131 . a molecular dynamics trajectory obtained at a temperature @xmath132 can be used to determine @xmath133 from the probability to occupy the ground state , and subsequently a second simulation at a higher temperature @xmath131 can sample high energy basins more efficiently . for all basins which are properly sampled in both molecular dynamics runs , the ratio @xmath129 can be evaluated with ( [ eq : zratio ] ) . although it should not depend on the particular basin that was used for its calculation , this ratio actually fluctuates around a mean value which can be used to determine @xmath134 from @xmath135 . then ( [ eq : pz ] ) , applied at the higher temperature @xmath131 , can be used to compute @xmath69 in the high energy range . moreover , in the intermediate energy range , the basins are properly sampled by the two trajectories at temperatures @xmath130 and @xmath131 , which gives two ways to evaluate @xmath69 for those basins , and thus provides a way to check the consistency of the method . ( @xmath136 minima , red ) and @xmath137 ( @xmath138 minima , blue ) _ left : _ probability of occupation of the inherent structures versus @xmath89 . _ right : _ density of inherent states energy ( in logarithmic scale ) , calculated from the simulation at @xmath130 ( red ) and calculated at @xmath131 using the ratio @xmath129.,title="fig:",width=294 ] ( @xmath136 minima , red ) and @xmath137 ( @xmath138 minima , blue ) _ left : _ probability of occupation of the inherent structures versus @xmath89 . _ right : _ density of inherent states energy ( in logarithmic scale ) , calculated from the simulation at @xmath130 ( red ) and calculated at @xmath131 using the ratio @xmath129.,title="fig:",width=266 ] figure [ fig:2 t ] shows the results for the ww domain from different molecular dynamics simulations , simulated respectively at @xmath139 and @xmath140 . for the simulations at @xmath131 , the ground state is not sampled in the finite interval of time of the simulations , but @xmath141 can nevertheless be obtained through the evaluation of @xmath142 deduced from @xmath143 for all values of @xmath89 for which the histograms of the left - hand side of figure [ fig:2 t ] overlap . the right - hand side of figure [ fig:2 t ] shows that the values of the density of inherent state energies computed from data at @xmath130 and @xmath131 are in rather good agreement in the whole energy range where the basins are sampled at the two temperatures . there is however a discrepancy between the two results , with a systematic deviation towards lower values of @xmath69 for high inherent state energies when the density of states is calculated with data sampled at high temperatures . this is counterintuitive as one could tend to believe that sampling the basins at low temperatures would , on the contrary , underestimate the density of basins at high energy . this systematic deviation , which has been observed in all our calculations could point out a limitation of the isl method as presently applied , i.e. with a calculation which is only valid if @xmath144 is the sum of a term that depends on @xmath89 and a term that depends on temperature ( see ( [ separation ] ) ) . when the specific heat is calculated with the value of @xmath69 extended in the high energies range with this method , the function is still very close to the value obtained with only the data at temperature @xmath130 and the agreement with the numerical value of @xmath57 is not improved . this shows that the discrepancy between the exact results and those deduced from the isl approach are not due to technical difficulties such as an insufficient sampling in some energy range , but that they are rather inherent to the method itself together with its assumptions . we applied the inherent structure landscape ( isl ) approach to four different proteins of varying size and secondary structure elements using a coarse - grained off - lattice protein model , and calculated their inherent structure density of states . using these densities , we derived the specific heat from the reduced inherent structure thermodynamics , and compared it to the value obtained from equilibrium molecular dynamics as a function of temperature . our results show that the isl approach can correctly capture the shape of the temperature dependence of the specific heat , including some characteristic features such as the hump observed at @xmath145 for the ww domain . this is remarkable since the result is deduced from molecular dynamics simulations at a single temperature , close to @xmath1 , and nevertheless predicts the main features of the specific heat in a large temperature range , including low temperatures for which very long simulations would be necessary to reach exhaustive sampling of the phase space . this shows that _ many features of protein thermodynamics are encoded in the inherent structure landscape . _ still , the approach is not perfect as we observed quantitative differences between @xmath58 and @xmath57 which are particularly significant for small protein domains . the deviations show a systematic trend , the specific heat being underestimated below @xmath1 and overestimated above . this lead us to reexamine the approximations that enter the construction of the reduced thermodynamics from inherent structures for the model that we considered . the first approximation assumes that the correction to the density of states due to the structure of the basin associated to a minimum @xmath146 depends on the energy level of the minimum only , and not of the individual minimum . an evaluation of @xmath146 using a harmonic approximation based on local normal modes ( section [ nma ] ) shows that this is only approximately correct . for a given @xmath89 , the values of @xmath147 are actually distributed around an average value , with fluctuations that grow for higher values of @xmath89 and that appear to be larger for small proteins . this could explain some of the discrepancies between the isl results and the equilibrium data . moreover , the calculation of @xmath147 indicates that the second assumption that the correction can be considered to be @xmath43-independent is certainly not valid . however , we showed that if @xmath148 splits into a temperature - dependent and an @xmath43-dependent part , which is the case in the harmonic approximation , most of the thermodynamic results deduced from a direct application of the isl approach are not affected . this is true in particular for the the specific heat . in view of the persisting quantitative differences between reduced inherent structure and equilibrium thermodynamics , we therefore conclude that the correction of the free energy in term of a harmonic approximation is not sufficient . it is likely that the nonlinear terms in the free energy associated to a basin can not be ignored , and play a significant role . this is not surprising because , especially in the high temperature range , the protein fluctuates by exploring many basins , and consequently can not be assumed to be well described by a harmonic approximation . + in future studies , it would be useful to analyze the role of the structure of the full basin on the thermodynamic results beyond the approximation by local normal modes around the minima . this is a true challenge owing to the complexity of the energy landscape . a starting point for such a study might be the examination of the distribution of first rank saddle points associated to the different minima on the potential energy surface . a second aspect which is suggested by the present work is to apply the isl approach for protein folding in the context of more complex energy landscapes that arise in more realistic potential energy functions . the results on the small proteins analyzed in this study show that the global separation of the probability density into two basins associated to folded and unfolded states is not a necessary requirement to construct the reduced thermodynamics . it appears therefore likely that the formalism remains useful in cases where the energy landscape is less biased towards the ground state than in the g - model . an application of the isl method to other protein models therefore appears to be desirable and promising . * * acknowledgements : * * we thank the anonymous referee for valuable suggestions . j - g.h . acknowledges financial support from the collge doctoral franco - japonais and the rgion rhne - alpes . the simulations were partially performed at the ple scientifique de modlisation numrique ( psmn ) in lyon . in this work , we analyze the properties of an off - lattice g - type @xcite in which the smallest building unit is a single amino acid . effective interactions between amino acids are based on the reference positions of the @xmath42-carbons of each residue . these interactions are color - blind in the sense that they do not distinguish between the type of amino acids . the potential energy of the system comprises five terms : @xmath149 + \sum_{i > j-3}^{non - native } \epsilon \left(\frac{c}{r_{ij}}\right)^{12}\nonumber\\ & & + \ \sum_{i=1}^{n-3}k_d \left(1-cos\left(2\phi_i-\frac{\pi}{2}\right)\right ) \end{aligned}\ ] ] here , @xmath150 denotes the euclidean distance between residues @xmath151 . the @xmath152 degrees of freedom of the system are most conveniently expressed via internal coordinates : @xmath153 euclidean bond distances @xmath154 along the backbone , @xmath155 bond angles formed between two consecutive bond directions , and @xmath156 dihedral angles measured between the normal vectors of planes spanned by atoms @xmath157 and @xmath158 ) . bond distance and bond angle interactions are modelled through harmonic forces with coupling strengths @xmath159 and @xmath160 respectively . the zero indices indicate that the quantities ( angles , distances ) are evaluated in the reference state , i.e. , the position from the nmr / crystallographic structure . the dihedral angle potential does not assume a minimum in the reference position defined by the experimentally resolved structure : it favors angles close to @xmath161 and @xmath162 irrespective of the secondary structure element ( helix , sheet , turn ) the amino acid belongs to . while such values can be found in @xmath43-helices , they statistically do not appear largely in other secondary structure elements . as a consequence , the reference state defined by an nmr or crystallographic structure can give rise to a competition between the helix for which the reference state is close to its minimum energy , and other parts which experience forces due to the angular constraint . the constraint was introduced as a source of additional frustration affecting the dynamics and thermodynamics of the model towards a more realistic representation @xcite . attractive non - bonded native interactions are modelled by steep ( 12,10)-lennard - jones potentials accounting for non - local contacts between residues due to side chains . the database of contacts is established as described in section [ section2 ] . in addition , non - native repulsive interactions are added to those residues which do not form a contact and which lie at least four residues apart . the dimensionless parameters used in this study are @xcite : @xmath163 , @xmath164 , @xmath165 , @xmath166 , @xmath167 . 99 n. nakagawa and m. peyrard ; proc . usa * 103 * , 5279 ( 2006 ) p.w . fenimore , h. frauenfelder , b.h . mcmahon and r.d . young ; proc . usa * 101 * , 14408 ( 2004 ) f.h . stillinger and t.a . weber ; phys . a * 25 * , 978 ( 1982 ) j. kim and t. keyes ; j. phys . b * 111 * , 2647 ( 2007 ) a. imparato , s. luccioli , a. torcini ; phys . lett . * 99 * , 168101 ( 2007 ) d.m . ming , m. anghel and m.e . wall ; phys . e * 77 * , 021902 ( 2008 ) z. guo and d. thirumalai ; j. mol . 263 * , 323 ( 1996 ) a. baumketner , j .- e . shea and y. hiwatari ; phys . rev . e * 67 * , 011912 ( 2003 ) n. nakagawa and m. peyrard ; phys . e * 74 * , 041916 ( 2006 ) j. karanicolas and c.l . brooks iii ; proc . usa * 100 * , 3954 ( 2003 ) d.k . west , d.j . brockwell , p.d . olmsted . s.e . radford and e. paci ; biophys . j. * 90 * , 287 ( 2006 ) http://www.rcsb.org w.l . delano ( 2002 ) ; _ the pymol molecular graphics system . _ ; http://www.pymol.sourceforge.org v. sobolev _ et al . _ ; bioinformatics * 15 * , 327 ( 1999 ) a. brnger , c.l . brooks iii and m. karplus ; chem . phys . lett . * 105 * , 495 ( 1984 ) n. nakagawa ; phys . lett . * 98 * , 128104 ( 2007 ) mathematica 5.0 , wolfram inc . n. rathore , t.a . knotts , j.j . de pablo ; biophys . j. * 90 * , 1767 ( 2006 )

model predictive control ( mpc ) is a consolidated control technique that can efficiently handle constraints on the process to be controlled . nevertheless , its application is not yet widespread in many domains where real - time computational constraints and requirements of certified solutions are of major concern , such as aerospace or automotive applications . there is a growing interest in both industry and academia for exploring parallel solutions to mpc problems ( @xcite ) , especially in light of the emerging many - core architectures , aiming to improve the computational efficiency of solving the underlying optimization problem . * contribution*. in this paper , we explore the use of parallelization techniques to efficiently solve a typical mpc problem for a linear discrete - time system , with a substantial computational speedup compared to nonparallel implementations . our proposed algorithm combines the use of alternating direction method of multipliers ( admms @xcite , @xcite ) to handle the coupling constraints that arise from the dynamics of the system and inexact solvers ( i.e. , solvers that guarantee feasibility and optimality only _ asymptotically _ with the number of iterations ) , such as the nesterov s dual fast gradient ( dfg ) method @xcite . in particular , the first step of the proposed algorithm is to split the original mpc problem over the length @xmath0 of the prediction horizon into @xmath1 independent subproblems ( _ time - splitting _ @xcite ) solved by @xmath1 parallel _ workers _ periodically exchanging information at predetermined synchronization points . then , the second step is to solve these subproblems in parallel using an inexact solver and guarantee , at the same time , that the solution of the original mpc problem is _ recursively feasible _ and the system is _ closed - loop stable_. the combination of parallelization and inexact solvers can result in infeasibility and closed - loop instability . we rely on an algorithm based on constraint tightening to overcome these issues . loosely speaking , constraint - tightening algorithms solve an alternative problem in which the constraints have been tightened by a certain amount to compensate for the accuracy loss ( and possible related infeasibility ) introduced by the solver . we rely on an _ adaptive _ tightening strategy to select an appropriate amount of tightening for our algorithm . every time new measurements are available from the plant , our algorithm chooses the amount of tightening required for each subproblem in order to compensate for the error introduced by the time - splitting combined with the inexact solver . * related work*. the _ time - splitting _ technique has been proposed in @xcite . in contrast to @xcite , we combine admm with inexact solvers and focus on the requirements for recursive feasibility and closed - loop stability of the original problem . other constraint - tightening schemes have been proposed in the literature ( outside the parallel framework ) . for example , the authors in @xcite propose an algorithm in which the amount of tightening is chosen offline to guarantee suboptimality and feasibility of the solution for all the initial states of the mpc problem . solutions based on adaptive constraint tightening have been recently proposed in @xcite , where the tightening parameter is chosen adaptively . compared to @xcite , our tightening update rule allows for a nonuniform amount of tightening ( the tightening varies along the prediction horizon ) . furthermore , thanks to the modular structure of our approach , the optimizer solves simpler problems of fixed dimension , which is independent from @xmath0 . as a consequence , an increase of @xmath0 does not affect the conditioning of the problem and the convergence of the solver . hence , our approach leads to a performance improvement even when forcing full serialization of the parallel operations ( i.e. , _ serialized _ mode @xcite ) . _ * outline*_. in the following , section [ sec : problem_formulation ] presents the initial problem formulation . section [ sec : subproblem_reformulation ] introduces the auxiliary subproblems and our proposed solver . section [ sec : tightening_improvements ] describes our strategy to select the tightening of each subproblem to handle the parallelization error . section [ sec : closed_loop_certification ] proposes an online update strategy of the tightening parameters that guarantees recursive feasibility , suboptimality , and closed - loop stability . section [ sec : evaluation ] presents numerical results using an academic example . finally , section [ sec : conclusion ] concludes the paper . _ * notation*_. for @xmath2 , @xmath3 is the euclidean norm and @xmath4_+$ ] is the projection onto non - negative orthant @xmath5 . given a matrix @xmath6 , @xmath7_i$ ] denotes the @xmath8-th row of a and @xmath7_{i , j}$ ] the entry @xmath9 of a. furthermore , @xmath10 is the vector of ones in @xmath11 and @xmath12 the identity matrix in @xmath13 . in addition , @xmath14 and @xmath15 denote the largest and the smallest ( modulus ) eigenvalues of the matrix a , respectively . @xmath16 denotes that @xmath17 is positive definite . consider the discrete - time linear system described below : @xmath18 where @xmath19 denotes the state of the system and @xmath20 denotes the control input . the sets @xmath21 and @xmath22 are simple proper convex sets ( i.e. , convex sets that contain the origin in their interior ) . our goal is to steer @xmath23 to the origin and satisfy the plant constraints . we use mpc to achieve these objectives . in this respect , consider the following finite - time optimal control problem : [ eq : initial_problem ] @xmath24 where @xmath25 and @xmath26 are more compact notations for @xmath23 and @xmath27 , respectively . for @xmath28 ( @xmath0 denotes the prediction horizon ) , the states and the control inputs are constrained in the polyhedral set described by , where @xmath29 , @xmath30 , @xmath31 . note that can include constraints on the state only , i.e. , @xmath32 , and/or constraints on the control inputs only , i.e. , @xmath33 . in , @xmath34 and @xmath35 . our problem formulation considers also a terminal cost @xmath36 associated with a terminal polyhedral set @xmath37 . through the remaining of the paper , we assume : [ ass : stabilizability ] the pair @xmath38 is stabilizable . [ ass : terminalset ] suppose assumption [ ass : stabilizability ] holds . given the gain @xmath39 obtained by the infinite - horizon linear quadratic regulator ( ih - lqr)characterized by the matrices @xmath6 , @xmath40 , @xmath41 , and @xmath42the following holds : @xmath43 in addition , the terminal penalty @xmath44 in the stage cost is defined by the solution of the algebraic riccati equation associated with the ih - lqr . in general , the mpc controller solves the optimization problem every time new measurements are available from the plant and returns an optimal sequence of states and control inputs that minimizes the cost . let the optimal sequence be defined as follows : @xmath45 only the first element of @xmath46 is implemented in closed - loop , i.e. , the control law obtained using the mpc controller is given by : @xmath47 and the closed - loop system is described by @xmath48 we aim to solve problem in parallel . hence , we exploit a similar approach as the one proposed in @xcite . specifically , as in @xcite , problem is decomposed along the length of the prediction horizon @xmath0 into @xmath1 independent subproblems to be solved by @xmath1 _ parallel workers _ @xmath49 ( @xmath50 ) . each @xmath49 is allowed to communicate with its neighbours @xmath51 and @xmath52 at predefined synchronization points . the decomposition is possible thanks to the introduction of @xmath0 auxiliary variables @xmath53 used to break the dynamic coupling that arises from . these @xmath54 can be seen as the global variables of the algorithm . in particular , each @xmath54 stores the local predicted state @xmath55 of each subproblem and exchanges this stored information to guarantee consensus between neighboring subproblems , i.e. , to ensure that the predicted state of the @xmath56-th subproblem , namely @xmath57 , is equal to the current state of the @xmath58-st subproblem , namely @xmath59 . specifically , by introducing the _ consensus constraints _ @xmath60 , defining @xmath61^t$ ] , @xmath62 $ ] , @xmath63 $ ] , @xmath64 problem becomes : [ eq : consensus_problem ] @xmath65 where , defining @xmath66^t$ ] : * @xmath67 , where @xmath68 . * @xmath69 , @xmath70 , . * @xmath71 , where @xmath72 . furthermore , @xmath73 and @xmath74 vary for each subproblem as follows : * @xmath75 $ ] and @xmath76 , @xmath77 . * @xmath78 $ ] and @xmath79 . note that we introduced a quadratic penalty in the cost of the form @xmath80 , according to the admm strategy @xcite . this penalty has no impact on the cost of the original problem , if the consensus constraints are satisfied . in the following , we introduce the subproblems that derive from problem . let @xmath81 ( @xmath28 ) and @xmath82 ( @xmath83 ) be the lagrange multipliers associated with the equality constraints and , respectively . then , let the augmented lagrangian with respect to the multipliers @xmath81 and @xmath84 be defined as follows : @xmath85.\end{aligned}\ ] ] hence , we obtain the following @xmath1 independent subproblems , called _ original _ subproblems , associated with the @xmath1 workers @xmath49 ( @xmath50 ) : @xmath86 figure [ fig : steps ] summarizes the main steps that lead to a suboptimal solution of the aforementioned problem and introduces the keywords used in the remaining of the paper . consider the mpc problem and refer to this problem as the _ original _ mpc problem . the first step ( _ parallelization _ , detailed in section [ subsec : parallelization ] ) is to rewrite the original problem in @xmath1 independent subproblems ( the _ original _ subproblems ) . the aim is to use a dual fast gradient ( dfg ) method to solve these subproblems in order to certify in terms of suboptimality and recursive primal feasibility , along with closed - loop stability the mpc solution . the use of the inexact solver will eventually cause a violation of the consensus constraints - introduced to define the subproblems . hence , the second step ( _ relaxation equality constraints _ detailed in section [ subsec : equality_relaxation ] ) is introduced to relax the consensus constraints by a quantity indicates that @xmath87 varies along the prediction horizon . this also holds for the later - defined @xmath88 , @xmath89 , and @xmath90 . ] @xmath87 , preventing the occurrence of consensus - constraint violations . we refer to these subproblems as the _ equality relaxed _ ( er ) subproblems . the set of inequality constraints of the er subproblems includes the inequality constraints of the original subproblems and the inequality constraints due to the relaxation of the consensus constraints - . the third step ( _ definition tightening parameters _ detailed in sections [ subsec : up_lagrange_multiplier ] , [ sec : tightening_improvements ] , and [ sec : closed_loop_certification ] ) is required to address the following remaining issues . first , the solution of each er subproblem computed by the dual fast gradient method might violate the inequality constraints , due to the termination of the solver after a finite number of iterations . hence , the constraints of the er subproblems must be tightened by a quantity @xmath88 proportional to the desired level of suboptimality @xmath90 chosen by the algorithm . second , due to the relaxation of the consensus constraints , the _ consolidated _ prediction , i.e. , the predicted evolution of the state computed ( a posteriori ) using the control sequence obtained by the independent subproblems , might deviate from the predicted _ local _ solution computed by the independent subproblems and , eventually , violate the inequality constraints of the original problem . hence , an additional tightening ( dependent on @xmath87 ) must be introduced on the subset of inequality constraints of the er subproblems that corresponds to the original inequality constraints . the proposed algorithm addresses the aforementioned issues by exploiting the _ inequality tightened _ ( it ) subproblems . the it subproblems differ from the er subproblems in the definition of the feasibility region , which is tightened by a quantity @xmath91 that depends on both @xmath88 and @xmath87 . the last step ( _ set solver accuracy _ ) selects a suboptimality level @xmath90 for each subproblem , that guarantees a primal feasible and suboptimal solution for the original mpc problem within a fixed number of iterations . in the following , we introduce the er subproblems and the proposed algorithm to solve them using @xmath1 parallel workers . furthermore , we introduce an initial formulation of the it subproblems and derive conditions on the choice of the relaxation and tightening parameters to guarantee primal feasible and suboptimal solutions for of each subproblem . our goal is to obtain a solution for problem by solving the independent subproblems in parallel using inexact solvers , such as the nesterov s dfg @xcite . in order to use our proposed solver ( introduced in section [ sec : alg1 ] ) , which relies on first - order methods , we introduce a reformulation of problem to take into account that the constraints and can not be satisfied at the equality due to the iterative nature of the proposed solver and its asymptotic convergence properties . in particular , introducing the relaxation parameters @xmath92 , for each subproblem ( @xmath28 ) , the former equality constraints - are replaced by the following inequality constraints : or , in a more compact notation : @xmath94 where @xmath95 and @xmath96 . in the remaining of the paper , we consider the following _ equality relaxed _ ( er ) subproblems : @xmath97 hence , let @xmath98^t\in \mathbb r^{p_{\xi_t}}_+$ ] be the lagrange multiplier associated with the new set of inequality constraints defined by , where @xmath99 , @xmath100 @xmath101 , @xmath102 , and @xmath103 are the multipliers associated with the constraints , , , , and , respectively . then , to handle the complicated constraints , define , for each subproblem , the dual function @xmath104 as follows : @xmath105 where @xmath106 . we refer to as the _ inner subproblem_. hence , we aim to solve the following dual subproblems ( _ outer subproblem _ ) in parallel to obtain a solution for problem : @xmath107 [ rem : modularity ] the size of the er subproblems remains unaffected if @xmath0 increases . intuitively , this _ modularity _ is an additional feature that can be exploited to preserve some favorable numerical properties of the problem ( e.g. , conditioning , lipschitz constant , etc . ) even when the algorithm is running in serialized mode . this section introduces algorithm [ alg : alg1 ] that we use to solve the mpc problem exploiting @xmath1 parallel workers @xmath49 . note that , at this stage , we can not yet ensure that the computed solution is feasible and suboptimal for problem . algorithm [ alg : alg1 ] relies on nesterov s dfg in which the inner problem is solved in an admm fashion , as explained below . specifically , at each iteration of the algorithm_{1:n}=x_0 $ ] and @xmath108 to initialize algorithm [ alg : alg1 ] . ] , @xmath49 computes a minimizer @xmath109 for @xmath110 ( steps 1 - 4 ) , i.e. , the algorithm returns a solution for each inner subproblem . in particular , our algorithm , in compliance with the admm strategy , first minimizes @xmath111 with respect to @xmath112 in parallel for each subproblem ( step 1 ) . then , using the information received by @xmath52 , i.e. , the updated value of @xmath113 ( synchronization step 2 ) , our algorithm computes in parallel for each subproblem the value of the global variable @xmath54 according to the following rule ( step 3 ) : @xmath114 note that this strategy allows to handle the coupling introduced by the 2-norm in the cost function of . then , ( synchronization step 4 ) @xmath49 receives ( sends ) the updated value of @xmath115 ( @xmath116 ) from @xmath52 ( to @xmath117 ) , respectively . finally , the worker @xmath49 computes the new values of the multipliers @xmath118 ( steps 5 - 7 ) . we compute offline ( for each subproblem ) the lipschitz constant @xmath119 associated with @xmath120 to perform the multipliers update : * @xmath121 . * @xmath122 , ( @xmath70 ) . * @xmath123 . [ sec : alg1 ] of algorithm [ alg : alg1 ] . ] note that our update rule is different from the one proposed in @xcite , where admm is used in combination with nesterov s fast gradient methods . at each iteration , the algorithm proposed in @xcite , first computes the exact minimizer @xmath112 and then updates @xmath124 and @xmath84 . our algorithm does not wait until the dfg returns a minimizer @xmath112 to update the multipliers @xmath124 and @xmath84 , but starts updating their values along with the dfg iterations , encouraging the information exchange between neighboring subproblems . this algorithm is also different from the one proposed in @xcite . in particular , the workers exchange the necessary pieces of information before the update of the lagrange multipliers and none of the dual variables is exchanged between the neighboring workers . furthermore , the information exchange between neighboring workers is unidirectional , i.e. , @xmath52 sends the updated information to @xmath49 , but @xmath49 does not send any updated information to @xmath52 . using an argument similar to the one of theorem 1 in @xcite , we can compute the primal feasibility violation and the level of suboptimality of the solution of each er subproblem returned by algorithm [ alg : alg1 ] . [ thm : thm1 ] ( @xcite ) let @xmath125 be strongly convex , the sequences ( @xmath126 , @xmath127 , @xmath128 ) be generated by algorithm [ alg : alg1 ] , and @xmath129 . then , an estimate on the primal feasibility violation for the er subproblem is given by the following : @xmath130_+\|\le \frac{8r_{t}{\max\{l_{\mu_t}\}}}{(k+1)^2}=:\eta_t,\ ] ] where @xmath131 . moreover , an estimate on primal suboptimality is given by the following : @xmath132 algorithm [ alg : alg1 ] terminates after a fixed number of iterations that depends on @xmath90 and @xmath133 @xcite : @xmath134 in order to guarantee the primal feasibility of each subproblem using algorithm [ alg : alg1 ] , we introduce @xmath1 auxiliary subproblems , namely the _ inequality tightened _ ( it ) subproblems , which differ from the er subproblems in the definition of the feasible region . in particular , each it subproblem can be defined as follows : @xmath135 where @xmath136 is the tightening parameter , which depends on the suboptimality level @xmath90 that the proposed algorithm can reach within @xmath137 iterations . according to @xcite , solving using algorithm [ alg : alg1 ] ensures , with a proper choice of @xmath88 , that the solution of is primal feasible and suboptimal for subproblem . to define an @xmath88 similar to the one introduced in @xcite , we must compute an upper bound for the optimal lagrange multiplier , namely @xmath138 , associated with the it subproblem . we use an argument similar to the one of lemma 1 in @xcite . in particular , we compute the aforementioned upper bound for @xmath138 according to the following lemma . [ lem : lemma_up_bound ] assume that there exists a slater vector @xmath139 such that @xmath140 . then , there exists @xmath141 , @xmath142 , @xmath143 , such that the upper bound for @xmath138 is given by @xmath144_j\}}},\ ] ] where @xmath145 [ 2\rho(\epsilon_{z_t}-\epsilon_t)\operatorname{\mathbf{1}}_{n}^t ] [ 2\rho(\epsilon_{z_{t+1}}-\epsilon_t)\operatorname{\mathbf{1}}_{n}^t]\big]^t\in \mathbb r^{p_t+2n}$ ] , and @xmath146 is the dual function for the original subproblem evaluated at @xmath147 . see appendix [ app : lemma1 ] . see appendix i in @xcite . lemma [ lem : lemma_up_bound ] does not only provide an upper bound for @xmath148 , but it also provides guidelines to select the values of @xmath149 and @xmath150 as a function of @xmath151 , which only depends on the primal variable @xmath152 . an alternative way to determine the relaxation parameters is to include @xmath149 and @xmath150 in the set of decision variables and penalize them in the cost function as it is usually done to handle soft constraints . this will , however , increase the number of decision variables in the problem formulation and it will have an impact on the original cost . , the blue color refers to subproblem handled by worker @xmath153 , etc . ) . ] the previous section showed how to choose the tightening parameter @xmath88 of each it subproblem to ensure that the @xmath154-th _ local _ solution , i.e. , the solution computed by the @xmath154-th it subproblem , is primal feasible for the @xmath154-th er subproblem . this section provides guidelines to improve the choice of the tightening parameter of each it subproblem in order to guarantee the primal feasibility of the _ consolidated _ solution , i.e. , the predictions obtained , starting from the initial state @xmath155 , using the control sequence @xmath156 where the elements of @xmath157 are computed by the independent it subproblems . in particular , when a new measurement is available from the plant , the subsystems compute in parallel @xmath158 @xmath159,@xmath160 and @xmath161 , respectively . according to the results of the previous section , the pair @xmath162 is primal feasible for the @xmath154-th subproblem , thanks to the introduction of the it subproblems . nevertheless , due to the relaxation introduced on the equality constraints - , there is a bounded mismatch between @xmath57 and @xmath59 ( @xmath28 ) . hence , starting from the initial state @xmath155 , when the control sequence @xmath163 is applied to compute the consolidated state prediction @xmath164 the feasibility of @xmath165 is no longer guaranteed . note , however , that @xmath166 , i.e. , @xmath157 is feasible . hence , no additional tightening is needed on the input constraints . in the following , section [ subsec : upbound ] defines an upper bound on the maximal feasibility violation of @xmath165 . this feasibility violation is a consequence of the local relaxations of the equality constraints . then , section [ subsec : improve_tightening ] introduces sufficient conditions to ensure the primal feasibility of the _ consolidated _ prediction and provides guidelines for the choice of the tightening parameters for each it subproblem . let @xmath168 and @xmath169 be defined by and , respectively . moreover , from and , the following holds : @xmath170 our goal is to characterize how far the _ consolidated _ predicted state is from the _ local _ predicted state . [ lem : upper_bound_global_prediction ] let the @xmath154-step - ahead _ consolidated _ prediction @xmath171 be defined by and assume that holds . then , there exists @xmath172 , such that the mismatch between @xmath171 and the state of the @xmath154-th subproblem @xmath173 is bounded , as follows : @xmath174 . see appendix ii in @xcite . [ rem : after_lemma_2_a ] according to lemma [ lem : upper_bound_global_prediction ] a possible choice of @xmath175 is the following : @xmath176 according to lemma [ lem : upper_bound_global_prediction ] , @xmath171 differs from @xmath173 by a quantity bounded by @xmath175 . thus , @xmath171 might violate the constraints of the @xmath154-th subproblem by as much as @xmath175 , in the worst - case scenario . in particular , we must ensure that @xmath177 . using the computed upper bound , the following holds : @xmath178 where @xmath179 indicates the absolute value of @xmath180 . recall that these mismatches are caused by the use of inexact solvers and that @xmath175 depends on @xmath87 . in the following , we provide guidelines to improve the choice of @xmath88 for each subproblem . furthermore , we provide a modified upper bound for the optimal lagrange multiplier associated with the tightened subproblems , which considers the additional tightening introduced by @xmath175 . [ lem : improve_epsilon_t ] consider the following it subproblems : @xmath181 for @xmath182 , where @xmath89@xmath183@xmath184 @xmath185^t$ ] . consider the assumptions of lemma [ lem : lemma_up_bound ] and the existence of @xmath175 for all @xmath83 according to lemma [ lem : upper_bound_global_prediction ] . then , for each subproblem , there exist @xmath141 , @xmath186 such that the upper bound for the optimal lagrange multiplier associated with the it subproblems is described by @xmath187_j\}}},~{\ensuremath{t\!=\!0,\ldots , n}},\ ] ] @xmath188 [ 2\rho ( \epsilon_{z_{t}}-\epsilon_t)\mathbf{1}_n^t ] [ 2\rho(\epsilon_{z_{t+1}}-\epsilon_t)\mathbf{1}_n^t]\big]^t\in \mathbb r^{p_t+2n}$ ] . see appendix [ app : lemma3 ] . see appendix iii in @xcite . [ rem : after_lemma_3 ] the choice of @xmath88 ( @xmath182 ) is not unique and depends on the choice of @xmath87 ( @xmath83 ) . for example , given @xmath175 in , a possible choice of @xmath87 ( @xmath83 ) is : @xmath189_{j , i}|\big\}}}}\!\bigg\}. \label{eq : choice_epsilon_z}\end{aligned}\ ] ] consequently , the tightening parameters are given by : @xmath190 for @xmath182 . this choice implies that first we select the relaxation parameters and then we _ adapt _ the tightening parameters on the original inequality constraints based on the choice of @xmath87 for all @xmath83 . an alternative is to fix @xmath88 for the inequality constraints and consequently compute @xmath87 . in general , the choice of the parameters strongly depends on the system - state matrix @xmath6 in . [ rem : offline_computation_parameters ] in the context of this work , algorithm [ alg : alg2 ] , described in the next section , adapts the above derived parameters at each problem instance . if we consider a fixed tightening scheme , such as the one proposed by @xcite , @xmath191 and @xmath87 can be computed offline ( for all the initial states in the region of attraction ) . in the following , we show that by using @xmath192@xmath193 is the control sequence obtained by solving the it subproblems and @xmath194 is the corresponding consolidated prediction the inequality constraints of the original mpc problem are satisfied . consequently , the predicted final state is in the terminal set of the original problem . if the desired level of suboptimality of algorithm [ alg : alg1 ] is chosen as : @xmath195 then , according to theorem [ thm : thm1 ] , there exists @xmath196^t$ ] such that @xmath197_+\| \leq\eta_t<\epsilon_t$ ] . using similar arguments as in @xcite , the following holds for @xmath182 : @xmath198_+<\epsilon_t\operatorname{\mathbf{1}}_{p_t+4n}.\end{aligned}\ ] ] hence , for all @xmath199 , the following holds @xmath200_j\bigg]_+\leq \epsilon_t.\end{aligned}\ ] ] consequently , exploiting the upper bound , for all @xmath199 , we have : @xmath201_j\leq \epsilon_t\end{aligned}\ ] ] which leads to @xmath202 , where @xmath203 is the @xmath154-step - ahead _ consolidated _ prediction computed using the solution to the it subproblem with tighening parameter @xmath89 . in summary , this section showed that there exists a choice of the relaxation and tightening parameters that guarantee a feasible consolidated prediction with respect to the original problem . in the following , we derive bounds for @xmath204 , i.e. , the cost obtained using @xmath192 , with respect to the optimal cost @xmath205 of the original problem . [ th : theorem2 ] assuming that there exist @xmath88 ( @xmath182 ) and @xmath87 ( @xmath83 ) selected according to lemma [ lem : improve_epsilon_t ] , then the following holds : @xmath206 where @xmath207_{j , i}|\big\}}}\alpha_t$ ] . see appendix [ app : proof_theorem_2 ] . see appendix iv in @xcite . theorem [ th : theorem2 ] established the level of suboptimality of the consolidated prediction with respect to the original problem . in particular , the sequence @xmath208 is suboptimal for the original problem and satisfies the original inequality constraints ( including those associated with @xmath209 ) . recall that for the update of @xmath210 , our algorithm requires a strictly feasible vector @xmath152 for . hence , every time new measurements are available from the plant , our algorithm must provide a strictly feasible solution ( not necessarily optimal ) for the first @xmath211 inequality constraints of each er subproblem . the following lemma provides guidelines to compute @xmath152 . [ lem : update_y ] let @xmath212 be defined as @xmath213=[(x_{0}^t~\bar u_{0,\gamma_0}^t)\ldots(\bar x_{n-1,\gamma_{n-1}}^t~\bar u_{n-1,\gamma_{n-1}}^t)~(\bar x_{n,\gamma_n}^t))]^t$ ] . then , a feasible @xmath214 at the next problem instance , is given by : @xmath215}}((a+bk_f)\bar x_{n,\gamma_n})^t]^t \end{aligned}\ ] ] see appendix [ app : proof_lemma_update_y ] . see appendix v in @xcite . we want to show that the cost decreases at each problem instance . using a similar argument as in @xcite , under assumption [ ass : terminalset ] on @xmath216 and ensuring that @xmath217 ( thanks to a proper choice of the tightening parameters , as the previous section showed ) , we can show that : @xmath218 where @xmath219 is the region of attraction . hence , from and , the following holds : @xmath220\!\!-\!\!{\ensuremath{\mathbf v}}_{0}(y_{0,\gamma_0})\end{aligned}\ ] ] where @xmath221 , using @xmath222 to represent the updated values of these parameters according to @xmath223 . the inequality above shows that the total cost decreases at each problem instance if @xmath216 is defined according to assumption [ ass : terminalset ] and if the @xmath0-step - ahead _ consolidated _ prediction lies in the terminal set . asymptotic stability of our controller follows if @xmath224 . hence , we can modify the update of @xmath88 and @xmath87 to ensure that is satisfied . [ rem : update_z ] a possible choice of @xmath87 @xmath225 to fulfill is the following : @xmath226 @xmath227_{j , i}|\big\}}}\bigg)}\bigg]^{-1}.\notag\end{aligned}\ ] ] consequently , @xmath88 can be selected according to to preserve the definition of the upper bound on the optimal lagrange multipliers given in lemma [ lem : improve_epsilon_t ] . algorithm [ alg : alg2 ] summarizes the main steps needed to obtain a stabilizing control law when the original mpc problem is solved in parallel using inexact solvers . in particular , note that , if the measured state is in @xmath216 , from assumption [ ass : terminalset ] , the state and the control constraints are automatically satisfied without solving the mpc problem in parallel . steps 17 - 23 are the only nonparallel ones of the algorithm ( algorithm [ alg : alg1 ] is instead fully parallelizable ) . the main reason is in the adaptive nature of the algorithm ( see also remark [ rem : offline_computation_parameters ] ) . algorithm [ alg : alg2 ] adapts @xmath88 and @xmath87 every time new measurements are available from the plant . a fully parallel algorithm [ alg : alg2 ] is possible using a fixed tightening strategy , in which @xmath88 and @xmath87 can be computed offline . we evaluated algorithm [ alg : alg2 ] using the lti system described in @xcite . the system ( sampled at @xmath228 s ) is described by : @xmath229 where @xmath230 , @xmath231 , and the quadruple @xmath232 is given by : @xmath233 the weighting matrices @xmath41 , @xmath42 , and @xmath234 in the cost and the ih - lqr gain @xmath235 are selected according to @xcite . we implemented our design in matlab ( to tune the controller and test the initial design ) and in c ( to run a performance analysis ) . in particular , in matlab , we used the parallel computing toolbox to assign the computation of algorithm [ alg : alg1 ] to 8 parallel workers , given a prediction horizon @xmath236 . furthermore , we relied on the mpt3 toolbox @xcite to compute @xmath216 and the optimal solution of problem . finally , we compared our design to @xcite . we considered the following scenario . the initial state of the system is @xmath237^t$ ] . the total number of complicated constraints for the original problem is 90 . we used and to initialize @xmath87 and @xmath191 , respectively . to update them , we relied on and . the selected @xmath155 caused @xmath238 and @xmath239 to saturate ( 12 active constraints ) . in this scenario , the state enters @xmath216 in 3 steps . figure [ fig : upp_bound_alpha ] shows the mismatch between the _ local _ prediction and the _ consolidated _ prediction for one problem instance . as figure [ fig : upp_bound_alpha ] depicts , the mismatch ( for both states ) is below the predicted upper bound @xmath175 for all the @xmath1 subproblems . table [ tab : iterations ] compares the proposed technique to @xcite . the table reports the upper bound @xmath240 on the number of iterations needed to achieve a suboptimal solution for problem and the level of suboptimality @xmath241 . the table lists only the first four subproblems , which are the most significant due to the presence of active constraints in these subproblems . the method in @xcite and our new proposed parallel algorithm produce a comparable behavior , thanks to an appropriate selection of the tightening parameters . the parallel algorithm , however , is able to achieve similar results to those in @xcite using a smaller number of iterations . the larger values of @xmath240 for @xcite are probably caused by the value of the lipschitz constant and by the problem conditioning , which affect the convergence requiring a higher accuracy for the solver . in particular , in our proposed framework , the dfg is applied to simpler problems characterized , in the worst case scenario , by a lipschitz constant @xmath242 and by a condition number @xmath243 . in @xcite , the dfg solves a larger problem characterized by a lipschitz constant @xmath244 and by a condition number @xmath245 . hence , the modularity of our approach has positive implications on important properties for the convergence of the solver . table [ tab : iterations ] lists the time required by the optimizer to return a suboptimal solution for problem . to measure the performance , we implemented both algorithms in c on a linux - based os . we noticed that given the small size of the problem , running algorithm [ alg : alg1 ] in parallel did not result in significant speedups compared to our algorithm running in serialized mode @xcite . nevertheless , in both cases , we registered a speedup ( 230x ) compared to @xcite . the modularity of the proposed algorithm is beneficial even for problems of small size , such as the one considered in this section for the comparison with @xcite . we expect the benefits to be even more pronounced when considering problems of larger size . .performance analysis of algorithm [ alg : alg1 ] and @xcite . results show the median of 11 experiments . [ cols="^ , < , < , < , < , < " , ] [ tab : iterations ] we proposed an algorithm tailored to mpc that guarantees recursive feasibility and closed - loop stability , when the solution of the mpc problem is computed using inexact solvers in a parallel framework . in particular , our algorithm combines admm and dfg methods and relies on an adaptive constraint - tightening strategy to certify the mpc law . our numerical analysis shows performance improvements compared to state - of - the - art nonparallel techniques @xcite . furthermore , our study shows that , for small - size problems , even if the solver is implemented in a serialized mode , there is substantial performance improvement with respect to the state of the art . we expect further benefits from the parallelization when the size of the problem increases . a scalability analysis of the proposed algorithm on many - core architectures is part of our ongoing work . this section contains the proof of lemma [ lem : lemma_up_bound ] presented in section [ subsec : up_lagrange_multiplier ] . assume that there exists a slater vector @xmath139 such that @xmath140 . then , there exists @xmath246 , @xmath142 , @xmath143 , such that the upper bound on @xmath138 is given by @xmath247_j\}}},\ ] ] where @xmath248 and @xmath146 is the dual function for the original subproblem evaluated at @xmath147 . the following inequality holds : @xmath249 where the last inequality takes into account that @xmath250 . define @xmath251 and @xmath252 . consequently , using the definition of @xmath253 and @xmath254 , the following holds : @xmath255^t\|.\ ] ] furthermore , recalling that @xmath256 , @xmath257 , and @xmath258 , the following holds : @xmath259^t\|\!\leq\![\lambda_{t,\epsilon_t}^{*t}~w_{t,\epsilon_t}^{*t}~v_{t+1,\epsilon_t}^{*t}]^t\!\!\operatorname{\mathbf{1}}_{p_t+2n}.\ ] ] hence , if we compute an upper bound for the vector @xmath260^t$ ] , we obtain an upper bound for @xmath148 . thus , from the inequality , it follows that : @xmath261 notice that choosing @xmath262 , @xmath263 , i.e. , according to the assumptions of the lemma , the elements of @xmath264 are all greater than zero . thus , using and , it follows : @xmath265_j\}\|\mu_{t,\epsilon_t}^*\| \le\notag \\ \label{eq : lemma_1_proof } & \le \begin{bmatrix } \lambda_{t,\epsilon_t}^ { * } & w_{t,\epsilon_t}^ { * } & v_{t+1,\epsilon_t}^{*}\end{bmatrix } \gamma_t.\end{aligned}\ ] ] consequently , the upper bound on the optimal lagrange multiplier is given by : @xmath266_j\}}.\end{aligned}\ ] ] this section contains the proof of lemma [ lem : upper_bound_global_prediction ] presented in section [ subsec : upbound ] . let the @xmath154-step - ahead _ consolidated _ prediction @xmath171 be defined by and assume that holds . then , there exists @xmath172 , such that the mismatch between @xmath171 and the state of @xmath154-th subproblem @xmath173 is bounded , as follows : @xmath267 in the following , we omit the dependence from @xmath88 to simplify the notation . the proof is constructive . for @xmath268 , @xmath269 . for @xmath270 , @xmath271 , which is the 1-step - ahead state computed by the local subproblem 0 , i.e. , the subproblem associated to worker @xmath272 . hence , the mismatch between @xmath273 and @xmath274 is simply given by @xmath275 for @xmath276 , the following holds : @xmath277 @xmath278 which proves the lemma . this section contains the proof of lemma [ lem : improve_epsilon_t ] presented in section [ sec : tightening_improvements ] . consider the following it subproblems : @xmath279 for @xmath182 , where @xmath89@xmath183@xmath184 @xmath185^t$ ] . given the assumptions of lemma [ lem : lemma_up_bound ] and the existence of @xmath175 for all @xmath83 according to lemma [ lem : upper_bound_global_prediction ] . then , for each subproblem , there exist @xmath141 , @xmath186 such that the upper bound on the optimal lagrange multiplier associated with the it subproblems is described by @xmath187_j\}}},~{\ensuremath{t\!=\!0,\ldots , n}},\ ] ] @xmath280 this lemma follows from lemma [ lem : lemma_up_bound ] applied to the subproblems . from inequality formulated for subproblem , the following must hold @xmath281 hence , in order to satisfy the inequality above , we can select the relaxation parameters @xmath87 and the tightening parameters @xmath88 according to the assumption of the lemma for @xmath182 , i.e. , the following must hold : @xmath282_{j , i}|\big\}}}\alpha_t+\epsilon_t\\ \text{\it ( ii)~ } & { \ensuremath{\max\limits_{j=1,\ldots , p_t}\!\!\!\big\{\sum_{i=1}^n|[c_{t}]_{j , i}|\big\}}}\alpha_t+\epsilon_t > 0\\ \text{\it ( iii)~}&\epsilon_{z_t},\epsilon_{z_{t+1}}\geq \epsilon_t\geq 0,\end{aligned}\ ] ] where @xmath152 is a strictly feasible solution for the original @xmath154-th subproblem . hence , there exists @xmath210 such that the upper bound on the optimal lagrange multiplier @xmath283 is defined as follows : @xmath284\}}}.\end{aligned}\ ] ] this section contains the proof of theorem [ th : theorem2 ] presented in section [ sec : closed_loop_certification ] . assuming that there exist @xmath88 ( @xmath182 ) and @xmath87 ( @xmath83 ) selected according to lemma [ lem : improve_epsilon_t ] , then the following holds : @xmath285 where @xmath207_{j , i}|\big\}}}\alpha_t$ ] . due to the tightening of the original inequality constraints , @xmath286 , given that , as proved in section [ subsec : improve_tightening ] , the feasible region of the tightened subproblems is inside the one of the original subproblems . recall that the consolidated prediction satisfies the equality constraints by construction . hence , the following holds : @xmath287\gamma_t\right\rangle\big]\\ \leq & \mathcal v^ * + 2\sum\limits_{t=0}^n\mathcal r_t\sqrt{p_t}(\epsilon_t+\!\!\!{\ensuremath{\max\limits_{j=1,\ldots , p_t}\!\!\!\big\{\sum_{i=1}^n|[c_{t}]_{j , i}|\big\}}}\alpha_t).\end{aligned}\ ] ] where @xmath288\gamma_t$ ] selects the first @xmath211 components of the vector @xmath89 . this section contains the proof of lemma [ lem : update_y ] presented in section [ sec : closed_loop_certification ] . let @xmath212 be defined as @xmath213=[(x_{0}^t~\bar u_{0,\gamma_0}^t)\ldots(\bar x_{n-1,\gamma_{n-1}}^t~\bar u_{n-1,\gamma_{n-1}}^t)~(\bar x_{n,\gamma_n}^t))]^t$ ] . then , @xmath214 at the next problem instance , is given by : @xmath289}}((a+bk_f)\bar x_{n,\gamma_n})^t]^t \end{aligned}\ ] ] we can use a similar argument as the one of lemma 4.2 in @xcite , recalling that @xmath290 for @xmath28 , @xmath291 for @xmath292 , @xmath293 for @xmath28 , @xmath294 for @xmath292 , and @xmath295 according to assumption [ ass : terminalset ] . 99 s. d. cairano , m. brand , and s. a. bortoff , `` projection - free parallel quadratic programming for linear model predictive control '' . _ international journal of control _ , vol . 86 , no . 8 , pp . 13671385 , 2013 . m. k " ogel and r. findeisen , `` parallel solution of model predictive control using the alternating direction multiplier method '' , proc . of the _ ifac conference on nmpc 1 , pp . 369374 , 2012 . s. boyd et al . `` distributed optimization and statistical learning via the alternating direction method of multipliers '' . in _ foundations and trends@xmath296 in machine learning _ , vol . 1 , pp . 1122 , 2011 . m. k " ogel and r. findeisen , `` fast predictive control of linear , time - invariant systems using an algorithm based on the fast gradient method and augmented lagrange multipliers '' , proc . of the _ ieee cca _ , pp . 780785 , 2011 . m. rubagotti , p. patrinos and a. bemporad , `` stabilizing linear model predictive control under inexact numerical optimization '' , in _ ieee trans . on automatic control _ , 59 , no . 6 , pp . 16601666 , 2014 . i. necoara , l. ferranti , and t. keviczky , `` an adaptive tightening approach to linear model predictive control based on approximation algorithms for optimization '' , _ optimal control applications and methods _ , doi : 10.1002/oca.2121 , http://dx.doi.org/10.1002/oca.2121 , to appear in 2015 . 99 s. d. cairano , m. brand , and s. a. bortoff , `` projection - free parallel quadratic programming for linear model predictive control '' . _ international journal of control _ , vol . 86 , no . 8 , pp . 13671385 , 2013 . m. k " ogel and r. findeisen , `` parallel solution of model predictive control using the alternating direction multiplier method '' , proc . of the _ ifac conference on nmpc 1 , pp . 369374 , 2012 . s. boyd et al . `` distributed optimization and statistical learning via the alternating direction method of multipliers '' . in _ foundations and trends@xmath296 in machine learning _ 1 , pp . 1122 , 2011 . m. k " ogel and r. findeisen , `` fast predictive control of linear , time - invariant systems using an algorithm based on the fast gradient method and augmented lagrange multipliers '' , proc . of the _ ieee cca _ , pp . 780785 , 2011 . m. rubagotti , p. patrinos and a. bemporad , `` stabilizing linear model predictive control under inexact numerical optimization '' , in _ ieee trans . on automatic control _ , vol . 59 , no . 6 , pp . 16601666 , 2014 . i. necoara , l. ferranti , and t. keviczky , `` an adaptive tightening approach to linear model predictive control based on approximation algorithms for optimization '' , in _ optimal control applications and methods _ , doi : 10.1002/oca.2121 , 2015 .

tail - biting trellises are perhaps the simplest instances of decoding graphs with cycles . a tail - biting trellis has a tanner graph @xcite with a single cycle and usually approximate algorithms are used for decoding , as exact algorithms are believed to be too expensive . these approximate algorithms iterate around the trellis until either convergence is reached , or for a preset number of cycles . to the best of our knowledge , no _ exact _ decoding algorithms other than the brute force algorithm have been proposed so far for the general case , though there are several _ approximate _ algorithms for maximum - likelihood decoding @xcite and exact algorithms for bounded distance decoding @xcite . the problem of maximum a - posteriori probability(map ) decoding is not addressed here . we propose an _ exact _ recursive algorithm , which exhibits very good average case behavior . the algorithm exploits the fact that a linear tail - biting trellis can be viewed as a coset decomposition of the group corresponding to the linear code with respect to a specific subgroup and is based on the @xmath0 algorithm @xcite . we also propose two approximate variants that always converge , and observe their performance on tail - biting trellises for the ( 24,12 ) extended golay code and two convolutional codes of rate 1/2 and memory of 4 and 6 respectively . the performance of the first approximate variant is indistinguishable from that of the exact algorithm in terms of bit error rate for the two convolutional codes , and it is guaranteed to update each node in the tail - biting trellis at most twice i.e it performs a computation equivalent to at most two rounds on the trellis . section [ related ] briefly mentions related work . section [ sec : background ] provides some background . section [ algo ] describes the algorithm , while section [ analysis ] analyses the algorithm . section [ approx ] describes the approximate algorithm and provides an analysis for its good performance . section [ simulations ] reports the results of simulations on an awgn channel and section [ conclusion ] concludes the paper . aji et al . @xcite have shown that iterative maximum - likelihood ( ml ) decoding on tail - biting trellises will asymptotically converge to exact maximum likelihood decoding for certain codes . they provide experimental evidence that practically ml decoding is achieved for the @xmath1 hamming code with five rounds of the tail - biting trellis . the presence of _ pseudocodewords _ sometimes results in sub - optimal decoding and it is also possible to have situations where the iterative message passing algorithm does not converge . several maximum likelihood decoding algorithms on tail - biting trellises have been proposed without a theoretical analysis @xcite , but with good experimental results . most of these are sub - optimal algorithms in that they may not produce the exact maximum - likelihood result on termination . anderson and hladik @xcite have given an algorithm that is optimal for bounded distance decoding . the @xmath0 algorithm @xcite has been used for maximum likelihood soft decision decoding on conventional trellises for block codes @xcite . in @xcite han propose the use of the @xmath0 algorithm for ml decoding of block codes on their conventional trellises and report significant experimental gains in decoding complexity for signal to noise ratios ranging from 5 db to 8 db . this algorithm has been analyzed in @xcite and shown to be efficient for many practical communication systems . in @xcite a modified algorithm is proposed which searches through error patterns instead of codewords and similar gains are reported . heuristic search algorithms are proposed in @xcite which combine previously proposed algorithms and are able to outperform other practical decoders . a tutorial paper on the application of the @xmath0 algorithm to soft decision decoding appears in @xcite . sorokine and kschischang @xcite propose a metric called the variable bias term that is used in an @xmath0 algorithm , which has low computational complexity . aguado and farrell @xcite discuss modified sequential algorithms on conventional trellises for block codes , which offer reduced complexity in comparison with the original stack algorithm @xcite for sequential decoding . han et al . @xcite propose a trellis based ml soft - decision decoder for convolutional codes which uses a stack and a metric that ensures ml decoding . we first present some background on tail - biting trellises . tail - biting trellises for convolutional codes were introduced in @xcite . minimal tail - biting trellises for block codes have been discussed in @xcite . a _ tail - biting trellis _ @xmath2 of depth @xmath3 is an edge - labeled directed graph with the property that the set @xmath4 can be partitioned into @xmath3 vertex classes @xmath5 such that every edge in @xmath6 is labeled with a symbol from the alphabet @xmath7 , and begins at a vertex of @xmath8 and ends at a vertex of @xmath9 , for some @xmath10 . we identify @xmath11 the set of time indices with @xmath12 , the residue classes of integers modulo @xmath3 . an interval of indices @xmath13 $ ] represents the sequence @xmath14 if @xmath15 , and the sequence @xmath16 if @xmath17 . every cycle in @xmath6 starting at a vertex of @xmath18 defines a vector @xmath19 which is an _ edge - label sequence_. we assume that every vertex and every edge in the tail - biting trellis lies on some cycle , that is the tail - biting trellises we are dealing with are _ reduced _ @xcite . the trellis @xmath6 represents a block code @xmath20 over @xmath7 if the set of all edge - label sequences in @xmath6 is equal to @xmath20 . let @xmath21 denote the code represented by a trellis @xmath6 . a linear tail - biting trellis , for an @xmath22 linear block code @xmath20 over @xmath7 can be constructed as a _ trellis product _ @xcite of the representation of the individual trellises ( called _ elementary _ trellises ) corresponding to each of the @xmath23 rows of the generator matrix @xmath24 for @xmath20 @xcite . let @xmath25 and @xmath26 be the component trellises . the set of vertices @xmath27 of the product trellis @xmath28 at time index @xmath29 , is just the cartesian product of the vertices of the component trellises . thus @xmath30= @xmath31 . consider @xmath32 , and interpret an element @xmath33 in this product , where @xmath34 are vertices and @xmath35 edge labels , as the edge @xmath36 where @xmath37 denotes addition in the field @xmath7 . if we define the @xmath38 section as the set of edges connecting the vertices at time index @xmath29 to those at time index @xmath39 , then the edge count in the @xmath38 section is the product of the edge counts in the @xmath38 section of the individual trellises . let @xmath40 be the rows of a generator matrix @xmath24 for the linear code @xmath20 . each vector @xmath41 generates a one - dimensional subcode of @xmath20 , which we denote by @xmath42 . therefore @xmath43 , and the trellis representing @xmath20 is given by @xmath44 , where @xmath45 is the trellis for @xmath46 . to specify the component trellises in the trellis product above , we will need to introduce the notions of linear@xcite and circular spans@xcite and elementary trellises @xcite . given a codeword @xmath47 , the _ linear span _ of @xmath48 , is the smallest interval @xmath13 \in i= \{1,2,\ldots n\ } , i\leq j$ ] which contains all the non - zero positions of @xmath48 . circular span _ has exactly the same definition with @xmath17 . note that for a given vector , the linear span is unique , but circular spans are not they depend on the runs of consecutive zeros chosen for the complement of the span with respect to the index set @xmath49 . for a vector @xmath50 over the field @xmath7 and a specified span @xmath13 $ ] , there is a unique _ linear elementary trellis _ representing @xmath51 @xcite . this trellis has @xmath52 vertices at time indices @xmath29 to @xmath53 mod @xmath3 , and a single vertex at other positions . consequently , @xmath45 in the trellis product mentioned earlier , is the elementary trellis representing @xmath42 for some choice of span ( either linear or circular ) . koetter and vardy @xcite have shown that any linear trellis , conventional or tail - biting can be constructed from a generator matrix whose rows can be partitioned into two sets , those which have linear span , and those taken to have circular span . the trellis for the code is formed as a product of the elementary trellises corresponding to these rows . we will represent such a generator matrix as @xmath54 $ ] , where @xmath55 is the submatrix consisting of rows with linear span , and @xmath56 the submatrix of rows with circular span . [ def : zerorun ] for a vector @xmath57 of circular span @xmath58 $ ] in @xmath56 , the interval @xmath59 mod @xmath60 mod @xmath61 $ ] is called the _ zero run _ of the vector . the path in the trellis corresponding to this vector shares all states at time indices in the zero run with the path corresponding to the all - zero codeword in the product trellis . + for example , consider the codeword @xmath62 with circular span @xmath63 $ ] . this has zero run @xmath64 $ ] . the elementary trellis corresponding to this vector has state cardinality profile @xmath65 . ( recall the time indices are numbered from 0 to @xmath66 where @xmath3 is the length of the code ) . as an example we display a tail - biting trellis for a binary @xmath67 hamming code . though this is not a minimal trellis for the code , it serves to illustrate some of the definitions above . the spans of the rows are shown alongside the rows . all spans @xmath13 $ ] with @xmath29 greater than @xmath68 are circular spans . [ ex : overlay2 ] let @xmath20 be a @xmath69 hamming code a with a product generator matrix @xmath70 defined as @xmath71 \begin{array}{c } \lbrack 1,6 \rbrack \\ \lbrack 3,7 \rbrack \\ \lbrack 6,2 \rbrack \\ \lbrack 7,4 \rbrack \end{array}\ ] ] the product tail - biting trellis for this generator matrix is given in figure [ fig : hamming - overlay ] . a subtrellis of a tailbiting trellis consists of a start node at time index zero and all edges and nodes which can be traversed in any cycle of the graph that begins and ends at this start node . let @xmath72 denote the minimum conventional trellis for the code generated by @xmath55 . clearly @xmath72 is a subtrellis of the tail - biting trellis . if @xmath73 is the number of rows of @xmath24 with linear span and @xmath74 the number of rows of circular span , the tail - biting trellis constructed using the product construction will have @xmath75 start states . each such start state defines a subtrellis whose codewords form a coset of the subcode corresponding to the subtrellis containing the all 0 codeword . the coset structure is well known and has been reported in @xcite . each vector in the circular span can be considered to be a coset leader . the set of zero runs , of the coset leaders determines the structure of the tail - biting trellis in the following way . if a coset leader has zero run @xmath13 $ ] then the subtrellis associated with that coset shares all states at time indices in the interval @xmath13 $ ] with the subtrellis corresponding to the subcode defined by vectors of linear span . further , we recall , the _ coset leader _ shares all states in the interval @xmath13 $ ] with the states corresponding to the all - zero codeword . the four subtrellises of the tail - biting trellis of figure [ fig : hamming - overlay ] are shown in figures [ fig : subtrellis0 ] , [ fig : subtrellis1 ] , [ fig : subtrellis2]and [ fig : subtrellis3 ] along with their associated coset leaders and zero runs . if subtrellises @xmath76 and @xmath26 share states from time indices @xmath29 to @xmath68 then the interval @xmath13 $ ] is called the _ merging _ interval of @xmath25 and @xmath26 . it is easy to see that two subtrellises do not share any states outside their merging interval . + a tail - biting trellis is said to satisfy the _ intersection property _ if the intersection of all the zero runs of the members of @xmath56 is non - empty . the tail - biting trellis for the hamming code given in example [ ex : overlay2 ] satisfies the intersection property as the interval @xmath77 $ ] is contained in the intersection of all the zero runs of @xmath56 . the decoding algorithm proposed here is different from the sub - optimal algorithms mentioned in section [ related ] , that go round and round the tail - biting trellis updating all the nodes of the trellis in every round . it makes one round of the tail - biting trellis and subsequently judiciously uses the information gathered to further update as few nodes as it can before it closes in on the most likely codeword . our algorithm has two phases . in the first phase a viterbi algorithm is performed on the tail - biting trellis . this phase performs computations at _ every _ node of the tail - biting trellis . in the second phase however , _ only one path _ is tracked at a time , this being the most likely path . the initial estimate of the most likely path is obtained from the first phase . this path is present in some subtrellis and is followed until the algorithm decides that some other path ( perhaps in another subtrellis ) looks more promising based on some metric . when such a situation is encountered , computation on this path is suspended and the more promising path is taken up . while this strategy at first glance looks like the stack algorithm @xcite for decoding convolutional codes , it differs from it because it has the property that it _ always _ delivers the optimal path as the metric used satisfies the property required by the @xmath0 algorithm . ( we will prove this property formally ) . for purposes of decoding we use the unrolled version of the trellis with start states @xmath78 and final states @xmath79 where @xmath73 is the number of subtrellises . an @xmath80 path is a path from start vertex @xmath81 to final vertex @xmath82 , and is consequently a codeword path in trellis @xmath45 , whereas an @xmath83 path for @xmath84 is a non codeword path as it starts in subtrellis @xmath45 and ends in subtrellis @xmath85 . for purposes of our discussion we term the label sequence along such a path as a _ semicodeword_. maximum - likelihood decoding for a tail - biting trellis is equivalent to finding the codeword closest to the received sequence measured in terms of a soft decision metric . assume that the channel is modeled as an additive white noise gaussian(awgn ) channel and that antipodal signaling is used for communication . a binary code digit 0 is mapped into @xmath86 and a 1 is mapped into @xmath87 where @xmath88 is the signal energy per bit entering the channel . for a discrete additive white gaussian noise(awgn ) channel we have @xmath89 where @xmath90 is the received signal at time @xmath91 , @xmath92 is the transmitted signal and @xmath93 is the value of a white gaussian noise random variable with variance @xmath94 where @xmath95 is the noise spectral density . without loss of generality we can assume that @xmath96 . the signal - to - noise ratio or snr is the quantity @xmath97 . the decoder uses the received vector @xmath98 to determine which codeword was transmitted . it forms an estimate @xmath99 of the codeword @xmath100 that was transmitted . a decoding error occurs if @xmath101 . the maximum likelihood decoding rule is to decode the received sequence @xmath98 to codeword @xmath102 whenever @xmath103 for all @xmath104 , where @xmath105 is the conditional probability of @xmath98 given @xmath102 . let @xmath106 be the signal vector corresponding to the codeword @xmath100 .. if @xmath107 is the euclidean distance between @xmath108 and @xmath98 , then the maximum likelihood decoding rule for decoding binary linear block codes transmitted over the awgn channel using antipodal signaling is to decode @xmath98 into codeword @xmath102 whenever @xmath109 for all @xmath110 . the decoding algorithm is thus cast as a shortest path problem in which each path is associated with a _ metric _ , and the problem is to find a codeword path with minimum metric . the @xmath0 algorithm is used to cut down the search space . it does so by using a node metric which is the sum of the length of the shortest path from the source to a node and an _ underestimate _ of the length of the shortest path from the node to the goal node to guide the search . as mentioned earlier , only one path is explored at a time and the algorithm derives it s advantage from the fact that if the estimates used are close to the actual values then the search space that yields the optimal path is greatly reduced . we give the algorithm below . the algorithm maintains two sets of vertices , @xmath111 and @xmath112 . the set @xmath111 is the set of _ closed _ nodes and represents nodes to which the shortest paths have been finalized . at any iteration , the set @xmath112 is the set of candidate nodes the best of which will be closed in the succeeding iteration . these are called the _ open _ or _ visited _ nodes . an operation of _ expanding _ a node consists of the following three steps : + 1 . getting all the immediate successors of the node . checking for each immediate successor if this successor has been visited before . if the successor has been visited then _ updating _ the minimum cost path to the successor by taking the minimum of the cost of the previous path and the cost of this one . all the expanded nodes are put into the closed set and the visited nodes are put into the open set . when the goal node is reached an optimal path has been found . + the following is a formal description of the algorithm . line 1 performs the initialization of the sets and the costs and paths . line 3 selects the vertex to be expanded . line 4 puts the selected vertex into the closed set and deletes it from the open set . line 5 detects if the algorithm has completed ; lines 6 through 9 perform an expansion of a node . they update the cost of an immediate successor as well as the best path to that successor and mark the successor as visited by putting it into the open set . + * algorithm @xmath113 * + * input :* a trellis @xmath114 where @xmath4 is the set of vertices , @xmath115 is the set of edges and @xmath116 for edge @xmath117 in @xmath115 , a source vertex @xmath118 and a destination vertex @xmath119 , and an estimate @xmath120 for the shortest path from @xmath121 to @xmath119 for each vertex @xmath122 . + * output :* the shortest path from @xmath118 to @xmath119 . + / * @xmath123 is the cost of the current shortest path from @xmath118 to @xmath121 and @xmath124 is a current shortest path from @xmath118 to @xmath121 * / + _ begin _ + @xmath125 ; + _ repeat _ + . let u be the vertex in @xmath112 with minimum value of @xmath126 . @xmath127 ; + . _ if _ @xmath128 _ then _ _ return _ @xmath129 ; + . _ for _ each @xmath130 _ do _ _ if _ @xmath131 _ then _ + . _ begin _ + . @xmath132 ; + _ if _ @xmath133 _ then _ append @xmath117 to @xmath124 to give @xmath134 ; + . @xmath135 ; + . _ end _ + . _ forever _ + _ end _ + the @xmath0 algorithm is guaranteed to output the shortest path if the following two conditions hold : let @xmath136 be the shortest path length from @xmath121 to @xmath119 in @xmath6 . let @xmath120 be any lower bound such that @xmath137 , and such that @xmath120 satisfies the following inequality , i.e , for @xmath121 a predecessor of @xmath138 , @xmath139 . if both the above conditions are satisfied , then the algorithm @xmath113 , on termination , is guaranteed to output a shortest path from @xmath118 to @xmath119 . + the algorithm proposed here is a variant of the @xmath0 algorithm , which at any given instant , is executing an @xmath0 algorithm on exactly one of the subtrellises , with perhaps suspended executions of the algorithm on a set of other subtrellises . the subtrellis on which the algorithm is curently executing , appears the best in its potential to deliver the minimal cost path . since the algorithm is not straightforward , we first give an informal explanation of how it works . the algorithm has two phases . the first phase performs a viterbi algorithm on the tail - biting trellis and examines surviving paths , called _ survivors _ here , at all states of the tail - biting trellis . the first phase is described below . let @xmath140 denote the initial vertex of edge @xmath141 . let @xmath142 denote the vertex entered via edge @xmath141 . + * algorithm first phase * + * input : * an unrolled tail - biting trellis with start nodes @xmath143 , final nodes @xmath144 for the @xmath73 subtrellises , and an edge cost @xmath145 associated with each edge @xmath141 of the tail - biting trellis . + * output : * the cost @xmath146 of a least cost path to each node @xmath138 from any start node . + * begin * + * for * each node @xmath138 in the tail - biting trellis initialize @xmath147 ; + * for * @xmath29 = 1 to @xmath3 * do * + * for * each vertex @xmath138 at time index @xmath29 * do * + @xmath148 + * for * @xmath68 = 1 to @xmath73 * do * + @xmath149 + * end * at the end of the first phase therefore we have a set of survivors at final nodes @xmath144 some of which may not correspond to codewords . the costs of these paths are taken as initial estimates for the second phase . we first informally describe the second phase below and then describe a recursive version in more detail . + * algorithm second phase * + * input : * the initial metrics @xmath150 computed in the first phase for the @xmath73 subtrellises and the costs @xmath146 of the survivors at all vertices @xmath138 of the tail - biting trellis . + * output : * the maximum likelihood path . + sort the metrics @xmath150 in increasing order ; if the lowest metric is that of a codeword path then output that path as the ml path and return , else go to next step . + @xmath151 = cost of lowest codeword survivor if there is one , otherwise , otherwise @xmath152 . + if any of the metrics @xmath153 is greater than @xmath151 then discard subtrellis @xmath85 from the set of participants in the second phase . + @xmath154-@xmath155 = set of all non - discarded trellises with non - codeword survivors ; + create a set @xmath112 of the initial vertices along with metrics , of all residual trellises , and let the start node @xmath118 of the @xmath0 algorithm be the start node of the residual trellis with a minimum initial metric ; + execute lines 2 to 11 of algorithm @xmath0 modifying statement in line 11 as * if * @xmath156 * then * @xmath157 and statement @xmath158 in line 5 by @xmath159 + if the open set @xmath160 becomes empty before a final node is reached , then the codeword with cost @xmath151 is output as the decoder s estimate of the transmitted codeword . + the algorithm above is therefore different from the standard @xmath0 algorithm in the following ways : 1 . it may switch from one subtrellis to another depending on which subtrellis the node with minimum metric is located in . each shared node in a subtrellis is regarded as a distinct node for purposes of the algorithm . thus there will be as many distinct copies of a given node of the tail - biting trellis as there are residual subtrellises sharing that node . 3 . before adding an element to the open set , we check to see that its metric is less than that of the best codeword survivor stored in @xmath151 . in the traditional algorithm there is no such check . if the open set @xmath112 becomes empty before a final node is reached then the codeword with cost @xmath151 is output . we need to define the estimate @xmath120 in line 3 of algorithm @xmath0 . recall that this has to be an _ underestimate _ of the path length from node @xmath121 to the final node if the ml path is to be output . the estimate we use for node @xmath138 in subtrellis @xmath85 is the difference between the _ initial _ metric for trellis @xmath85 computed in the first phase and the cost of the _ survivor _ at node @xmath138 in the first phase . we will prove later that this is _ indeed _ an underestimate and therefore guarantees that the ml path is output on termination . we implement the open set @xmath112 as a _ heap _ @xcite . this ensures that the minimum element can be retrieved in _ constant _ time and that whenever an element is _ inserted _ into the heap , restructuring it in order to preserve the property of constant time access to the minimum element , has complexity logarithmic in the size of the heap . + we now describe the second phase of the algorithm more formally beginning with the notation used variable @xmath161 is the estimate obtained for the shortest path from the start to the final node in subtrellis @xmath45 in the first phase . variable @xmath162 is the estimate for the shortest path from node @xmath138 to node @xmath82 in subtrellis @xmath45 which is computed when an update occurs at node @xmath138 . this is the difference between the initial estimate at @xmath81 in trellis @xmath45 , and the cost of the survivor at node @xmath138 in the first phase . variable @xmath163 is a pointer to a structure representing a node in the trellis ; @xmath164 is the state , @xmath165 indicates which trellis that state belongs to ; @xmath166 stores the current metric which is the sum of the length of the path from the start node in trellis @xmath165 to @xmath167 and the estimate of the path length from @xmath167 to the final node in that trellis . variable @xmath168 is a pointer to the successor of a node ; @xmath169 and @xmath170 have meanings that can be deduced from 3 above . variable @xmath171 refers to the time index and takes on values from 0 to @xmath66 where @xmath3 is the length of the code . variable @xmath172 is a unique number associated with a subtrellis . function @xmath173 inserts a node into the heap ; function @xmath174 extracts the node with minimum value of metric from the heap . function @xmath175 returns a boolean value which is true if the heap is empty and false otherwise . variable @xmath176 represents the actual cost of the path from the start state of a subtrellis that ends at the node @xmath177 . variable @xmath178 represents the cost of the survivor in the first phase at that node . variable @xmath179 is the updated metric at a successor of a node in a trellis using function @xmath180 , which is called when that node is closed using @xmath181 . variable @xmath182 is the sequence of nodes representing the winning path at the state @xmath183 . variable @xmath151 is the cost of the lowest cost @xmath80 path in the first phase . variable @xmath184 is used to detect whether the winning path is the one identified in the first or second phase . it is initialized to 0 . if the heap becomes empty without reaching a final node in the second phase then the lowest cost @xmath80 path is output as flag remains 0 . else the path that first reaches a final node in the second phase is the winning path . * function * @xmath185 + / * begin with @xmath186 residual trellises whose metrics have been sorted in increasing order , and with variable @xmath151 which stores the metric of the best codeword survivor*/ + * begin * + / * first create a heap @xmath187 with these @xmath186 metrics ; each element of the heap is a record containing the trellis number , the node , the time index , and the metric*/ + * for * @xmath188 * to * @xmath186 * do * + @xmath189 + * endfor * + @xmath190 ; + * while * @xmath191 * false * * and * @xmath192 * do * + @xmath193 + @xmath194 /*add @xmath167 to the set of closed nodes*/ + @xmath195 / * expand @xmath167*/ + * endwhile * + * if * @xmath196 * then * output the codeword with metric @xmath151 ; * return * + * end * + * function * @xmath197 + 1 . * begin * + 2 . * if * @xmath198 * then * @xmath199 ; output @xmath182 ; * return * + 3 . * else * + 4 . * for * each successor @xmath168 of @xmath183 * do * + 5 . @xmath200 + 6 . * if * @xmath201 * then * @xmath202 + 7 . @xmath203 + 8 . * else * + 9 . * if * @xmath204 + 10 . * then * @xmath205 + 11 . * endif * + 12 . * endif * + 13 . * endfor * + 14 . * endif * + 15 . * end * + * function * @xmath206 ; + * begin * + @xmath207 + @xmath208 + * if * @xmath209 * then * + @xmath210 / * update the current shortest path to @xmath211*/ + @xmath212 / * update the cost of the current shortest path to node 2*/ + @xmath213/ * update the metric at @xmath211 ; @xmath214 is the cost of the survivor in the first phase*/ + * endif * + * end * we first prove that on termination the algorithm always outputs the optimal path each survivor at a node @xmath121 has a cost which is a lower bound on the cost of the least cost path from @xmath215 to @xmath121 in an @xmath216 path passing through @xmath121 . assume that @xmath121 is an arbitrary node on an @xmath216 path and that path @xmath217 is the survivor at @xmath121 in the first phase . there are two cases . either @xmath217 is a path from @xmath215 to @xmath121 or @xmath217 is a path from @xmath218 to @xmath121 , @xmath219 . if the latter is the case , then the cost of @xmath217 is less than the cost of the path from @xmath215 to @xmath121 ; hence the cost of the survivor at @xmath121 is a lower bound on the cost of the least cost path from @xmath215 to @xmath121 . [ properties ] the quantity @xmath220 defined in the algorithm satisfies the following two properties : 1 . @xmath221 2 . @xmath222 where @xmath117 is an edge . 1 . @xmath223 + also @xmath224 , from which the result follows . 2 . to prove : @xmath225 + lhs @xmath226 + @xmath227 + if survivor at @xmath138 is survivor at @xmath121 concatenated with edge @xmath117 , then + lhs @xmath228 + @xmath229 + on the other hand if survivor at @xmath138 is not a continuation of the survivor at @xmath121 , + @xmath230 + @xmath231 + or , @xmath232 + or , @xmath233 + therefore , @xmath225 lemma [ properties ] , and the fact that all estimates on trellises on which execution is suspended are underestimates , assures us that _ if the final node is reached in any subtrellis then this is indeed the shortest path in the tail - biting trellis or in other words the ml codeword . we first make a few observations about the algorithm . during any point in the second phase , the algorithm is exploring some path in a candidate subtrellis called the _ current _ trellis even though it may do so in discontinuous steps . this path is called the _ current path _ in that subtrellis . the metric which it uses to decide whether to continue on the current path on the current trellis , say @xmath45 , or forsake it in favour of another path either in the current trellis or on another candidate trellis is initially @xmath161 . we have the following lemma specifying how the metric changes along the path . [ update ] during the second phase , if the current path updates a node @xmath138 using function @xmath180 , where the survivor in the first phase was not in the current subtrellis then the metric becomes @xmath234 where @xmath235 is the difference between the cost of the least cost path ending at @xmath138 in the current trellis and the survivor at @xmath138 during the first pass . we know that @xmath236 and @xmath237 the metric is just the sum of the two lefthand sides of the previous two equations . thus if the survivor is the current path then @xmath238 and the lemma follows . if the survivor is not the current path then the metric is increased by the difference between the length of the current path up to @xmath138 and the survivor at @xmath138 . a critical node on a path in a subtrellis is one at which the metric for a subtrellis reaches its final value(i.e . the actual cost of the path ) . [ lem : critical ] during the second phase , once a critical node is closed in a subtrellis , the algorithm goes on to reach the final node in that subtrellis without switching trellises , and outputs an ml path . the critical node was closed because it had the minimum metric . the metric represents the _ actual _ cost of the path at a critical node . this is no greater than the metrics of all other visited nodes which are _ underestimates _ of the costs of all other paths . thus once a critical node is closed , the metric does not change along the continuation of this winning path to the final node . therefore line 6 of function @xmath181 is always true at some successor andno trellis switching takes place . the following properties hold for the metric . let @xmath239 denote the metric in subtrellis @xmath29 at node @xmath240 : [ predecessor ] let an @xmath241 path be the winner at @xmath82 in the first phase and let it win over an @xmath80 path at node @xmath242 . then @xmath243 and @xmath244 for any proper predecessor @xmath245 of @xmath242 . since the @xmath241 path was the overall winner at @xmath82 its length will be the metric at the start node of trellis @xmath45 and by lemma [ update ] , the metric on the path in @xmath45 will rise by the appropriate amounts @xmath246 at each node @xmath68 where the path was overtaken by a path from some other subtrellis . when it reaches node @xmath242 , which is a critical node , the metric will reach its final value , namely @xmath247 . since @xmath245 is a predecessor of @xmath242 and the metric _ rises _ at @xmath242 , @xmath244 . for each shortest path in a subtrellis @xmath29 , the nodes where it was overtaken by paths originating at the start nodes of other subtrellises in the first phase , are the nodes where its metric will rise during the second phase . these nodes are called _ rising points_. thus the node at the final rising point in a subtrellis is the critical node . [ share ] let subtrellises @xmath45 and @xmath85 share a node @xmath240 and between them , let @xmath45 be the first to close the node in the second phase . then @xmath248 . since @xmath45 is the first to close the node it closes it either before @xmath85 was first opened or after . if the former was the case , then @xmath249 . if the latter was the case the least current metric of @xmath85 is greater than the metric @xmath239 of @xmath45 from which the result follows as the metric can only increase . [ segment ] for nodes @xmath242 and @xmath245 let @xmath250 be a path segment in the merging interval of @xmath45 and @xmath85 and let @xmath251 . then @xmath252 . since at @xmath242 , @xmath251 and thereafter all updates to the metrics in trellises @xmath45 and @xmath85 until node @xmath245 is reached will be identical as the survivors at those node in the first phase will be the same for both trellises @xmath45 and @xmath85 , @xmath252 . we next show that any path from an arbitrary start node to any final node represents a vector in a vector space . for the sake of simplicity we restrict our arguments to binary codes . [ vectorspace ] the set of all labels from an arbitrary start node to any final node is a vector space . assume that each of the @xmath74 vectors in the submatrix @xmath56 of the generator matrix is of the form @xmath253 $ ] where @xmath254 has circular span @xmath255 $ ] , where @xmath256 stands for the sequence of symbols from the first , up to and including the @xmath257 symbol and is called the _ head _ , and @xmath258 stands for the sequence of symbols from positions @xmath68 to @xmath66 and is called the _ tail _ ; @xmath259 represents the run of zero symbols in between the head and the tail , spanning the appropriate number of codeword indices . ( this run may be empty if @xmath260 . ) let @xmath261 be the vectors of @xmath56 . then the matrix @xmath262 defined as @xmath263 $ ] , where @xmath264 consists of @xmath265 rows of the form @xmath266,[{{\mathbf 0}},{{\mathbf t_i } } ] , 1\leq i \leq c$ ] , ( where the number of zeroes in @xmath259 makes up a total of @xmath3 elements for the row ) generates the set of labels of all paths from any start node to any final node . this set has @xmath267 elements . this can be verified from the product construction . the set of elements of this vector space consists of _ semicodewords _ and codewords . each semicodeword is the label of an @xmath83 path @xmath84 . the matrix @xmath262 corrresponding to the matrix @xmath70 for the hamming ( 7,4 ) code of example [ ex : overlay2 ] is displayed below . + @xmath268\ ] ] it can be observed that the semicodeword 1100110 formed by adding rows 1 and 3 of @xmath262 traces a path from start vertex @xmath269 to final vertex @xmath270 in the tail - biting trellis of figure [ fig : hamming - overlay ] . [ not - close ] the algorithm will not close any node whose metric exceeds the cost of the ml path . the lemma follows from lines 6 and 7 of function @xmath181 and the observation that calling function @xmath181 on a node is equivalent to closing the node . the test ensures that only nodes with metric value less than the current metric are closed . since the current metric is a lower bound on the cost of the ml path the lemma follows . we use a result of tendolkar and hartmann @xcite stated below . [ tendolkar ] let @xmath187 be the parity check matrix of the code and let a codeword @xmath100 be transmitted as a signal vector @xmath106 . let the binary quantization of the received vector @xmath271 be denoted by @xmath272 . let @xmath273 and @xmath274 . then ml decoding is achieved by decoding a received vector @xmath98 into the codeword @xmath275 where @xmath276 is a binary vector that satisfies @xmath277 and has the property that if @xmath278 is any other binary vector such that @xmath279 then @xmath280 where @xmath281 is the inner product . a direct consequence of lemma [ tendolkar ] is the following result . [ ml - codeword ] if the all - zero codeword is the ml codeword for an error pattern @xmath276 then@xmath282 for any non - zero codeword @xmath48 . since the space explored by the algorithm , namely the space of semicodewords and codewords is a vector space , we can analyse the algorithm assuming that the ml codeword is the all 0 codeword . [ semicodewords ] assume the all 0 codeword is the ml codeword . let @xmath276 be the binary quantization of the received vector . for the error pattern @xmath276 the second phase of the decoding algorithm will close the start nodes of only those subtrellises whose initial metric corresponds to a semicodeword @xmath283 satisfying @xmath284 we first note that at the start of the second phase the metrics at the start nodes of all residual subtrellises correspond to the costs of vectors in the vector space of codewords and semicodewords , i.e. the vector space defined by the generator matrix @xmath262 . from lemma [ vectorspace ] we have @xmath285 where @xmath286 is the parity check matrix corresponding to the matrix @xmath262 . from lemma [ tendolkar ] maximum likelihood decoding on the set of semicodewords will initially choose @xmath283 , a semicodeword , which satisfies the inequality of the lemma and the algorithm will close the start node of the subtrellis with that initial metric . as the algorithm proceeds with updating metrics it may close start nodes of other subtrellises . however by lemma [ not - close ] it will never close the start node of any trellis @xmath85 whose initial metric exceeds that of the ml codeword , which implies that the all-0 codeword is more likely than the semicodeword survivor in @xmath85 , thus implying equation [ equ : weights ] . the properties of the algorithm proved in this section will be used to explain the good performance of the approximate algorithms described in the following section . recall that each shared node is treated as a distinct node in the second phase of the algorithm . we now propose an approximate variant of the exact algorithm which closes a shared node _ at most once _ in the second phase . we term this algorithm @xmath287 . + assume we replace line 5 of function @xmath181 by + * if * @xmath288 * then * @xmath200 * else * continue + what this ensures is that each shared node is closed _ at most once _ , that is , by at most one subtrellis , in the second phase . therefore the total number of viterbi updates in the first phase and expansions in the second phases is at most @xmath289 where @xmath4 is the number of states in the tail - biting trellis . since a node is closed by at most one subtrellis , it is conceivable that a shared node that is on the ml path is closed by a subtrellis that does not contain the ml codeword . in such a case the result produced will not be the ml codeword . we now analyse the conditions under which this happens . the symbols are the same as those defined for lemma [ semicodewords ] . the following theorem gives the conditions under which the approximate algorithm produces a non - ml output . recall that the intersection property requires that the intersection of all the zero runs of vectors in @xmath264 be non - empty . [ explanation ] if the tail - biting trellis satisfies the intersection property , the approximate algorithm produces a non - ml output for error patterns @xmath276 satisfying equation [ equ : weights ] whenever @xmath283 is a semicodeword which is formed as a linear combination of rows of @xmath262 that contain at least one non - zero multiple of a vector from @xmath55 . let us assume that the all - zero codeword is the ml codeword but that it is not the output of the approximate algorithm @xmath287 . therefore some trellis say @xmath45 must close a node @xmath240 on the all 0 path ( so that @xmath290 never gets to close it , as only one closure is allowed , and therefore can not output the all 0 path ) . clearly node @xmath240 must be in the merging interval of @xmath290 and @xmath45 . since @xmath45 is a residual trellis(otherwise it would have not participated in the second phase ) , let the survivor at @xmath82 in the first phase be an @xmath241 path that overtakes the @xmath291 path at node @xmath242 , in other words , @xmath242 is the critical node for trellis @xmath45 . * case 1 . * suppose node @xmath242 is a predecessor of node @xmath240 . by lemma [ predecessor ] , @xmath292 , and since @xmath242 is a critical node , by lemma [ lem : critical ] , @xmath45 would have gone on to win in the exact algorithm and therefore the all - zero codeword could not have been the ml codeword giving a contradiction . * case 2 . * suppose node @xmath242 is a successor of @xmath240 _ within _ the merging interval of @xmath45 and @xmath290 . by lemma [ predecessor ] @xmath243 . since @xmath259 is the ml codeword @xmath293 implying that @xmath294 . since subtrellis @xmath45 closed node @xmath240 , by lemma [ share ] , @xmath295 . by the property of the metric @xmath296 implying that @xmath297 . since @xmath242 is in the merging interval of @xmath290 and @xmath45 by lemma [ segment ] @xmath298 giving a contradiction . therefore we conclude that if subtrellis @xmath45 closes @xmath240 and @xmath242 is a successor of @xmath240 , then @xmath242 can not be in the merging interval of @xmath45 and @xmath290 . + we thus conclude that @xmath242 is beyond the merging interval of @xmath290 and @xmath45 , and hence the @xmath299 path does not touch the all - zero path . since the intersection property is satisfied , any path which is a linear combination of vectors of @xmath264 alone must have at least one node on the all - zero path . hence the semicodeword corresponding to the @xmath299 path can not be formed as a linear combination of rows only in @xmath264 and therefore it is formed as a linear combination of vectors with at least one member of @xmath55 . theorem [ explanation ] and lemmas [ ml - codeword ] and [ semicodewords ] provide an explanation of the experimental observation that decoding differences between the exact and the approximate algorithm are infrequent , so much so , that the bit error rate curves are practically indistinguishable . lemma [ semicodewords ] tells us that in order for a subtrellis to be opened it must contain a semicodeword satisfying equation [ equ : weights ] , being the most likely semicodeword among the possible candidates . theorem [ explanation ] establishes the condition that if a node on the all - zero path is closed by some trellis @xmath45 other than @xmath290 when the all - zero codeword was transmitted , then the initial metric of @xmath45 must be that of a semicodeword of pretty high weight ( because it is a linear combination of vectors which contain at least one vector in @xmath55 ) . further , the error @xmath276 which caused the cost of this high weight semicodeword to drop significantly enough to satisfy equation [ equ : weights ] , should not cause the weight of any non - zero _ codeword _ to drop by an amount enough to violate equation [ equ : ml ] . since semi - codewords share prefixes and suffixes with codewords , such events may be quite infrequent . + one could get an even better approximation by allowing a node to be closed at most twice . we have experimented with this and observe that the bit error rate for this approximation is indistinguishable from that of the exact algorithm at all values of signal to noise ratio for all the three codes on which we have run the simulations . the significance of this is that the time complexity can be explicitly bounded by the complexity of at most three computations for each node of the tail - biting trellis , one update in the first viterbi decoding phase and at most two expansions in the second phase . we now estimate the time complexity of the approximate algorithm . the following bound on the complexity of the viterbi algorithm is well known@xcite . [ viterbi ] the complexity of the first phase of the decoding algorithm is @xmath300 where @xmath115 is the number of edges in the tail - biting trellis . the next lemma is a statement of a well known result on heap data structures@xcite . [ heap ] each insertion into the heap has complexity @xmath301 where @xmath187 is the number of elements in the heap . [ approx1 ] the algorithm @xmath287 has complexity bounded by @xmath302 where @xmath4 is the number of states in the tail - biting trellis . the number of vertices that are updated is at most @xmath289 as each vertex is expanded at most once in the second phase . each time a vertex is expanded it results in computations on every edge leaving it and at most a constant number of elements being visited and inserted into the heap @xmath112,(as this number is bounded by the field size assumed to be a constant ) . the complexity of each insertion phase is @xmath303 where @xmath187 is the size of the heap . since this size is proportional to @xmath4 the complexity of the second phase is @xmath304 . the sorting operation at the end of the first phase has complexity @xmath305 where @xmath18 is the number of states at time index @xmath306 . the complexity is dominated by the @xmath302 term and hence the theorem . to reduce the overheads , the heap is implemented as @xmath307 separate heaps if there are @xmath307 residual trellises , with a separate heap of pointers , each element of which points to the root of a distinct subtrellis heap . the individual heap sizes are small in practice and the algorithm is practically linear in the size of the trellis . in the next section we present results from profiling the program which bear out the claim that the overheads of heap operations are negligible . an argument similar to that in theorem [ approx1 ] estalishes the complexity of algorithm @xmath308 as @xmath302 . we next look at the space complexity of the algorithm . + [ lem : space - complexity ] the space requirement for algorithm @xmath287 is @xmath309 bits . the algorithm requires @xmath310 space to store the estimates at each state in the first phase . the additional space required to store the heap is also @xmath310 as each expanded node can put at most all its successors on the heap . the bit vectors that store trellis membership are of size @xmath18 where @xmath18 is the number of start nodes of the tail - biting trellis . the space requirements for the bit vectors is therfore @xmath311 bits . the space requirements for storing the current cost at each node is @xmath310 . this follows from the fact that each shared node is closed at most once . this means that at most one copy of a shared node updates its succesors . this in turn means that each successor has at most one update along each of its incoming edges . since the number of incoming edges is a constant which is at most the size of the field , a constant number of costs are associated with each node in the tail - biting trellis from which the result follows . we have coded the exact and approximate algorithms and show the results of simulations on minimal tail - biting trellises for the 16 state tail - biting trellis @xcite for the extended ( 24,12 ) golay code on an awgn channel with antipodal signaling , and tail - biting trellises for two rate 1/2 convolutional codes with memory 6 , circle size 48 ( which is the same as the ( 554,744 ) convolutional code experimented with in @xcite , and memory 4 , circle size 20 ( which is the same as the ( 72,62 ) convolutional code used in @xcite respectively . we show the variation of both , the average as well as the maximum number of node computations ( counting viterbi updates in the first phase and expansions in the second phase ) with the signal to noise ratio for our exact algorithm , and compare this with the number of viterbi updates needed for the brute force approach . note that this number is indicative of the _ time complexity _ of the algorithm . the results are encouraging and are displayed in tables [ tab : expansions1 ] , [ tab : expansions2 ] and [ tab : expansions3 ] respectively for the golay code and the two convolutional codes . _ on the average , the number of updates to get the exact ml result requires fewer than two computations at each node of the tail - biting trellis at all values of signal to noise ratio , one in the first pass and one in the second_. the maximum number of node computations for the algorithm @xmath287 is obviously bounded by twice the number of nodes in the tail - biting trellis . we also display the bit error - rate performance of the approximate algorithms closing nodes at most once for the first approximation @xmath287 , and at most twice for the second approximation , @xmath308 in figures [ fig : golay - ber ] , [ fig : conv1-ber ] , [ fig : conv2-ber ] and and find that there is virtually no difference in the bit error rates for the second approximation and the exact ml algorithm . _ thus we get virtually ml performance for an explicit linearly bounded update complexity at all values of signal to noise ratio_. * average node computations * + 0.0 & 285 & 602 & 245.2 + 0.5 & 294 & 688 & 235.3 + 1.0 & 311 & 709 & 225.7 + 1.5 & 273 & 637 & 217.7 + 2.0 & 271 & 580 & 210.6 + 2.5 & 256 & 576 & 204.8 + 3.0 & 289 & 643 & 200.1 + 3.5 & 242 & 557 & 197.2 + 4.0 & 192 & 480 & 195.1 + 4.5 & 152 & 423 & 193.8 + 5.0 & 135 & 396 & 193.0 + * average node computations * + 0.0 & 13064 & 22311 & 4414.1 + 0.5 & 15698 & 24958 & 4051.4 + 1.0 & 13161 & 20369 & 3738.5 + 1.5 & 12926 & 18981 & 3487.9 + 2.0 & 9948 & 16162 & 3330.0 + 2.5 & 7492 & 11700 & 3233.5 + 3.0 & 5743 & 11175 & 3175.0 + 3.5 & 3354 & 7163 & 3138.2 + 4.0 & 2781 & 6447 & 3115.0 + 4.5 & 1526 & 5104 & 3099.5 + 5.0 & 1059 & 4693 & 3088.2 + * average node computations * + 0.0 & 701 & 1437 & 426.9 + 0.5 & 784 & 1447 & 405.4 + 1.0 & 824 & 1554 & 384.9 + 1.5 & 749 & 1426 & 367.6 + 2.0 & 623 & 1214 & 353.5 + 2.5 & 563 & 1179 & 342.7 + 3.0 & 578 & 1162 & 334.6 + 3.5 & 503 & 984 & 329.5 + 4.0 & 412 & 918 & 326.2 + 4.5 & 292 & 718 & 323.7 + 5.0 & 241 & 660 & 322.3 + we have proposed an exact algorithm for ml decoding on tail - biting trellises and also experimented on two approximate variants . the average time complexity of the exact algorithm is seen to be quite low . the approximate variants perform as well as the exact one in terms of the bit error rate at an explicitly bounded update complexity equivalent to two , or sometimes three rounds on the tail - biting trellis . the algorithm does not suffer from the effects of limit cycles or pseudocodewords which current iterative algorithms are subject to . profiling measurements carried out on the program are displayed in table [ tab : profile ] . the execution time was averaged over 10,000 runs of the decoder . the percentage of execution time taken up by each of the five major operations in the decoding process , namely , the initializations of all the arrays , the first pass , the sorting operation at the end of the first pass , the second pass , and the heap operations is displayed . it can be observed that heap operations incur an overhead of only 11 % of the program running time at 0 db and are negligible for higher values of signal to noise ratios . & * initializations * & * phase 1 * & * sorting * & * phase 2 * & * heap operations * + & & & & & + 0.0 & 14.75% & 34.09% & 5.88% & 31.38% & 11.12% + 1.0 & 8.34% & 46.95% & 8.34% & 24.94% & 6.61% + 2.0 & 6.07% & 58.84% & 10.20% & 17.62% & 2.74% + 3.0 & 3.21% & 69.02% & 10.79% & 12.68% & 0.92% + 4.0 & 1.13% & 74.23% & 11.49% & 9.62% & 0.23% + 5.0 & 0.52% & 75.70% & 11.59% & 8.425% & 0.09% + the results of simulations on the extended ( 24,12 ) golay code , a rate 1/2 , memory 6 convolutional code with a circle size of 48(which is the same as the ( 554,744 ) convolutional code used for experiments in @xcite and a rate 1/2 memory 4 convolutional code with a circle size of 20(which is the same as the ( 72,62 ) rate 1/2 convolutional used for experimentation in @xcite ) have been reported . it is seen that the second approximate variant has a bit error rate which is indistinguishable from that of the exact algorithm for all values of signal to noise ratio . + * acknowledgement * the authors gratefully acknowledge discussions with aditya nori . they would also like to thank the anonymous referees for their comments which greatly improved the presentation of the paper . + 99 l.e . aguado and p. g. farrell , on hybrid stack decoding algorithms for block codes , _ ieee trans . inform . theory _ , january 1998 , pp 398 - 409 . a. aho , j.e . hopcroft and j.d . ullman , _ data structures and algorithms _ , addison - wesley , reading , ma . 1983 . s. aji , g. horn , r. mceliece and m. xu , iterative min - sum decoding of tail - biting codes , _ proceedings of information theory workshop _ , killarney , ireland , june 22 - 26 , pp . 68 - 69 . anderson and s.m . hladik , an optimal circular viterbi decoder for the bounded distance criterion , _ ieee transactions on communications _ , * 50*(11 ) , november 2002 . anderson and s.m . hladik , tail - biting map decoders,_ieee journal in selected areas in communication _ , * 16*(2 ) , february 1998 . calderbank , g.d . forney , jr . , and a. vardy , minimal tail - biting trellises : the golay code and more , _ ieee trans . inform . theory _ , * 45*(5 ) , july 1999 , pp . 1435 - 1455 . r.v . cox and c.v . sundberg , an efficient adaptive circular viterbi algorithm for decoding generalized tailbiting convolutional codes , _ ieee transactions on vehicular technology _ * 43*(1 ) , february 1994 , pp 57 - 68 . kaustubh deshmukh , priti shankar , amitava dasgupta , and b. sundar rajan . on the many faces of block codes . in _ symposium on theoretical aspects of computer science(stacs ) _ , pages 5364 , 2000 . l. ekroot and s. dolinar , @xmath0 decoding of block codes , _ ieee trans . _ * 44 * ( 9 ) , september 1996 , pp 1052 - 1056 . han , c.r.p . hartmann and c .- c . chen , efficient priority - first search maximum - likelihood soft - decision decoding of linear block codes , _ ieee trans . inform . theory _ * 39 * ( 5 ) september 1993 , pp 1514 - 1523 . han , c.r.p . hartmann , and k.g . mehrotra , decoding linear block codes using a priority first search : performance analysis and suboptimal version , _ ieee trans . inform . theory _ * 44*(7 ) , november 1998 , pp 3091 - 3096 . han , a new treatment of priority - first search maximum - likelihood soft - decision decoding of linear block codes , _ ieee trans . inform . theory _ * 44*(7 ) november 1998 , pp 3091 - 3096 . n . chen and h. -b . wu , a maximum - likelihood soft - decision sequential decoding algorithm for binary convolutional codes , _ ieee trans . * 50 * ( 2 ) , february 2002 , pp 173 - 178 . han , c.r.p . hartmann and k.g . mehrotra , decoding linear block codes using a priority - first search : performance analysis and suboptimal version , _ ieee trans . inform . theory _ * 44*(3 ) , may 1988 , pp 1233 - 1246 . f. jelinek , a fast sequential decoding algorithm using a stack , _ ibm j. res . devel _ * 13 * , 1969 , pp 675 - 685 . r. koetter and a. vardy , on the theory of linear trellises , _ information , coding and mathematics _ ( m. blaum , editor ) , boston : kluwer , may 2002 . kschischang and v. sorokine , on the trellis structure of block codes , _ ieee trans . inform . theory _ , * 41*(6 ) , nov 1995 , pp . 1924 - 1937 . kschischang and v. sorokine , a sequential decoder for linear block codes with a variable bias term metric , _ ieee trans . inform . theory _ january 1998 , pp 410 - 411 . kudryashov , decoding of block codes obtained from convolutional codes , _ problemy peredachi informatsii_,*26*(2 ) , pp 18 - 26 , april - june 1990(in russian ) . english translation , plenum publishing corporation , oct 1990 . mceliece , on the bcjr trellis for linear block codes , _ ieee trans . inform . theory _ , * 42 * ( 4 ) , july 1996 , pp 1072 - 1092 . j.h.ma and j.k.wolf , on tail - biting convolutional codes , _ ieee trans . commun._,*34*,february 1986 , pp 104 - 111 . nilsson , principles of artificial intelligence , tioga publishing co. , palo alto , ca , 1980 . p. shankar , a. dasgupta , k. deshmukh and b.s . rajan , on viewing block codes as finite automata , theoretical computer science , * 290*(2003 ) 1775 - 1797 . p. shankar , p.n.a . kumar , k. sasidharan and b.s . rajan , ml decoding of block codes on their tail - biting trellises , in _ proc . 2001 ieee int . symposium on information theory _ , ieee press , 2001 , pp . c.shih , c.r . wulff , c.r.p . hartmann and c.k . mohan , efficient heuristic search algorithms for soft - decision decoding of linear block codes , _ ieee trans . inform . theory _ , november 1998 , pp 3023 - 3038 . g.solomon and h.c.a . van tilborg , a connection between block and convolutional codes , _ siam j. appl . _ , * 37 * , october 1979 , pp 358 - 369 . m. tanner , a recursive approach to low complexity codes , _ ieee trans . inform . theory _ * 27 * , pp 533 - 547 , 1981 . tendolkar and c.r.p . hartmann , generalization of chase algorithms for soft decision decoding of binary linear codes , _ ieee trans . inform.theory,_*30*(5 ) , september 1984,pp 714 - 721 .

while string theory has had remarkable successes over the last several years , accelerated by the revolutions in understanding its non perturbative properties , it is still very much the case that we do not yet know what the theory is . we can not state unambiguously what the basic degrees of freedom are ( it is highly context dependent in a way that depends upon the dynamics themselves ) , and even the backgrounds in which the theory propagates are themselves open to interpretation . for example , in some descriptions and situations , the theory contains gravity , and in others , it does not . from some perspectives there are open strings present , and from others , only closed . ironically , several of these frustrating ( from the point of view of finding simple definitions ) features are also among the theory s most powerful positive traits , allow an ever widening range of applications of the theory to diverse problems , often of a strongly coupled nature . while applications continue , it is still important to try to get to grips with what the theory is . at the very least , this is important from a pragmatic standpoint , since perhaps a useful definition or characterization of string theory might be put to use as , for example , a diagnostic device in identifying when a physical problem may have some aspect of it that is amenable to solution by string theory methods . more generally , if string theory ultimately plays some fundamental role in the understanding of physics beyond the standard model , and/or in cosmology and other origins questions about the universe at large , a more profound understanding of the nature , power , and scope of the theory would seem to be highly desirable . at best , to date , as a result of various dualities , we know that it is probably part of some larger physical framework which itself is only string theoretic in various corners of its parameter space . this physical setting , called m theory , remains profoundly mysterious well over a decade after the first clear glimpses of it@xcite . historically , problems pertaining to such essential matters of understanding in physics are greatly illuminated by having a rich set of examples that are simple , but yet complex enough to contain all the important phenomena in question . for the problems outlined above , it would be rather excellent to have the simplest possible string theories that still contain some of the marvellous non - perturbative physics we know and love , and be able to follow them as they connect to each other in ways that are entirely invisible in perturbation theory . further icing on the cake would be to have the physics all captured in terms of relatively familiar structures for which there is an existing technology for its study . this is the subject of this paper ( and a follow up to appear later@xcite ) , at least in part . the simplest known strings with tractable non perturbative physics that contain a rich set of phenomena ( such as holography and open closed dualities ) are the minimal strings@xcite , and in particular ( where non perturbative physics is concerned ) the type 0 strings ( formulated in refs.@xcite and refs.@xcite , and recognized as type 0 strings in ref.@xcite ) . the non perturbative formulation of the type 0a and type 0b strings can be done rather beautifully in terms of certain integrable systems , as we will review later : type 0a has the korteweg de vries system while type 0b has zakharov shabat . the non perturbative physics of each is formulated in terms of associated non linear ordinary differential equations often called `` string equations '' . while much of the language of the two is similar , these are very different systems , and except for various accidental ( from the perspective of those separate formulations ) perturbative coincidences ( and a non trivial non perturbative equivalence for one model see later ) , the physics of each are quite separate indeed . this paper builds on all of these results , taking them much further . we have found that there is a larger framework into which the type 0a and type 0b string theories can be naturally embedded and within which they are connected as parts of a larger theory . we found further that the two string theories are merely special points in a much larger tapestry of possibilities . when perturbation theory is examined , other special points suggest themselves , and they turn out to be just as `` stringy '' as the original type 0 theories , deserving to be thought of as string theories as well . we begin the program of trying to identify some of these theories , with some success . we also find that the larger framework provides natural definitions of regimes of the type 0 string theories that are hard to define using perturbation theory , and we will report more fully on non perturbative aspects in a follow up paper@xcite . in this sense , we have a precise analogue of m theory . we have a larger physical framework that is not itself a string theory , but that can be readily specialized to yield string theories as special limits . we can move between different theories in a quite natural way , which is nonetheless outside the framework of any of its daughter string theories . we find this encouraging and exciting . at the base of our infinite family of string equations , organizing much of this remarkable structure , is a non linear differential equation known as painlev iv . this well known equation from the classical mathematics literature , part of a celebrated family of six equations , has two arbitrary constants , usually denoted @xmath7 and @xmath8 . ( actually , two copies of painlev iv turn up in our story , intertwined in an interesting way . ) it turns out that the type 0a and 0b points in the tapestry of theories occur at the vanishing of one or other of these constants for one of the copies of painlev iv . the vanishing of the constants of the second painlev iv hint at interesting new special points . after reviewing crucial aspects of the type 0a and 0b string theories in section [ sec : type0 ] , we unpack the dispersive water wave hierarchy and present the infinite family of equations we propose as the string equations in section [ sec : dww ] . section [ sec : painleve ] highlights the role of painlev iv . in section [ sec : connectab ] we show how the structures of section [ sec : type0 ] arise as special points in this larger framework , while section [ sec : beyond ] is a detailed study of the rich properties of the string equations and the types of solutions available . we organize and classify a great deal of the physics that appears , and notice in section [ sec : square ] that much of the physics can be organized in terms of a square . the square is reminiscent of the main square organizing the moduli space of @xmath9 ( two dimensional ) string theories , discovered in ref.@xcite , and we contemplate a possible connection , perhaps induced by renormalization group flow@xcite to the @xmath0 context of the work in question . the possible relation between the squares helps us make a conjecture about the nature of two new special points we find : they might be type iia and iib minimal string theories . ( note that these are type ii theories in the sense of the structure of the gso projection used to formulate them . there is no spacetime supersymmetry@xcite . ) in section [ sec : new ] we carry out a comparison of the new structures we found to some continuum computations for one loop partition functions . from this we strengthen aspects of our type ii suggestion . we conclude in section [ sec : conclusion ] with a brief summary and discussion . we will start with a brief review of the type 0 string theories coupled to the @xmath10 superconformal minimal models and show how these theories can be elegantly described within the framework of an integrable hierarchy of partial differential equations ( pdes ) accompanied by an hierarchy of ordinary differential equations ( odes ) . these models exhibit novel and interesting physics , all of which and more will be seen to be embedded in the dww system that we will describe in the next section . this review will help establish our notation and the framework upon which we can readily build the more general structure . we begin with the following ordinary differential equation ( known in the old days as a `` string equation '' ) @xmath11 this equation ( or , really , family of equations ) and its properties have been studied in several papers . it was first derived and studied as a fully non perturbative definition of a string theory in refs.@xcite , and evidence that it defines a type 0a string theory was presented first in ref@xcite . further properties of the equation , in particular concerning how branes and fluxes are encoded by it and the underlying integrable ( kdv ) system , were presented in refs.@xcite . here @xmath12 is a real function of the real variable @xmath13 , a prime denotes @xmath14 , and @xmath15 and @xmath16 are real constants . the quantity @xmath17 is defined by @xmath18 where the @xmath19 $ ] are polynomials in @xmath12 and its @xmath13derivatives , called the gelfand dikii polynomials@xcite . they are related by a recursion relation ( defining a recursion operator @xmath20 ) @xmath21 and fixed by the value of the constant @xmath22 and the requirement that the rest vanish for vanishing @xmath23 . some of them are : @xmath24 the @xmath25th model is chosen by setting all the @xmath26 to zero except @xmath27 and @xmath28 this number is chosen so that the coefficient as @xmath29 . if we had instead chosen @xmath30 , we would have chosen the coefficient of @xmath31 to be unity . ] of @xmath31 in @xmath32 is set to @xmath33 . the function @xmath12 defines the partition function @xmath34 of the string theory @xmath35 @xmath36 where @xmath37 is the coefficient of the lowest dimension operator in the world sheet theory . so @xmath12 is a two point function of the theory . from the point of view of the @xmath25th theory , all the other @xmath38 represent couplings of closed string operators @xmath39 . it is well known@xcite that the insertion of each operator is captured in terms of the integrable kdv hierarchy of flows describing how @xmath40 evolves in @xmath26 : @xmath41 for the @xmath25th model , equation ( [ streqn0a ] ) , which has remarkable properties@xcite , is known to furnish a complete non perturbative definition of a family of spacetime bosonic string theories @xcite . the models are actually type 0a strings @xcite , based upon the @xmath10 superconformal minimal models coupled to super liouville theory . as superconformal theories , they have central charge @xmath42 the asymptotic expansions for the first two @xmath25 are : @xmath43 @xmath44 @xmath45 @xmath46 it should be noted that the solution for @xmath47 can be numerically and analytically shown to match onto the solution for @xmath48 , providing a unique@xcite non perturbative completion of the theory . ( see figure [ fig : plot ] for an example of a solution found using numerical methods . ) type 0a solution showing how the perturbative regimes at large @xmath49 are smoothly connected . section [ sec : dww ] discusses a function @xmath50 ( @xmath51 ) , which has a number of different classes of behaviour distinguished by choice of boundary condition . the type 0a theory has class @xmath52 in the @xmath53 perturbative regime and class @xmath54 in the @xmath55 perturbative regime . here we have set @xmath56 and @xmath57.,title="fig:",width=453 ] + as instructed in equation ( [ 0afreeenergy ] ) , integrating twice the asymptotic expansions ( such as those in equations ( [ 0aexpnsk=1 ] ) and ( [ 0aexpnsk=2 ] ) ) furnishes the free energy @xmath58 , and it can be seen to define a perturbative expansion in the dimensionless string coupling @xmath59 for all models , in the @xmath60 regime , @xmath15 represents@xcite the number of background zz d branes @xcite in the model , with a factor of @xmath15 for each boundary in the worldsheet expansion . these are point like branes localized at infinity in the liouville direction @xmath61 , deep in the strong coupling region . in the @xmath62 regime , @xmath15 represents the number of units of rr flux in the background , with @xmath63 appearing when there is an insertion of pure rr flux @xcite . since there is a unique non perturbative solution connecting the two regimes , the string equation ( [ streqn0a ] ) supplies a non - perturbative completion of the theory that is a very clear example of a geometric transition between these two distinct ( d - branes @xmath64 rr fluxes ) spacetime descriptions of the physics . the function @xmath12 is the potential in the hamiltonian @xmath65 of the well known ( in the inverse scattering literature ) associated sturm liouville problem connected to the integrable kdv hierarchy . the wavefunctions of that problem define the partition functions of fzzt@xcite d branes stretched along the liouville direction @xmath61 , ending at a finite @xmath66 set by the eigenvalue . the zero energy problem is interesting@xcite , since there the fzzt d branes stretch to infinity , and the hamiltonian s factorization , @xmath67 where @xmath68 , is highly convenient . the function @xmath69 ( its definition in the equation before is termed a miura map in the integrable literature ) satisfies@xcite an infinite hierarchy of equations sometimes called the painlev ii hierarchy since the equation at @xmath70 is the painlev ii equation . the asymptotic expansion of @xmath71 generated by these equations is in terms of worldsheets involving zz d branes ( or fluxes ) and fzzt d branes . the entire problem defines a toy supersymmetric quantum mechanics problem within which the celebrated bcklund transformations of the kdv system can be made manifest . in this language the zz d branes are identified with the number of threshold bound states ( formally , zero velocity solitons ) of the system , and the bcklund transformations change their number by an integer . these more recently established features@xcite , together with the earlier identification@xcite of the role of the kdv flows in organizing the close string operators , show how the integrable model and inverse scattering technology of the mathematical physics literature comes to life in organizing the open and closed string content of minimal string theory . type 0b string theory coupled to the @xmath72 superconformal minimal models @xcite is described succinctly by the following string equations@xcite : @xmath73 where the @xmath74 and @xmath75 are polynomials of functions @xmath76 and @xmath77 ( and their derivatives ) , and @xmath16 and @xmath78 are real constants . the differential polynomials satisfy the following recursion relations @xmath79 where a prime denotes @xmath80 . some of them are : @xmath81 the function @xmath82 defines the partition function of the theory @xmath35 @xmath83 the @xmath84th model is chosen by setting all @xmath85 to zero except @xmath86 and @xmath87 , analogous to what was done in the previous section concerning the 0a case . note that these models have an interpretation as type 0b strings coupled to the @xmath88 superconformal minimal models only for even , lack of modular invariance of the partition function rules out the interpretation as type 0b strings coupled to superconformal matter @xcite . ] writing @xmath89 , we again have a set of models connected to the @xmath10 superconformal minimal models , this time type 0b . as in the 0a case , from the point of view of the @xmath25th theory , all the other @xmath26 represent coupling to closed string operators @xmath39 . again the insertion of each operator can be expressed in terms of the zakharov shabat@xcite hierarchy of flows , the underlying integrable system in this case : @xmath90 where @xmath91 . the asymptotic expansions of the string equations ( [ streqn0b ] ) for the first even @xmath92 are : @xmath93 @xmath94\ , \quad ( x \rightarrow \infty ) \nonumber\\ { \widetilde w}(x ) & = & \frac{\nu q\sqrt{2}}{4|x|^{1/2 } } - \frac{\nu^2 q^2}{4 |x|^2 } + \frac{\nu^3}{|x|^{7/2}}\frac{5\sqrt{2}}{64 } q\left(1 + 4q^2\right)+\cdots \quad ( x \rightarrow -\infty)\end{aligned}\ ] ] @xmath95 @xmath96 upon integrating twice , the asymptotic expansions in equations ( [ 0bexpnsm=2 ] ) and ( [ 0bexpnsm=4 ] ) furnish the free energy perturbatively as an expansion in the dimensionless string coupling , given by the same expression as before in equation ( [ 0astrcoupling ] ) . for these models , in the @xmath62 regime , @xmath78 represents the number of background zz d branes in the model , with a factor of @xmath78 for each boundary in the world sheet expansion , while in the @xmath97 regime it counts the number of units of rr - flux in the background@xcite . the asymptotic expansions in the two directions can be argued in ref.@xcite to match onto each other analytically in a particular ( t hooft ) limit . for the case @xmath70 , the full non perturbative solution is known since it can be mapped directly to the solution known for the @xmath70 type 0a case , as will be discussed below in section [ sec : connectingab ] . for later reference , we briefly discuss the structure of these solutions with increasing @xmath84 . as argued in ref.@xcite the @xmath98 expansions are deformations of the solutions of the equation with @xmath99 . however things get interesting for @xmath100 . as before , the @xmath101 solution is a deformation of the solution with @xmath99 . for @xmath102 , @xmath99 allows for the trivial solution @xmath103 , but trying to deform this for @xmath104 leads only to a complex solution . additionally , @xmath99 allows for two nontrivial solutions with @xmath105 and @xmath106 . these two are related by @xmath107 and are interpreted as @xmath108 symmetry breaking solutions . in the interpretation of ref.@xcite , this is due to the presence of r both these solutions have a real extension to the case with @xmath109 . the requirement of matching a @xmath102 solution to the @xmath101 solution picks out one of these for @xmath110 and the other for @xmath111 . ( the @xmath112 solution listed above is for @xmath110 . ) for higher @xmath84 , more such symmetry breaking solutions arise . we will see how this is organized explicitly in section [ sec : beyond ] . it turns out that the simplest case of @xmath113 ( @xmath70 ) , the 0a and 0b theories are non perturbatively related in a very special way . in this case the conformal model is trivial ( _ i.e. _ @xmath114 ) and we simply have the pure world sheet supergravity sector . the strings are unencumbered by a spacetime embedding ( not counting the ubiquitous liouville direction , @xmath61 ) . the string equation for the @xmath70 0a theory , equation ( [ streqn0a ] ) with @xmath115 , is @xmath116 on the other hand , the string equation ( [ streqn0b ] ) for the @xmath70 0b theory can be written succinctly as @xmath117 notice that the perturbative expansion for the @xmath70 0a theory as @xmath118 ( equation ( [ 0aexpnsk=1 ] ) ) looks similar to the perturbative expansion for the @xmath98 0b theory as @xmath119 ( equation ( [ 0bexpnsm=2 ] ) ) up to a ( non universal ) sphere term , once one identifies @xmath15 with @xmath78 . the two expansions are just offset by various powers of @xmath120 . in fact there exists a non perturbative map between the two equations @xcite that can be seen as follows . first define a function @xmath121 _ via _ @xmath122 for which the string equation for the 0a theory ( [ streqn0ak=1 ] ) becomes @xmath123 after rescaling using @xmath124 equation ( [ 0ak1to0bm2 ] ) becomes the string equation for the 0b theory ( [ streqn0bm=2 ] ) , but with the sign of @xmath125 reversed . the physics of pure 0a supergravity and pure 0b supergravity ( _ i.e. _ @xmath114 ) are non perturbatively related , with the brane and flux perturbative regions exchanged . this non perturbative connection between the 0a and 0b theories ( for @xmath70 ) follows , mathematically , from the fact that the same basic string equation appears at the base of the two separate ( kdv _ vs _ zs ) hierarchies of equations . this connection is partly in the spirit of a much larger set of connections that we are reporting on in this paper . we find that the kdv and zs structures are embedded in a much larger structure , the dispersive water wave hierarchy of equations , and find a class of connections ( of a different sort ) between 0a and 0b for all @xmath25 and see that they define two special corners of a larger tapestry of physical theories . the standard dispersive water wave ( dww ) hierarchy@xcite , which will play a central role in the new physics we uncover , is a two component system . it is described by : @xmath126 \quad , \ ] ] where @xmath127 , @xmath128 ( note that here and in the rest of the paper , for any function @xmath129 , @xmath130 will denote @xmath131 ) , and we adopt a matrix notation : @xmath132=\left(\begin{array}{c } l_n[u , v]\\ k_n[u , v]\end{array}\right ) \ .\end{aligned}\ ] ] here , @xmath133 is the recursion operator for the dww hierarchy . the operator @xmath134 can be written as the quotient of two hamiltonian operators @xmath135 and @xmath136 , @xmath137 where @xmath135 and @xmath136 are given by @xmath138 the @xmath139 obey the recursion relation @xmath140 which follows immediately from ( [ dwwpde ] ) . the first few @xmath141 and @xmath142 are as follows : @xmath143 the normalization of @xmath144 is chosen so as to reproduce ( [ dwwpde ] ) for @xmath145 . + the dww hierarchy can be reduced to a one component system by demanding that one of the two independent functions vanish . if we set @xmath146 , we actually reduce to the kdv hierarchy : two operations of the ( reduced ) dww recursion operator give @xmath147 where @xmath148 is the recursion operator of the kdv hierarchy , shown in equation ( [ kdvrec ] ) . thus we obtain a reduction of the even flows of the original hierarchy ( [ dwwpde ] ) to @xmath149 which is the kdv system in @xmath150 with independent variable @xmath125 , and the even times of dww map to the times of the kdv @xmath151 . ( compare with equation ( [ kdvflow ] ) ) . + in addition , the recursion relation ( [ lrec ] ) reduces to @xmath152 which is exactly the kdv recursion relation ( [ kdvrec ] ) . the relative normalizations are @xmath153 and @xmath154 so we conclude that @xmath155 . moreover , with @xmath156 , it is immediate that @xmath157 . the other obvious reduction , @xmath158 , reduces the system to the burgers hierarchy . we do not explore if there are any string theory consequences of that in this paper , since @xmath50 is used to define the partition function of our theories in all our examples . integrable hierarchies of partial differential equations ( pdes ) can be reduced to ordinary differential equations ( odes ) through an additional condition on the variables . in our context , these odes are sometimes to be thought of as defining string theories ; they are the `` string equations '' of a family of theories , forming an hierarchy themselves . as outlined around equation ( [ kdvflow ] ) , string equations can be thought of as supplying the initial conditions for the partition function , and then the pdes of flows describe how the partition function evolves as a function of the operators that couple to the @xmath159@xcite . the original example of all this@xcite was a hierarchy of equations that have the painlev i equation as the non trivial equation at their base , indexed by an integer @xmath25 . they defined the bosonic @xmath160 string theories coupled to the @xmath161 conformal minimal models . it was later realized@xcite that another rich family of string equations ( those in equations ( [ streqn0a ] ) ) can be obtained by imposing certain scaling relations on the variables of the kdv system ( note however that the equations were originally derived@xcite directly from matrix model constructions analogous to the original route ) . with this in mind , we explore a similarity reduction of the dww hierarchy , expecting to obtain new string equations at the end of the day . we follow the approach originally used to derive the string equations of type 0a ( [ streqn0a ] ) for the kdv hierarchy @xcite . to that end , assign @xmath162 mass dimension 1 . the dimensions of the other terms uniquely follow from ( [ dwwpde ] ) and are @xmath163=\frac{1}{2}$ ] , @xmath164=-\frac{1}{2}$ ] and @xmath165=-\frac{1}{2}(n+1)$ ] thus we can write down two callan symanzik equations expressing the scaling symmetry , @xmath166 using ( [ dwwpde ] ) and ( [ lrec ] ) we can rewrite these equations , @xmath167 defining , @xmath168 we can rewrite ( [ cs1 ] ) , @xmath169 where we have used the fact that we will take @xmath170 and the other @xmath87 to be independent of @xmath125 . equation ( [ cs21 ] ) can readily be integrated . moreover , solving ( [ cs21 ] ) for @xmath171 and substituting the result into ( [ cs22 ] ) yields an expression , which , after multiplying by @xmath172 , can also be integrated . the results are our new coupled string equations , @xmath173 where we have introduced two integration constants , @xmath174 and @xmath15 . we stress that the simplest possibility is for @xmath174 and @xmath15 to be independent of @xmath125 _ and _ @xmath175 , though only independence of @xmath125 is strictly necessary . + the @xmath84th model is chosen by setting all @xmath175 equal to zero except for @xmath170 and @xmath87 which is chosen to be a numerical factor to fix the normalization . we choose to parameterize @xmath87 as @xmath176 in order to make direct contact with some recent literature which discusses this system in a much different ( mathematical ) context@xcite . let us focus on the case @xmath177 , which forms the bottom of the hierarchy of string equations from which all others follow using the recursion relations . the string equations in this case reduce to the painlev iv equation , an important equation from the mathematical literature . its appearance at the bottom of the ladder of string equations we re presenting here is significant . note that the entire family of string equations can be generated from this @xmath177 case by use of the recursion operator , and so structures at this level will be reflected at higher @xmath84 , even while the complexity of the equations increases . also notable is that this is the first time that a role for this equation has been uncovered in this context of non perturbative string theory , and it takes its place alongside the painlev i and ii equations whose roles ( mentioned earlier ) have been established in this context already . in fact , part of the motivation that led to the discoveries upon which we report here was the question as to the further role of the painlev equations in such systems . painlev iv emerged naturally as a candidate equation to play a role and this led to our studying of the dww system that we found connected to painlev iv in the literature@xcite . let us see more explicitly how the equation emerges . remarkably , it will naturally appear in two different ways@xcite . the string equations for @xmath177 are : @xmath178 where we have used the relation @xmath179 . solving the first of these for @xmath162 gives @xmath180 and substituting this into the second yields a second order ode in @xmath181 which , under the change of variables @xmath182 becomes @xmath183 y - \nu^2 \frac{1}{2 } \frac{\beta^2_1}{y } \quad .\ ] ] @xmath184 and @xmath185 are constants related to @xmath174 and @xmath15 in the dww string equations through @xmath186 setting @xmath187 , and dropping the subscripts on the constants , gives @xmath188 which is the fourth painlev equation @xmath189 in standard form , and @xmath190 we will see in the next section , specific constraints yielding the 0a and 0b theories that require @xmath191 and @xmath192 respectively . notably , these are precisely the values for which the constants @xmath7 and @xmath8 in the standard form of painlev iv vanish . in fact , there is another natural appearance of the painlev iv equation in this system@xcite , at @xmath177 . there is a natural generalization@xcite of the miura map ( that we saw for kdv in section [ sec : type0 ] ) to the dww system , defining new variables @xmath193 and @xmath194 : @xmath195 now , as we saw above , @xmath196 satisfies painlev iv with constants @xmath7 and @xmath8 given in equation ( [ eq : painleve ] ) . well , additionally , @xmath197 satisfies a copy of painlev iv ( equation ( [ eq : painleve ] ) ) also , but with constants related to our physical parameters by @xmath198 we take this seriously , not the least because the variables described by the miura map in the case of kdv ( type 0a ) were seen to be physically very natural , pertaining as they do to the fzzt and zz d branes . ( see the end of section [ sec:0a ] for a brief review . ) we expect therefore ( but this needs more exploration ) that this dww miura map also leads to rich physics . the cases @xmath199 and @xmath200 imply vanishing of @xmath7 and @xmath8 and may well have some special significance in this context . ( we will , for example , find special solutions for all @xmath84 corresponding to @xmath199 points . it is also interesting to note that the equations together point to the values @xmath201 , @xmath202 , values which do feature prominently in what is to follow . ) having introduced the dww hierarchy , we now show how both the type 0a and type 0b string theories coupled to the @xmath10 superconformal minimal models can be found embedded in this system . we show that by placing appropriate constraints on the full system of string equations , one can recover the respective string equations for both of the type 0 theories . quite beautifully , these constraints require one of the two integration constants @xmath203 to freeze to particular values . the remaining unfixed constant then acts as the parameter that counts the number of zz branes or units of r r flux in each theory , depending on which asymptotic region ( positive or negative large @xmath125 ) is under consideration . it was seen in section ( [ sec : dww ] ) that setting @xmath181=@xmath204 reduces the dww hierarchy to the kdv hierarchy . we therefore expect that this constraint also reduces our new string equations to the 0a string equations . that this indeed occurs can be seen as follows . equation ( [ lk ] ) gives , @xmath205&=\sum_{n=0}^{\infty}\frac{1}{2}(n+1)t_n^{\mathrm{dww}}l_n \\ & = \sum_{n=0}^{\infty}\frac{1}{2}(2n+1)t_{2n}^{\mathrm{dww}}l_{2n } \\ & = 4\sum_{j=0}^{\infty}(j+\frac{1}{2})t_{2j}^{\mathrm{dww}}p_{j } \\ & = \mathcal{r}[w ] \end{split}\ ] ] where we have used that @xmath206=@xmath207=0 $ ] and @xmath208=@xmath209=@xmath210=4p_n[w]$ ] ( see below ( [ kdvrec2 ] ) ) . the last equality holds provided that we make the identification , @xmath211 finally , we see that when @xmath156 and @xmath212 , equation ( [ dwwstring2 ] ) exactly reduces to equation ( [ streqn0a ] ) , _ i.e. _ our new string equations encode 0a string theory coupled to the @xmath10 superconformal minimal models . for even flows it is easy to show that @xmath156 is only consistent with the other string equation ( [ dwwstring1 ] ) if @xmath174 is frozen : @xmath213 so one of the parameters in the original dww equations becomes fixed when recovering type 0a coupled to the @xmath10 superconformal minimal models , while the other parameter @xmath15 counts the number of branes or units of rr flux in the type 0a theory . we will see this behaviour again in the case that we recover the type 0b theory . consider the following redefinition of the dww variables @xmath214 to the zs variables @xmath215 : @xmath216 the recursion relation ( [ lrec ] ) becomes , @xmath217 or , @xmath218 where we have defined , @xmath219 these are precisely the recursion relations of the zs hierarchy ( [ zsrec ] ) . moreover , the @xmath220 and @xmath221 just defined actually agree with those presented in ( [ zspoly ] ) . it suffices to check @xmath145 : from ( [ dwwpolys ] ) we have @xmath153 and @xmath222 which imply @xmath223 and @xmath224 , as expected . + finally , we may ask how we can produce the 0b string equations ( [ streqn0b ] ) from our new string equations ( [ dwwstring1 ] ) and ( [ dwwstring2 ] ) . the answer turns out to be simple and elegant : all we must do is set @xmath225 equation ( [ dwwstring1 ] ) then requires , @xmath226 which further implies , @xmath227 the @xmath87 required here to consistently produce the equations of @xcite are identical to the values we determined earlier ( [ tn ] ) . so , upon identifying @xmath228 , we have exactly produced the 0b string equations . + again notice how the consistency of our constraint @xmath229 with ( [ dwwstring2 ] ) forces one of our parameters to vanish @xmath230 leaving the parameter @xmath231 to count the number of zz branes or rr fluxes in the type 0b theory . finally , we remark that the partition function of the 0b theory is determined _ via _ @xmath232 so that @xmath150 encodes the partition function for both 0a and 0b . we have seen that constraining the dww string equations ( [ dwwstring1 ] ) appropriately leads to the 0a and 0b theories ( coupled to the superconformal @xmath10 series ) , respectively . the constraints take the system with two free parameters @xmath203 and define two special points : 0a with @xmath233 free ) and 0b with @xmath234 free , @xmath235 . we can also consider the fully unconstrained system with both parameters @xmath203 unfixed , and general @xmath236 . interestingly , we get _ multiple _ asymptotic expansions for the variable @xmath50 . the structure of the equations and the corresponding asymptotic expansions gets richer as @xmath84 increases . since in both cases , asymptotic expansions of @xmath50 gave us , upon integrating twice , an expansion of a partition function for a string theory , we look again for it to define an interesting partition function in the new cases we will encounter . while this is an assumption , we shall see it bear fruit presently . in this section we will first list the asymptotic expansions for the first few @xmath84 obtained from the corresponding string equations . we will explain the organizational rules we use to group these expansions into various classes . a careful analysis of the patterns we uncover in what follows allows us to extrapolate to higher @xmath84 and predict the structure of the expansions for any @xmath84 . we will see that a subset of these , when appropriately combined , reproduce the type 0 expansions that we have already encountered . in addition , we obtain _ completely new _ expansions which have not been presented in the literature before . our key observation here is that these also resemble perturbative sectors of string theories ( either with branes or fluxes present ) . we take these seriously as new string theories and our task after this section will be to identify what string theories they might be . the number of expansions grows large as @xmath84 increases ( we will see later that the number of expansions is @xmath237 ) . to deal with this proliferation of expansions , we classify them into classes whose members are related to one another by simple symmetries . the classes themselves are distinguished by a number of salient features , many of which we explore in what follows . we choose to define the classes based on their behavior at order @xmath238 ( this is equivalent to the leading behavior in @xmath239 , the sphere level of closed string perturbation theory , as we will see later ) . since dww is a two component system , we must consider the leading behavior of both functions , @xmath181 and @xmath162 . we adopt the following classification scheme : + @xmath240 + we postpone the detailed study of @xmath181 to subsequent work . here we mention its leading behavior only to complete the classification ; in what follows , we focus exclusively on @xmath162 and its asymptotic expansions . the details of sections [ sec : expandone][sec : expandfour ] , being a list of examples that we found instructive , might be a little dry on first reading and so the reader is encouraged to skip to section [ sec : general ] for the general case . the string equations for this case were already written in equations ( [ dwwn=1 ] ) . solving the first of these for @xmath162 gave equation ( [ dwwn=1v ] ) , and substituting into the second yields a scalar second order ode in @xmath242 ( equivalent to painlev iv ) , which can be used to produce the expansions . asymptotic expansions for @xmath242 can then be used to yield asymptotic expansions for @xmath50 using equation ( [ dwwn=1v ] ) . we obtain three classes of expansions for @xmath50 , @xmath243 upon integrating twice ( following what we learned from the type 0 theories in sections [ sec:0a ] and [ sec:0b ] ) , one can obtain the free energy for a genus expansion of a string theory which allows us to identify the string coupling to be @xmath244 . the other expansions within each class can be obtained by the following symmetry operation : @xmath245 since @xmath246 and @xmath247 contain only even powers of @xmath15 , the are invariant under this map ; however , @xmath248 . thus there are two expansions in the @xmath249 class , and , together with @xmath246 and @xmath247 , these comprise four total @xmath177 expansions . the string equations are : @xmath250 where again @xmath251 . solving the first of these for @xmath162 gives , @xmath252 substituting this into the second yields a scalar fourth order ode in @xmath181 , which can then be used to produce the expansions for @xmath162 . in this case , there are four relevant classes of expansions @xmath253 here , we see that the string coupling is @xmath254 . the other expansions within each class can be obtained by the following operations , @xmath255 and compositions thereof . some of the @xmath256 are invariant under one or both of these maps . altogether , there are nine distinct expansions . here are the four classes of expansions together with the number of distinct expansions within each class and the maps that lead to them : @xmath257 the string equations are too long to be written down explicitly here and so we omit them . four classes of expansions are produced in this case , and one sees expansions in class 5 appearing for the first time here . @xmath259 here the string coupling turns out to be @xmath260 the other expansions within each class can be obtained by the following operations , @xmath261 and any compositions of these maps . a quick calculation shows that there 16 different expansions @xmath262 our explicit string equations are rather complicated and so we will not list them here . the following five classes of expansions are obtained : @xmath264 here , the string coupling is @xmath265 . the other expansions within each class can be obtained by the following operations , @xmath266 and any arbitrary composition of those maps . a quick calculation shows that there are 25 distinct expansions . here are the five classes of expansions together with the number of distinct expansions within each class and the maps that lead to them : @xmath267 the expansions displayed above for @xmath177 to @xmath100 exhibit a rich structure which we explore shortly . first we briefly review the interpretation given to the parameter @xmath15 of the 0a theory ( and also to @xmath78 of the 0b theory ) . in the @xmath37 ( or @xmath125 ) @xmath268 regime , @xmath15 represents the number of background zz d branes in the model , with a factor of @xmath15 for each boundary in the worldsheet expansion . since an orientable surface with odd ( even ) euler characteristic must contain an odd ( even ) number of boundaries , @xmath15 must be raised to an odd ( even ) power if @xmath269 is . in addition , the power of @xmath15 must be less than or equal to the power of @xmath269 . on the other hand , in the @xmath37 ( or @xmath125 ) @xmath270 regime , @xmath15 represents the number of units of rr flux in the background , with @xmath63 appearing when there is an insertion of pure rr flux . so in this case both @xmath15 and @xmath269 should appear with even powers . + in applying these observations to our dww expansions , we immediately notice the remarkable fact that the various expansions have powers of the parameters which somehow allow for interpretations as counting branes or fluxes . this is by no means guaranteed , and indeed its occurrence was one of our main motivations for in depth study of the system . the presence of two parameters , however , leads to a few subtleties . for example , in some expansions an interpretation in terms of branes is only possible if one of the two parameters is set to zero . with keep such observations in mind as we begin the study of the various expansions . finally we note that the asymptotic direction ( _ i.e. , _ positive or negative @xmath125 ) of each expansion can be fixed by requiring that once we fix the value of @xmath271 , the expansion must be real ( which is an important constraint since @xmath162 encodes the free energy ) . the value of @xmath271 , in turn , can be fixed using the values listed in equation ( [ tn ] ) since we must reproduce the 0a theory . with all of these observations we are ready to begin analyzing our expansions . * @xmath272 contains powers of @xmath15 consistent with those of a parameter counting branes . this remains true for any value of @xmath174 . * @xmath272 contains powers of @xmath174 consistent with those of a parameter counting fluxes . however , for arbitrary @xmath15 , @xmath269 appears with odd powers , inconsistent with our requirements for a description of fluxes , as mentioned above . this problem is avoided if we set @xmath57 since this forces the odd powers of @xmath269 to vanish . * in particular , setting @xmath192 with @xmath271 given by equation ( [ tn ] ) reduces this expansion to the @xmath101 flux - expansion in @xmath174 ( or @xmath78 ) for the type 0b theory seen in equations ( [ 0bexpnsm=2],[0bexpnsm=4 ] ) for @xmath273 respectively . * alternatively , setting @xmath274 with the same value of @xmath271 reduces these expansions to those of the type 0a for @xmath101 . ( we listed them in equations ( [ 0aexpnsk=1 ] and [ 0aexpnsk=2 ] ) for @xmath275 respectively is related to kdv hierarchy index @xmath25 by @xmath92 . ] . ) * with the values of @xmath271 need to reduce to 0a and 0b , one obtains real expansions in this class _ only _ if @xmath276 . hence we fix this class of expansions to be @xmath277 asymptotic expansions . * this class of expansions only exists for even @xmath84 . * for even @xmath84 , @xmath246 contains powers of @xmath15 and @xmath269 consistent with those of a parameter counting fluxes . this is true for all values of @xmath174 . for odd @xmath84 , the powers of @xmath15 are still consistent with the flux interpretation , but there are odd powers of @xmath269 which are inconsistent with fluxes . these odd powers can be removed by setting @xmath278 . * for even @xmath84 , @xmath246 contains powers of @xmath174 and @xmath269 consistent with a parameter counting fluxes . this is true for all values of @xmath15 . for odd @xmath84 , the powers of @xmath174 are consistent with those of a parameter counting branes . in this interpretation , there are no contributions from surfaces with only one boundary . * setting @xmath279 with @xmath271 as chosen in equation ( [ tn ] ) , reduces the @xmath246 expansions to the type 0a expansions for @xmath102 . ( we listed them in equations ( [ 0aexpnsk=1 ] ) and ( [ 0aexpnsk=2 ] ) for @xmath275 respectively . ) * only even powers of @xmath271 and @xmath125 appear , so the requirement of reality does not fix the direction of these expansions . * consistency with the 0a expansions forces us to consider the expansions in this class as @xmath280 expansions . we note the possibility that these might appear as @xmath277 expansions outside of the simple type 0a context we ve seen so far type expansions ( for @xmath281 ) in both asymptotic directions for @xmath125 . clearly there are analogous rational solutions for the full dww equations that have @xmath246 asymptotic expansions that generalize the known cases . we have constructed large families of them , and leave their study for a later publication . ] . * these expansions vanish when @xmath282 . ( this is a likely special point(s ) in parameter space . we got a first hint of this point in section [ sec : painleve ] where the second copy of painlev iv has @xmath283 . ) * this class of expansions appears for all @xmath84 . * @xmath249 contains powers of @xmath15 consistent with those of a parameter counting branes . this is true only for @xmath278 . * @xmath249 contains powers of @xmath174 consistent with those of a parameter counting branes . this is true only for @xmath57 . * we notice that associating one boundary to each factor of @xmath174 _ and _ @xmath15 also produces a consistent worldsheet expansion . we might speculate about whether , in general , these expansions might capture @xmath174 and @xmath15 simultaneously counting branes . * setting @xmath192 with @xmath284 reduces these expansions to the @xmath102 brane - expansions for the type 0b theory for @xmath98 as seen in equation ( [ 0bexpnsm=2 ] ) . hence we fix the expansions in this class to be @xmath280 expansions . * at @xmath100 , the value @xmath285 with @xmath192 renders this expansion complex for @xmath112 . this fits in nicely with the structure of expansions observed in the 0b case , reviewed in that case did not have a real deformation for @xmath104 . the @xmath249 class is exactly the analogue of this trivial solution . ] in section [ sec:0b ] . * @xmath286 causes these expansions to vanish ( we got a first hint of this point in section [ sec : painleve ] where the second copy of painlev iv has @xmath283 . ) * this class of expansions exists for all @xmath84 , but is real as an @xmath102 expansion only for @xmath98 mod @xmath287 . * the @xmath247 class of expansions has not , to our knowledge , made a previous appearance in the literature , as it does not appear until encountering the dww system . * @xmath247 contains powers of @xmath174 consistent with those of a parameter counting branes . this remains true for any value of @xmath15 . * @xmath247 contains powers of @xmath15 consistent with those of a parameter counting fluxes . however , for arbitrary @xmath174 , @xmath269 appears with odd powers , inconsistent with our requirements for a description of fluxes , as mentioned above . this problem is avoided if we set @xmath278 since this forces the odd powers of @xmath269 to vanish . * the direction of @xmath247 is not immediately determined by the consistency conditions we have used so far . compatibility with the type 0 theories at @xmath98 requires @xmath288 which renders @xmath247 real for @xmath280 . on the other hand , compatibility with the type 0 theories at @xmath100 requires @xmath289 which renders @xmath247 real for @xmath277 . * we will later provide evidence in favor of @xmath247 existing for @xmath290 . * in general , for the type 0 choices ( [ tn ] ) for @xmath271 , @xmath247 remains real for @xmath101 when @xmath291 mod @xmath287 and becomes complex when @xmath113 mod @xmath287 . * in the special case @xmath98 , @xmath247 can be made real by setting @xmath278 . * @xmath292 contains powers of @xmath15 consistent with those of a parameter counting branes . this is true only for @xmath278 . * @xmath292 contains powers of @xmath174 consistent with those of a parameter counting branes . this is true only for @xmath57 . * as for @xmath249 , we notice that associating one boundary to each factor of @xmath174 _ and _ @xmath15 also produces a consistent worldsheet expansion . we might speculate that , in general , these expansions might capture @xmath174 and @xmath15 simultaneously counting branes . * these expansions do not exist for @xmath293 . * with @xmath192 and @xmath294 this class reduces to the @xmath102 brane expansion in @xmath174 seen for the 0b theory at @xmath100 in equation ( [ 0bexpnsm=4 ] ) . ( recall that there @xmath295 . ) * these are the non trivial broken symmetry solutions obtained in the 0b theory as one increases @xmath84 . we reviewed this at the end of section [ sec:0b ] . * as @xmath84 increases , further expansions in this class arise for every odd @xmath84 , which we generically label @xmath296 . these are distinguished by the different values of @xmath297 in ( [ classes ] ) , but since their behavior is identical for our purposes we often group them together . * for odd @xmath84 , reality imposes no restrictions on the direction of @xmath296 . * for @xmath100 , reality requires that we fix @xmath292 to be an @xmath298 expansion . for the subsequent even @xmath84 , some of the @xmath296 are real for @xmath299 , while the remaining are real for @xmath300 . we can extrapolate the pattern observed for the first few @xmath84 and make predictions for the structures that should appear at higher @xmath84 . the first observation is that there are @xmath237 expansions in all ( taking into account the various expansions related by symmetries in each class ) at each @xmath84 . the counting can be broken down as follows . .the number and types of expansion classes for @xmath50 , as a solution to the string equations ( [ dwwstring1 ] ) and ( [ dwwstring2 ] ) , with increasing @xmath84 from 1 to 8 . see text for further discussion . [ cols="^,^,^,^,^,^,^,^,>",options="header " , ] as table 1 indicates , we have multiplied the torus terms by unknown normalizations @xmath301 , @xmath302 and @xmath303 , which we will later attempt to fix . notice that we have distinguished two possibilities for @xmath304 : @xmath305 and @xmath306 , to match the @xmath307 choices . we now construct the three possible @xmath308 from these torus terms . the known theories : 0a and 0b @xmath309 \quad \ \quad ( x > 0 ) \nonumber\\ z^{(\rm 0a , 0b)}_{\rm even } & = & \frac{1}{2}\left[a c^2 + \left(\frac{n+3}{2n}\right)b\gamma^{2}\right]\;\;\quad\quad\quad\quad\quad \quad\quad\quad ( x < 0)\end{aligned}\ ] ] 2 . the unknown theories : @xmath310 and @xmath311 @xmath312 \quad \quad ( x > 0 ) \nonumber\\ z^{(\rm \widetilde{a } , \widetilde{b})}_{\rm even } & = & \frac{1}{2}\left[\left(\frac{n-1}{2n}\right)\tilde{a}c^2 - \tilde{b}\gamma^2\right ] \quad\qquad\qquad\qquad\quad\qquad \qquad\quad ( x < 0)\end{aligned}\ ] ] 3 . the unknown theories : @xmath313 and @xmath314 @xmath315 \quad , \quad ( x > 0 ) \nonumber\\ z^{(\rm \widehat{a } , \widehat{b})}_{\rm even } & = & \frac{1}{2}\left[\hat{a}\left(\frac{n+1}{12n}\right ) - \frac{1}{n}\hat{a}c^2 - \hat{b}\gamma^2\right ] \qquad \qquad\qquad \qquad\quad \quad\quad ( x < 0)\end{aligned}\ ] ] we will now see what conditions are required to deduce the special parameter values , @xmath274 and @xmath192 , for type 0a and 0b respectively . we begin by observing that the continuum partition functions listed in section [ continuum ] are independent of the sign of @xmath125 . therefore , we will impose this as a constraint on every @xmath316 that we construct , @xmath317 we impose this condition by equating the two expressions in equation ( [ zevenknown ] ) . interestingly , the @xmath84dependence factorizes completely leading to @xmath318 while it is indeed possible for @xmath174 and @xmath15 to inherit @xmath84 dependence from dependence on @xmath87 , the simplest possibility is that these parameters are independent of @xmath84 . that the @xmath84 dependence factors out of equation ( [ relncgammaaa ] ) allows this simplicity to be realized . this suggests that we also impose @xmath319 substituting ( [ relncgammaaa ] ) back into ( [ zevenknown ] ) we obtain , @xmath320 finally , we impose one more condition , @xmath321 which , together with condition 2 , forces us to conclude @xmath322 condition 3 is motivated by the remarkable fact that @xmath323 _ is _ independent of @xmath84 , which is nontrivial given the form of equation ( [ pfnaamins2 ] ) , and also by the observation that this constraint correctly produces the required parameter values in this case . this is as far as we can go without some extra information . fortunately , for the type 0 theories , we actually know @xmath324 and @xmath325 : the factor of half is because of the doubling of the free energy the 0b theory relative to the 0a theory and the negative sign is because it is @xmath326 which defines the two point function for these theories . ( recall the observation at the end of section ( [ sec:0breduc ] ) ) . thus we see that we obtain the correct parameter values @xmath327 and @xmath57 for the theories under consideration . moreover , we correctly obtain the known value to be the @xmath112 contribution to the 0b theory leading us to @xmath328 . this means that the @xmath112 contribution to the 0b theory is actually zero . while @xmath249 is real as a @xmath112 expansion for @xmath329 , it is complex for @xmath330 . the resolution is that the zero contribution is to be understood as coming from the _ trivial _ @xmath331 solution of the 0b theory obtained by setting @xmath332 in the 0b string equations . it is interesting that our procedure using @xmath249 should give parameter values that are consistent with the _ trivial _ solution . ] of @xmath333 . next , we turn to exploring where our new conditions take us in investigating the unknown corners . using equation ( [ zevenunknown2 ] ) , we see that condition 1 gives , @xmath335 conditions 2 and 3 give , @xmath336 and so result means that the contribution to the torus terms coming from @xmath249 is zero . in a sense , this result is the analog of what was seen for the type 0b theories . our argument again is that the torus term for @xmath102 at the @xmath337 corner comes from the trivial @xmath338 solution , analogous to the 0b case . ] we have @xmath339 remarkably enough , everything seems to work as before , except now we obtain new values for our parameters and a new expression for @xmath340 . we have yet to make an numerical prediction for @xmath340 as we still must fix the normalizations . we will return to this point in next section [ sec : new - theories ] . condition 1 gives : @xmath342 conditions 2 and 3 imply , @xmath343 this result is not as encouraging . @xmath316 does not depend on @xmath344 which implies that it does not depend on theory @xmath345 at all . we find this unacceptable and therefore conclude that conditions 2 and 3 are incompatible constraints in this case . we nevertheless take seriously the possibility that theories like this may exist , but with @xmath84dependent parameters and an @xmath84dependent @xmath316 . we will briefly consider this possibility in section ( [ sec : ndep ] ) . in the above we have explored the crucial and quite constraining assumption that @xmath316 is independent of the index @xmath346 . theories @xmath347 and @xmath348 seemed particularly suited to produce such a @xmath316 . in an effort to better understand what these theories may be , we next ask which continuum theories are capable of producing an @xmath84independent @xmath340 . looking at the various partition functions listed in section [ continuum ] , it is easy to see that only the following choices give rise to an @xmath84independent partition function : + 1 . @xmath349 in equation ( [ pfnaamins2 ] ) with @xmath350 and @xmath351 . and @xmath78 labeling the super minimal models must obey : @xmath352 ; @xmath353 ; if both are odd , they are coprime and if both are even , then @xmath354 and @xmath355 must be coprime . there is also a standard restriction that if @xmath356 and @xmath78 are even , then @xmath357 must be odd . it follows that if @xmath350 then @xmath351 . ] these theories are type 0 string theories coupled to the @xmath10 @xmath1 models and are already described at the known corners . @xmath358 in equation ( [ pfnsadmins1 ] ) with @xmath359 and @xmath360 .. the choice @xmath361 would suggest that the @xmath362 @xmath2 model exists at at @xmath70 ; however , all combinations of real expansions at @xmath70 have been exhausted in describing the type 0 @xmath1 @xmath363 model . we therefore expect the @xmath362 model to appear at @xmath364 where more combinations of expansions exist since @xmath247 is real . thus we choose @xmath365 with @xmath366 . ] these theories are the type 0 strings theories coupled to the @xmath367 @xmath2 models with @xmath368 . @xmath369 in equation ( [ pfnaaminstypeii2 ] ) with @xmath350 and @xmath351 . these are the type ii strings coupled to the @xmath10 @xmath1 models with @xmath370 . @xmath371 in equation ( [ pfnsadminstypeii2 ] ) with @xmath350 and @xmath351 . these are the type ii strings coupled to the @xmath10 @xmath2 models with @xmath370 . this is the same partition function as case @xmath372 above . @xmath371 in equation ( [ pfnsadminstypeii1 ] ) with @xmath359 and @xmath373 . these theories are the type ii string theories coupled to the @xmath374 @xmath2 models with @xmath375 . + we seek to identify @xmath376 with one of the above partition functions , but to do so we will need some way of determining @xmath377 and @xmath378 . we consider the possibility that the theories @xmath379 and @xmath380 are type iia and type iib string theories , respectively . as we will see , this will uniquely fix the normalizations and allow us to identify the resulting @xmath381 as that of case ( 5 ) above , _ i.e. _ type ii string theories coupled to the @xmath374 @xmath2 models . pursuing the apparent similarity with the moduli space of @xmath382 theories@xcite a little further , ( recall our discussion in section [ sec : square ] , and figure [ fig3 ] ) we can study the partition functions of those theories ( compactified on a circle ) to get clues as to the possible relative normalizations . we list them here@xcite : @xmath383 an important clue is to be found in the @xmath384 behaviour of each theory . this is the physics of the field theory sector ( kaluza klein states ) that propagate on the circle , and the relative normalizations are of interest to us . type 0b has twice as much energy in this sector as type 0a , and type iia has double that of type iib . now , as expected , the type 0 theories exchange under the t duality operation @xmath385 , as do the type ii theories . duality vertically should take the type ii theories to the type 0 theories , but this needs to be done on a circle twisted by @xmath386 . the partition functions for iia / b on such a circle is : @xmath387 notice that the field theory term of type iia is now actually half that of type iib . now we can compare the overall type ii normalization to the type 0 by doing the vertical t duality . under @xmath385 we see that type iib field theory term matches that of the type 0a theory , while for type iia , the field theory term is 1/4 that of the type 0b value . in summary , we see that comparing squares would suggest that @xmath388 and @xmath389 , and the chain of relationships above gives the relative normalization of 0b and 0a as 2 to 1 , while that of ( twisted ) iib and iia is 2 to 1 , so @xmath390 . finally , t duality ( with a twist ) between 0a and iib gives them the same relative normalization , implying , @xmath391 in this case we get therefore that for our putative type ii theories , @xmath392 which is equal to that of the type 0 case , but with an opposite sign . this is the same as the partition function ( [ pfnsadminstypeii1 ] ) for the type ii theories coupled to the @xmath374 @xmath2 super minimal models . interestingly , equation ( [ gammaad ] ) implies that @xmath393 for iib and @xmath278 for iia , which nicely mirrors the values @xmath394 for 0a and @xmath57 for 0b . further work is required to strengthen our conjecture that we have here the type iia and iib theories coupled to the superconformal minimal models . as stated , an explicit definition of @xmath0 type ii strings ( separately for types a and b , and not just our suggestion for @xmath316 given in section [ sec : type - ii - z - even ] ) does not yet seem to exist in the literature , and our attempts to directly define them so far are incomplete . note that we have assumed that the relative normalizations of the partition functions that follow from @xmath382 really descend to the @xmath0 case , and while reasonable , this needs to be proven . the explicit @xmath25 dependence of the individual partition functions that we get by taking the dww one loop expressions seriously have a ( so far ) unilluminating form at positive @xmath125 that we have not been able to check against a continuum computation . on the other hand the negative @xmath125 result neatly mirrors the type 0 case . we list the results here , and leave this line of investigation for the future and @xmath395 theories are again type 0a and type 0b string theories , respectively , perhaps coupled to an @xmath2 modular invariant . this would again fix the relative normalization of the partition functions such that @xmath396 and @xmath397 , yielding @xmath398 , which unfortunately differs from the continuum partition function for the @xmath367 @xmath2 super minimal models by a sign . to match , we would have to assume that @xmath399 encodes the partition function instead of @xmath326 , which contradicts our earlier establishment of the @xmath1 type 0 points . ] : @xmath400 this should be compared with the type 0 cases displayed in equation ( [ pfns0a0b ] ) . our main successes so far have involved theories in which @xmath340 and the values for @xmath15 and @xmath174 have been independent of @xmath84 . while these are the simplest possibilities , there is little reason to suspect that these theories are the only types amenable to description by dww . we observed in section [ sec : unknown - tilde ] that there might be new theories , _ e.g. , _ the ones we called @xmath401 and @xmath402 , whose description will most likely require such @xmath84dependence for @xmath403 . unfortunately , the availability of continuum calculations to compare to is limited , so obtaining convincing evidence for any particular proposal is difficult . nevertheless , there is much more evidence to accumulate from the differential equations themselves . we hope to report upon this issue in later work . we briefly mention one way that @xmath84dependence might emerge naturally . the @xmath316 that we have constructed from the torus terms of our perturbation expansions depend on three variables : @xmath84 , @xmath174 , and @xmath15 . we must however impose sign independence which reduces the total number of independent variables describing @xmath340 to two . in special cases the algebra conspires and we find that @xmath316 actually depends on no variables at all . this behavior is to be compared to @xmath316 as computed from the continuum partition functions . for generic @xmath356 and @xmath78 , @xmath316 is a function of two variables , but in special cases this dependence completely vanishes . this motivates us to attempt to express for @xmath84 , @xmath174 , and @xmath15 as functions of @xmath356 and @xmath78 . equation ( [ z0a0b ] ) indicates that @xmath404 is inversely proportional to @xmath84 . the form of equation ( [ pfnaamins2 ] ) then suggests that we take @xmath405 . since @xmath406 in this case , we can predict @xmath407 now equating @xmath404 of equation ( [ z0a0b ] ) with the quantity in equation ( [ pfnaamins2 ] ) and using the relation @xmath408 , which follows from sign independence , gives @xmath409 note that the models already studied had @xmath350 which gave us @xmath57 . we can perform the same exercise for @xmath410 which we have argued describes type ii strings coupled to @xmath2 @xmath374 models . again we have @xmath411 , which holds for @xmath359 and @xmath412 . equating @xmath413 of equation ( [ relncgammaad ] ) with the quantity in equation ( [ pfnsadminstypeii1 ] ) and using @xmath414 gives , @xmath415 for the models we ve previously considered , @xmath359 , so we correctly reduce to the condition @xmath278 . note here that we would run into problems if we considered @xmath416 because then @xmath174 would become imaginary . as discussed at length in the introduction , we find very significant this rich framework into which we can embed so many string theories and discover how they intertwine with each other . we suspect that there may be many more theories to be found in this framework , and that we have only just begun to learn how to extract and identify the various theories using the limited comparisons we can do to existing continuum computations . we have been able to gather evidence for a square of connected theories , firmly establishing the top corners as type 0a and 0b theories coupled to the @xmath1 modular invariant @xmath10 minimal models , and finding several strong pieces of evidence that the special points at the bottom are string theories , the ( non supersymmetric@xcite ) type iia and iib theories coupled to the @xmath2 modular invariant @xmath374 minimal models . see figure [ square2 ] . notice that the @xmath249 side of the square is generically understood as defining the theory when the number background branes , _ i.e. , _ @xmath174 or @xmath15 , is set to zero . this is the case for type 0b ( already established in ref . @xcite ) and is inherited by the type iia theory as well . away from @xmath278 @xmath249 is complex for @xmath25 even , and one of the symmetry breaking @xmath417 may supply the @xmath102 physics instead . ( for @xmath25 odd , @xmath249 is real for @xmath418 and so may well furnish the @xmath102 physics in those cases . ) it is also important to note that for @xmath25 odd , @xmath247 is no longer real ( with the exception of @xmath278 at @xmath70 ) . this suggests that this regime of perturbation theory at @xmath101 ( containing fluxes for type iia and branes for type iib ) is ill defined at those values of @xmath25 , even while the opposite regime at @xmath102 exists . this may suggest non perturbative instabilities for those values of @xmath25 . an instructive prototype of this possibility is the original bosonic family of string theories defined non perturbatively in terms of a painlev i hierarchy of string equations@xcite . at large @xmath125 , the leading behaviour for the two point function was @xmath419 for the @xmath25th model , and for positive @xmath125 the physics was the @xmath161 conformal minimal models coupled to liouville theory . for even @xmath25 , the the @xmath102 regime gives complex values , signalling the non perturbative problems . something analogous may be going on here for odd @xmath25 ( except @xmath70 , @xmath278 ) for the type ii @xmath2 @xmath374 models . we will report further non perturbative aspects in a follow up paper where we discuss numerical and analytic studies of solutions that connect different perturbative regimes@xcite . + there are many more interesting issues to understand . while we compared to our proposal for a continuum definition of the even sector partition function of the type ii theories , a direct continuum definition of the type iia and iib string theories would be valuable to have , in order to provide another check on our conjecture for the new points . whether or not they are type ii string theories , it is clear that these new theories are of interest , and may ( as already stated ) be only the first in a very large family of new theories that our string equations define . they are all nicely interconnected with the known type 0 theories , and have a rich non perturbative sector . we will explore much of this non perturbative physics in a follow up paper@xcite . we have seen several signs that the organizing square is inherited from the organizing square of theories seen at @xmath382 in ref.@xcite . we suspect this inheritance arises physically by rg flow , and it would be of value to explore this further ( there are bosonic investigations in the literature@xcite ) . in particular , the special points with @xmath420 , where our partition functions vanish ( also suggested by the underlying painlev iv structures we saw in section [ sec : painleve ] , which are also worth understanding better ) , may well be related to the supersymmetric points identified in ref.@xcite . in summary , we have found a rich laboratory of solvable string theory models with several non perturbative connections between them by realizing them as special points of a larger physical system , the theory of dispersive water waves . in some respects it is an analogue of what we would like to see in studies of m theory . it will be interesting to learn whether the larger framework of dispersive water waves can yield any new insights about the non perturbative nature of string theories and related theories . this work was supported by the us department of energy . ri thanks tameem albash , nikolay bobev , arnab kundu , hubert saleur , and nicholas warner for useful conversations . cvj thanks the aspen center for physics for hospitality while some of this work was carried out . c. m. hull and p. k. townsend , `` unity of superstring dualities , '' _ nucl . * b438 * ( 1995 ) 109137 , http://www.arxiv.org/abs/hep-th/9410167[hep-th/9410167 ] . p. k. townsend , `` the eleven - dimensional supermembrane revisited , '' _ phys . * b350 * ( 1995 ) 184187 , http://www.arxiv.org/abs/hep-th/9501068[hep-th/9501068 ] . e. witten , `` string theory dynamics in various dimensions , '' _ nucl . * b443 * ( 1995 ) 85126 , http://www.arxiv.org/abs/hep-th/9503124[hep-th/9503124 ] . r. iyer , c. v. johnson , and j. s. pennington , `` non perturbative string theory from water waves ( working title ) , '' _ _ to appear , _ _ ( 2010 ) . d. j. gross and a. a. migdal , `` nonperturbative two - dimensional quantum gravity , '' _ phys . * 64 * ( 1990 ) 127 . e. brezin and v. a. kazakov , `` exactly solvable field theories of closed strings , '' _ phys . _ * b236 * ( 1990 ) 144150 . m. r. douglas and s. h. shenker , `` strings in less than one - dimension , '' _ nucl . * b335 * ( 1990 ) 635 . t. r. morris , `` 2-d quantum gravity , multicritical matter and complex matrices , '' . fermilab - pub-90 - 136-t . s. dalley , c. v. johnson , and t. r. morris , `` multicritical complex matrix models and nonperturbative 2- d quantum gravity , '' _ nucl . * b368 * ( 1992 ) 625654 . s. dalley , c. v. johnson , and t. r. morris , `` nonperturbative two - dimensional quantum gravity , '' _ nucl . phys . _ * b368 * ( 1992 ) 655670 . c. v. johnson , t. r. morris , and a. watterstam , `` global kdv flows and stable 2-d quantum gravity , '' _ phys . _ * b291 * ( 1992 ) 1118 , http://www.arxiv.org/abs/hep-th/9205056[hep-th/9205056 ] . s. dalley , c. v. johnson , t. r. morris , and a. watterstam , `` unitary matrix models and 2-d quantum gravity , '' _ mod . lett . _ * a7 * ( 1992 ) 27532762 , http://www.arxiv.org/abs/hep-th/9206060[hep-th/9206060 ] . c. crnkovic , m. r. douglas , and g. w. moore , `` physical solutions for unitary matrix models , '' _ nucl . * b360 * ( 1991 ) 507523 . t. j. hollowood , l. miramontes , a. pasquinucci , and c. nappi , `` hermitian vs. anti - hermitian 1-matrix models and their hierarchies , '' _ nucl . * b373 * ( 1992 ) 247280 , http://www.arxiv.org/abs/hep-th/9109046[hep-th/9109046 ] . i. r. klebanov , j. m. maldacena , and n. seiberg , `` unitary and complex matrix models as 1-d type 0 strings , '' _ commun . * 252 * ( 2004 ) 275323 , http://www.arxiv.org/abs/hep-th/0309168[hep-th/0309168 ] . m. noumi , `` painlev equations through symmetry , '' _ american mathematical society ( translations of mathematical monographs , volume 223 ) . _ ( 2004 ) . n. seiberg , `` observations on the moduli space of two dimensional string theory , '' _ jhep _ * 03 * ( 2005 ) 010 , http://www.arxiv.org/abs/hep-th/0502156[hep-th/0502156 ] . d. j. gross and i. r. klebanov , `` one - dimensional string theory on a circle , '' _ nucl . phys . _ * b344 * ( 1990 ) 475498 . e. hsu and d. kutasov , `` the gravitational sine - gordon model , '' _ nucl . * b396 * ( 1993 ) 693707 , http://www.arxiv.org/abs/hep-th/9212023[hep-th/9212023 ] . j. e. carlisle , c. v. johnson , and j. s. pennington , `` baecklund transformations , d - branes , and fluxes in minimal type 0 strings , '' _ j. phys . _ * a40 * ( 2007 ) 1245112462 , http://www.arxiv.org/abs/hep-th/0501006[hep-th/0501006 ] . j. e. carlisle , c. v. johnson , and j. s. pennington , `` d - branes and fluxes in supersymmetric quantum mechanics , '' http://www.arxiv.org/abs/hep-th/0511002[hep-th/0511002 ] . i. m. gelfand and l. a. dikii , `` asymptotic behavior of the resolvent of sturm - liouville equations and the algebra of the korteweg - de vries equations , '' _ russ . math . surveys _ * 30 * ( 1975 ) 77113 . m. r. douglas , `` strings in less than one dimension and the generalized kdv hierarchies , '' _ phys . _ * b238 * ( 1990 ) 176 . t. banks , m. r. douglas , n. seiberg , and s. h. shenker , `` microscopic and macroscopic loops in non - perturbative two- dimensional gravity , '' _ phys . lett . _ * b238 * ( 1990 ) 279 . a. b. zamolodchikov and a. b. zamolodchikov , `` liouville field theory on a pseudosphere , '' http://www.arxiv.org/abs/hep-th/0101152[hep-th/0101152 ] . v. fateev , a. b. zamolodchikov , and a. b. zamolodchikov , `` boundary liouville field theory . i : boundary state and boundary two - point function , '' http://www.arxiv.org/abs/hep-th/0001012[hep-th/0001012 ] . j. teschner , `` remarks on liouville theory with boundary , '' http://www.arxiv.org/abs/hep-th/0009138[hep-th/0009138 ] . v. periwal and d. shevitz , `` unitary matrix models as exactly solvable string theories , '' _ phys . rev * 64 * ( 1990 ) 1326 . v. periwal and d. shevitz , `` exactly solvable unitary matrix models : potentials and correlations , '' _ nucl . _ * b344 * ( 1990 ) 731746 . h. airault , `` rational solutions of painlevequations , '' _ stud . _ * 61 * ( 1979 ) 3153 . v. e. zakharov and a. b. shabat , `` integration of nonlinear equations of mathematical physics by the method of inverse scattering . ii , '' _ funct . * 13 * ( 1979 ) 166174 . t. r. morris , `` multicritical matter from complex matrices , '' _ class grav . _ * 9 * ( 1992 ) 18731881 . b. kuperschmidt , `` mathematics of dispersive water waves , '' _ comm . * 99 * ( 1985 ) 5173 . c. v. johnson , `` string theory without branes , '' http://www.arxiv.org/abs/hep-th/0610223[hep-th/0610223 ] . d. kutasov and n. seiberg , `` noncritical superstrings , '' _ phys . * b251 * ( 1990 ) 6772 . h. ginsparg and g. w. moore , `` lectures on 2-d gravity and 2-d string theory , '' http://www.arxiv.org/abs/hep-th/9304011[hep-th/9304011 ] . p. di francesco , p. h. ginsparg , and j. zinn - justin , `` 2-d gravity and random matrices , '' _ phys . * 254 * ( 1995 ) 1133 , http://www.arxiv.org/abs/hep-th/9306153[hep-th/9306153 ] . p. di francesco , h. saleur , and j. b. zuber , `` generalized coulomb gas formalism for two - dimensional critical models based on su(2 ) coset construction , '' _ nucl . phys . _ * b300 * ( 1988 ) 393 . a. cappelli , `` modular invariant partition functions of superconformal theories , '' _ phys . _ * b185 * ( 1987 ) 8288 . m. bershadsky and i. r. klebanov , `` partition functions and physical states in two - dimensional quantum gravity and supergravity , '' _ nucl . * b360 * ( 1991 ) 559585 .

quantum walks have been a particularly fruitful field of research in quantum information going back to ideas from feynman @xcite , meyer @xcite , and aharonov @xcite . one focus of recent work on quantum walks is their application to spatial search algorithms . it is well - known that grover s algorithm @xcite for searching on unstructured databases offers a quadratic improvement in search time over classical models . grover s algorithm is not designed to search physical systems where only local operations are possible and quantum walk algorithms have been employed for this purpose . in developing such algorithms , the major consideration is the search time attempting to match the quadratic improvement over the classical case offered by grover s algorithm . the first spatial search algorithm to do this , using the standard model of quantum walks , was developed by shenvi , kempe and whaley @xcite for searches on a hypercube . while there exist discrete - time searches on @xmath0-dimensional cubic lattices which are faster than classical searches for @xmath1 @xcite , effective continuous - time quantum searches only exist for @xmath2 @xcite or else they require additional memory ( in the form of spin degrees of freedom ) in order to improve their search time in lower dimensions @xcite . in @xcite , we demonstrated that effective searches over two - dimensional lattices may be achieved in an arguably simpler way which does not require extra degrees of freedom , and could , therefore , be viewed as more efficient . this is achieved through the choice of a different lattice , specifically , a honeycomb lattice which is the underlying lattice structure of carbon atoms in the material graphene . the association with graphene is important as , although we first study a purely theoretical problem in quantum information , the use of a graphene lattice also offers a potential physical realisation . we thus envisage using the quantum walk and quantum search algorithm framework to investigate the effect of perturbations on the dynamics on graphene and other carbon structures . this offers not only the possibility for demonstrating two - dimensional continuous - time quantum searching , but also paves the way for looking for novel effects in the material graphene . in this paper , we offer a detailed account of the theory and numerical results thus expanding on the findings as presented in @xcite . generalisations of the results in @xcite have been given in @xcite , where a approach to solving the problem has been taken through encoding extra degrees of freedom into crystal lattices . recent experiments on artificial microwave graphene @xcite have shown that additional site perturbations , as discussed in sec . [ sec : alternativemarking ] , can be used to create a search protocol . searching on honeycomb lattices in discrete - time setting has been considered in @xcite ; however , no improvement over discrete time searches on cubic lattices was found . the paper is structured as follows : in section [ sec : quantumwalksandsearching ] we give an introduction to the formalism of continuous - time quantum walks and the construction of quantum walk search algorithms . we also explain why previous search algorithms struggled on lower - dimensional lattices and why graphene offers a solution . section [ sec : graphene ] will detail the relevant properties of graphene and set - up the notation we shall use throughout . sections [ sec : threebondperturbation ] & [ sec : threebondperturbation_analysis ] contain our main analytical results , where we detail the specifics of our search algorithm and offer an analysis of the search running time and success probability . we shall then show in section [ sec : communication ] how this search protocol can be adapted to demonstrate novel communication setups . sections [ sec : alternativemarking ] & [ sec : alternativenanostructures ] contain numerical work demonstrating the possibility of using alternative methods of marking to create search behaviour and other carbon nanostructures . we conclude with a review and discussion of our results in section [ sec : discussion ] . continuous - time quantum walks ( ctqw ) , first defined in @xcite , are the quantum analogue of continuous - time markov chains . they are defined purely on the state space , that is , the hilbert space @xmath3 , spanned by the states @xmath4 which represent the @xmath5 site of the lattice . thus , the time - evolution of such systems is defined by the schdinger equation @xmath6 where @xmath7 is the probability amplitude at the @xmath8 vertex of a system described by the state vector @xmath9 , and @xmath10 is the hamiltonian describing the connectivity of the lattice . note that we are using a dimensionless description setting e.g. @xmath11 . the dynamics of a quantum walk over a network is defined by the nature of the interaction between connected sites . therefore , the hamiltonian is generally constructed from the adjacency matrix of the underlying lattice . the adjacency matrix @xmath12 is defined as @xmath13 typically , the hamiltonian is chosen as @xmath14 . the parameter @xmath15 determines the coupling strength between connected sites and the parameter @xmath16 is an on - site energy that only enters the dynamics in a trivial way and thus can be set to a desired value . if @xmath17 and @xmath16 is equal to the valency of the lattice @xmath18 is the discrete laplacian . this form of hamiltonian is closely related to the tight - binding model for condensed matter systems @xcite . as first explained in @xcite , a quantum walk is transformed into a search protocol by introducing a localised perturber state , forming an avoided crossing in the spectrum of the search hamiltonian between an unperturbed eigenstate and the localised perturber . thus , initialising the system in the unperturbed eigenstate involved in the crossing and allowing the system to evolve in time , one finds that the system rotates into the localised perturber state . the first ctqw search over @xmath0-dimensional cubic lattices @xcite introduced the localised perturber state as a projector onto a single site @xmath19 , resulting in the search hamiltonian @xmath20 here , @xmath21 is a parameter governing the strength of interactions between sites in the lattice . this parameter is chosen carefully @xcite such that the perturber state @xmath22 is brought into resonance with the ground state of the unperturbed hamiltonian , the uniform superposition @xmath23 . let us assume for the moment that all other unperturbed eigenstates are energetically sufficiently separated from the ground state we will come back to this assumption later . then there will be an avoided crossing of two eigenvalues corresponding to the perturber state and the ground state . perturbation theory estimates that the energy splitting of these resonant states is proportional to the overlap of the localised perturber state @xmath22 and the uniform ground state @xmath23 , which scales as @xmath24 and that the corresponding eigenstates are of the form @xmath25 . by preparing the system initially in @xmath23 and allowing it to evolve for a set period of time @xmath26 , a measurement of the system will result in the state @xmath22 being measured with high probability . this explains the speed - up of the quantum walk search . the grover algorithm and its speed - up can be understood in analogous terms and one can show that all other states are indeed energetically very well separated . in the present case this separation only holds above a critical dimension @xmath27 . a simple argument for the critical dimension can be obtained by comparing the scaling behaviour of the energy splitting @xmath28 with the energy separation of the first excited state from the ground state in the unperturbed lattice . for a cubic lattice we have a quadratic dispersion relation which allows us to estimate the energy separation @xmath29 for the first excited state in a @xmath0-dimensional lattice ) . for @xmath30 one then has @xmath31 as @xmath32 and a detailed analysis indeed proves that the quantum walk search works with optimal speed - up @xcite . at the critical dimension @xmath33 the two energy scales scale in the same way the detailed analysis shows that the search still works but the dynamics is more complicated due to the interference of excited states . while there is still a speed - up it is only almost optimal ; the optimal search time @xmath34 gets multiplied with a logarithmically increasing factor . for the experimentally relevant regimes of 2- and 3-dimensional cubic lattices , all states participate in the dynamics as any avoided crossing gets dissolved completely as the number of sites grows . as a result no speed - up over classical searches is found . the above estimate offers a simple way how to reduce the critical dimension . if one can construct a search around a uniformly distributed state at an energy @xmath35 where the dispersion relation is conic ( linear in @xmath36 ) , then in @xmath0 dimensions @xmath37 . in @xcite this was implemented using a modified dirac hamiltonian . however , this requires the addition of a spin degree of freedom , essentially a doubling of memory , with the added complication that it is not immediately clear how such a system would be physically realised . instead , our solution here is to change the lattice from a square ( cubic ) to a graphene lattice . this change of topology automatically implies the first step in our solution , as one of the important electronic features of graphene is the conic dispersion relation around the dirac energy . this arises naturally from the tight - binding descriptions of graphene . note that graphene has the critical dimension @xmath38 . indeed our construction as presented in @xcite has an almost optimal speed - up with logarithmic corrections that need to be evaluated in a detailed analysis that goes beyond the simple perturbative description given above . graphene is a single layer of carbon atoms arranged in a honeycomb lattice . the lattice is bipartite with two sublattices , labelled @xmath39 and @xmath40 , and a unit cell containing two carbon atoms . the spatial and reciprocal lattices are shown in figure [ fig : both_lattices ] . the primitive vectors describing the lattice are @xmath41 , such that the position of a unit cell in the lattice is given by @xmath42 . we use dimensionless units in space where the distance between nearest neighbor sites is @xmath43 . the reciprocal lattice shows two important points , the dirac points @xmath44 and @xmath45 at the two inequivalent corners of the brillouin zone . , translation vectors @xmath46 and unit cell ( dashed lines ) . right : reciprocal lattice with basis vectors @xmath47 , symmetry points @xmath48 , @xmath44 , @xmath45 , @xmath49 and first brillouin zone ( hexagon).,title="fig : " ] , translation vectors @xmath46 and unit cell ( dashed lines ) . right : reciprocal lattice with basis vectors @xmath47 , symmetry points @xmath48 , @xmath44 , @xmath45 , @xmath49 and first brillouin zone ( hexagon).,title="fig : " ] [ fig : reciprocal_lattice ] the energy spectrum of electrons in graphene was first derived by wallace @xcite when considering the band structure of graphite using a tight - binding hamiltonian @xmath50 where @xmath51 is the on - site energy ( which we will identify as the energy of the dirac points ) and @xmath15 the hopping strength ( both dimensionless in our setting ) . the tight - binding model for graphene and the derivation of the solution are well - known @xcite and give rise to the dispersion relation @xmath52 shown in figure [ fig : infinite_graphene_lattice ] for an infinite graphene lattice . as there are two atoms per unit cell the spectrum has two branches , the upper branch being the conduction band and the lower the valence band , which meet at the corners of the brillouin zone , the @xmath44-points . the energy at the @xmath44-points is @xmath16 which we name the dirac energy . it is around these points , that the behaviour of the spectrum is conical , that is , @xmath53 with a reduced density of states , a necessary feature for the creation of the search dynamics . as the lattice possesses a translational symmetry the hamiltonian can be solved using linear superpositions of bloch functions over both sublattices . as a basis we use the orthonormal states @xmath54 to denote states on either the @xmath39 or @xmath40 sublattice in the cell at position @xmath55 . for the majority of what follows ( except in section [ sec : alternativenanostructures ] ) , we will focus on finite - sized lattices with assumed periodic boundary conditions along the axes of both primitive vectors so that the topology of our lattice is a torus ; that is , our wavefunction is of the general form @xmath56 . our boundary conditions imply that the state vector must satisfy @xmath57 where @xmath58 denote the period of the lattice . thus , the wavefunctions on the torus take the bloch function form @xmath59 \ , , \label{eq : wavefn}\end{aligned}\ ] ] where @xmath60 is the number of sites in the lattice , @xmath61 is the momentum , and @xmath62 is a relative phase contribution dependent upon whether the state belongs to the conduction or valence bands ; it can be calculated by explicitly working through the tight - binding model . application of the periodic boundary conditions results in the following quantised momenta @xmath63 where @xmath64 and @xmath65 . in the following and whenever we consider quantum walk dynamics on a torus , the number of cells in each direction is generally chosen to be the same , that is , @xmath66 . this choice is purely for simplifying the notation ; alternative torus dimensions are possible as are other choices of boundary conditions corresponding to alternative carbon structures ( e.g. nanotubes or a graphene sheet ) as will be demonstrated in section [ sec : alternativenanostructures ] . for our choice of torus dimensions , we find there are momenta equal to the @xmath44-points and , consequently , eigenstates with energies equal to the dirac energy when both @xmath67 and @xmath68 are some multiple of 3 . in fact , using the quantised momenta in eq . obtained for periodic boundary conditions , we find that there are four degenerate eigenstates with an energy that coincides exactly with the dirac energy when @xmath67 and @xmath68 are both multiples of 3 . these four states , known as the dirac states , can be constructed to live only on one of the sublattices , and are given by @xmath69 where @xmath70 for states on the @xmath39-sublattice or @xmath71 for states on the @xmath40-sublattice . in this section we will describe how a site is marked and will derive our optimal search starting state . our approach will be to analyse the system s spectrum and its dynamics in a reduced hamiltonian model involving only the relevant states from the avoided crossing . in the next section we will then validate this approach with a more detailed analysis , using the results obtained here as an initial guide . as already established , we introduce the localised perturber state in a region of the spectrum with a conic dispersion relation and thus a low density of states . simply altering the on - site energy in eq.([hgamma ] ) as done in @xcite does , however , not work here . as the on - site energy @xmath16 and the dirac energy are equal , one finds that the perturbation only interacts with the dirac states in the limit of zero perturbation strength , returning the unperturbed lattice . therefore , we choose an alternative perturbation method : namely , we modify the coupling strength between the site we wish to mark and its neighboring vertices . we focus in this section on changing the coupling to all three neighboring vertices of a particular site equally . our choice of perturbation matrix @xmath72 to mark the @xmath39-type vertex @xmath73 is then @xmath74 where the state @xmath75 is the symmetric superposition over the three neighbors of the perturbed site @xmath76 this leaves us with the search hamiltonian @xmath77 where @xmath21 is a free parameter . in what follows , we always set the on - site energy @xmath78 . considering our search hamiltonian , we can see that setting @xmath79 corresponds to a coupling strength of @xmath80 from the perturbed vertex and its nearest - neighbors ; our perturbation essentially removes the marked vertex @xmath73 from the lattice . note that vacancies are a common , naturally occurring , defects in graphene lattices @xcite . in order to establish the critical value of @xmath21 , we numerically calculate the spectrum of @xmath81 as a function of @xmath21 , plotted in figure [ fig:3bond_spectrum ] for a torus of dimensions @xmath82 . as @xmath72 is a rank-2 perturbation , we see in figure [ fig:3bond_spectrum ] two perturber states working their way through the spectrum to an avoided crossing around @xmath83 when @xmath84 , that is , when the perturbed vertex is removed from the underlying lattice . 12x12_3p_spectrum_colour ( 54.25,26.75 ) as well as establishing a critical value for @xmath21 , we also find from this figure the states involved in the avoided crossing : the four dirac states at the dirac energy and the two perturber states which form our perturbation matrix @xmath72 . however , with further consideration , we may reduce the number of states involved further . as we have removed the site @xmath85 from the lattice it can no longer interact with the rest of the lattice and the corresponding state drops out . also , by direct calculation one finds that the @xmath40-type dirac states do not interact with the perturbation , that is , @xmath86 . thus , they remain an eigenstate of the search hamiltonian and do not interact with the perturbation . we are left with three states taking part in the avoided crossing : @xmath87 . we reduce our search hamiltonian in this three state basis at the critical point @xmath84 to obtain the following reduced hamiltonian describing the local dynamics at the avoided crossing @xmath88 with @xmath89 and @xmath90 . this reduced hamiltonian has eigenvalues @xmath91 , and eigenvectors @xmath92 using the eigenvectors of the reduced hamiltonian , we can construct a search starting state which is a superposition of dirac states @xmath93 allowing our search starting state @xmath23 to evolve under the reduced hamiltonian we find @xmath94 so that our system rotates from our dirac superposition @xmath23 to a state which is localised on the neighbors of the perturbed vertex @xmath75 in a time @xmath95 . thus , we find a @xmath96 search time , a polynomial improvement over the classical search time . we plot in figure [ fig : search_prob ] a numerically calculated search for a lattice with @xmath97 sites . the system has been prepared in @xmath23 and allowed to evolve under the full search hamiltonian from eq . ( [ eq : full_search_hamiltonian ] ) at the critical value @xmath84 . one finds that the system localises on the neighbors of the marked vertex and , for this particular system size , the probability for a measurement of the system to return one of the neighbors is around @xmath98 at its peak . this is orders of magnitude higher than the average probability to measure other vertices in the lattice , @xmath99 , which for this system size is around @xmath100 . the localised state interacting with the spectrum under the full hamiltonian has a tail into the rest of the lattice ; this leads to a loss of probability to be found at the neighboring vertices below @xmath101 . it is worth emphasising that the marked vertex plays no role in the search - it is removed from the lattice . rather , we force the system to localise on the nearest - neighbors , essentially making the marked vertex conspicuous by its absence . this feature differentiates our algorithm from previous searches , which have generally concentrated on localising directly on the marked vertex . an approach considering the neighboring vertices was detailed in @xcite , however , the focus was here still on localising on a single site but with classical post - processing steps considering the tail of the localised state extending into the neighborhood around the marked vertex . at this point , there are a number of issues which need to be mentioned . the first is that our starting state in eq . ( [ eq : starting_state ] ) contains information about the marked vertex in the form of the relative phase between the dirac states . however , this phase can only take three different values . therefore , we have three possible optimal starting states for an @xmath39-type perturbation and the same number for @xmath40-type perturbations ; there are thus in total six possible optimal states . as not all of these states are orthogonal , we only find an increase of necessary runs for a successful search by a factor of 4 ; this additional overhead is independent of @xmath60 and , therefore , does not affect the overall time complexity of the search . the particular representation of the dirac states in eq . ( [ eq : k_state ] ) was chosen in such a way as to make the calculations and conceptualisation of the system dynamics easier . in reality , constructing starting states which exist on one sublattice only , experimentally is extremely difficulty ; rather , in an experiment , one is likely to excite a superposition of dirac states , reducing the success probability , on average , by a factor of @xmath102 . our search is based on the conic dispersion relation in the spectrum and the @xmath103 scaling between successive energy levels in the linear regime , giving rise to the @xmath96 search time found in our reduced model . however , previous searches in continuous - time using a modified dirac hamiltonian @xcite or in discrete - time @xcite have found logarithmic corrections to the search time . as we have seen in eq . ( [ eq : search_time ] ) the search time is related to the energy gap at the avoided crossing . in the inset of figure [ fig:3bond_spectrum ] we have numerically calculated the scaling of the gap in the spectrum of the full search hamiltonian and we indeed find evidence for a logarithmic correction . the lack of a logarithmic term in the search time for our reduced model is due to the neglecting contributions from the rest of the spectrum . in what follows we establish a more accurate estimate of the running time and success probability by working with the full hamiltonian . we also justify the findings of our reduced model : while this model is insufficient to estimate the finer details , such as logarithmic correction terms , the accurate calculations show that our search algorithm indeed takes place mainly in a two - dimensional subspace spanned by the dirac states and a state localised on the neighbors of the marked vertex . we follow here closely our derivation as presented in @xcite which builds on ideas given in @xcite for a similar a similar calculation for regular rectangular lattices . we focus first on the search success amplitude , that is , @xmath104 where @xmath105 is the search time , that is , the time our search probability reaches a maximum , and our search hamiltonian @xmath106 is given by eq . ( [ eq : full_search_hamiltonian ] ) with @xmath84 , that is,@xmath107 here , @xmath108 , @xmath109 are the eigenstates and eigenenergies of @xmath106 . we assume , without loss of generality , a specific starting state @xmath110 where the marked vertex is chosen such that @xmath111 leading to @xmath112 in the following , we suppress the sublattice superscript as the analysis is the same regardless on which sublattice the perturbation lives . we also denote @xmath113 the positive eigenenergies of @xmath114 from eq . ( [ eq : graphene_unperturbed_eigenenergies ] ) and set @xmath83 . for eigenstates @xmath115 with eigenenergies @xmath116 that are not in the unperturbed graphene spectrum ( @xmath117 for all points @xmath61 in the dual lattice ) , we may rewrite eq . ( [ eq : specific_search_hamiltonian ] ) in the form @xmath118 where @xmath119 and the phase of @xmath115 is chosen such that @xmath120 . at this point , we can remove several states from the summation in eq . ( [ eq : amplitude ] ) . note that the basis state associated with the marked vertex , @xmath121 , is itself an eigenstate of the search hamiltonian with @xmath122 , and also , @xmath123 so that @xmath121 does not contribute to the time - evolution . we may also remove eigenstates of @xmath124 which are at the same time eigenstates of @xmath114 with the same energy . this can be seen in the following way . we first consider an unperturbed eigenstate @xmath125 such that @xmath126 . let us assume that there is an eigenvector @xmath115 of the search hamiltonian with the same eigenenergy @xmath109 , that is , @xmath127 . considering the matrix element @xmath128 , we find @xmath129 . as @xmath121 is an eigenvector of the search hamiltonian we know that @xmath130 . this leaves us with @xmath131 . as the unperturbed eigenstate @xmath125 is simply a bloch state we know @xmath132 . thus , we obtain @xmath133 . it is then clear that eigenstates of the search hamiltonian whose eigenenergies remain in the spectrum of @xmath134 do not play a role in the time - evolution of the search . we now derive an eigenvalue condition for those perturbed eigenvalues which are in the spectrum of @xmath124 . using the orthogonality of eigenstates @xmath135 , eq . ( [ eq : eigenstate_representation ] ) leads to @xmath136 by expressing @xmath137 in terms of the eigenstates of the _ unperturbed _ walk hamiltonian @xmath138 , we may write this as a quantization condition @xmath139 with @xmath140\ , , \ ] ] were @xmath60 is the total number of sites in the lattice . ( [ fe ] ) is written as two summations to incorporate the fact that the spectrum of @xmath114 as well as @xmath124 is symmetric around @xmath141 . choosing @xmath115 to be normalised @xmath142 , eq . ( [ eq : eigenstate_representation ] ) also implies @xmath143 which allows @xmath144 to be rewritten as @xmath145 we may now rewrite the amplitude in eq . in the form @xmath146 where we have used the adjoint of eq . ( [ eq : eigenstate_representation ] ) . note again that eigenstates of @xmath124 , which are also in the spectrum of @xmath114 , do not contribute to the time - evolution of the search ; it may be seen from the definition of @xmath147 in eq . ( [ fe ] ) that @xmath148 , where @xmath109 is in the spectrum of @xmath114 . as we saw in the previous section , the avoided crossing is formed by two perturbers approaching symmetrically either side of the dirac states with energy @xmath83 . thus , we concentrate on evaluating the contribution to the time - evolution from these perturbed eigenstates of @xmath124 either side of the dirac point and we label these states @xmath149 . in order to evaluate these contributions we calculate the perturbed eigenenergy @xmath150 and also derive a leading - order expression for @xmath151 ( since the spectrum is symmetric , we have @xmath152 ) . using the definition of @xmath147 in eq . ( [ fe ] ) , we estimate @xmath153 by separating out the dirac points , where @xmath154 , from the summations and then taylor expanding the remaining terms at @xmath141 , to find @xmath155 where the sums @xmath156 are given by @xmath157^{n } } + \frac{1}{\left[-\epsilon\left(\underline{k}\right)\right]^{n } } \right ] \,.\ ] ] as the unperturbed spectrum is symmetric only those @xmath156 with even @xmath68 are non - zero ; thus from now on we focus only on @xmath158 where @xmath159 . we stated earlier that the spectrum of graphene is well - approximated by a conic dispersion relation around the dirac points . thus , it is from around these points that the major contributions to the @xmath158 sums arise . by applying the linear approximation from eq . ( [ eq : linear_dispersion ] ) and the momenta quantum numbers from eq . ( [ eq : quantised_momenta ] ) we may approximate the @xmath158 summations as @xmath160 \nonumber \\ & + & o(1 ) \ , . \label{eq : i_epstein}\end{aligned}\ ] ] here @xmath161 is the epstein zeta - function @xcite @xmath162 for a real positive definite real symmetric @xmath163 matrix @xmath164 . for our purposes we use @xmath165 which describes the spectrum close to the dirac points . the linear approximation around both dirac points @xmath166 and @xmath167 is the same and , thus , the matrices @xmath168 and @xmath169 are equal . our summation in eq . ( [ eq : i_epstein ] ) is over the rectangular regions @xmath170 and @xmath171 of the lattice @xmath172 . both are centered on @xmath173 and have side lengths proportional to @xmath174 , however the center , @xmath173 , corresponding to the relevant dirac point , is omitted from the summation . convergence of the epstein zeta function is well - known for @xmath175 @xcite , and leads to the bounds @xmath176 for @xmath177 . a sharp estimate for @xmath178 is given in appendix [ appendix_a ] . in order to calculate an estimate for @xmath150 , we truncate the taylor expansion of @xmath147 in eq . ( [ eq : exp ] ) at the @xmath179 term ( the first term in the summation ) , and apply the eigenvalue condition @xmath180 . that is , we solve @xmath181 and obtain the approximation @xmath182 . this solution also leads to the estimate @xmath183 . we consider next whether our solution for @xmath150 lies inside the radius of convergence of the taylor expansion . we note that each term in @xmath184 is smaller than the corresponding term in @xmath185 for @xmath186 , and so it follows that @xmath187 for @xmath186 . this property of the epstein zeta function and the estimate from eq . ( [ eq : ikest ] ) imply @xmath188 i.e. the infinite sum in eq . ( [ eq : exp ] ) converges for @xmath189 . thus , for large @xmath60 , our solution lies within the radius of convergence of our taylor expansion in eq . ( [ eq : exp ] ) . in order to show that the dominant contributions to the search come from @xmath149 , we need to establish the leading order error term . all the @xmath158 sums are positive so that , given the sign of all the terms in the taylor expansion in eq . ( [ eq : exp ] ) , the true value of @xmath190 has to be smaller than the estimate we have obtained . thus , we write the true value of @xmath150 as @xmath191 with @xmath192 . one may rewrite @xmath193 , using the true value of @xmath150 , to give @xmath194 we follow the same arguments as used for the calculation of the radius of convergence to obtain an upper bound for the summation . this together with the already established fact that @xmath195 is inside this radius for sufficiently large @xmath60 , we get the following inequality : @xmath196 so @xmath197 . thus , we obtain @xmath198 it also follows that @xmath199 this shows that our perturbed eigenstates @xmath149 have an @xmath200-overlap with the starting state and are , therefore , the relevant states to be considered in the time - evolution of the algorithm . using the definitions of @xmath115 and @xmath144 in eqs . ( [ eq : eigenstate_representation ] ) , ( [ eq : ra ] ) , the inner product of the starting state and the perturbed eigenvectors can be expressed as @xmath201 where @xmath202 is the overlap of the starting state with the marked vertex state . applying our previous approximations for @xmath150 and @xmath151 , that is , for our perturbed eigenstates closest to the dirac point , we find @xmath203 thus , our starting state is indeed a superposition of the perturbed eigenstates , @xmath149 as assumed in the previous section . this makes it possible to investigate the running time and success amplitude of the algorithm in more detail , that is , @xmath204 it is clear from our earlier results for @xmath205 and @xmath179 that our algorithm localises on the neighbor state @xmath75 in time @xmath206 with probability amplitude @xmath207 . this confirms the logarithmic correction observed numerically and displayed in the inset of fig . [ fig:3bond_spectrum ] . it has been demonstrated in @xcite that discrete - time search algorithms can be modified to create a communication protocol . we show here that a communication setup can be established also in the continuous - time search algorithm with minor changes due to the subtleties of our search . we use the same unperturbed walk hamiltonian as before , that is , @xmath208 and the same type of perturbation matrix as in eq . ( [ eq:3bond_perturbation_matrix ] ) , but now at two different sites in the lattice . explicitly , our communication hamiltonian is @xmath209 where the perturbation matrices @xmath210 mark the source and target sites , respectively . as prescribed in @xcite , the communication protocol operates by preparing a state localised on the source perturbation , in our case on the neighbors of the source site , and allowing the system to evolve under our communication hamiltonian . the system then localises on the neighboring vertices of the target site . the effect of the two perturbation matrices @xmath210 amounts to disconnecting the two vertices from the underlying lattice . as noted before , our search algorithm has several different optimal starting states depending on the position of the marked vertices . therefore , we divide our communication analysis into three different cases : the two sites i ) are on the same sublattice and have the same optimal search starting state ; ii ) are on the same sublattice but _ do not _ share the same optimal search starting state ; iii ) are on different sublattices . both of the first two cases can be treated by applying the reduced hamiltonian method demonstrated earlier , but communication between different sublattices is not tractable by this method and so we focus on numerics in this case . for signal transfer between two sites on the same sublattice , we will assume , without loss of generality , that our communication takes place between two sites on the @xmath39-sublattice , @xmath211 ; here , the subscript @xmath212 or @xmath213 denotes the source or target vertices respectively . using arguments as employed in sec . [ sec : threebondperturbation ] , we can reduce the number of relevant states . ultimately , the search dynamics takes place in the subspace spanned by the basis @xmath214 , the @xmath39-type dirac states , and @xmath215 , the uniform superpositions over the neighbors of the source and target vertices . this basis leads to the reduced hamiltonian @xmath216 where @xmath217 and @xmath218 . in the first case considered we assume that there are two perturbations on the graphene lattice located at the points @xmath219 and @xmath220 , chosen such that @xmath221 . this implies that a search for either vertex , using our search algorithm from secs . [ sec : threebondperturbation ] & [ sec : threebondperturbation_analysis ] , would use the same optimal starting state . as the phases are equal , we drop the subscript on the perturbation coordinates in what follows . fixing our phases and diagonalising the reduced hamiltonian , we find it has eigenvalues @xmath222 and @xmath223 with eigenvectors @xmath224 using these eigenstates , we may rewrite the source neighbor state as @xmath225 placing the system in the source state @xmath226 and allowing the system to evolve under the reduced hamiltonian , one finds @xmath227 we note that , in the last line above , the term in the brackets involving only the dirac states is actually the optimal search starting state for both perturbed vertices , defined in eq . ( [ eq : starting_state ] ) . the system thus oscillates between the states localised on the neighbors of the perturbed vertices , @xmath226 and @xmath228 , in a time @xmath229 , via their optimal search starting state . figure [ fig : equivalent_communication_torus ] shows the system evolving under the full communication hamiltonian , eq . ( [ eq : communication_hamiltonian ] ) . the initial state used for the time evolution shown in figure [ fig : equivalent_communication_torus ] is the true localised state , that is , we run the quantum search with a single perturbation located at vertex @xmath219 until it reaches maximum success probability , and then apply the second perturbation to the vertex @xmath220 . the figure confirms the behaviour expected from the reduced model calculation . the communication mechanism essentially works in the same way as the quantum search algorithm , as it can be viewed as one marked vertex ` finding ' another . the initial localised source state decays back towards the search starting state , and the system then searches for the target state . in our second case of signal transfer we analyse the behaviour of a communication system where we have two perturbed vertices on the same sublattice , but with the restriction that @xmath230 . as such , the two marked sites can not interact via the same search starting state . however , the optimal search starting states are not orthogonal so that signal transfer is still possible , but the resulting interference effects make the analysis slightly more complicated . we rewrite the coordinates of the target in terms of the source , @xmath231 and @xmath232 . again , fixing our phases and diagonalising the reduced hamiltonian in eq . ( [ eq : reduced_communication_hamiltonian ] ) , we find it has eigenvalues @xmath233 and @xmath234 with eigenvectors @xmath235 using these eigenstates , one may write the source perturbation as @xmath236 the full expression for the time - evolution is rather cumbersome ; therefore , we only show the terms and prefactors we are interested in , namely @xmath237 and @xmath238 @xmath239{\left| \ell_{s } \right\rangle } \label{eq : simulated_nonequivalent_communication } \\ & - \frac{1}{2}\left[\cos\left(\lambda^{+}_{\sqrt{3}}t\right)-\cos\left(\lambda^{+}_{1}t\right)\right]{\left| \ell_{t } \right\rangle } \nonumber \ , .\end{aligned}\ ] ] we can see here that the prefactors do not depend upon the coordinates of either the source or the target sites ; the transport signal between sites on the same sublattice but with different optimal search starting state is thus independent of the position of the source or target site . in figure [ fig : both_nonequivalent_communication ] we show the system evolved under the full communication hamiltonian . again , the initial state is the true localised state on the nearest - neighbors of the source vertex , obtained by running the search algorithm using one marked vertex until it reaches its peak success probability . the time - evolution is radically different to the previous communication case ; the behaviour here is erratic with uneven peaks of probability at the two perturbations involved in the protocol . however , there are still significant probability revivals . the transport behaviour from eq . ( [ eq : simulated_nonequivalent_communication ] ) , calculated using the reduced hamiltonian , is also shown in figure [ fig : both_nonequivalent_communication ] . the probability at time @xmath240 has been scaled to match that shown in figure [ fig : both_nonequivalent_communication ] . our calculated behaviour has the same signal pattern as the numerically calculated behaviour from the full hamiltonian , although over a shorter period of time . as our reduced model only makes use of the dirac states and the perturber states , we lose the contribution to the time - evolution from the rest of the spectrum giving rise to logarithmic corrections as discussed in sec . [ sec : threebondperturbation_analysis ] . this leads to differences in the overall time scales for the reduced model and the full hamiltonian . this also supports our findings that signal transfer between all non - equivalent vertices on the same sublattice is the same . the communication protocol is again set up by a search mechanism in reverse , where one vertex finds another . the slightly erratic behaviour that emerges here is due to interference between the two separate search mechanisms which interact due to the non - zero inner product of the three possible optimal search starting states for vertices on the @xmath39-sublattice . in our final case of signal transfer , we consider communication between sites on different sublattices which we can not treat in the reduced hamiltonian model . it has been demonstrated previously that perturbations to one sublattice do not interact with the dirac states which live on the other sublattice . therefore , when attempting to reduce the communication hamiltonian as done before , we merely find that it decouples into two non - interacting reduced hamiltonians , each describing a search protocol on one sublattice . the interaction between the two sublattices is here facilitated due to interactions via the bulk of the spectrum . in what follows , we focus on numerics and inspect the behaviour for systems involving the same source perturber but different targets . the numerics show a finite number of different signal patterns , two examples are shown in figure [ fig : both_differentsublattice_communication ] . we can see from these examples that the communication mechanism takes place over a much longer timescale with a superimposed oscillatory dynamics . the increase in timescale can be attributed to the weak nature of the interaction between the two perturbations due the fact that the localised states live mainly on one sublattice . the signal pattern in figure [ fig : both_differentsublattice_communication ] thus decouples into a fast oscillation between one site , say the source site , and the delocalised lattice state and a slow time scale on which a small amount of probability amplitude escapes into the other sublattice due to the weak interaction . this process continues until the recurrence probability at the target perturbation reaches the same peak as the initial localised source state , and then the behaviour reverses . in the previous sections , we have discussed a method of marking a vertex through modifying the hopping potential from the perturbed site to all three of its nearest - neighbors . here , we will discuss alternative methods of marking a vertex which still keep the perturber interaction in the conic dispersion region of the spectrum . this can be done by altering the hopping potential in different ways as discussed above ; this approach is necessarily a rank-2 perturbation and these two perturber states must meet at the dirac energy . we will discuss several ways of applying such a perturbation in the next paragraph . another approach is based on coupling extra sites to the lattice ; this set - up will be discussed at the end of this section . focussing on hopping potential perturbations , we may consider perturbing the bonds to any number of nearest - neighbors to any strength . we have seen previously that a symmetric three - bond perturbation successfully creates a search . here , we demonstrate a search based on perturbing the hopping potential from a given site to only one of its nearest - neighbors . that is , we use the same search hamiltonian as in eq . ( [ eq : full_search_hamiltonian ] ) with the perturbation matrix @xmath241 and eigenstates @xmath242 note that it no longer makes sense to speak of a single marked vertex . rather , our perturbation marks both vertices @xmath243 simultaneously ; thus , it may be more accurate to view our perturbation matrix as marking a single cell of the lattice . for this type of perturbation , the avoided crossing used to generate search behaviour is not necessarily at @xmath84 , see figure [ fig : singlebond_spectrum ] , where the spectrum of our search hamiltonian as a function of @xmath21 for a @xmath244 cell torus is shown . as stated before , our single bond perturbation is also a rank-2 matrix , and so again there are two perturber states approaching the spectrum from the negative and positive regions of the spectrum . inspecting the region around the dirac energy , see inset in figure [ fig : singlebond_spectrum ] , we see that avoided crossings are formed by four states in total , the two blue ( solid ) curves and the two red ( dashed ) curves . at @xmath245 there is an exact crossing between the red and blue curves indicating that these states are orthogonal . singlebond_spectrum ( 54.26,26.75 ) we can make this avoided crossing picture clearer by breaking down the eigenstates of the search hamiltonian in terms of the symmetries of the lattice . in particular , we use the rotation operator @xmath246 , which is a rotation of @xmath247 about the mid - point of the perturbed bond . considering the action of @xmath248 on the marked vertices and the eigenstates of the perturbation matrix , one finds @xmath249 , @xmath250 and @xmath251 . the inset of figure [ fig : singlebond_spectrum ] shows the states nearest to the dirac energy , where the same colour indicates the same @xmath248 parity . it now becomes clear that the states near the dirac point actually form two avoided crossings , one for each parity with a minimum energy gap at @xmath245 . without going into details , we can again find optimal starting states using a reduced hamiltonian approach ( one for each of the avoided crossings ) ; the evolution of the system using these optimal states is shown in figure [ fig : singlebond_search ] for a @xmath244 cell lattice . we find there is a significant localisation on the marked vertices and their nearest - neighbors , with the probability of being found on either of the marked vertices peaking at around @xmath252 . we can also see that the probability of being found on each of the nearest - neighbors peaks at around @xmath253 , resulting in a total probability of being found on the marked vertices and their nearest - neighbors of approximately @xmath254 . the success probability fluctuates , as the probability amplitude oscillates between the marked vertices and their nearest - neighbors . this is due to the probability amplitude being constrained in the local area by the increased hopping potential between the two marked vertices . although our demonstration is for a @xmath244 cell lattice , numerical investigations show that the search behaviour remains the same as the lattice size increases with critical value fixed at @xmath245 . we also find that the gap at the two avoided crossings scales as @xmath255 , in the same way as for the three - bond perturbation shown in figure [ fig:3bond_spectrum ] . as the search time is inversely proportional to the energy splitting at the avoided crossing , it also gives an estimate of the running time of the search @xmath256 . we now turn to coupling additional sites to the lattice as a way of introducing a perturbation as also considered in @xcite ; such a treatment is in many ways closer to experimental realisations as defects due to additional add - on atoms are quite common in graphene @xcite . we focus here on the idealised case where an additional site is coupled to a single lattice vertex . we use the perturbation matrix @xmath257 for coupling the additional site @xmath258 to the @xmath39-type vertex @xmath73 and @xmath21 is a free parameter related to the on - site energy of the additional vertex . the choice of the coupling terms is such that the binding energy between the additional site and the lattice vertex is the same as the internal couplings in the lattice . thus , our search hamiltonian is of the form @xmath259 in figure [ fig : singlesite_spectrum ] , the spectrum of the search hamiltonian is given as function of @xmath21 . as our free parameter @xmath21 only changes a single term , our spectrum only has a single perturber state . one finds a clear avoided crossing around the dirac energy ( @xmath141 ) when @xmath260 , that is , when the on - site energy of the additional site matches the on - site energy of the lattice vertices . singlesite_spectrum ( 11.75,28.25 ) in the reduced hamiltonian picture , we write eq . [ eq : additional_site ] in terms of a basis consisting of the dirac states and our perturber state , @xmath258 . only three states are involved in the search , that is , @xmath261 , and @xmath10 reduces , up to a different prefactor , to the same @xmath262 matrix found in eq . ( [ redh ] ) for the three - bound perturbation . thus , our previous reduced hamiltonian analysis holds for this case and the optimal search starting states are the same ; the time - evolution for the additional site search is shown in the inset in figure [ fig : singlesite_spectrum ] . the only differences between this case and the three - bond marking are that the system localises on the additional site , and also the change in prefactor leads to a search time of @xmath263 . we note that other , more realistic types of perturbations involving additional sites coupling to a lattice site and its neighbors , can be shown to result in effective search protocols . also , the single additional site perturbation described here can , like the three - bond perturbation , be used to setup a communication protocol , with the same signal patterns and behaviours as found previously . the initial state is in this case completely localised on the additional site . so far we have described the dynamics of searches on graphene lattices with periodic boundary conditions , that is , graphene on a torus . here we consider more realistic boundary conditions such as nanotubes and graphene sheets . for the dynamics on nanotubes , we move to periodic boundary condition along one axis only and impose dirichlet boundary conditions along the other directions . the properties of nanotubes are well - known @xcite , and it has been shown that the band structure of armchair nanotubes , that is , nanotubes with armchair boundaries , always allows for an energy at the dirac energy regardless of the nanotube diameter ; we will focus on these types of nanotubes . we are interested in searching on finite length nanotubes , where the band structure becomes discrete ; the length of the nanotube are in addition chosen such that there exists an eigenenergy at the dirac energy . an example of the cell we use to construct finite armchair nanotubes is shown in figure [ fig : armchairnanotube_example ] . we choose the finite length of the nanotube to be along the horizontal axis and we close the underlying graphene lattice into a nanotube along the vertical axis . we denote the basis states of sites in our nanotube as @xmath264 , where @xmath67 indicates the @xmath265 @xmath266-type vertex in the horizontal direction in the @xmath267 cell . it is simple to see that , due to the periodic boundary conditions along the circumference of the tube , the vertical component of the eigenstates must be bloch states and the horizontal component of the amplitudes must be sinusoidal in nature . that is , the eigenstates of the finite nanotube are standing waves along its length . by working through the tight - binding model for this system , one can show that @xmath268 is the only point where there exists an eigenenergy equal to the dirac energy . it is also possible to show that the spectrum includes dirac points when the number of sites along the length of the nanotube @xmath269 , where @xmath270 is an integer . as there is only one potential dirac point for finite armchair nanotubes , it follows that there are only two dirac states ( one from the bonding and the anti - bonding regions of the spectrum ) . as we have @xmath271 at the dirac point , the bloch wave around the circumference of the nanotube is simply a uniform superposition . another important feature of the dirac states on the nanotube , which we have not encountered previously , is the existence of nodal points where the amplitude of the eigenstate is 0 . we find these nodal points occur at every third site along the horizontal axis . we focus on applying our three - bond perturbation from sections [ sec : threebondperturbation ] & [ sec : threebondperturbation_analysis ] to the armchair nanotube . in our new labelling our perturbation takes the form @xmath272 we have assumed that we are perturbing an @xmath39-type vertex on an even horizontal coordinate , so that the perturbed site and its nearest - neighbors remain within one nanotube cell . while it is easy to re - write this perturbation matrix for other sites we restrict ourselves to this form for simplicity . note even in this form it can still be expressed in terms of a marked site , @xmath273 , and a state which lives on the neighbors , @xmath274 . our search hamiltonian for the nanotube is the same as in eq . ( [ eq : full_search_hamiltonian ] ) . inspecting the spectrum of the search hamiltonian as a function of the free parameter @xmath21 where the perturbation is not located on a nodal point of a dirac state , one finds an avoided crossing around the dirac energy when @xmath84 . this spectrum is very similar to the one shown in figure [ fig:3bond_spectrum ] for the search on the torus and will be omitted . if the perturbation is located on a nodal site the picture is different , however ; one finds that there is an exact crossing at the dirac energy when @xmath84 and searching is not possible . recall that the effect of the perturbation at the critical point is to completely remove the site from the lattice , if the site already has zero amplitude then the perturbation will not interact with the dirac state . similar to the previous section , we numerically reduce the full search hamiltonian in a basis consisting of the two dirac states and the state living purely on the neighbors , @xmath75 . the perturbed site , @xmath273 , is decoupled from the lattice . through this process , we find two possible starting states , one for each sublattice . the starting states are weighted superpositions of the dirac states over the sublattice containing the marked vertex . the nodal points are excluded from this treatment . figure [ fig : armchair_search ] shows the system evolution for two searches using the numerically found optimal starting states . one search has the marked site located in the centre of the nanotube , the other has the marked site positioned near the edge of the nanotube . the dimensions of the nanotube have been chosen so that there are 320 sites , comparable to the searches shown in earlier sections . comparing these searches to the three - bond perturbation search on the torus in figure [ fig : search_prob ] , we see that there is a marked difference in behaviour and success probability induced by relaxing the periodic boundary conditions along one axis . using a perturbation located near the edge , the lower image in figure [ fig : armchair_search ] , we see a reduction in success probability by a factor of 2 - 3 and strong fluctuations in peak height when compared to the search on the torus . searching in the bulk , shown in the lower image of figure [ fig : armchair_search ] , displays behaviour closer to searches on a graphene torus , but again with significant fluctuations at each peak . we propose that this effect is due to the reflection of probability amplitude from the edges of the nanotube . this is supported by the changes in the interference pattern in the signal as the perturbation is moved across the lattice . numerics demonstrate that additional site perturbations and communication protocols can also be used on nanotube lattices . we now move to working on graphene sheets , that is , removing the periodic boundary conditions along both axis . a detailed account of a specific version of this set - up together with experimental results has been presented in @xcite for graphene sheets consisting of armchair boundaries only . note that such a configuration can not be achieved on rectangular sheets ; the simplest configuration has the form of a parallelogram , see @xcite for further details . the advantage of such a configuration is that the form of the boundary does not admit so - called ` edge - states ' . in the following , we will look at the influence of these edge - states in more detail by considering rectangular graphene sheets , such as in figure [ fig : finite_lattices_example ] . in addition to armchair edges , here along the vertical axis , these sheets have boundaries formed by bearded edges ( left ) and zigzag edges ( right ) along the horizontal axis . these boundaries support edge states , i.e. states localised along these edges , with an eigenenergy close to the dirac energy @xcite . in the following , we will investigate , how the existence of these edge states influences the search . as in the case of the finite armchair nanotube , the imposition of dirichlet boundary conditions at an edge generates sinusoidal eigenstates . as a result , there are no extended ( bulk ) eigenstates at the dirac energy due to the inability to equate the quantised momenta with the necessary points in @xmath61-space . we thus need to find other extended eigenstates in the sea of edge states near the dirac point to undertake a search in this set - up . we mark sites using the triple - bond perturbation such as in eq . ( [ eq : full_search_hamiltonian ] ) . throughout this section we choose the dimensions in terms of primitive cells of the graphene sheets , @xmath275 ; a bearded edge sheet thus consists of @xmath276 sites and a zigzag sheet is formed of @xmath277 sites . the spectra of the search hamiltonians for the bearded lattice is shown in figure [ fig : spectrum_bearded_3p ] constrained to the energy region of interest . the spectrum for zigzag edges is very similar and is not show here . one can clearly see an avoided crossing around the dirac energy at @xmath84 , the critical value for this type of perturbation . there are several states very close to the dirac energy ; these are the edge states which are all non - degenerate . therefore , it is not possible for us to construct a superposition of degenerate eigenstates which is optimal for searching . rather , our initial starting state must be a single eigenstates of the unperturbed hamiltonian @xmath278 . using the search hamiltonian from eq . ( [ eq : full_search_hamiltonian ] ) and fixing @xmath84 , we proceed by allowing the system to evolve after being prepared in one of the unperturbed eigenstates . one finds for both types of lattice , that the edge states near the dirac energy fail to produce localisation behaviour ; the probability at the neighboring vertices of the marked site do not rise above noise levels . rather , search behaviour only begins to emerge when we use the first delocalised , non - edge state as our initial state . we also note that we _ can not _ search for sites near the zigzag or bearded edges , where the edge states exist . only as we move further into the bulk of the lattice or along an armchair - type edge , the localisation behaviour returns . note that the edge states can be viewed as a kind - of one - dimensional system and , as we saw from the scaling argument towards the end of sec . [ sec : quantumwalksandsearching ] , one - dimensional systems imply an energy spacing between successive energy levels of @xmath279 . thus , the perturber state interacts with many states in a dense part of the spectrum and the search fails . in figure [ fig : bearded_searches ] we show the search behaviour arising from marked sites placed at different positions on a bearded graphene lattice ( zigzag lattice types display similar behaviour ) . the results are similar to those found for the nanotube searches . there is a slight increase in variation of signal pattern with position , and an increase in success probability maxima as we move towards the centre of the lattice from either of the bearded edges . we have shown that continuous - time quantum search can be done effectively on a two - dimensional graphene lattice without the use of internal degrees of freedom . this is achieved by making use of the conical ( linear ) dispersion relation in the graphene spectrum . the search succeeds in time @xmath280 with probability @xmath281 . this is the same time complexity found in @xcite for discrete - time searches and in @xcite for continuous - time searches . to boost the probability to @xmath282 , @xmath283 repetitions are required giving a total time @xmath284 . amplification methods @xcite may be used to reduce the total search time further . our main result focusses on perturbations which involve altering the hopping potential from a marked site to all three of its nearest - neighbors equally . we have also demonstrated other types searches based on perturbing the hopping potential in a cell and the adding extra sites . we have shown that search mechanisms can be utilised for the purposes of signal transfer . our findings point towards applications in directed signal transfer , state reconstruction , or sensitive switching . this opens up the possibility of a completely new type of electronic engineering using single atoms as building blocks of electronic devices . our results demonstrate that a range of nanostructures constructed from graphene could be used to this end . we thank julian bhm , ulrich kuhl and fabrice mortessagne for discussions on how the ideas of @xcite could be adapted to more physical systems , which lead to some of the work on additional perturbation types and graphene flakes . we give here some details regarding the derivation of eq . ( [ eq : i2est ] ) and the logarithmic divergence of @xmath179 . it is clear that the dominant contributions to the @xmath158 summations come from the vicinity of the dirac points . approximating the spectrum close to the dirac points , one has @xmath286 + o(1 ) \ . \end{split}\ ] ] here the sums over integers @xmath287 and @xmath288 is over a rectangular region @xmath170 of the lattice @xmath172 which is centered at @xmath173 and has side lengths proportional to @xmath174 the center @xmath173 , corresponding to the relevant dirac point , is omitted from the sum . as stated in the main text , for @xmath289 the corresponding sums converge which proves eq . ( [ eq : ikest ] ) . + for @xmath290 we will establish constant @xmath291 and @xmath246 such that @xmath292 which then directly leads to eq . ( [ eq : i2est ] ) . to establish @xmath291 note that because each term in the sum ( [ logineq ] ) is positive its value decreases by restricting it to a square region @xmath293 , @xmath294 which is completely contained in @xmath170 . up to an error of order one the sum over a square region can in turn be written as a sum over eight terms of the form @xmath295 for fixed @xmath287 we can find @xmath296 such that @xmath297 we may choose @xmath298 for some constant @xmath299 , so @xmath300 which diverges as @xmath301 . establishing @xmath246 and the corresponding logarithmic bound from above , and thus proving eq . ( [ eq : i2est ] ) , follows the same line by first extending the sum to a square of side length @xmath302 that completely contains @xmath170 and then establishing @xmath303 with @xmath304 . d. aharonov , a. ambainis , j. kempe , and u. vazirani . quantum walks on graphs , in _ proceedings of the thirty - third annual acm symposium on theory of computing _ , ( stoc 01 , acm , new york , ny , usa , 2001 ) , pages 50 - 59 . l. k. grover . a fast quantum mechanical algorithm for database search , in _ proceedings of the twenty - eighth annual acm symposium on theory of computing _ , ( stoc 96 , acm , new york , ny , usa , 1996 ) , pages 212 - 219 . a. ambainis , j. kempe , and a. rivosh . coins make quantum walks faster . in _ proceedings of the sixteenth annual acm - siam symposium on discrete algorithms _ , ( society for industrial and applied mathematics , philadelphia , pa , usa , 2005 ) , pages 1099 - 1108 . j. boehm , m. bellec , f. mortessagne , u. kuhl , s. barkhofen , s. gehler , h.j . stckmann , i. foulger , s. gnutzmann , and g. tanner . experiments on ` quantum ' search and directed transport in microwave artificial graphene , _ phys . _ , * 114*:0 110501 , 2015 . a. ambainis , a. backurs , n. nahimovs , r. ozols and a. rivosh . search by quantum walks on two - dimensional grid without amplitude amplification . in _ theory of quantum computation , communication , and cryptography _ , ( springer berlin heidelberg , berlin , germany , 2013 ) , pages 87 - 97 .

the study of the formation , composition , mass and kinematics of galaxy halos are among the most active areas of modern astrophysical research . the formation of galaxy halos are now widely believed to be due to hierarchical structure formation ( e.g. , freeman & bland - hawthorn 2002 ) where galaxies are produced by the combination of a monolithic collapse ( eggen et al . 1962 ) and the accretion of galactic fragments ( searle and zinn 1978 ) . for the halo of the milky way , all aspects of the halo can probed by studying the nature of the remnants of disrupted dwarf galaxies . numerous tidal streams and dwarf galaxies have been discovered within the galactic halo in the last 20 years ( e.g. , ibata et al . 1994 ; ibata et al . 2001 ; vivas et al . 2001 ; grillmair 2006 ; newberg et al . 2009 ) . the most well - studied of these is the sagittarius dwarf galaxy ( sgr , ibata et al . 1994 ) and its associated tidal stream . the sgr streams have been traced on large scales using blue horizontal branch ( bhb ) stars ( newberg et al . 2003 ) , m - giants ( majewski et al . 2003 ) , and main - sequence turn off ( msto ) stars ( e.g. , belokurov et al . 2006 ; koposov et al . the sgr system has also been studied using rr lyrae ( rrl , vivas & zinn 2006 ; miceli et al . 2008 ; sesar et al . 2010 ) . other halo streams and structures discovered recently include a virgo stellar stream ( vss ; vivas & zinn 2006 ; vivas et al . 2008 ) , a virgo overdensity ( vod , newberg et al . 2002 ; newberg et al . 2007 ) , an overdensity in pisces ( sesar et al . 2007 ; kollmeier et al . 2009 ) and a monoceros stream ( newberg et al . 2002 ; majewski et al . additionally , a cetus stream has been discovered in the south ( newberg et al . 2009 ; koposov et al . 2012 ) and also evidence for an overdensity dubbed the hercules - aquila cloud ( belokurov et al . 2007 ) . although the sgr stream system has been the focus of much study , since the structure is extends completely around the galaxy , it is yet to be fully mapped in velocity , metallicity and distance . furthermore , the cause of the bifurcation in sgr stream stars discovered by belokurov et al . ( 2006 ) is yet to be explained . additionally , newberg et al . ( 2003 ) proposed the existence of a stream of stars associated with the sgr dwarf galaxy at galactocentric distances of @xmath3 . these authors photometrically identified candidate bhb stars in sdss data and found evidence for an overdensity of stars with @xmath4 . this overdensity was found using sdss data covering @xmath5 , @xmath6 . newberg et al . ( 2003 ) attributed this feature to a trailing arm of the sgr stream , yet also noted a possible link to the globular cluster ngc 2419 . their analysis showed evidence for a feature visible across @xmath7 along the sgr plane . newberg et al . ( 2007 ) continued this work and found additional evidence for what they called the sgr trailing tail using bhb candidates selected from sdss data release 5 ( dr5 ; alderman - mccarthy et al . more recently , ruhland et al . ( 2011 ) found additional evidence for this overdensity using candidate bhb stars selected from the larger sdss dr7 dataset . as with newberg et al . ( 2003 , 2007 ) , these authors found a stream of bhb candidates to be located at heliocentric distances of 60 - 80 kpc with @xmath8 significance . within their analysis the feature was seen to span @xmath9 of the sky in the region @xmath10 . ruhland et al . ( 2011 ) also compared this extended stream with five sets of published numerical simulations of the sgr stream . they found that none were able to explain the existence of sgr stream stars in the location observed . current evidence for an outer sgr stream / tidal - tail remains based purely on photometrically selected bhb stars . as demonstrated by sirko et al . ( 2004 ) and ruhland et al . ( 2011 ) , it is not possible to select a 100% pure set of bhb stars using sdss photometry alone . depending on the colour cuts used one has to accept varying levels of contamination blue straggler ( bs ) stars which far out number bhb stars at the greatest detection completeness . with strict colour cuts , that remove a large fraction of the bhb stars , the level of bs star contamination can be reduced to 30% ( ruhland et al . however , to retain a large number of bhb stars , a 50% or greater contamination level must be accepted ( brown et al . 2005 ; bell et al . 2010 ) . in contrast to photometric selection , it is possible to select bhb stars with much less contamination by combining colour selection with high s / n spectra because bhb stars have significantly lower surface gravity than the bs stars ( yanny et al . the sdss segue - i and segue - ii projects ( yanny et al . 2009 ) undertook spectroscopy of @xmath0 300,000 star with an aim to identifying bhb stars and trace galactic halo structure . however , while radial velocities are available for stars to @xmath11 ( yanny et al . 2009 ) , the s / n required to accurately measure @xmath12 limits the survey to bhb stars with @xmath13 ( @xmath14 kpc ) . almost all of the bhb candidates discovered by newberg et al . ( 2003 , 2007 ) and ruhland et al . ( 2011 ) are beyond this limit and do not have measurements of @xmath12 to confirm their nature . additionally , while the sdss segue surveys cover @xmath15 , the fields cover a patchwork over the northern sky , rather than just the sgr stream region where this feature is located . like bhb stars , rrl stars also exist on horizontal branch ( hb ) . however , unlike bhb stars , rrl exhibit a significant level of intrinsic photometric variability . based on their characteristic variability rrl can be cleanly separated from other stars . type - ab rrl stars ( rrab ) have absolute magnitudes of @xmath16 with uncertainties of 6% ( catelan et al . this makes them excellent distance indicators . although , significant numbers of repeated observations are required to provide accurate average magnitudes . to date , a few tens of thousands of rrl are known in dense regions near the galactic bulge and in the magellanic clouds due to microlensing surveys ( soszyski et al . 2009 , pietrukowicz et al . recently we discovered @xmath0 10,000 rrab over a large fraction of the sky ( drake et al . 2013 ; dr13 ) . however , even with the dr13 rrab s most of the galactic halo remains unexplored at heliocentric distances @xmath17 . in this paper we outline our search , discovery , and calibration of rrl at distances up to 100 kpc and beyond . we will compare our discoveries with results from past surveys and undertake an analysis of the distant hb stars uncovered . the catalina sky survey began in 2004 and uses three telescopes to cover the sky between declination @xmath18 and @xmath19 in order to discover near earth objects ( neos ) and potential hazardous asteroids ( larson et al . the three telescopes are each considered sub - surveys . these consist of the catalina schmidt survey ( css ) , the mount lemmon survey ( mls ) and the siding spring survey ( sss ) . in addition to avoiding high declinations , the galactic plane region is avoided by between 10 and 15 degrees due to reduced source recovery in crowded stellar regions . all of the survey images are taken in sequences of four observations separated by ten minutes , and all the observations are unfiltered to maximize throughput . photometry is carried out using the aperture photometry program sextractor ( bertin & arnouts 1996 ) . all the catalina data is also analyzed for transient sources by the catalina real - time transient survey ( crts , drake et al . 2009 ; djorgovski et al . 2011 ) . in paper - i ( dr13 ) , we concentrated on analyzing rrab data from the css sub - survey . in this paper we mainly work with data observed by the mls 1.5 m telescope . this data predominately covers ecliptic latitudes @xmath20 . for this telescope each image from the @xmath21 catalina ccd camera covers 1.1 @xmath22 on the sky . observations are taken 21 nights per lunation avoiding bright time . typical exposures of 30 seconds to reach @xmath23 . in total , the mls source catalog consists of 155 million sources . as noted in dr13 , the mls data has the same photometric sensitivity as the css data since all observations are taken unfiltered with the same type of ccd camera . all images are processed with the same software . from dr13 , the colour transformation from catalina photometric system @xmath24 magnitudes to johnson @xmath25 is given by : @xmath26 the average b - v colour of rrl is about 0.3 mag with stars varying between about 0.1 and 0.5 as they pulsate ( e.g. , clube et al . 1969 ; stepien 1972 ; cacciari et al . 1987 ; layden 1997 ; nemec 2004 ) . as the mls data has no colour information , we adopt the average values in our analysis . this gives rise to an maximum uncertainty of @xmath27 mags in the @xmath25 magnitudes . combining this with the photometric uncertainty , we expect a colour - based dispersion of @xmath28 mags in our rrl photometry . to find rrl candidates we follow the analysis undertaken in dr13 . that is , we first selected variable sources using a welch - stetson variability index @xmath29 ( welch & stetson 1993 ) . 3.1 million mls sources were selected as variable candidates using @xmath30 . all sources were processed for periodicity using the lomb - scargle ( ls , lomb 1976 ; scargle 1982 ) periodograms . in total @xmath31 thousand sources were found to exhibit significant periodicity with a false alarm probability @xmath32 . however , the bulk of these detections are due to the systematic sampling of the data and occur at integer day frequencies . here , as with dr13 , we are primarily interested in rrab s since they have well known magnitudes and characteristic light curves that are not easily mistaken for other periodic variables . after excluding the period aliases near 0.5 and 1 day , 3087 of the objects found to have periods between 0.34 and 1.4 days . this period region was deliberately chosen to be broader than the rrab range in order to include sources found by the ls technique at aliases of their true periods . as with the css sources of dr13 , each of the rrab candidate was run through the afd software ( torreabla et al . 2012 ) to select the best period from among the best 15 given by the ls and aov software ( schwarzenberg - czerny 1989 ) . briefly , this process involves fourier fitting and iteratively rejecting bad data to determine the best period based on reduced @xmath33 values . non - rrab sources are rejected using the m - test statistic ( kinemuchi et al , 2006 ; equation 8) , as well as the fourier fit order . this initial selection resulted in the detection of 1125 rrab candidates . upon reviewing the light curves of the rrab candidates we discovered a number of faint rrab in the mls data with @xmath34 . such rrl were found to be mainly concentrated in the gemini constellation . in figure [ mlsrrab ] , we plot the phased light curves of eight of these faint stars . these distant sources are very important for defining distant streams within the galactic halo . from simulations , based on adding artificial rrab light curves to our data , we discovered that our detection sensitivity was @xmath35 for rrab s with @xmath36 . to improve our completeness we investigated the light curves of mls sources with sdss spectroscopic types a0 and f5 that had @xmath37 . in addition , we used sdss photometry to select stars with the same colour range as the rrab stars that we had discovered . these stars were matched with mls sources at lower variability and periodicity thresholds ( @xmath38 and @xmath39 ) to select rrab candidates . to better constrain the variability , we combined the mls photometry with shallower css data . likewise , we searched the css data for additional faint rrab candidates by selecting objects with rrab colours in sdss photometry . furthermore , since our analysis in dr13 showed that @xmath40 of rrab s were missed because the selection based on ls periods , we reprocessed the css and mls photometry using the aov period - finding software . we inspected the light curve of each rrl candidate to assess variability and then carried out period finding for all the new variable candidates . in total , approximately six thousand additional sources were searched . each of the new sources exhibiting some sign of periodicity were searched for improved periods using our fourier - fitting afd process . in a number of cases the rrl were found to exhibit blazhko phase and amplitude variations ( blazhko 1907 ) . in total the additional searches yielded an 219 extra mls rrl and 2051 css rrab s . of these only 425 were previously known . as distant rrl are important for defining halo streams we retain the 17 clear c - type rrl ( rrc ) that were discovered in mls data beyond 70 kpc . a number of these have sdss spectra . after removing non - rrl sources the final set of mls rrl consists of 1,207 stars and is presented in table 1 . the new rrab s found in css data are given in table 2 . in figure [ mlshist ] , we present the magnitude distribution of mls and the css rrl . the mls histogram shows clear bumps near @xmath41 and @xmath42 . the first bump is easy to understand as it is clearly due to the sagittarius leading and trailing arms ( as seen in dr13 ) . the second peak is due to the more distant rrl that are concentrated near the galactic anti - center . we hereafter call this the gemini stream . in section [ sagstream ] , we will investigate the origin of this feature . the absolute magnitudes of rrab s are given by catelan & corts ( 2008 ) : @xmath43}_{\rm zw84}+ 0.948,\ ] ] where @xmath44}_{\rm zw84}$ ] is the metallicity in the zinn & west ( 1984 ) scale . the average metallicity for rrab s with sdss spectra found by dr13 was @xmath45= -1.55 $ ] . the comparison given in dr13 , between distant halo rrab s ( @xmath46 kpc ) and the full set of rrab s showed little difference in the distribution or average metalicity . however , css rrab s with sdss metallicities are stars with @xmath47 , corresponding to distance of @xmath48 kpc . brighter foreground field rrab s are likely to have higher metallicities . the mls rrab s should have a similar metallicity distribution to the larger sample of css sources discovered in dr13 . therefore , we adopt the average css metallicity and assume rrab s have magnitude @xmath49 . this value is close to the value of 0.6 that is often adopted ( e.g. , keller et al . 2008 , sesar et al . the rrl v magnitudes were corrected for extinction using schlegel et al . ( 1998 ) reddening maps . the dispersion in the metallicity from dr13 is 0.3 dex , which corresponds to a variation of 0.07 magnitudes . the distances to individual sources are determined using : @xmath50 here we correct the average rrab @xmath51 magnitudes to static values @xmath52 using values derived from a polynomial fit to the amplitude corrections given by bono et al . ( 1995 ) . combining the uncertainties from the photometric calibration and colour variation of 0.09 , with the variation in metallicity we derive an overall uncertainty of 0.11 magnitudes , corresponding to a @xmath53 uncertainty in distance . as demonstrated by dr13 , faint rrl have larger uncertainties in their average magnitude . however , these uncertainties should generally not exceed 0.2 magnitudes ( @xmath54 in distance ) . the faintest mls rrab in our dataset have @xmath55 corresponding to heliocentric distances of 120 kpc . in figure [ dist ] , we plot the distribution of the distances of mls rrl and mark the locations of known halo features as well as the gemini stream . we matched the mls rrab s with past rrl discoveries using the simbad and the international variable star index ( vsx ; watson et al . 2006 ) databases , as well as those presented by dr13 . in total , there were 605 matches with past surveys of which 316 came from css data ( dr13 ) and 289 were discovered in other surveys . as there is overlap between css and the other previously known rrab s , there was css photometry for 540 of the mls rrab s . of the 289 simbad and vsx rrab s , 278 of the rrab have known periods given by vsx . in figure [ other ] , we compare the periods derived from mls data with previous determinations . the figure shows that the agreement between the css and mls periods is much better than with other sources . assuming the differences in periods are normally distributed , the mls - css matches have @xmath56 , while the mls - vsx source matches have @xmath57 . as the level of agreement is good in both cases , we have confidence in the mls rrab periods . the sekbo survey discovered 2016 candidate rr lyrae in a survey covering 1675 @xmath22 along the ecliptic ( keller et al . like the mls survey , the sekbo survey covered a band within @xmath58 of the ecliptic . however , sekbo rrl candidate selection was based on the two colours observed simultaneously by the macho camera and variability was determined from between 2 and 8 epochs of images . in contrast , the mls rrab candidates are observed unfiltered and have an average of 130 observations ( sampling ranges from 42 to 342 measurements ) . we matched the mls rrab with the sekbo rrl candidates and found only 103 matches . however , this is not unexpected since the sampling of mls is highly concentrated within few degrees of the ecliptic with few observations having been taken near the @xmath59 limit . for example , 70% of the mls rrab s are within @xmath60 of the ecliptic , compared to 27% of the sekbo rrl candidates . to extend the comparison , we also matched the sekbo candidates with rrab s from the dr13 sample and the sss rrab s ( torrealba et al . this yielded an additional 538 rrab matches . because of overlap between the css , mls and sss rrab catalogs there were 540 unique rrab stars matching sekbo rrl candidates . clearly this number falls far short of the number of sekbo candidates . we compared average v magnitudes for the catalina sources with those of the sekbo sources . in figure [ sekbo ] , we present the v magnitude differences between the measurements . the average difference between magnitudes is -0.02 magnitudes with @xmath61 magnitudes , corresponding to a @xmath62 uncertainty in source distances . the internal dispersion for css rrab magnitudes from overlapping fields is @xmath63 mags and thus does not significantly contribute to the observed dispersion . in order to investigate the sekbo candidates not found in the catalina rrab catalogs , we matched the remaining 1481 unmatched sekbo objects to the sdss dr8 spectroscopic catalog . we found 40 matches , of these 35 had measured @xmath12 values and 23 had @xmath64 consistent with rrl ( dr13 ) . we also matched the sekbo sources with sdss dr8 photometry . of the 276 sources with sdss photometry , approximately one third had sdss colours beyond the limits observed for dr13 rrab s . next , we matched the 1481 sekbo sources to photometry from mls , css and sss surveys . among the matches , we sub - selected the objects exhibiting significant variability based on the @xmath30 and also objects with ls periodic significance @xmath32 . in total , we find matches for 90% of the sekbo objects , including 88% of the sources that were not selected as rrab s . combining the known catalina rrab s and other sources with apparent variability , we find that 74.5% of the 1833 sekbo sources we cover exhibit significant variability . in table 3 , we present the numbers of matches with each of the catalina surveys , as well as the number of objects selected by our variability and periodicity significance thresholds . for each of the 1314 sekbo rrl candidates with catalina photometry , that were not in our rrab catalogs , we determined a ls period ( regardless of their @xmath29 variability ) . we inspected both the observed and phase folded light curves of each source and determined a classification based on this photometry . as there are 2759 catalina light curves matching the 1314 sekbo candidates , the majority of these sources have multiple light curves . we discovered 551 additional rrl candidates , 140 eclipsing binary candidates ( mainly wuma ) and 140 other types of variable sources ( including qso , cvs , @xmath65 scuti variables and objects where the classification was unclear ) . this discovery of many non - periodic sources is not unexpected since a number of these are already known , including qsos and cataclysmic variables . for example , sekbo 106646.2532 ( css080623:140454 - 102702 ) is a cv that was discovered in outburst by crts in 2008 ( kato et al . 2012 ) . additionally , based on sdss dr8 spectroscopy , sekbo-105832.627 is a cv , sekbo-117146.1412 is a qso at z=1.02 , and sekbo-096514.1360 is an eclipsing wd - md binary system ( rebassa - mansergas et al . 2010 ) . in order to obtain the best periods for the clear rrl , and to find rrl that may have been missed during our inspection , we ran every light curve through the afd software ( torrealba et al . for 307 of the 551 rrl candidates we found good fourier fits and periods consistent with rrab s and rrc s . in table 4 , we present all the new sekbo rrl found using catalina data ( including rrl candidates where the period is uncertain ) . this table contains 263 rrab , 282 rrc , and 11 rrd candidates . combining the new rrab s with the 540 from catalina catalogs we find that @xmath66 were missing from our catalogs in good agreement with the simulations presented in dr13 . the overall sample suggests that 24% of the rrl are rrc stars whereas keller et al . ( 2006 ) estimated that @xmath67 of the sekbo candidates would be rrc s . however , we note that pietrukowicz et al . ( 2012 ) found @xmath68 rrc among their @xmath69 bulge rrl . nevertheless , it is possible that some fraction of the rrc candidates we identify may be w uma variables ( due to the similarity of their light curves ) . the uncertain separation of rrc s and w uma types is the main reason why rrc s are often excluded as distance indicators . our results suggest that @xmath70 of the sekbo rrl candidates are likely rrl based on the 1833 objects covered by catalina data . this is @xmath71 smaller than the @xmath72 non - rrl contamination estimated by prior et al . ( 2009a ) . however , we note that the prior et al . ( 2009a ) sample was based on an average of just nine photometric measurements for @xmath73 ( @xmath53 ) of the sekbo stars . prior et al . ( 2009b ) followed 21 additional sekbo rrls overlapping the sgr tidal stream . more recently , akhter et al . ( 2012 ) followed 137 sekbo candidates and found 57 to be rrls , although they found a high fraction with colours matching the colours of rrl candidates from sdss data . from catalina photometry we have an average of 121 observations per light curve ( 250 per object ) covering 90% of the sekbo sources . our analysis suggests that @xmath74 of the sekbo sources are either not variable or have significantly less variability than expected for rrl . nevertheless , with @xmath75 likely rrl from the sekbo survey , the results shows that large numbers of rrl can be found using very small numbers of observations and colour selection . the outer halo rrl discovered in this analysis overlap with tidal stream first identified by newberg et al . ( 2003 ) using photometrically selected bhb candidates . as noted above , additional bhb candidates have since been found is sdss dr5 ( newberg et al . 2007 ) and sdss dr7 data ( ruhland et al . furthermore , ivezic et al . ( 2004 ) discovered many rrl candidates overlapping the sgr stream plane . the most distant sources coincide with the newberg et al . ( 2003 ) bhb candidates . however , the ivezic et al . ( 2004 ) rrl sample was based on photometric selection combined with limited variability information . of the 1,269 ivezic ( 2004 ) rr lyrae candidates selected in the sdss stripe-82 region , only 483 ( 38% ) were eventually confirmed by sesar et al . it seems likely the a similar fraction of the sources found in this marginally significant ( @xmath76 ) detection by ivezic et al . ( 2004 ) were indeed rrl . recently , additional evidence for a distant group of rrl stars was found by sesar et al . ( 2012a ) . these authors discovered eight rrab between @xmath77 and @xmath78 at heliocentric distances @xmath79 kpc . this region lies on the edge of the fields where bhb candidates were earlier discovered by newberg et al . sesar et al . ( 2012a ) selected potential rrab candidates based on variable palomar transient factory ( ptf ) sources with rrl colours based on sdss dr8 photometry . using simulations sesar et al . ( 2012a ) estimated that in their worst - case scenario their search recovered 95% of their simulated sources , with at most one rrab star being missed . six of the sesar et al . ( 2012a ) rrab s were independently discovered in our searches , while the remaining two were missed due to poor sampling . for all six known rrab , we find periods matching those given by sesar et al . ( 2012a ) to better than @xmath80 . in addition to the six ptf rrab , we have discovered seven rrab within the region and distant range noted by sesar et al . we find 16 more rrab s in this region at heliocentric distances @xmath81 kpc . as we missed two of the ptf rrab s it is possible that there are even more rrab s within this region . in figure [ ptf ] , we present the locations of the rrab s found within the ptf praesepe fields . among the set of eight rrab s , sesar et al . ( 2012 ) found tentative evidence for two separate groups of rrl based on radial velocities . one group of four rrab s was noted as having radial velocities in the galactic standard of rest of @xmath82 , while the other had @xmath83 . although sesar et al . ( 2012 ) note that there is a 37% chance that these stars are drawn from the same gaussian velocity distribution based on the shapiro & wilk ( 1965 ) sw statistic , they assume that their eight sources are representative of the overall velocity distribution and via numerical simulations find a @xmath84 chance that the sources would exhibit the observed velocities . these authors also discovered that the metallicities of the two groups were consistent to within uncertainties . in total , we identify 103 rrab s at heliocentric distances beyond 70 kpc . four of these rrab s appear near @xmath85 , @xmath86 kpc . these stars are associated with the pisces stream ( sesar et al . 2007 ; watkins et al . 2009 ) . however , we see no evidence for variations in distance . in agreement with the results of kollmeier et al . ( 2009 ) . this supports evidence that this system is not part of a tidal stream itself . of the remaining stars , 82 rrab s lie within the range @xmath87 and @xmath88 . these objects exhibit a range of distances and thus comprise significant evidence for a stream of stars , suggesting association with a tidally disrupted galaxy . to better understand the nature of the relation between the distant rrab s and the sgr stream we selected the rrab sources within @xmath89 of the sgr plane ( as defined by majewski et al . 2003 ) . based on koposov et al . ( 2012 ) and dr13 . most sgr stream stars , including rrab s , lie within these limits . we combine rrab s from dr13 along with new css rab s and mls rrab s in the range @xmath90 within the sgr streams region . we transform the rrab distances to galactocentric values and plot the sources in figure [ distg ] . in this plot the distant stream of rrab s becomes clearer , as does the leading and trailing sgr streams that are observed between @xmath91 and @xmath92 , respectively . among the set of distant mls rrab s , the four most distant objects are found within the @xmath93 region @xmath94 , @xmath95 . these rrab s have distances @xmath96 kpc , their average magnitudes varying by @xmath97 mags and are consistent with no difference at their observed brightness . this concentrated group is significantly fainter than any of the other mls rrab s , suggesting a possible association between them that varies from the other sources . however , measurements of radial velocities and metallicities are required before we can rule out association with the other rrab s . after excluding the group of four very distant rrab s , and four outlier - rrab s near @xmath98kpc with @xmath99 we determine the slope of the outer stream by selecting the rrab s with @xmath100 kpc , in the range @xmath101 . a simple linear fit gives @xmath102 . the slope of this line suggests that the distant stream does not meet with the so - called trailing sgr stream lying in the region @xmath103 . in order to better visualize the direction and slope of the streams , in figure [ polar ] , we present a polar plot of the same rrab sources in the sgr plane coordinate system . in contrast to figure [ distg ] , here we plot average static star magnitudes for the rrab fainter than @xmath104 ( @xmath105 ) . clearly a more accurate description of path of this outer stream requires more rrab in the region @xmath106 , which is not covered by the mls photometry . as noted earlier , many new css rrab were found in this region . however , as shown in figure [ mlshist ] , and dr13 , rrab s fainter than @xmath107 ( @xmath108 kpc ) are at the limit of css data . discoveries in css data covering this region will be likely biased to the brighter , nearer sources . to further explore the relationship with the sgr streams , we plot the rrab s in the sgr x - y plane system of majewski et al . ( 2003 ) in figure [ sgrxy ] . here we mark the proposed path of the gemini stream . we also plot the law & majewski ( 2010 ) n - body model of the sgr stream system . the model is reasonable match to the structure of the inner rrab data . the agreement is increased if the source distances are reduced by @xmath109 . however , as found by ruhland et al . ( 2011 ) , the gemini stream is not explained by this model or other models of the sgr stream system . nevertheless , it appears that if the path of sgr leading debris was extended , it might better match the location of the sources . indeed , law & majewski ( 2010 ) found that if the sgr dwarf had been orbiting 2 gyr longer , a much longer leading arm would be present . nevertheless , their simulations did not produced sources corresponding to the distances and locations of the bhb candidates and rrl . it is apparent that the sgr trailing arm model is a very poor match to the gemini stream , both in extent and location . to determine the significance of the gemini stream it is necessary to compare the source density with that expected from halo models . we adopt the halo density model from sesar et al . ( 2010 ) for our comparison . the local density of rrl is determined using the nth nearest neighbour method , we find the 8th nearest star to the point where we want to calculate the density and calculate the area that bounds these stars . we use 8 stars motivated by ivezic et al . ( 2005 ) where they found that this number gives the best results for the underlying density on their improved bayesian method . we also saw that this number was a good combination of precision and computational efficiency . we compute the observed number density within our grid ( @xmath110 ) , and density from the sesar et al . ( 2010 ) halo model ( @xmath111 ) . to visualize the over - densities on the sgr plane , we produce a @xmath112 grid on the mawejeski et al . ( 2003 ) sgr coordinates at @xmath113 ( for the x - y plot ) and @xmath114 ( for the @xmath115-@xmath116 plot ) . in figure [ sgrgab ] , we plot the resulting density ratios in the x - y plane of the sgr system and in figure [ sgrgabx ] , we plot the densities in the @xmath115-@xmath117 system . the densities here can be readily contrasted sources in figure [ sgrxy ] . the main density features are the leading and trailing sgr arms ( marked by a and b , respectively ) . the gemini stream ( marked as c ) , largely has a higher ratio density ratio than the trailing stream system ( marked by b ) . another possible feature is seen near @xmath118 , @xmath119 kpc ( marked @xmath120 ) . this has much lower significance than the other features , yet is lactated where models predict the sgr trailing arm should cross the galactic plane and meet with the trailing arm @xmath121 . since the gemini feature , like much of the sgr stream , is offset from the sgr plane , we have collapsed the source z positions to visualize the stream . unlike the gemini stream , the leading and trail sgr arms a and b are all ready well known ( law and majewski 2010 , and references therein ) . to quantify the significance of the gemini stream , we estimated the number of stars expected by the model at the position of the stream . for simplicity , we selected the volume that bounds the stream in the sgr coordinate system . that is @xmath122 , @xmath123 and @xmath124 . based on the sesar et al . ( 2010 ) model density , within this volume we would expect to find 373 rrl assuming 100% detection efficiency . however , only 106 rrl were found . assuming pure poisson uncertainties , this suggests the gemini area has @xmath125 under - density . as our detection efficiency is close to @xmath126% based on dr13 this short fall can not be explained pure by missing rrl . the detection of no distant rrl across most of the mls survey area strongly suggests that the halo is much less dense than expected by the model . this result is in agreement with the results of watkins et al . ( 2009 ) and sesar et al . ( 2010 ) based on halo rrl discovered in sdss stripe-82 . this confirms that the halo density declines more rapidly than suggested by the juric et al . ( 2008 ) halo model . as a second model comparison , we compare the rrl densities with the watkins et al . ( 2009 ) halo model . their results suggest a break in density occurs @xmath127 kpc . matching the model to the data we find that the watkins et al . ( 2009 ) model requires normalization by a factor of 11 to match our data . this is in good agreement with the factor of 10 found by sesar et al . ( 2010 ) . in figure [ sgrgabwat ] , we plot the observed radial density compared with the sesar et al . ( 2010 ) model and the normalized watkins et al . ( 2009 ) model . the observations are found to be in good agreement with the watkins et al . ( 2009 ) model , although the break in density appears to occur nearer to 50 kpc when averaged over a large area . however , this is because the densities are enhanced by rrl in the sgr leading and trailing arms . based on the watkins et al . ( 2009 ) model we expect to find 50 rrl in the region selected above . the number of gemini stream rrl we find is thus @xmath128 larger than expected from this model . to investigate differences between the gemini rrab sources and the overall population , one can infer the oosterhoff type based on the rrab period - amplitude relationship ( smith et al . 2011 , and references therein ) . however , since average colours were used to transform the mls light curves , we need to account for the effect of colour variation of the rrab light curves . in dr13 we found that rrab amplitudes were systematically reduced by 0.15 due to pulsational colour variations within the broad bandpass of css images compared to v - band . we also found that the css oosterhoff type - i ( ooi ) rrab s exhibit a well - defined amplitude limit that is 0.1 mags higher than the zorotovic et al . ( 2010 ) period - amplitude relationship . to separate ooi rrab candidates from oosterhoff type - ii s ( ooii s ) , we correct the mls amplitudes to values by adding 0.15 mags . next we define a set of rrab s that are between 0.1 and 0.25 mags above than the average ooi period - amplitude relationship . these objects are a mixture of oosterhoff types ( although some may be oosterhoff - intermediate sources ) that we remove from consideration . we then investigated the variation in mls rrab s oosterhoff types with distance . in figure [ mlsotype ] we plot the period - amplitude distribution of the mls rrab s . by dividing these sources into a nearer sample consisting of rrab with @xmath129 and the more distant rrab @xmath130 we see a division in types . in figure [ mlsotype ] , we also plot the period - amplitude distribution for the gemini rrab s . we also over - plot the zorotovic et al . ( 2010 ) ooi line and an ooii line that is offset by @xmath131 in @xmath132 to match the css observations . the nearer rrab exhibit a mixture of oosterhoff types when compared to the zorotovic et al . ( 2010 ) period - amplitude relationships . this is significant evidence that the rrab s in the range @xmath129 do not come from a single population . in contrast , the distant mls rrab s all lie near the ooi line , suggesting that they come from a single population that is more metal - rich . in figure [ mlshbd ] , we show the locations of gemini stream rrab s after removing the mls rrab s with ambiguous oosterhoff type . the css rrab s with @xmath129 have also been included , but are not separated by oosterhoff type since their amplitudes and periods are less certain than the mls rrab s . this figure shows that the gemini rrab s are spread across the width of sgr stream system . it also shows that the ooii rrab s selected among the nearer rrl set ( @xmath129 ) are distributed across the region . the nearer ooi group ends around @xmath133 where the stream divides between the nearer and further groups . analysis of this figure suggests that there may be two or more overlapping populations : one that follows a steep distant gradient and is predominantly ooi , and another that has a shallower gradient and is an ooii population . with respect to this figure , we once again note that the mls rrl are naturally concentrated toward the ecliptic because of sampling . so it is not possible to make inferences about changes in density across the sgr stream . while the sdss photometry has little of the repeated photometry required to unambiguously identify rrab s , it is deeper than both css and mls data ( reaching hb stars to @xmath134 ) . additionally , sdss data does cover most of the gemini stream observed in the mls data as well as the main region where mls data does not overlap the sgr stream . therefore , it is possible to use this data to bridge the gap in between the depth of css data and coverage of mls photometry . unlike previous authors whom searched for bhb stars covering the sgr tidal streams , here we sought to select both bhb and potential rr lyrae stars . based on the colours of the distant mls rrab s and prior bhb work , we investigated the observed colours of rrl in sdss photometry and selected sdss dr8 stars within @xmath135 , @xmath136 , @xmath137 , sdss object type=6 ( star ) , and @xmath138 . based on dr13 we know that the bulk of rrab s are located near @xmath139 , @xmath140 and @xmath141 . however , rrab s with these colours are outnumbered by main sequence turnoff stars ( msto ) by a factor of @xmath142 ( koposov et al . 2012 ) . using maps of sdss source density , we sub - selected stars in the range @xmath143 for @xmath144 . this selection retains most of the bhb candidates of ruhland et al.(2011 ) as well as a 22% of the distant mls rrab sample and many bs stars . this colour selection reduces the initial number of sdss point sources in our selection from 1.65 million to 81552 . by selecting stars in the sgr stream region , with @xmath145 , the number reduces to 23507 hb candidates . to match the mls photometry we transform the sdss photometry to @xmath25 magnitudes using ivezic et al . ( 2007 ) . as found with newberg et al . ( 2003 , 2007 ) and ruhland et al . ( 2011 ) we found the spatial distribution of sdss hb candidates suffers from a significant crowding and background that is best viewed in source density . in figure [ hess ] , we provide the hb candidates in the form of the hess ( point - density ) diagram . this figure shows evidence that the density of hb candidates closely follows the fit to the distant gemini stream of rrab s , in agreement with the distribution of bhb candidates from ruhland et al . ( 2011 ) and others . here the scaling has been set to match that in figure [ polar ] where the main streams are visible . in addition to the rrl , the paths expected for bs stars that pass the colour cuts and form a shadow that is @xmath146 mag fainter than the hb stars is also shown . to obtain another view of mls rrl in the gemini stream in relation to the sdss hb candidates , we divided the sdss sources into three groups . these were bright sources , with @xmath147 , intermediate - brightness sources , with @xmath148 , and faint sources , with @xmath149 . in figure [ mlshb ] , we present the locations of these sources . the bright sources were selected to show the nearby hb stars as well as the bs stars in the monoceros stream ( near the galactic anti - center at @xmath150 limit of the sdss coverage ) . the intermediate - brightness sources were selected to be indicative of the hb stars in the leading arm of the sgr stream . the faintest sources were selected to include the hb candidates in the gemini stream . however , this also include bs stars that mirror the distribution of hb stars along the sgr leading arm . the figure clearly shows the overdensity associated with the sgr streams system . the separation into two streams is not clear here since there are far fewer bhb stars than the msto stars used by belokurov et al . ( 2006 ) . we matched the entire mls rrl catalog with the sdss dr8 spectroscopic catalog and found 89 matches . a much larger sample of rrl with spectra is given in dr13 . however , here our main purpose was analysis of the distant halo rrl . we found that 16 of these matches are sources beyond 70 kpc . as noted earlier , many of the faint rrl candidates were found based on their sdss spectra . the 16 rrl with sdss spectra include 12 rrab s and four rrc s . to separate the radial velocities from the velocities due to pulsation , we use the sdss observation times and the fourier fits to derive the phase at which the rrl spectrum was observed . as noted in dr13 , for sdss spectra , radial velocities are determined by averaging both balmer and metallic lines ( mainly ca lines ) . we follow dr13 by applying sesar et al . ( 2012b ) velocities corrections for pulsation based corrections derived from both hydrogen and metallic lines . the average velocity correction for the 12 rrab is @xmath151 , in good agreement with the uncertainty derived from 905 rrab spectra in dr13 ( @xmath152 km / s ) . we adopt this level of uncertainty for all the sdss rrab spectra . for the four rrc stars which pulsate in the first overtone mode the velocity of the pulsation is much smaller than for the rrab s . based on the rrc s observed by liu & janes ( 1989 ) and jones et al . ( 1988 ) , the amplitude is expected to be approximately @xmath153 . to account for this factor , for these rrl we increase the observed sdss radial velocity uncertainties by an additional @xmath154 . following law & majewski ( 2010 ) we transform the radial velocities to the galactic standard of rest assuming a solar peculiar motion of ( u , v , w ) = ( 9 , 12 + 220 , 7 ) km / s in the galactic cartesian coordinate system . in figure [ mlsvel ] , we plot the radial velocities for the outer halo mls rrab s with spectra , the css rrab s with @xmath155 kpc , and the 10 ptf rrab s found by sesar et al . ( 2012a ) . we also plot the velocities predicted by law & majewski ( 2010 ) n - body simulations . the figure shows that two of the rrab s observed near @xmath156 appear to be associated with the leading arm of the sgr stream . however , for the other gemini stream rrab s , the velocities , like the distances , are not explained by the law & majewski ( 2010 ) n - body simulations . in this regard , sesar et al . ( 2012a ) noted that the rrab velocity measurements they found suggested the rrab s belonged to two distinct groups . additionally , sources associated with our group - d of figure [ sgrxy ] ( occurring near @xmath157 ) have radial velocities that appear to vary rapidly with right ascension . these velocities are also inconsistent with the law & majewski ( 2010 ) model , suggesting that they do not belong to the sgr leading arm . the association between these feature - d rrab s and the gemini stream rrab s is unclear , although values are similar . as noted by newberg et al . ( 2003 ) , the outer halo bhb candidates they discovered reside near the unusual globular cluster ( gc ) ngc 2419 . this system is located at @xmath158 , @xmath159 and distance @xmath160 kpc ( harris 1996 , 2010 edition ) . the corresponding coordinates in the majewski et al . ( 2003 ) sgr coordinate system are @xmath161 , @xmath162 , making it well within the limits of the sgr stream system ( koposov et al . the galactocentric radial velocity is given by newberg et al . ( 2003 ) as @xmath163 , and baumgardt et al . ( 2009 ) find an internal velocity dispersion of @xmath164 . ngc 2419 is noted as being one of the most metal - poor gcs ( @xmath45 \sim -2.1 $ ] ; mucciarelli et al . the cluster is notably old , with age 12.3 gyr according to forbes & bridges ( 2010 ) . the system has the highest luminosity ( @xmath165 mag ) of any gc with a galactocentric distance @xmath166 kpc , apart from the likely sgr dwarf - associated gc , m54 ( cohen et al . the half - light radius of this cluster is 19 pc making it significantly more extended in the @xmath167 versus @xmath168 plane than other gcs with @xmath166 kpc ( mackey & van den berg 2005 ) . indeed , the exceptional nature of ngc 2419 relative to outer halo gcs led van den berg & mackey ( 2004 ) to suggest that the object is the stripped core of a former dwarf spheroidal galaxy ( dsph ) . based on abundance studies cohen et al . ( 2010 ) and cohen & kirby ( 2012 ) also found that the ngc 2419 appears like no other globular cluster , but rather the core of an accreted dwarf galaxy . from figures [ distg ] & [ mlsvel ] we see that the location and radial velocity of ngc 2419 are a relatively good match for the gemini tidal stream . the average velocities and metallicities of the sesar et al . ( 2012a ) cancer group b ( @xmath169 ) and @xmath45 = -2.1 \pm 0.4~dex$ ] ) and the metallicity of the cancer group b is in reasonable agreement with the values expected for a stream from ngc 2419 . however , the velocities and metallicities do not provide a strong enough association to link these sources . another important point to consider in the possible association between the gemini rrl and ngc 2419 is the proximity of the sources . some of the rrl have @xmath170 near the right ascension of the ngc 2419 ( @xmath158 ) . thus , if one was to assume that the gemini tidal stream proceeds in the direction of the sgr stream ( as suggested by sdss hb candidates ) , the stream stars would have to be dispersed across the entire @xmath171 of the sgr system between the mls rrab s and ngc 2419 . sources within this gap are not covered by mls observations because of the coverage limits of the survey . it is thus remain unclear whether there is any link between the gemini stream and ngc 2419 . also , as noted earlier , our data shows that the most distant gemini rrab s are ooi type stars while ngc 2419 is well known to be ooii type . indeed , mucciarelli et al . ( 2012 ) found that ngc 2419 exhibits very little spread in @xmath45 $ ] ( @xmath172 dex ) suggesting it could not be linked with metal - rich rrl in the gemini structure . we have performed a periodicity analysis of 3.1 million variable star candidates selected from photometry taken by the mls survey and uncovered 1,207 rrl ( of which 538 are new ) . comparison of the periods for the @xmath173 previous known rrl shows that the sources are accurately measured . we have also discovered 2040 new rrab stars in a re - analysis of css photometry . using catalina surveys photometry we have determined the nature of 90% of the sekbo ( keller et al . 2006 ) rrl candidates and find that 60% are likely to be rrl . our analysis of the sekbo rrl candidates revealed , selection of a pure set of rrab s with accurate average magnitudes requires many observations . the importance of repeated observations for characterizing variable star types was also recently , demonstrated by sesar et al . ( 2010 ) , who used rrab light curves to study over - densities that they had earlier been attributed to rrl in sesar et al . they discovered that a number of the over densities attributed to rrl , based on photometric selection coupled with a small number of observations , were in fact due to intrinsically fainter @xmath65 scuti stars as well as non - variable sources . overall they found that only 70% of their initial candidates were rrl stars . similarly , ivezic et al . ( 2005 ) found that although it was possible to completely colour - select a small fraction ( 6% ) of rrl based on sdss photometry , if one wanted to select 60% , a 72% non - rrl contamination rate would result . similar levels of contamination make it equally difficult to trace halo structures using photometrically selected bhb stars . in our analysis we also found a significant group of rrl with average @xmath25 magnitudes @xmath174 . by combining these sources with css rrab s , and photometrically - selected hb candidates , we find strong evidence for a tidally disrupted stellar stream that crossing @xmath175 of the sky at galactocentric distances from 70 to 110 kpc . this result confirms the existence of a stream first noted by newberg et al . comparison with halo density models shows that the feature is significant and that the halo density declines rapidly beyond 30 - 50 kpc , as previously noted by watkins et al . ( 2009 ) and sesar et al . however , since these results , like those of watkins et al . ( 2009 ) and sesar et al . ( 2010 ) , are based on a thin slice through the halo , caution has to be taken when interpreting the extent of density regions outside the observed fields . although the gemini rrl overlap with the sgr stream system , we find that the large distances are inconsistent with existing sgr models . this result is in agreement with ruhland et al . ( 2011 ) and a recent sample of rrl discovered by sesar et al . ( 2012a ) . furthermore , we find that the radial velocities of the rrl are inconsistent with simulations of sgr stream . however , we note that models of the sgr streams system as a whole remains poorly constrained by observations . we have investigated the possible relationship between the gemini tidal stream and ngc 2419 as first proposed by newberg et al . the most recent analyses of ngc 2419 shows significant evidence for it being the nuclear remnant of a disrupted dwarf galaxy ( van den berg & mackey 2004 ; mackey & van den berg 2005 ; cohen et al . 2010 , 2011 ; forbes & bridges 2010 ; cohen & kirby 2012 ) . although , we find that the distances of many of the rrls and hb candidates are consistent with ngc 2419 , the available velocities and location of the gemini stream are insufficient agreement to link the two structures . furthermore , the most distant of the rrab s discovered appear to be metal rich sources and would thus be inconsistent with stars observed in ngc 2419 . nevertheless , given the location ngc 2419 within the halo , and with 10 degrees of the sgr streams plane , it seems possible that a stream associated with ngc 2419 could join it to the sgr system . this may in part account for the significant diversity in metallicity observed for varying sgr stellar streams ( law & majewski 2010 ) . furthermore , as the photometric - selected hb candidates and rrab s cover the sgr stream system and exhibit a distance gradient , it is possible that there is a second galaxy remnant associated with the sgr stream . such a source might explain the origin of the two intersecting streams of the sgr system that has now been well delineated with msto stars by belokurov et al . ( 2006 ) and kosopov et al . however , even with 40% uncertainties in the distances to msto stars ( newby et al . 2011 ) , the gemini stream rrab s are twice as distant as expected for msto stream stars ( koposov et al . 2012 ) . alternately , the gemini stream may originate from the remnant of another disrupted dwarf galaxy that lies beyond the gemini stream stars and is yet to be discovered . the gemini stream leads into the galactic plane beyond 100kpc . a highly extincted system in the galactic plane would be very difficult to detect . future photometric and spectroscopic observations of the hb stars within @xmath58 of ngc 2419 could confirm whether there truly is a tidal stream of rrl associated with ngc 2419 . for example , the seemingly unique mg and k abundance patterns in ngc 2419 found by mucciarelli et al . ( 2012 ) would chemically tag stars originating from this source , even in the presence of overlapping tidal streams . if the gemini tidal stream does follow the path expected from the sdss hb candidates , additional deep photometric observations undertaken by projects such as lsst ( abell et al . 2009 ) should reveal numerous additional rrl along this tidal stream . moreover , if these stars are associated with ngc 2419 , we predict that the rrl will mostly be type oosterhoff ii . crts and csdr1 are supported by the u.s . national science foundation under grants ast-0909182 and cns-0540369 . the css survey is funded by the national aeronautics and space administration under grant no . nng05gf22 g issued through the science mission directorate near - earth objects observations program . j. l. p. acknowledges support from nasa through hubble fellowship grant hf-51261.01-a awarded by the stsci , which is operated by aura , inc . for nasa , under contract nas 5 - 26555 . support for m.c . and is provided by the ministry for the economy , development , and tourism s programa inicativa cientfica milenio through grant p07 - 021-f , awarded to the milky way millennium nucleus ; by proyecto basal pfb-06/2007 ; by proyecto fondecyt regular # 1110326 ; and by proyecto anillo act-86 . sdss - iii is managed by the astrophysical research consortium for the participating institutions of the sdss - iii collaboration funding for sdss - iii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , and the u.s . department of energy office of science . the sdss - iii web site is http://www.sdss3.org/.

it has been recognized shortly after the discovery of pulsars that neutron stars have velocities much in excess of those of any other normal stellar populations in our galaxy ( gunn & ostriker 1970 ; minkowski 1970 ) . however , it is only in the last few years that a significant body of evidence has come into place to support the view that type ii supernovae are nonspherical and neutron stars receive large kick velocities at birth . lyne and lorimer ( 1994 ) analyzed pulsar velocities in light of new proper - motion measurements ( harrison et al . 1993 ) and up - to - date pulsar distance scale ( taylor & cordes 1993 ) , and concluded that pulsars were born with a mean speed of @xmath9 km s@xmath7 , much larger than previously thought . more recent studies of pulsar velocities have adopted more sophisticated satistical methods and included better treatment of selection effects and uncertainties ( lorimer et al . 1997 ; hansen & phinney 1997 ; cordes & chernoff 1997 ) . they all yielded a result for the pulsar birth velocity in qualitative agreement ( although not in quantitative details ) with that of lyne & lorimer ( 1994 ) . in particular , cordes and chernoff ( 1997 ) found that the three - dimensional space velocities of their sample of 47 pulsars have a bimodal distribution , with characteristic speeds of @xmath10 km s@xmath7 and @xmath11 km s@xmath7 ( corresponding to @xmath12 and @xmath13 of the population ) . they also estimated that pulsars with velocities greater than @xmath14 km s@xmath7 may be underrepresented owing to selection effects in pulsar surveys [ the uncertainty in the high - velocity end of the pulsar velocity distribution function has also be emphasized by hansen & phinney ( 1997 ) ] . concrete evidence for the existence of pulsars with velocities of @xmath15 km s@xmath7 has come from the observation of the guitar nebula pulsar ( b2224 + 65 ) , which produces a bow shock when plowing through the interstellar medium ( cordes , romani & lundgren 1993 ) . in addition , studies of pulsar - supernova remnant associations have uncovered a number of pulsars having velocities much greater than @xmath14 km s@xmath7 ( frail et al . 1994 ) , although in many cases the associations are not completely secure . compelling evidence for supernova asymmetry and pulsar kicks also comes from the detection of geodetic precession in the binary pulsar psr 1913 + 16 ( cordes et al . 1990 ; arzoumanian et al . 1996 ) , and the orbital plane precession in the psr j0045 - 7319/b star binary ( lai et al . 1995 ; kaspi et al . 1996 ) and its fast orbital decay ( which indicates retrograde rotation of the b star with respect to the orbit ; see lai 1996 ) . these results demonstrate that binary break - up ( as originally suggested by gott , gunn & ostriker 1970 ; see iben & tutukov 1996 ) can not be _ solely _ responsible for the observed pulsar velocities , and that _ natal kicks are required_. in addition , evolutionary studies of neutron star binary population imply the existence of pulsar kicks ( e.g. , deway & cordes 1987 ; fryer & kalogera 1997 ; fryer et al . 1998 ; see also brandt & podsiadlowski 1995 ) . finally , there are a large number of direct observations of nearby supernovae ( e.g. , cropper et al . 1988 ; trammell et al . 1993 ; mccray 1993 ; utrobin et al . 1995 ) and supernova remnants ( e.g. , morse , winkler & kirshner 1995 ; aschenbach et al . 1995 ) in radio , optical , and x - ray bands which support the notion that supernova explosions are not spherically symmetric . the origin of the pulsar velocities is still unknown . two classes of mechanisms for the _ natal kicks _ have been suggested . the first class relies on convective instabilities in the collapsed stellar core and within the rebounding shock ( e.g. , burrows & fryxell 1992 ; burrows et al . 1995 ; janka & mller 1994 , 1996 ; herant et al . the asymmetries in the matter and temperature distributions associated with the instabilities naturally lead to asymmetric matter ejection and/or asymmetric neutrino emission . numerical simulations indicate that the local , post - collapse instabilities are not adequate to account for kick velocities higher than a few hundred km s@xmath7 ( janka & mller 1994 ) . a variant of this class of models therefore relies on the global asymmetric perturbations seeded in the presupernova cores ( goldreich et al . 1996 ; see also bazan & arnett 1994 ) . clearly , the magnitude of kick velocity depends on the degree of initial asymmetry in the imploding core ( burrows & hayes 1996 ) . due to various uncertainties in the presupernova stellar models , it is not clear at this point whether sufficiently large initial perturbations can be produced in the precollapse core ( lai & goldreich 1998 ) . in this paper we focus on the second class of models in which the large pulsar kick velocities arise from asymmetric neutrino emission induced by strong magnetic fields . since @xmath16 of the neutron star binding energy ( a few times @xmath17 erg ) is released in neutrinos , tapping the neutrino energy would appear to be an efficient means to kick the newly - formed neutron star . magnetic fields are naturally invoked to break the spherical symmetry in neutrino emission . but the actual mechanism is unclear . we first briefly comment on previous works on this and related subjects . a number of authors have noted that parity violation in weak interactions may lead to asymmetric neutrino emission from proto - neutron stars ( chugai 1984 ; dorofeev et al . 1985 ; vilenkin 1995 ; horowitz & piekarewicz 1997 ) . however , their studies are largely unsatisfactory for a number of reasons : they either failed to identify the most relevant neutrino emission / interaction processes or the relevant physical conditions in proto - neutron stars , or stopped at estimating the magnetic field effects on neutrino opacities . chugai ( 1984 ) and vilenkin ( 1995 ) ( see also bezchastnov & haensel 1996 ) considered neutrino - electron scattering and concluded that the effect is extremely small ( e.g. , to obtain @xmath18 km s@xmath7 would require a magnetic field of at least @xmath8 g ) . however , neutrino - electron scattering is less important than neutrino - nucleon scattering in determining the characteristics of neutrino transport in proto - neutron stars . similarly , dorofeev et al . ( 1985 ) considered neutrino emission by urca processes in strong magnetic fields . but as shown by lai & qian ( 1998 ) , in the bulk interior of the neutron star , the asymmetry associated with neutrino emission is cancelled by the asymmetry associated with neutrino absorption . concerning the parity violation effect in proto - neutron stars , horowitz & li ( 1997 ) recently suggested that the asymmetry in neutrino emission may be enhanced due to multiple scatterings of neutrinos by nucleons which are slightly polarized by the magnetic field . they estimated that a field strength of a few times @xmath19 g is adequate to account for kick velocities of a few hundred km s@xmath7 . however , the paper by horowitz & li ( 1997 ) only discusses the idealized situations of scattering media , and ignores various essential physics that is needed in a proper treatment of neutrino transport in proto - neutron stars . in particular , it does not consider the effect of neutrino absorption which one might suspect to wash out the cumulative effect from multiple scatterings . our own study of the parity violation effect ( lai & qian 1998 ) was flawed in the treatment of scattering terms in the neutrino transport equation . as shown by arras & lai ( 1998 ) , detailed balance requires that there be no cumulative effect from multiple scatterings in the bulk interior of the proto - neutron star where local thermodynamic equilibrium applies to a good approximation . enhancement of neutrino emission asymmetry from multiple scatterings obtains only after neutrinos thermally decouple from proto - neutron star matter , and therefore is insignificant . bisnovatyi - kogan ( 1993 ) attributed pulsar kicks to asymmetric magnetic field distribution in proto - neutron stars . using the fact that neutron decay rate can be modified by the magnetic field , he inferred that neutrino emissions from opposite sides of the neutron star surface are different . however , neutron decay is not directly relevant for neutrino emission from a newly - formed neutron star . roulet ( 1997 ) ( whose paper was posted while this paper was being written ) considered the relevant neutrino absorption processes . but his study was restricted to calculating the neutrino cross sections as functions of neutrino energy ( corresponding to 2.1 of this paper ) , and therefore was insufficient for addressing the issue of asymmetric neutrino emission . finally , it is worthwhile to mention a speculative idea on pulsar kicks which relies on nonstandard neutrino physics . kusenko & segr ( 1996 ) suggested that asymmetric @xmath20 emission could result from the mikheyev - smirnov - wolfenstein flavor transformation between @xmath20 and @xmath0 inside a magnetized proto - neutron star because a magnetic field changes the resonance condition for the flavor transformation . with a @xmath20 mass of @xmath21 ev , they claimed that magnetic fields of @xmath22 g can give the pulsar a kick velocity of a few hundred km s@xmath7 . however , their treatment of neutrino transport was oversimplified : e.g. , they ignored that neutrinos of different energies have different resonance surfaces . furthermore , a simple geometric effect and realistic proto - neutron star conditions easily reduce their estimated kick velocity by an order of magnitude . most likely , a magnetic field strength of @xmath23 g is needed to produce the observed average pulsar kick velocity via their mechanism ( see qian 1997 ) . similarly strong magnetic fields are required in variants of the kusenko & segr mechanism , e.g. , those considered by akhmedov et al . ( 1997 ) ( which relies on both the neutrino mass and the neutrino magnetic moment to facilitate the flavor transformation ) . as should be clear from the brief review in 1.1 , in spite of quite a number of suggestions , there is at present no consensus on the magnitudes of the magnetic field induced asymmetry in neutrino emission from proto - neutron stars and the resulting kick velocities . in this paper we study the effect of asymmetric magnetic field topology on pulsar kicks . since the energy levels of @xmath2 and @xmath3 in a magnetic field are quantized , the @xmath0 and @xmath5 absorption opacities near the neutrinosphere and the neutrino - matter decoupling region depend on the local magnetic field strength . if the _ magnitude _ of magnetic field in the north pole is different from that in the south pole ( the field does not need to be ordered ) , then asymmetric neutrino flux may be generated . here we ignore the effects of magnetic fields on the proto - neutron star structure through the equation of state these are secondary effects as far as neutrino transport is concerned . the organization of this paper is as follows . in 2 we calculate the @xmath0 and @xmath1 absorption cross sections as functions of neutrino energy in strong magnetic fields . section 3 summarizes the basic features of neutrino transport near the neutron star surface . in 4 we derive and evaluate the `` rosseland mean '' neutrino opacities which are directly related to the local neutrino flux . we then consider in 5 the change in neutrino flux due to the modified opacities , and estimate the kick velocity resulting from an asymmetric magnetic field topology . in 6 we briefly consider how the emergent neutrino spectra may be modified by the strong magnetic field . in 7 we present our conclusion . unless noted otherwise , we shall use units in which @xmath24 , @xmath25 , and the boltzmann constant @xmath26 are unity . in this section we consider how magnetic fields modify the ( total ) neutrino absorption ( @xmath27 and @xmath28 ) cross sections for specific neutrino energies . we present our calculation of the @xmath0 absorption cross section in detail . the @xmath1 absorption cross section can be obtained in a similar manner . to begin with , it is useful to review various energy scales involved in the problem . near the neutrinosphere and the neutrino - matter decoupling region , the matter density is typically @xmath29@xmath19 g@xmath30@xmath31 , and the temperature is about several mev . the electrons are extremely relativistic and highly degenerate , with fermi energy @xmath32mev , where @xmath33 is the electron fraction ( the number of electrons per nucleon ) and @xmath34 . the neutrons and protons have fermi energies @xmath35^{2/3}~$]mev and @xmath36mev , respectively , and thus they are essentially nondegenerate . in strong magnetic fields the transverse motion of the electron is quantized into landau levels . the electron energy dispersion relation is e_e=^1/2 , n=0 , 1 , 2 , where @xmath37 is the longitudinal momentum ( along the magnetic field ) , and @xmath38 is the landau level index . the dimensionless field strength @xmath39 is defined as b , with b_c = m_e^2c^3e = 4.41410 ^ 13 g. the quantization effect on the proton is extremely small due to the large proton mass , and is neglected throughout this paper . since the cyclotron radius ( characterizing the size of the landau wavefunction ) @xmath40^{1/2}= [ \hbar/(m_e c)]b^{-1/2}=386\,b^{-1/2}$ ] fm is much greater than the range of the weak interaction [ @xmath41 fm , where @xmath42 gev is the mass of w boson ] , one can use the @xmath43 theory with contact interaction . locally ( at the point of interaction ) the landau wavefunction resembles a plane wave . thus we expect that the spin - averaged matrix element is the same as in the case without magnetic fields . this expectation has been borne out by detailed calculations ( e.g. , fassio - canuto 1969 ; matese & oconnell 1969 ) . therefore we only need to consider the effect of magnetic field on the electron phase space . since the neutrino energy ( a few to tens of mev ) of interest is much less than the nucleon mass , we first consider the case where the nucleon mass is taken to be infinity , an approximation adopted in the standard zero - field calculations ( e.g. , tubbs & schramm 1975 ) . the cross section for @xmath44 can be written as a sum over the ( spatial and spin ) states of the electron ( per unit volume ) ^_b(e_)=a_e(1-f_e)(e_+q - e_e ) , where we have defined ag_f^2 ^2_c ( c_v^2 + 3c_a^2 ) , and the other symbols have their usual meanings : @xmath45 is the universal fermi constant , @xmath46 refers to the cabbibo angle , @xmath47 and @xmath48 are the coupling constants , @xmath49 mev is the neutron - proton mass difference , @xmath50 is the @xmath0 energy , and @xmath51 is the pauli blocking factor . the electron occupation number @xmath52 is f_e=1 + 1 , where @xmath53 is the electron energy , @xmath54 is the electron chemical potential , and @xmath55 is the matter temperature . for zero field , the sum becomes @xmath56 , and we have ^_b=0(e_)=a^2(e_+q ) ^1/2(1-f_e ) , with @xmath57 . for @xmath58 , knowing that the degeneracy of the landau level ( per unit area ) is @xmath59 , with @xmath60 for @xmath61 and @xmath62 for @xmath63 , we have ^_b(e_)=a _ -^_n g_nm_e^2b2(1-f_e ) ( e_+q - e_e)=am_e^2b2 ^ 2 _ n=0^n_max g_n e_ep_z , n(1-f_e ) , [ absb]where @xmath57 , and @xmath64^{1/2}$ ] . the upper limit @xmath65 for the sum is the maximum value of @xmath38 for which @xmath66 is meaningfully defined , i.e. , n_max = int , [ nmax]where int@xmath67 $ ] stands for the integral part of @xmath68 . henceforth , @xmath65 is similarly defined , and we shall not explicitly write out the lower and upper limits for the sum . figure 1 depicts @xmath69 for @xmath70 and @xmath71 as functions of @xmath50 for typical conditions ( @xmath72 , @xmath73 mev , and @xmath74 mev ) in a proto - neutron star . note that ideally , one should compare @xmath69 and @xmath71 at the same matter density . for @xmath75 ( so that @xmath76 ) , the matter density is related to @xmath54 via @xmath77 in practice , we have found that for a given @xmath54 , the dependence of @xmath78 on the field strength is rather weak for the range of parameters of interest in this paper . thus in all of our calculations here and below , we compare the cross sections at a given @xmath54 . a prominent feature of @xmath69 shown in fig . 1 is the presence of `` resonances , '' which occur when @xmath79 ( cf . eq . [ [ absb ] ] ) , i.e. , at the neutrino energies for which the final state electron lies in a `` stationary '' landau level ( with zero longitudinal momentum ) . these resonances reflect the divergence of the density of states of electrons at these stationary landau levels . the cross section for @xmath80 can be similarly calculated . one only needs to replace @xmath81 with @xmath82 and @xmath54 with @xmath83 in the expressions for @xmath0 absorption to obtain the desired expressions for @xmath1 absorption . the numerical results for @xmath1 absorption cross section are shown in fig . 2 for the same physical parameters as used in fig . 1 . the divergent cross sections at the `` resonances '' shown in figs . 1 and 2 are clearly unphysical . they are smoothed out by the thermal motion of the neutron . to the leading order , the width of the resonance is @xmath84 , corresponding to the doppler shift of the neutrino energy in the neutron s rest frame . the recoil energy , of order @xmath85 , is much less effective in smoothing out the resonances , and will be neglected . with this approximation , the smoothed cross section can be written as ^_b(e_)v_n^2 dv_nf_nd_n _ b^(e_ ) , [ aver]where @xmath86 and @xmath87 is the velocity distribution function for the neutron . if we replace the factor @xmath51 [ which depends on @xmath88 in @xmath89 with the average value @xmath90 , which can be approximately taken as @xmath51 evaluated at @xmath57 , then the angular integral in eq . ( [ aver ] ) can be carried out analytically . for @xmath91 , we have ^_b=0(e_)1-f_ev_n dv_nf_n13 ^ 2e _ ( e_e^2-m_e^2)^3/2|^e_e = e_(1+v_n)+q _ e_e = e_(1-v_n)+q , [ abs0aver]which is reduced to ^_b=0(e_)(e_+q)^21-f_e[abs0aver0]if we neglect the electron mass @xmath92 . for @xmath58 , we find has an integrable singularity at @xmath93 . ] ^_b(e_)1-f_ev_n dv_nf_nm_e^2b2 ^ 2e _ _ n g_n^1/2|^e_e = e_(1+v_n)+q _ e_e = e_(1-v_n)+q . [ absaver]we could have approximately included the neutron recoil effect by replacing @xmath81 with @xmath94 in eqs . ( [ abs0aver])([absaver ] ) . but we choose to leave out this effect since as mentioned previously , it is ineffective in smoothing out the resonances in @xmath69 . we have numerically evaluated eqs . ( [ abs0aver ] ) and ( [ absaver ] ) , and the results are depicted in fig . 1 together with the cross sections before averaging . similar results for the @xmath1 absorption cross section are shown in fig . as expected , we find that for @xmath91 , the averaged cross section @xmath95 is almost identical to @xmath71 ( see eq . [ [ abs0aver0 ] ] ) . however , for @xmath58 , the thermal averaging is essential in obtaining the smoothed physical absorption cross sections . the oscillatory behaviors of @xmath96 in figs . 1 and 2 can be demonstrated analytically . using the poisson summation formula ( see e.g. , ziman 1979 ) : _ n=0^f(n+12)= _ s=-^(-1)^s_0^f(x)e^2i s xdx , we can rewrite the sum in eq . ( [ absaver ] ) as @xmath97^{1/2}\nonumber\\ & & = { 2\over 3}(e_e^2-m_e^2)^{3/2 } + ( 2m_e^2b)^{3/2}{\rm re}\sum_{s=1}^\infty { 1-i\over 4\pi i s^{3/2}}\exp\left[2\pi i s \left({e_e^2-m_e^2\over 2m_e^2b } \right)\right ] . \label{summ}\end{aligned}\ ] ] the first term on the right - hand side of eq . ( [ summ ] ) corresponds to the zero - field result , and the second term ( involving the sum of many oscillatory terms ) corresponds to the correction due to the quantization effect . the sum is clearly convergent , and thus for small @xmath98 , the averaged cross section @xmath96 is reduced to the zero - field value @xmath95 . for sufficiently large @xmath98 at a given @xmath50 , or , equivalently , for sufficiently small @xmath50 at a given @xmath98 , only the ground state landau level is filled by the electron , i.e. , @xmath99 ( see eq . [ [ nmax ] ] ) . in this limit , @xmath100 , and only the first term in eq . ( [ absb ] ) needs to be retained . we then have ^_b(e_)am_e^2b2 ^ 2 ( e_+q)^1/2(1-f_e ) [ sigbapp]for @xmath0 absorption . a similar expression for @xmath1 absorption is obtained by replacing @xmath81 with @xmath82 and @xmath54 with @xmath83 in eq . ( [ sigbapp ] ) . the physical conditions which figs . 1 and 2 are based on ( @xmath74 mev , @xmath72 , and @xmath73 mev , corresponding to a matter density of @xmath101 g @xmath31 ) are typical of the neutrinosphere of the proto - neutron star . in fig . 3 we plot the @xmath0 and @xmath1 absorption cross sections for a number of slightly different physical parameters , specified by @xmath102 and @xmath55 ( all with @xmath72 ) . the cross section for elastic neutral - current scattering @xmath103 , @xmath104 ( e.g. , tubbs & schramm 1975 ) , is also shown for comparison . absorption dominates the transport of @xmath0 for high energies , with the ratio of absorption to scattering cross section per nucleon @xmath105 approaching 5 for @xmath106 . for smaller energies ( @xmath107 ) , absorption is severely suppressed by the pauli blocking of the final electron states , and scattering becomes more important . in the case of @xmath1 , pauli blocking is negligible , and scattering almost always dominates the transport , with the ratio of scattering to absorption cross section per nucleon @xmath108\simeq 2\,(0.1/y_e)$ ] for @xmath109 mev . as we can see from figs . 1 - 3 , for a given set of physical parameters of the medium ( @xmath110 or @xmath78 ) , the magnetic field effect on the absorption opacity is more prominent for lower - energy neutrinos than for higher - energy ones . this is because for a given field strength , as the neutrino energy increases , more and more landau levels become available for the electron in the final state ( see eq . [ [ nmax ] ] ) . consequently , the electron phase space and hence the cross section should approach the zero - field values according to the correspondence principle . we found that typically when the number of landau levels filled by the electron ( @xmath65 ) is greater than 23 , the magnetic field effect on the absorption opacity becomes negligible . the `` critical '' neutrino energy is approximately given by e_^(crit)m_e(2bn_max+1)^1/2 6(n_max3)^1/2 ( b10 ^ 15g)^1/2 mev . for @xmath111 , one can approximate @xmath96 by the zero - field value @xmath95 with good accuracy ( within a few percent ) . alternatively , we can define a threshold magnetic field , @xmath112 , above which the @xmath0 and @xmath1 absorption opacities are modified ( by more than a few percent ) by the magnetic field : b_thb_c 310 ^ 15(e_10 mev)^2 ( 3n_max ) g , where @xmath113 mev is the average neutrino energy . in the deeper layers of the neutron star , the typical neutrino energy is higher . therefore we expect the magnetic field effect on the absorption opacities to be smaller there . in addition , for @xmath0 , absorption becomes unimportant as compared to scattering in the high density region ( although not necessarily in the deepest interior , where @xmath0 can be degenerate ) . thus , in order to assess the importance of the magnetic field effect on the neutrino emission , we only need to focus on the surface region , the region between the neutrinosphere and the neutrino - matter decoupling layer . we note that because the absorption opacities in the magnetic field for different neutrino energies oscillate around the the zero - field values , without a detailed calculation of the mean opacity ( appropriately averaged over the neutrino energy spectrum ) one can not even answer the qualitative question such as whether the magnetic field increases or decreases the local neutrino energy flux . in 36 , we study how the modified absorption cross sections affect the emergent @xmath0 and @xmath1 energy fluxes and spectra from the proto - neutron star . as discussed in 2.4 , the quantization effect of magnetic fields on neutrino absorption is important in the region only near the neutron star surface ( between the neutrinosphere and the neutrino - matter decoupling layer ) . in this section we consider neutrino transport in this relatively low density region . because @xmath0 and @xmath1 with typical energies of 10 mev have comparable absorption and scattering opacities ( see fig . 3 ) , the emergent @xmath0 and @xmath1 energy fluxes and spectra are determined essentially in the same region where these two neutrino species just begin to thermally decouple from the proto - neutron star matter . the @xmath0 and @xmath1 occupation numbers in this region are then approximately given by @xmath114 + 1 } , \label{occnue}\ ] ] and @xmath115 + 1 } , \label{occnueb}\ ] ] respectively , where @xmath116 , and @xmath117 is the @xmath0 chemical potential . from these occupation numbers , we obtain @xmath118 and @xmath119 where for example , @xmath120 and @xmath121 are the number and energy densities , respectively , for the @xmath0 . we can estimate the magnitude of @xmath122 as follows . the net deleptonization rate of the proto - neutron star is @xmath123 where @xmath124 is the initial electron fraction of the proto - neutron star , @xmath125 is the proto - neutron star mass , @xmath126 is the nucleon mass , @xmath127 is the neutrino diffusion timescale , and @xmath128 is the proto - neutron star radius . the sum of the @xmath0 and @xmath1 luminosities is @xmath129 from eqs . ( [ nuno])([nulum ] ) , we have @xmath130 where the numerical value is obtained for @xmath131 and @xmath132 km . therefore , to a good approximation , the @xmath0 and @xmath1 in the decoupling region have fermi - dirac energy distributions with zero chemical potentials . let @xmath133 be the specific intensity of neutrinos of a given species , where @xmath134 is the spatial position , and @xmath135 specifies the direction of propagation . the general transport equation takes the form i_e=_e^i_e^(fd ) + ^_e 14d i_e - i_e , [ inu]where @xmath136 , and the subscript `` @xmath137 '' indicates that the relevant physical quantities depend on the neutrino energy @xmath137 . the opacities @xmath138 and @xmath139 are related to the corresponding cross sections as , e.g. , @xmath140 and @xmath141 for @xmath0 . note that we have included the effect of stimulated absorption for neutrinos ( e.g. , imshennik & nadezhin 1973 ) by introducing the corrected absorption opacity @xmath142 $ ] ( @xmath55 is the matter temperature ) . hereafter , we shall suppress the superscript `` @xmath143 '' in @xmath138 , and the notation @xmath144 should be understood as having included the correction due to stimulated absorption . the quantity @xmath145 in eq . ( [ inu ] ) is the equilibrium neutrino intensity , given by the fermi - dirac distribution i_e^(fd)=e^3c^2h^31(e / t)+1 . note that the time derivative term @xmath146 on the left - hand - side of eq . ( [ inu ] ) has been neglected . this corresponds to an instantaneous redistribution of emission and absorption sources ( i.e. , an instantaneous redistribution of matter temperature ) the timescale for the redistribution is of order the mean free path divided by @xmath25 , much smaller than the neutrino cooling time of the star . the zeroth order moment of the transport equation can be written as _ e=_e^ c(u_e^(fd)-u_e ) , [ dfnu]where @xmath147 is the ( spectral ) neutrino energy flux , @xmath148 is the energy density , and @xmath149 is the corresponding equilibrium energy density . with the eddington closure relation , the first moment of eq . ( [ inu ] ) becomes _ e =- c3_e^(t)u_e , [ fnu]where @xmath150 . equations ( [ dfnu ] ) and ( [ fnu ] ) completely specify the neutrino radiation field . here we are concerned with the total neutrino energy flux emitted from a local surface region of the neutron star . the magnetic field strength in this local region is assumed to be constant . since the scale heights for physical quantities such as the matter density and temperature are much smaller than the local radii , we adopt the plane - parallel geometry . integrating eq . ( [ fnu ] ) over the neutrino energy , we find that the total neutrino energy flux ( of a specific species ) @xmath151 is given by f =- c3_rdudr , [ ftot]where @xmath152 is the total neutrino energy density , and @xmath153 is the `` rosseland mean opacity '' defined as = de(1/_e^(t))(d u_e / dr)d e(du_e / dr ) . [ ross]since neutrino emission and absorption rates are extremely fast even in the regions near the proto - neutron star surface , a radiative equilibrium , @xmath154 , @xmath20 , and @xmath155 is much weaker , the overall radiative equilibrium in the region near the proto - neutron star surface mainly involves matter and @xmath0 and @xmath1 . since the coupling between matter and @xmath0 is comparable to that between matter and @xmath1 , we further assume that an approximate radiative equilibrium holds separately for @xmath0 and @xmath1 . ] approximately holds , i.e. , _ e^ ( u_e^(fd)-u_e)de0 . > from eq . ( [ dfnu ] ) , we have @xmath156 , or @xmath157 is approximately constant . if we introduce the rosseland mean optical depth measured from the stellar surface via @xmath158 , with @xmath157 constant , eq . ( [ ftot ] ) can be integrated to yield cu=3f(_r+23 ) , [ profile]where we have used the boundary condition @xmath159 at the surface ( @xmath160 ) . equation ( [ profile ] ) resembles the standard eddington profile . note that outside the decoupling region the neutrino energy density u may not be equal to @xmath161 specified by the matter temperature @xmath55 . similarly , @xmath162 may differ substantially from @xmath163 ( see 6 ) . however , we can still approximate @xmath162 by the fermi - dirac function at the neutrino temperature @xmath164 , which is equal to the matter temperature at the decoupling layer . equation ( [ ross ] ) can then be rewritten as = 307 ^ 4t_^5de1_e^(t ) e^4 ( e / t_)^2 . [ rosst]with the cross sections given in 2 , we can calculate the rosseland mean opacity @xmath165 in magnetic fields using eq . ( [ rosst ] ) . note that in eq . ( [ rosst ] ) @xmath166 depends on the local matter temperature @xmath55 . only in the regions not far from the decoupling layer , do we have @xmath167 . figure 4 shows the ratio of the rosseland mean opacity @xmath165 of @xmath0 to the zero - field value @xmath168 as a function of the magnetic field strength for typical conditions near the proto - neutron star surface . we choose these conditions to be @xmath169 mev , and @xmath170 mev , all with @xmath72 . figure 5 shows similar results for @xmath1 with similar physical parameters ( @xmath171 mev ; note that the results are insensitive to @xmath54 ) . we see that deviation of @xmath165 from the zero - field value by more than @xmath172 generally requires field strength @xmath173 . the behavior of the ratio @xmath174 as a function of @xmath39 is certainly not obvious a priori . since the rosseland mean integral is dominated by neutrinos with the smallest opacities , and since @xmath175 scales linearly with @xmath98 in the regime where only the ground landau level is filled ( see eq . [ [ sigbapp ] ] ) , we expect that for sufficiently large @xmath98 , the rosseland mean opacity increases linearly with increasing @xmath98 . this is indeed what is indicated in figs . 4 and 5 . for `` intermediate '' field strength , the @xmath175 curve for @xmath0 ( see fig . 1 ) has a significant portion which is below the zero - field value at low neutrino energies . for such intermediate fields , we would expect @xmath165 to be less than the zero - field value . this explains the non - monotonic behavior of @xmath174 as a function of @xmath98 shown in fig . 4 . for @xmath1 , the non - monotonic behavior is much less pronounced ( see the curve for @xmath176 mev in fig . 5 ) since the energy range in which @xmath177 is smaller . imagine a proto - neutron star with an asymmetric magnetic field topology , e.g. , the north pole and the south pole have different _ magnitudes _ of magnetic field ( say , by a factor of a few ) . what is the difference in the neutrino fluxes from these two regions due to the different absorption opacities of @xmath0 and @xmath1 , and what is the resulting kick velocity of the neutron star ? we recall that the fractional asymmetry @xmath178 in the total neutrino emission ( of all species ) required to generate a kick velocity @xmath179 is @xmath180 \,[e_{\rm tot}/(3\times 10^{53}~{\rm erg})]$ ] , where @xmath181 is the total neutrino energy radiated from the neutron star ( the neutron star mass has been set to @xmath182 ) . to quantitatively understand how the magnetically - modified opacities change the local neutrino flux , we will have to evaluate these opacities at different depths ( corresponding to different physical conditions ) below the neutron star surface , and to perform complete cooling calculations which take into account these opacities . however , figs . 4 and 5 clearly indicate that , even near the stellar surface , where one expects the modification to the opacities to be most prominent , and even with the most favorable conditions for the magnetic field to be effective in changing the opacities ( i.e. , low density and low temperature ) , a @xmath172 change in the rosseland mean opacity ( for @xmath0 or @xmath1 ) requires a magnetic field with @xmath183 , or @xmath184 g. in the following , we shall content ourselves with merely estimating the change in the local neutrino flux from the modified opacities . we first consider the case in which the matter temperature throughout the neutron star is unaffected by the magnetic field . one might imagine this as a possibility if convection dominates energy transfer between different parts of the star . under the neutrino - matter decoupling layer the specific neutrino energy density @xmath162 is the same as the fermi - dirac function @xmath163 specified by the matter temperature . the emergent neutrino flux is then inversely proportional to the rosseland mean optical depth @xmath185 from the decoupling layer to the stellar surface ( see eq . [ [ profile ] ] ) . since @xmath185 for @xmath0 or @xmath1 at the decoupling layer is of order a few , the flux @xmath157 is inversely proportional to @xmath153 evaluated near the decoupling region . if we use the curve for @xmath74 mev and @xmath186 mev in fig . 4 ( which has the most pronounced magnetic correction ) , we find that the fractional change in the @xmath0 flux is @xmath187 at @xmath70 ( if we used the curve for @xmath188 or 20 mev , which more closely represents the physical condition at the decoupling region , the modification @xmath189 would be smaller by a factor of a few or more ) . similarly , using the curve for @xmath74 mev in fig . 5 to maximize the magnetic correction , we find @xmath190 at @xmath70 . since the fluxes of other types of neutrinos are unchanged and @xmath191 depend on the @xmath192-@xmath193 scattering cross sections , which can be affected by the magnetic field . however , the layers are located much deeper than those of @xmath0 and @xmath5 ( since @xmath192-@xmath193 scattering opacity is small ) , the modification to the @xmath192-@xmath193 scattering opacity due to electron landau levels is therefore smaller than the modification to the absorption opacities of @xmath4 and @xmath5 considered in this paper . ] , and since different neutrino species more or less carry away the same amount of energy , the change in the total neutrino flux is only @xmath194 at @xmath70 . if we now imagine that the north pole of the neutron star has @xmath195 while the south pole has a field strength a factor of a few smaller , then , taking into account the geometric reduction ( @xmath196 ) , the flux asymmetry is about @xmath197 , just enough to give a kick velocity of @xmath21 km s@xmath7 to the neutron star . with @xmath153 evaluated at more realistic conditions , we expect the resulting @xmath178 to be even smaller . the assumption that the matter temperature is unaffected by the magnetic field ( and therefore spherically symmetric ) is unlikely to be valid . in reality , when a region of the star ( say the north pole ) emits more flux , it also cools faster . with reduced temperature , this region will tend to emit less flux . in the extreme case , one would find @xmath198 , with @xmath199 the neutrino energy density in the stellar core , and @xmath185 the rosseland mean optical depth from the core to the surface . since in the bulk interior of the star , modification to @xmath153 due to the magnetic field is negligible , we would conclude that the change in the neutrino flux is much smaller than what is estimated in the last paragraph . to summarize , the modifications to the absorption opacities of @xmath0 and @xmath5 due to quantized landau levels give rise to rather small change in neutrino fluxes from the neutron star . to generate appreciable kick velocities ( @xmath6 km s@xmath7 ) from this effect requires the difference of field strengths at the two opposite poles to be at least @xmath8 g ( and possibly much larger ) . we now consider how the oscillatory behaviors ( see figs . 13 ) of the absorption opacities of @xmath0 and @xmath1 may manifest themselves in the radiated neutrino energy spectra . for an absorption dominated medium , neutrinos with energy @xmath137 decouple from the matter and start free - streaming at a column depth ( in g @xmath200 ) @xmath201 . thus the neutrino energy spectrum is given by @xmath202 with @xmath203 evaluated at the depth @xmath204 . in the presence of significant scattering , decoupling occurs at depth @xmath205 with the effective opacity @xmath206 , whereas the neutrinosphere from which neutrinos start free - streaming is located at @xmath207 . near the decoupling region we have @xmath208 , and thus from eq . ( [ fnu ] ) the neutrino spectral energy flux is given by f_e~_e^eff_e^(t ) cu_e^(fd)(y~1/_e^eff ) . the above qualitative argument can be made more rigorous by directly solving eqs . ( [ dfnu ] ) and ( [ fnu ] ) . consider a plane - parallel atmosphere and define the ( energy - dependent ) optical depth via @xmath209 . equations ( [ dfnu ] ) and ( [ fnu ] ) can be combined to give u_e=_e^2 [ u_e - u_e^(fd ) ] , with _ e^2=3_e^_e^(t ) . with a constant @xmath210 , the above equation can be solved exactly . for the boundary condition @xmath211 at @xmath212 , the neutrino energy flux at the surface is given by f_e=_e/31 + 2_e/3_0^d_e_ecu_e^(fd)(_e)(-_e_e ) cu_e^(fd)(_e_e^-1 ) . [ fnu2 ] figure 6 shows the spectral energy fluxes of @xmath0 calculated using eq . ( [ fnu2 ] ) for different field strengths ( @xmath70 and @xmath213 ) . the matter temperature profile is obtained from eq . ( [ profile ] ) assuming @xmath214 . for a total energy flux of @xmath215 ( where @xmath216 is the effective temperature ) , we have @xmath217^{1/4}$ ] . we choose @xmath218 mev in these examples . for simplicity , we have neglected the variations of the opacities as a function of @xmath219 . thus the temperature at decoupling ( @xmath220 ; see eq . [ [ fnu2 ] ] ) is given by @xmath221^{1/4}$ ] , where @xmath222 and we take the relevant opacities to have the corresponding values for @xmath72 , @xmath73 mev , and @xmath74 mev . this simplification should not introduce qualitative changes in the spectrum . as seen from fig . 6 , the spectrum is pinched at the high energies because of the @xmath223-dependence of the opacities . the magnetic field introduces modulations in the spectrum . for lower magnetic field strengths , the modulations occur at the lower energy band . as the magnetic field gets stronger , the modulations shift to the higher energy part of the spectrum . the distortions and modulations of the @xmath0 and @xmath1 energy spectra by strong magnetic fields may offer an interesting possibility to limit the magnetic field strength in proto - neutron stars through the detection of neutrinos from a galactic supernova . we note , however , that the @xmath0 spectra shown in fig . 6 are based on very approximate calculations . they serve only as illustrative examples of the magnetic field effect on the neutrino spectra . more detailed calculations are needed to obtain realistic spectra which include the magnetic field effect . in this paper , we have studied the issue of whether asymmetric magnetic field topology in a proto - neutron star can induce asymmetric neutrino emission from the star through the modifications of the neutrino absorption opacities by the magnetic field . these modifications arise from the quantized landau levels of electrons and positrons produced in strong magnetic fields . by calculating the appropriate mean neutrino opacities in the magnetic field , we demonstrate that this mechanism is rather inefficient in generating a kick to the neutron star : to obtain appreciable kick velocities ( @xmath6 km s@xmath7 ) , the difference in the field strengths at the two opposite poles of the star must be at least @xmath8 g. this work was started while d.l . held the richard c. tolman postdoctoral fellowship at caltech . additional support was provided by nasa grant nag 5 - 2756 and nsf grant ast-9417371 , as well as a start - up fund for new faculty at cornell university . is supported by the david w. morrisroe postdoctoral fellowship at caltech .

the past few years have witnessed significant progress in the building of realistic inflation models within string theory @xcite . the most popular approach has used the interaction potential between branes and antibranes as the source of energy driving inflation ; hence the interbrane separation plays the role of the inflaton . early attempts suffered from the assumption that moduli could be fixed by fiat , without specifying the stabilization mechanism . however , explicit examples have shown that properly fixing the moduli can interfere with the flatness which is required for the inflaton potential @xcite ( ) . flux compactifications in type iib string theory have provided a robust way of stabilizing the dilaton and complex structure moduli of the calabi - yau manifold @xcite , with the added benefit of providing warped throats @xcite which can be useful for tuning the scale of inflation or of the standard model . nonperturbative effects have been plausibly invoked to stabilize the remaining khler moduli @xcite . this last step , unfortunately , generically spoils whatever flatness of the inflaton potential which was accomplished by warping , and thus requires additional fine tuning of parameters . one can introduce extra parameters associated with corrections to the superpotential , which must be tuned to a part in 100 in order to obtain the minimum amount of 60 e - foldings of inflation @xcite . alternatively , it is possible to tune parameters already present within the model but at the level of a part in 1000 @xcite . the fine - tuning problem of the model has been considered by several authors @xcite . shift symmetry of the lagrangian has been proposed as a possible solution @xcite . it has also been pointed out that if lyman @xmath1 constraints are viewed with skepticism , the degree of tuning required is considerably relaxed @xcite . of course , this observation also applies to other models of inflation . in this paper we reexamine the fine - tuning issue , starting within the context of the original potential , without introducing additional corrections . a closer consideration of the potential between the brane and antibrane reveals the importance of a parameter ( which we call @xmath0 ) which was set to a certain value in . it is the distance which the brane will travel before annihilating with the antibrane in the warped throat , assuming that the brane starts somewhere closer to the the interior of the calabi - yau at an initial position @xmath2 . the other important parameters governing the shape of the inflaton potential are the warp factor of the throat , which we denote as @xmath3 , and the khler modulus at its stabilized value , @xmath4 . the warp factor is tunable through the ratio of ns - ns and rr fluxes @xmath5 : @xmath6 @xcite . the vev of the khler modulus @xmath4 is determined by the nonperturbative superpotential @xmath7 invoked by ref . @xcite , where @xmath8 in the case of gluino condensation in an @xmath9 gauge theory sector which can naturally arise in these compactifications . if the parameter @xmath0 is set to zero as in , then indeed it is necessary to introduce additional sources of dependence on the inflaton into the potential , in order to cancel the unwanted contribution to the inflaton mass coming from the khler modulus . however , in this work we point out that for a range of values @xmath10 of the combination @xmath11 , it is possible to achieve 60 or more e - foldings of inflation without any additional corrections to the potential . fortunately this restriction on @xmath12 applies only to the simplest model . we will show that the restriction disappears when one considers more general examples in which corrections to the khler and superpotentials are taken into account . another crucial parameter which we consider is the number of mobile mobiles branes , @xmath13 , generalizing the simplest case of only a single brane playing the role of the inflaton . we refer to this possibility as multibrane inflation . within the aforementioned range of @xmath12 values , we observe a very interesting phenomenon , whereby starting with a sufficiently large number @xmath13 dynamically leads to a sufficiently flat inflaton potential . at large @xmath13 , the branes start out being confined in a metastable minimum of the potential . but as successive branes tunnel out of this minimum , it becomes shallower , until finally all the remaining branes can roll together into the throat . this is illustrated in figure [ fig1 ] . the potential is often very flat when the branes start rolling , leading to a long period of inflation . this is a novel , dynamical mechanism for flattening the inflaton potential , which appears to be quite specific to brane - antibrane inflation . before looking at the dynamics of inflation with stacks of branes and antibranes , we consider under what conditions multiple branes are compatible with the stabilization of the extra dimensions , _ i.e. , _ the khler modulus , @xmath4 , in supergravity language . for each extra brane we have to introduce a corresponding antibrane such that the tadpole cancellation condition is statisfied . adding antibranes lifts the local minimum of the potential by an additional amount , given by the sum of their tensions , reduced by the warp factor at their location in the bottom of the klebanov - strassler throat . there is a critical number of antibranes that can be added such that the potential still has a local minimum along the @xmath4 direction and therefore the khler modulus remains stable . the stronger the warping , the larger the number of antibranes that can be added . in our analysis we will add antibranes such that the khler modulus remains stabilized and heavy . thus it does not act like a second inflaton field , since it changes very little during the slow - rolling of the brane - antibrane separation @xmath14 . not all models are suitable for constructing such a multi - brane setup . we will give here two examples , one which works and one which does not . the original kklt @xcite setup does not accommodate a large number of mobile branes ; just one additional antibrane is sufficient to lift the vacuum from anti - desitter ( ads ) to minkowski ( or desitter , ds ) . the barrier that prevents tha khler modulus from rolling has a height roughly equal to the depth of the ads minimum . addind a second brane lifts the potential by an amount that is roughly equal to the height of the barrier , which destroys the local minimum . in figure [ fig1n ] we take the parameters of the potential and the antibrane tension as given by @xcite . one sees that adding a second brane results in a very shallow minimum and three branes destabilizes the khler modulus . the potential of the khler modulus is given by v = aae^-a ( 13aa e^-a + w_0 + a e^-a)+nd^3 with @xmath15 , @xmath16 , @xmath17 @xmath18 . -0.25 cm and @xmath19 antibranes , showing there can be at most 2 antibranes without destabilizing the khler modulus . [ fig1n ] , scaledwidth=45.0% ] the situation is dramatically improved if one uses a racetrack superpotential as in ref . the minimum is already minkowski and adding antibranes in a highly warped throat lifts the vacuum to ds without destabilizing the khler modulus . the potential is : & & v= e^-2(a+b)6 ^ 2(b b e^a+ a a e^b ) + & & + where we choose the same parameters as in @xcite , @xmath20 and @xmath21 . with these values the critical number of branes turns out to be @xmath22 , illustrated in fig . [ fig2n ] , scaledwidth=45.0% ] in the remaining analysis , we will not be concerned with the particular model of khler modulus stabilization , beyond the assumption that it admits the required number of additional branes . it will also be found that a large number is not required , so the racetrack model mentioned here is sufficient to establish the existence of a viable model . our starting point for the dynamics of inflation is the lagrangian for the mobile brane position @xmath14 in the framework . to simplify the analysis , we will assume that all moduli ( notably the khler modulus @xmath4 ) are heavy except for the inflaton , and that the 4d effective cosmological constant vanishes at the end of inflation . the lagrangian can be parametrized as [ lag ] l = 6r^2 ^ 2 - tr^2(1 + ( -_0)^4)^-1 where @xmath23 @xcite , @xmath24 is related to the tension @xmath25 of the 3-brane or antibrane by @xmath26 @xcite , @xmath27 is the warp factor , @xmath14 is the 3-brane position , and @xmath0 is the location of the antibrane . to keep our work self - contained , this expression is derived in appendix a , in just the same way as in except for the omission of two approximations : ( 1 ) we do not assume that @xmath28 , and ( 2 ) we do not taylor - expand the term @xmath29 . let us first comment on point ( 1 ) above . in and certain other investigations of brane - antibrane inflation , it was assumed that inflation takes place sufficiently far from the antibrane to justify the approximation @xmath30 . on the other hand , the @xmath0 parameter was kept in @xcite and @xcite . physically , it may be important to allow for nonzero @xmath0 because of the possibility that the @xmath14-dependence coming from the @xmath31 factor in the potential will tend to stabilize the brane at some point other than the bottom of the warped throat , in the absence of the brane - antibrane interaction . an example where this is particularly clear is that of two symmetrically placed warped throats , each with its own antibrane @xcite . in that case the potential ( translated to our notation ) was taken to be v = tr^2(2 - ( -_0)^4 - ( + _ 0)^4 ) from symmetry considerations it is obvious in this model that there exists a metastable or unstable equilibrium point exactly midway between the two throats , at @xmath32 , which can not coincide with the bottom of either throat . if we simply remove one of the throats , the result corresponds to our starting point ( [ lag ] ) . the presence or absence of the @xmath0 parameter explains differences in the amount of required fine - tuning of the potential found in different investigations . to elucidate this , recall that the @xmath14-dependence in the volume modulus @xmath33 is the origin of the @xmath34 problem pointed out in the mass of the inflaton is of order @xmath35 in planck units , ruining the slow roll condition . it was noticed by @xcite that it is possible to tune parameters of a model with a potential similar to that in ( [ lag ] ) so that the slow - roll conditions are satisfied , even without adding any extra @xmath14-dependent corrections to the superpotential , by varying only the parameters which already appear . however , the degree of tuning needed was more severe than in where corrections to the superpotential were invoked . one way of understanding this is to realize that if @xmath36 and only corrections of order @xmath37 are added , then the potential automatically has @xmath38 , so that one of the slow roll parameters @xmath39 is small without any tuning . then only the @xmath34 parameter ( @xmath40 ) needs to be tuned . in our potential however , both slow - roll parameters must be tuned , for generic parameters in the potential . in @xcite it was found that 60 e - foldings of inflation could be obtained by tuning the warp factor to one part in 2000 , whereas only needed one part in 100 tuning . although this sounds contrary to the spirit of the present paper , it was due to the fact that no systematic study of the effect of varying @xmath0 was done in @xcite ; instead only one fixed value was considered . we note that the behavior of the potential as @xmath41 is classically correct , since by derivation it is of the form @xmath42 where @xmath43 satisfies @xmath44 in the 6d calabi - yau manifold . although the full brane position must be specified by 3 complex @xmath14 coordinates , we are keeping the angular ones fixed and following only the radial coordinate for the brane - antibrane separation . we also comment on our choice of avoiding the second approximation ( point ( 2 ) above ) , by not taylor expanding @xmath45 . we choose not to do so because there is no advantage in making this expansion it only leads to an unphysical divergence in the potential energy as the brane - antibrane distance shrinks . the unexpanded form has good behavior as @xmath46 , including the vanishing of the potential at the minimum . we will be interested below in the generalization to more than one brane rolling . the lagrangian for @xmath13 branes and @xmath13 antibranes is [ lagm ] l = 6r^2_i=1^n_i^2 - ntr^2(1 + _ i=1^n(_i-_0)^4)^-1 where now the volume modulus is r = 2- _ i=1^n_i^2 later we will specialize to the situation where all the branes are coincident , @xmath47 . then the potential takes the form @xmath48 . the optimal potential for achieving slow roll inflation is one where both the first and second derivatives vanish simultaneously at some critical point @xmath49 . although this requires fine tuning , we will discuss a mechanism where the tuning can automatically occur through a dynamical process . to characterize the tuning , we introduce new parameters a & = & n_0 ^ 4 , b^2 = 2_0 ^ 6 , x = _ 0 such that the potential is proportional to [ hatv ] v(x ) = 1(b^2/ a - x^2)^2 ( 1 + a(x-1)^4)^-1 the conditions for the potential to have a flat point , as defined above , are [ crit ] a = ( 6x-1)(x-1)^4 , 5x^2(x-1)^4 = b^2 these equations can be solved analytically , as the solution of a cubic equation , but the expression is not very enlightening . instead we give an approximate description , by first considering @xmath12 as a free parameter , and determining as a function of @xmath12 the critical values @xmath50 and @xmath51 which satisfy ( [ crit ] ) . for consistency , we should also require that @xmath52 ; otherwise the potential blows up at @xmath53 , before the brane reaches the antibrane . the range of values where @xmath54 ( obviously also required for a physically sensible solution ) and @xmath55 turns out to be rather narrow , [ range ] 0.259 < b < 0.286 , and in this region @xmath50 is to a good approximation a linear function of @xmath12 : [ ac ] a_c 3b - 0.77 one can also approximate the critical value @xmath51 by a quadratic form , @xmath56 . we will show later that the narrowness of the interval ( [ range ] ) is an accident , in the sense that physically well - motivated perturbations to the potential can remove this restriction . therefore we will ultimately not regard it as a fine tuning . however the special value @xmath50 for a given value of @xmath12 clearly does represent a fine tuning of the parameters . it is this condition which will be addressed by a dynamical mechanism . namely , by starting with a large enough number @xmath13 of branes , one can initially have @xmath57 ( but still @xmath58 ) . the potential always has a metastable minimum from which branes can tunnel in this case . after each tunneling event , the value of @xmath59 diminishes by a discrete amount . at some critical number of branes , @xmath60 , the curvature changes so that the branes roll instead of being trapped . it becomes a quantitative question , depending on the values of @xmath4 and @xmath27 , whether the potential is sufficiently flat to give enough inflation . we will show that in fact it is almost always flat enough to give at least 60 e - foldings of inflation . to find out how likely it is that adequate inflation will occur , we have explored the parameter space @xmath61 , subject to the aforementioned restrictions and the assumption that @xmath13 starts out being larger than the critical value , where all the branes start to roll . for each value of @xmath62 , the quantities @xmath0 and @xmath60 are determined . the fields start from rest at the value where the potential has a local minimum with @xmath63 branes , on the assumption that the nontunneling branes do not move during the tunneling event . the equations of motion are = -3h_i - v_i , = h for the canonical momentum _ i = 12r^2_i and number of e - foldings @xmath64 , where the hubble parameter is given by @xmath65 . since all the fields move together , we need consider only one equation for the inflatons , which we do numerically . the result is that , for @xmath12 satisfying ( [ range ] ) , there is a high probability of getting enough inflation , for random values of @xmath66 selected from within a reasonable range , @xmath67 , @xmath68 . scanning uniformly over this region , we find that @xmath69 of tries result in more than 55 e - foldings of inflation when @xmath70 , and an even greater percentage occurs for other values of @xmath12 . the results are illustrated for @xmath70 in figure [ fig2 ] , showing the number of e - foldings , @xmath71 ( omitting values where @xmath72 for clarity ) versus @xmath73 and @xmath74 . in figure [ fig2 ] it can be seen that @xmath71 is constant along lines of constant @xmath75 . this can be understood by rewriting the potential in terms of the canonically normalized field @xmath76 : v = tna^3 ( 1 - 16 ^ 2)^-2 ( 1 - 36(n / a)^3(-_0)^4)^-1 where @xmath77 . thus for fixed @xmath12 , the shape of the potential , power spectrum , and duration of inflation depend only upon combination of parameters in @xmath78 , and not on the orthogonal combination @xmath79 . we can therefore restrict our exploration of parameter space to a one - dimensional scan along the direction of @xmath78 . moreover , we can show that the number of branes @xmath13 depends in a very simple way on the value of @xmath78 : @xmath80 . this follows from eliminating @xmath0 from the definitions of @xmath59 and @xmath12 . therefore the critical number of branes is given by [ nc ] n_c = a_c b^-4/3 ( 2)^2/3 ^ 1/3 and for a fixed value of @xmath12 , we can treat the parameters @xmath78 and @xmath60 interchangeably . of course ( [ nc ] ) ca nt be satisfied exactly , because @xmath60 has to be an integer . the accidental discrepancies between the optimal value of @xmath60 and the closest integer below this value will affect how long inflation actually lasts . = 0.5 in figure [ fig3 ] we show the correlation between number of e - foldings and the critical number of branes , scanning uniformly in the interval @xmath81 of parameter space . although the two quantities are generally correlated , there are order - of - magnitude variations in @xmath71 over narrow ranges of @xmath60 . figure [ fig3 ] demonstrates that both features can be matched fairly well by an empirical relationship . let @xmath82 be the discrepancy between the ideal value of @xmath60 and the closest integer below this value . then we find that [ emp ] n_e 10 n_c the factor @xmath83 , is understandable since if @xmath84 , the potential is tuned to be perfectly flat at some position @xmath85 . = 0.5 the factor @xmath86 in ( [ emp ] ) can be partly understood in terms of assisted inflation @xcite . with @xmath13 fields rolling simultaneously , the kinetic term is multiplied by @xmath13 relative to one field . we should rescale the fields by @xmath87 , and the slow - roll parameters by @xmath88 . one therefore expects an enhancement in the length of inflation by some power of @xmath60 . it would of course be satisfying to understand why that power is @xmath89 in the present model . we have not yet succeeded in analytically deriving this result . however it is not surprising that @xmath71 fails to be enhanced by a full factor of @xmath60 as would be the case for ordinary assisted inflation , because the potential contains @xmath90 in the khler modulus . only if the potential was independent of @xmath13 would we be sure that @xmath91 . in that case , transforming to the canonically normalized field would cause @xmath92 and the slow roll parameters would go like @xmath93 . the surprise is that there is any assisted inflation effect at all . the reason for it is that at the critical value of @xmath13 , the potential is flatter than it would generically be for an arbitrary function of @xmath90 . the above results were for @xmath70 . allowing @xmath12 to vary , ones finds that smaller @xmath12 tends to give more inflation with a smaller number of branes . this dependence is shown in figure [ fig4 ] . the main observation , though , is that the great majority of all cases yield more than @xmath94 e - foldings of inflation . in addition to having a long enough period of inflation , we must obtain a spectrum of density perturbations in agreement with observational constraints from the cosmic microwave background . to compute the power spectrum , we use the following expression @xcite which applies to a set of noncanonically normalized fields with lagrangian @xmath95 : p(k ) = v75 ^ 2 m_p^2 n_e^i n_e^j g^ij = n_c v h^275 ^ 2 m_p^2 r6 ^ 2 these quantities are evaluated at horizon crossing of the relevant wave number , when @xmath96 . to normalize the spectrum , we evaluate @xmath97 for modes which crossed the horizon 55 e - foldings before inflation ended , corresponding to a scale of inflation around @xmath98 gev . the consistency of this assumption will be verified . the cobe normalization implies that @xmath99 for these modes . this allows us to normalize the scale of the potential by fixing the value of the parameter @xmath24 which is proportional to the unwarped brane tension . the results are expressed in terms of the scale of inflation relative to the planck scale . in figure [ fig5 ] we plot @xmath100 as a function of the number of branes . ( in the previous section it was explained that , just like @xmath71 , spectral properties depend only on the same combination of @xmath4 and @xmath27 which determine @xmath60 . ) from the figure one sees that the inflationary scale ranges between @xmath101 and @xmath102 , roughly consistent with our assumption of @xmath103 e - foldings of inflation . the string scale , which determines the brane tension , must be in about the same range . if the number of branes is small enough , the inflationary scale can be sufficiently high for the tensor contribution to the cosmic microwave background to be observable in future experiments . = 0.5 the scalar spectral index is given by @xmath104 . we have numerically evaluated it over the relevant range of parameters . like the other quantities , it is most conveniently displayed as a function of the number of branes . the deviation @xmath105 at 55 e - foldings before the end of inflation is shown as a function of @xmath60 in figure [ fig6 ] . there is a wide range of possible values , @xmath106 . = 0.5 it is also interesting to consider the running of the spectral index , @xmath107 , for which wmap has given a hint of a larger - than - expected negative value , @xmath108 @xcite . although the error bars are large enough to be consistent with no running , the suggestion of large running has attracted attention . generically , within the slow - roll formalism , one expects @xmath109 , which is usually smaller than the measured central value . our numerical survey shows that it is possible to have rare instances of running as negative as @xmath110 , even though inflation lasts long enough . however , as shown in figure [ fig7 ] , it is still true that @xmath109 . the rare cases of large running correspond to occasional extreme values of the spectral index . we have checked that the spectral index and running are in good agreement with the predictions of slow roll inflation , @xmath111 , @xmath112 , where @xmath113 , @xmath114 and @xmath115 , in units where @xmath116 . the derivatives must be taken with respect to the canonically normalized fields , @xmath117 . in all cases , @xmath118 , whereas @xmath119 and @xmath120 . the running is completely dominated by @xmath121 . = 0.5 it is interesting to compare the predicted values of the running and the spectral index with experimental constraints from combined wmap and sloan digital sky survey lyman @xmath1 data @xcite the allowed region is superimposed on the predictions in figure [ fig8 ] . although most of the predicted points are allowed , we note that the few examples with large running have too large a deviation of @xmath122 from unity . thus if large running should be experimentally confirmed in the future , it will rule out the models under consideration . the previous analysis shows that within a narrow range of the parameter @xmath123 , a long period of inflation will always ensue , as long as the initial number of mobile branes exceeded the critical number @xmath60 . but since @xmath12 looks fine - tuned , this is not a complete solution to the problem of naturalness for the inflaton potential . could this be an accident of the particular form of the simplest potential ? to answer this , we explore the effect of perturbing the model by corrections which are expected to be present , but which we have so far ignored for the sake of simplicity . at least two kinds of modifications to the basic model are generically expected : corrections to the superpotential and to the khler potential . a simple example of the former was considered in @xcite , where the superpotential for the khler modulus @xmath124 and inflaton was taken to be w = w_0 + g(t)(1 + ^2 ) leading to a contribution to the potential of the form [ deltav ] v = + ^2 + ^4 ( 2- ^2)^2 with @xmath4-dependent coefficients @xmath125 . this should be added to the brane - antibrane potential which we have been using so far . in the case of @xmath13 branes , we take @xmath126 . to simplify the analysis , we consider examples where @xmath125 are chosen in such a way that the full potential continues to have a global minimum at @xmath127 , where @xmath128 is assumed to vanish : v ~n we have included the factors of @xmath13 appropriate to the multibrane case , when all branes are coincident . the second kind of correction involves the khler potential , which for @xmath13 branes was taken to be @xmath129 @xcite . this was always understood to be an approximate expression , good when @xmath130 , but subject to higher order corrections in @xmath131 , for example @xmath132 . the denominator in @xmath133 would then be replaced by ( 2- n^2)^2 ( 2- n(^2 + d^4))^2 as shown in appendix b , the khler corrections also give contributions to the numerator of the f - term potential , whose taylor series starts with a term of order @xmath134 . for small @xmath14 , we can absorb this leading term into the @xmath135 coefficient of ( [ deltav ] ) . the kinetic term @xmath136 also gets @xmath137 corrections , which can be inferred from eq.([kmetric ] ) , but these are negligible whenever the branes are slowly rolling . by the above reasoning , the two kinds of corrections enter into the rescaled potential ( [ hatv ] ) as [ newhatv ] v ( c(x^2 - 1)^2 + ( 1 + a(x-1)^4)^-1 ) by experimenting with this generalized potential , one finds that the new coefficient @xmath138 allows for the removal of the lower bound on @xmath12 in ( [ range ] ) . moreover , negative values of @xmath139 can remove the upper bound . it is easy to find values of @xmath138 and @xmath139 for which the potential has the desired behavior as a function of @xmath13 , as in figure [ fig1 ] , regardless of the value of @xmath12 . for example @xmath140 , @xmath141 yields the potentials shown in figure [ fig9 ] , at @xmath142 and @xmath143 . it is important that the new parameters @xmath138 and @xmath139 are independent of the number of branes . thus @xmath59 continues to be the only parameter which changes due to a brane tunneling event . = 0.5 as in the previous section , we scanned over parameters for which @xmath78 varies , hence the number of branes , while keeping @xmath144 fixed , for the two examples with @xmath142 and @xmath143 . figure [ fig10 ] shows the number of e - foldings versus number of branes . the general growth of @xmath145 is evident , as it was for the simple model of the previous section . = 0.5 figure [ fig11 ] shows the correlation of @xmath146 with @xmath122 . it shows that the small-@xmath12 model tends to have a red spectrum , with @xmath147 , and negligible running . the large-@xmath12 model also has @xmath147 for most realizations , but for rare cases with @xmath148 , it is possible to get large deviations from a flat spectrum with @xmath149 . these are accompanied by sizable running , but as in figure [ fig8 ] , such points occur outside of the experimentally allowed region . the overall conclusion is that the generalized models have similar qualitative behavior to the simplest class which was studied in sections iii - iv . in this work we have presented a very general means whereby a long period of inflation , with a suitably flat spectrum of density perturbations , can be automatically obtained after tunneling of branes , from a metastable position in a calabi - yau manifold , into a warped throat containing antibranes . it depends only on the fact that the brane - antibrane potential generically has such a metastable minimum when the number of branes is sufficiently large , and the possibility of starting at a place in the string theory landscape where the number of mobile branes exceeds the threshold for having a minimum . once a certain critical number of branes @xmath60 is reached , the minimum disappears and the branes start rolling slowly to the true minimum . they automatically start from the same position , which is guaranteed to be close to the flattened region of the potential . in this scenario , an arbitrarily long time will be spent in a metastable de sitter phase resembling old inflation , before the slow rolling period ( similarly to ref . each tunneling event results in a small bubble of open universe , within which another period of quasi - de sitter expansion will occur . the process repeats itself until the critical number of branes is reached . for us , these long periods of old inflation are irrelevant we have focused entirely on the final phase of conventional inflation associated with the smooth motion of the branes until they annihilate with antibranes in the throat , hopefully leading to the reheating of the universe @xcite , and perhaps the production of cosmic superstrings @xcite . we numerically scanned over large ranges of the logarithm of relevant combinations of parameters which govern the shape of the inflaton potential , and found that large fractions ( @xmath150 ) of the trials resulted in inflation with no need for tuning , whenever the flattening mechanism could occur . this seems to be a quite promising approach for ameliorating the need for fine - tuning within the framework of the string - theory landscape @xcite . for convenience of numerical integration , we chose a form of the potential which has good behavior as the branes reach the bottom of the throat . this form , which differs from the commonly used taylor - expanded form , is in fact the result of a direct perturbative computation , and does not require any special assumptions . however , the perturbative derivation does assume that @xmath151 in order to be valid . one might therefore wonder how sensitive our results are to the behavior of the potential when this approximation is breaking down . we have investigated what happens if the taylor expansion @xmath152 is used in place of @xmath153 , and we also checked different ways of regulating the singularity , like @xmath154 . for all these alternatives , we found that although the specific parameter values where a flat potential occurs may be somewhat sensitive to the behavior of the potential near @xmath0 , the mechanism of multibrane inflation nevertheless works in the same way as we have presented . we therefore believe it to be a robust result , which is qualitatively quite different from field theory models of inflation . although the dynamical flattening of the potential requires a rather special range of parameters in the simplest model , we showed that this restriction can be lifted once corrections to the superpotential and the khler potential are taken into account . we exhibited an example of such corrections that works , without systematically exploring the parameter space of such corrections . subjectively , it seems that the parameter space ( referred to as @xmath155 in section v ) where multibrane inflation works will be large , given that it was easy to find a successful example , but a more complete investigation could be made . moreover , such corrections should in principle be computed from a fully string theoretic starting point , which might restrict the allowed ranges of @xmath155 . another issue which deserves further study is the precise relationship between the radial coordinate @xmath156 of the brane within the warped throat , and the field @xmath14 which is used to describe the inflaton position in the supergravity description . the exact correspondence between these fields was not needed in because the main dependence on @xmath14 was taken to be through the khler modulus factor ( and superpotential corrections ) , rather than a combination of these with the brane - antibrane potential . we have simply identified the two fields in our analysis , but the actual relation could be more complicated , possibly introducing additional parameters into the lagrangian . we thank c. burgess and f. quevedo for collaboration during the early stages of this work . we also thank h. firouzjahi , s. kachru , r. kallosh , l. kofman , a. linde , i. neupane , e. silverstein , l. susskind , and h. tye for helpful comments and discussions . jc and hs are supported by grants from nserc of canada and fqrnt of qubec . we consider the background of @xmath157 space with the @xmath158 metric : ds_5 ^ 2 = ( -dt^2+d*x*^2 ) + dr^2 where @xmath159 is the radius of the @xmath158 space created as the near - horizon geometry of a stack of @xmath160 @xmath161 branes . there is also an @xmath162 background field created by the stack of branes : a_x^0 , x^1 , x^2 , x^3 = _ r0()^4 if one now writes the action for a probe brane is this background , one finds : s_d3 & = & -t_3d^4 + & + & t_3d^4a_x^0 , x^1 , x^2 , x^3 we will use the convention that @xmath163 , but the rest of the coordinates will describe the location of the brane , including fluctuations ( local bendings ) , @xmath164 . we make here a further assumption , namely that only its radial location varies . _ ^ax^m_^bx^ng_nm = _ ^ax^0_^bx^0g_00 + _ ^ax^i_^bx^ig_ii + _ ^ar_^brg_rr the metric will look like : ( cccc g_00 + g_rr()^2 & g_rr & g_rr & g_rr + g_rr & g_xx + g_rr()^2 & g_rr & g_rr + g_rr & g_rr & g_yy + g_rr()^2 & g_rr + g_rr & g_rr & g_rr & g_zz + g_rr()^2 ) = ( cccc g_00 & 0 & 0 & 0 + 0 & g_xx & 0 & 0 + 0 & 0 & g_yy & 0 + 0 & 0 & 0 & g_zz ) + ( g_rr ) we can apply the usual formula : ( a+b ) = a(1+a^-1b ) a ( 1 + tr ( a^-1b ) ) writing the metric at the location of the brane as : @xmath165 we obtain : ( _ ^ax^m_^bx^ng_nm)= -()^8(1-()^4 f^ ) the factor @xmath166 in the bracket comes from the product g^g_rr = ( ) ^4f^ using these results in the @xmath167 action we obtain : [ dbi+cs ] s_d3 & = & -t_3d^4()^4 + & & t_3d^4 ( ) ( ) ^4_t^1 ^ 2 ^ 3 expanding the square root term we obtain for a @xmath161 brane ( sign @xmath168 ) : s_d3 = t_3d^4 f^ while for a @xmath169 ( sign @xmath170 ) : [ d3_bar_action ] s_3 & = & t_3d^4 f^ + & - & 2t_3d^4()^4 the extra piece in the case of the @xmath169 comes from the non - bps nature of the antibrane , since the graviton exchange and the rr exchange forces now add instead of canceling ( for 3-branes the dilaton is constant and contributes no force ) . the potential energy for the @xmath169 in the gravitational + rr field will make it favorable for the brane to move towards @xmath171 , that is , the bottom of the @xmath158 throat . for the setup in fig.([kklmmt_setup ] ) we have to add the piece corresponding to the graviton + rr exchange between the @xmath161 and the @xmath169 . , scaledwidth=40.0% ] however in this setup we have to include the effect of the @xmath161 in addition to that of the stack . the derivation of the background perturbed by the presence of the @xmath161 placed away from the stack can be found in appendix b of . it starts by writing the metric in the form : ds_10 ^ 2 & = & h^-(-dt^2+d*x*^2)+ h^(dr^2+_abdy^ady^b ) + h&= & placing the extra @xmath161 brane at @xmath172 and solving for the perturbed harmonic function h(r)= + h(r ) which satisfies the laplace equation in the space transverse to the brane , which is spanned by @xmath156 and @xmath173 : ^2_(6)h(r ) = c^(6 ) ( - _ 1 ) which amounts to solving : _ ( - _ 1 ) which give @xmath174 . a similar equation is satisfied by the unperturbed @xmath175 . the constants @xmath176 and @xmath177 depend on the total `` charge '' , more precisely the total tension of the brane or stack of branes , so they are related by : c = and since the solution for the unperturbed @xmath178 is already known , we can immediately find the solution for @xmath179 h(r ) = = we can use the perturbed metric to calculate the action for the @xmath169 using the new function @xmath180 . now modify the setup and place at @xmath181 a stack of @xmath182 @xmath169 s with @xmath183 and an equal number of @xmath161 s placed at @xmath184 , as shown in figure [ modified_setup ] . we can simply generalize the analysis above to come to the result : h(r ) = + _ j=1^m there is no interaction between the @xmath161 s or between the @xmath169 s in the second stack since they are mutually bps . we can now go back to eq . ( [ dbi+cs ] ) and replace the expression @xmath185 by @xmath186 . the relevant term is the second term from eq . ( [ d3_bar_action ] ) , which in the limit @xmath187 becomes : 2t_3d^4 = 2t_3d^4 d^4\{2t_3(1-_j=1^m ) } where @xmath181 is the location of the stack of antibranes . the above result is for one single @xmath169 in the stack . the final form of the effective action should be multiplied by @xmath182 , the number of antibranes in the stack . , scaledwidth=40.0% ] the khler potential is assumed to have the following form : k=-3(r)= -3(t+-f(_i,_i ) ) and we will also assume that @xmath188 decomposes into a sum of functions , each one depending on one field @xmath189 : [ kpot ] k=-3(t+-_if_i(_i,_i ) ) the simplest form of the functions @xmath190 is f_i(_i,_i ) = _ i_i but we can consider more complicated expressions , for example : f_i(_i,_i ) = _ i_i - ( _ i_i)^2 allowing for the more complicated expression for @xmath191 changes the structure of the singularities of the khler potential and of the f - term . in the general case the khler metric derived from the khler potential of eq . ( [ kpot ] ) has the form : the f - term is calculated using the usual formula : v_f = e^k(k^bd_bw-3|w|^2 ) where = -3(r ) the covariant derivatives are expressed as : d_tw & = & _ tw+w_tk = _ tw - w + d__iw & = & _ _ iw+w__ik = w + & = & _ + _ k = _ - + & = & _ + _ k = using the above expression for the inverse khler metric one obtains the following terms : k^td_tw & = & ( _ - ) ( _ tw - w ) ( r+ + + ) + k^_id__iw & = & ( _ -)w ( + + ) + k^td_tw & = & ( _ tw - w ) ( + + ) + k^_jd__jw & = & w ( + + ) g. r. dvali and s. h. h. tye , `` brane inflation , '' phys . lett . b * 450 * , 72 ( 1999 ) [ arxiv : hep - ph/9812483 ] . c. p. burgess , m. majumdar , d. nolte , f. quevedo , g. rajesh and r. j. zhang , `` the inflationary brane - antibrane universe , '' jhep * 0107 * , 047 ( 2001 ) [ arxiv : hep - th/0105204 ] . j. garcia - bellido , r. rabadan and f. zamora , `` inflationary scenarios from branes at angles , '' jhep * 0201 * , 036 ( 2002 ) [ arxiv : hep - th/0112147 ] . c. p. burgess , p. martineau , f. quevedo , g. rajesh and r. j. zhang , `` brane antibrane inflation in orbifold and orientifold models , '' jhep * 0203 * , 052 ( 2002 ) [ arxiv : hep - th/0111025 ] . r. blumenhagen , b. kors , d. lust and t. ott , `` hybrid inflation in intersecting brane worlds , '' nucl . b * 641 * , 235 ( 2002 ) [ arxiv : hep - th/0202124 ] . n. jones , h. stoica and s. h. h. tye , `` brane interaction as the origin of inflation , '' jhep * 0207 * , 051 ( 2002 ) [ arxiv : hep - th/0203163 ] . k. dasgupta , c. herdeiro , s. hirano and r. kallosh , `` d3/d7 inflationary model and m - theory , '' phys . d * 65 * , 126002 ( 2002 ) [ arxiv : hep - th/0203019 ] . o. dewolfe , s. kachru and h. l. verlinde , `` the giant inflaton , '' jhep * 0405 * , 017 ( 2004 ) [ arxiv : hep - th/0403123 ] . j. j. blanco - pillado _ et al . _ , `` racetrack inflation , '' jhep * 0411 * , 063 ( 2004 ) [ arxiv : hep - th/0406230 ] . e. silverstein and d. tong , `` scalar speed limits and cosmology : acceleration from d - cceleration , '' phys . d * 70 * , 103505 ( 2004 ) [ arxiv : hep - th/0310221 ] . m. alishahiha , e. silverstein and d. tong , `` dbi in the sky , '' phys . d * 70 * , 123505 ( 2004 ) [ arxiv : hep - th/0404084 ] . x. g. chen , `` multi - throat brane inflation , '' phys . d * 71 * , 063506 ( 2005 ) [ arxiv : hep - th/0408084 ] ; `` inflation from warped space , '' arxiv : hep - th/0501184 . k. becker , m. becker and a. krause , `` m - theory inflation from multi m5-brane dynamics , '' nucl . b * 715 * , 349 ( 2005 ) [ arxiv : hep - th/0501130 ] . s. kachru , r. kallosh , a. linde , j. maldacena , l. mcallister and s. p. trivedi , `` towards inflation in string theory , '' jcap * 0310 * , 013 ( 2003 ) [ arxiv : hep - th/0308055 ] . s. b. giddings , s. kachru and j. polchinski , `` hierarchies from fluxes in string compactifications , '' phys . d * 66 * , 106006 ( 2002 ) [ arxiv : hep - th/0105097 ] . i. r. klebanov and m. j. strassler , `` supergravity and a confining gauge theory : duality cascades and @xmath194sb - resolution of naked singularities , '' jhep * 0008 * , 052 ( 2000 ) [ arxiv : hep - th/0007191 ] . s. kachru , r. kallosh , a. linde and s. p. trivedi , `` de sitter vacua in string theory , '' phys . d * 68 * , 046005 ( 2003 ) [ arxiv : hep - th/0301240 ] . de alwis , `` effective potentials for light moduli , '' arxiv : hep - th/0506266 . c. p. burgess , j. m. cline , h. stoica and f. quevedo , `` inflation in realistic d - brane models , '' jhep * 0409 * , 033 ( 2004 ) [ arxiv : hep - th/0403119 ] . m. berg , m. haack and b. kors , `` loop corrections to volume moduli and inflation in string theory , '' phys . rev . d * 71 * , 026005 ( 2005 ) [ arxiv : hep - th/0404087 ] . j. p. hsu , r. kallosh and s. prokushkin , `` on brane inflation with volume stabilization , '' jcap * 0312 * , 009 ( 2003 ) [ arxiv : hep - th/0311077 ] . j. p. hsu and r. kallosh , `` volume stabilization and the origin of the inflaton shift symmetry in string theory , '' jhep * 0404 * , 042 ( 2004 ) [ arxiv : hep - th/0402047 ] . h. firouzjahi and s. h. h. tye , `` closer towards inflation in string theory , '' phys . b * 584 * , 147 ( 2004 ) [ arxiv : hep - th/0312020 ] . h. firouzjahi and s. h. tye , `` brane inflation and cosmic string tension in superstring theory , '' jcap * 0503 * , 009 ( 2005 ) [ arxiv : hep - th/0501099 ] . r. kallosh and a. linde , `` landscape , the scale of susy breaking , and inflation , '' jhep * 0412 * , 004 ( 2004 ) [ arxiv : hep - th/0411011 ] . o. dewolfe and s. b. giddings , `` scales and hierarchies in warped compactifications and brane worlds , '' phys . rev . d * 67 * , 066008 ( 2003 ) [ arxiv : hep - th/0208123 ] . s. b. giddings and a. maharana , `` dynamics of warped compactifications and the shape of the warped landscape , '' arxiv : hep - th/0507158 . n. iizuka and s. p. trivedi , phys . d * 70 * , 043519 ( 2004 ) [ arxiv : hep - th/0403203 ] . a. r. liddle , a. mazumdar and f. e. schunck , `` assisted inflation , '' phys . rev . d * 58 * , 061301 ( 1998 ) [ arxiv : astro - ph/9804177 ] . a. mazumdar , s. panda and a. perez - lorenzana , `` assisted inflation via tachyon condensation , '' nucl . b * 614 * , 101 ( 2001 ) [ arxiv : hep - ph/0107058 ] . y. s. piao , r. g. cai , x. m. zhang and y. z. zhang , `` assisted tachyonic inflation , '' phys . d * 66 * , 121301 ( 2002 ) [ arxiv : hep - ph/0207143 ] . m. sasaki and e. d. stewart , `` a general analytic formula for the spectral index of the density perturbations produced during inflation , '' prog . * 95 * , 71 ( 1996 ) [ arxiv : astro - ph/9507001 ] . h. v. peiris _ et al . _ , `` first year wilkinson microwave anisotropy probe ( wmap ) observations : implications for inflation , '' astrophys . j. suppl . * 148 * , 213 ( 2003 ) [ arxiv : astro - ph/0302225 ] . u. seljak _ et al . _ , `` cosmological parameter analysis including sdss ly - alpha forest and galaxy bias : constraints on the primordial spectrum of fluctuations , neutrino mass , and dark energy , '' phys . rev . d * 71 * , 103515 ( 2005 ) [ arxiv : astro - ph/0407372 ] . n. barnaby and j. m. cline , `` creating the universe from brane - antibrane annihilation , '' phys . d * 70 * , 023506 ( 2004 ) [ arxiv : hep - th/0403223 ] . n. barnaby , c. p. burgess and j. m. cline , `` warped reheating in brane - antibrane inflation , '' jcap * 0504 * , 007 ( 2005 ) [ arxiv : hep - th/0412040 ] . l. kofman and p. yi , `` reheating the universe after string theory inflation , '' arxiv : hep - th/0507257 . s. sarangi and s. h. h. tye , `` cosmic string production towards the end of brane inflation , '' phys . b * 536 * , 185 ( 2002 ) [ arxiv : hep - th/0204074 ] . n. t. jones , h. stoica and s. h. h. tye , `` the production , spectrum and evolution of cosmic strings in brane inflation , '' phys . lett . b * 563 * , 6 ( 2003 ) [ arxiv : hep - th/0303269 ] . l. pogosian , s. h. h. tye , i. wasserman and m. wyman , `` observational constraints on cosmic string production during brane inflation , '' phys . d * 68 * , 023506 ( 2003 ) [ arxiv : hep - th/0304188 ] . e. j. copeland , r. c. myers and j. polchinski , `` cosmic f- and d - strings , '' jhep * 0406 * , 013 ( 2004 ) [ arxiv : hep - th/0312067 ] . t. damour and a. vilenkin , `` gravitational radiation from cosmic ( super)strings : bursts , stochastic background , and observational windows , '' phys . rev . d * 71 * , 063510 ( 2005 ) [ arxiv : hep - th/0410222 ] . n. barnaby , a. berndsen , j. m. cline and h. stoica , `` overproduction of cosmic superstrings , '' jhep * 0506 * , 075 ( 2005 ) [ arxiv : hep - th/0412095 ] . l. susskind , `` the anthropic landscape of string theory , '' arxiv : hep - th/0302219 . t. banks , m. dine and e. gorbatov , `` is there a string theory landscape ? , '' jhep * 0408 * , 058 ( 2004 ) [ arxiv : hep - th/0309170 ] .

dense plasmas have recently gained growing interest due to their relevance for the interior of giant planets as well as for laser interaction with matter and inertial confinement fusion scenarios . examples of recent experimental studies include the ultrafast thermalization of laser plasmas @xcite or free electron laser excited plasmas @xcite , inertial confinement fusion experiments at the national ignition facility @xcite and magnetized z - pinch experiments at sandia @xcite . questions of fundamental theoretical importance are the conductivity and heat conduction , the energy loss of energetic particles ( stopping power ) in such a plasma , e.g. @xcite or the temperature equilibration of the electronic and ionic components @xcite . despite recent advances in modeling and computer simulations a fully selfconsistent treatment of these , in general , highly nonequilibrium electron - ion plasmas has not been possible so far due to the requirement of the simultaneous account of electronic quantum and spin effects together with the ( possibly ) strong ionic correlations . the main problem here are the vastly different time scales of electrons and ions resulting from their different masses . a possible solution of this dilemma is a multi - scale approach that has been proposed by ludwig _ it takes advantage of the weak electron - ion coupling that allows for a linear response treatment of the electrons . this idea has been used by graziani _ _ to decouple the electron kinetic equation using an stls ( singwi - tosi - land - sjlander ) scheme @xcite or a recently derived extension @xcite . the key of this multiscale approach is to absorb the fast electron kinetics into an effective screened potential @xmath0 of the heavy ions with charge @xmath2 where the screening is provided by the electrons via a proper dielectric function @xmath3 , e.g. @xcite @xmath4 taken in the static limit . we note that also the effect of streaming electrons maybe important in warm dense matter . then the frequency argument becomes @xmath5 which leads to wake effects , that are well investigated theoretically and experimentally for dusty plasmas , e.g. @xcite and quantum plasmas , e.g. ref . @xcite and references therein . however , this is beyond the scope of this paper . a similar but even simpler approach that is based on a quantum hydrodynamic model ( qhd ) and replaces the linear response potential of the ions by a simplified expression that is derived from linearized qhd ( lqhd ) , @xmath6 . the first expression for @xmath7 was derived by shukla and eliasson ( se ) who predicted an attractive interaction between ions , even in the absence of streaming , @xmath8 @xcite . comparisons with density functional theory revealed that this is incorrect @xcite . this also underlined the limitations of qhd models to weakly correlated dense plasmas @xcite . a recent overview and more references can be found in ref . @xcite . recently , modified expressions for @xmath7 were derived by akbari moghanjoughi ( am ) @xcite and stanton and murillo ( sm ) @xcite . the latter is particularly interesting because it is also applicable to finite temperatures , in contrast to the se and am potentials that neglect thermal effects . the question arises whether the am and sm potentials are more accurate than the se potential and whether they allow for an extension of the applicability limits of qhd . it is the goal of the present paper to perform such an analysis . to quantify the accuracy of these three potentials we use , as a basis , the standard static potentials of weakly correlated electrons that can be derived from quantum kinetic theory @xcite by linearization in the external perturbation this directly leads to the static potential ( [ pot_stat ] ) involving the random phase approximation ( rpa ) for the dielectric function and the mermin dielectric function , respectively . both include kinetic effects , finite temperature and , in the latter case , also collisions . being rooted in a kinetic approach , the corresponding static potentials are , by construction , more accurate than the potentials @xmath7 and , thus , allow for a rigorous test of the validity of the latter . the main conclusions of our analysis are the following : 1 . ) the se potential is qualitatively different from the kinetic results . 2 . ) the sm potential exhibits very good agreement with the rpa . the am potential is in agreement with the sm potential , at @xmath9 , differing only in the notation and 4 . ) the sm potential is in good agreement with the rpa result even at elevated temperature . this confirms the correctness of the recently reported @xcite correction factor @xmath1 that has to be included in front of the bohm term of the qhd equations for fermions . the paper is organized as follows : we first recall the rpa and mermin dielectric functions in sec . [ s : mermin ] and discuss their basic properties such as the reproduction of friedel oscillations . after this we recall the se , am and sm potentials in sec . [ s : potential ] and present numerical results that compare them to the potential @xmath0 computed from the rpa and mermin dielectric functions . the paper concludes with a brief discussion of the physical origin of the factor @xmath1 . the main steps of its derivation are outlined in an appendix . inset shows an enlargement of the potential around its zero values to resolve the friedel oscillations . ] the mean - field result for the electron dielectric function is given by the random phase ( rpa or hartree or quantum vlasov ) approximation , @xmath10 , where correlation effects are neglected , and damping of collective oscillations is entirely due to landau damping . temperature effects are straightforwardly included by using the corresponding fermi function . for an overview on the final expression , see the appendix of ref . @xcite . the effective ion potential ( [ pot_stat ] ) computed from the rpa dielectric function in the zero temperature limit is shown in figs . [ fig : rpa_friedel ] and [ fig : rpa_friedel2 ] for different densities . the zero temperature rpa potential reproduces the friedel oscillations , as is illustrated in fig . [ fig : rpa_friedel2 ] by comparison to the large @xmath11 asymptotics , @xmath12 , ( red dashed line ) . while the oscillations are very shallow and hardly of relevance for the thermodynamics and transport properties in warm dense matter , the reproduction of the correct periodicity is a useful consistency and accuracy test of the fourier transformation in eq . ( [ pot_stat ] ) . for later reference we also provide the static long wavelength limit of the effective ion potential which is nothing but the yukawa potential ( [ pot_stat ] ) @xmath13 with the familiar inverse yukawa screening length , @xmath14 , that interpolates between the debye and thomas - fermi expressions , in the non - degenerate and zero temperature limits , respectively , @xmath15 here @xmath16 is the thomas - fermi wave number , and @xmath17 is the fermi integral of order @xmath18 . note that the potential ( [ yukawa ] ) depends via @xmath19on density and temperature . in the warm dense matter regime we expect the electrons to be weakly to moderately coupled so that correlation effects are , in general , not negligible . the simplest quantum dielectric function ( df ) which takes collisions into account in a conserving fashion via a relaxation time approximation is the mermin df @xcite @xmath20}{\omega+i\nu[\epsilon_{\rm rpa}(k,\omega+i\nu)-1]/[\epsilon_{\rm rpa}(k,0)-1]},\end{gathered}\ ] ] which involves the finite temperature rpa df and the electron collision frequency @xmath21 . in ref . @xcite various choices for the collision frequency were studied , so here we will just use a typical set of values to highlight the main trends . ( top ) and @xmath22 ( bottom ) and @xmath23 . the zoom into the small potential values confirms the correct oscillation frequency at large distances that is expected for friedel oscillations , as seen from the comparison to the ( red ) dashed curve . here @xmath24 corresponds to the distance indicated by a vertical dashed line . the grey box shows that the oscillation period of the full potential agrees with the known asymptotics.,title="fig : " ] + ( top ) and @xmath22 ( bottom ) and @xmath23 . the zoom into the small potential values confirms the correct oscillation frequency at large distances that is expected for friedel oscillations , as seen from the comparison to the ( red ) dashed curve . here @xmath24 corresponds to the distance indicated by a vertical dashed line . the grey box shows that the oscillation period of the full potential agrees with the known asymptotics.,title="fig : " ] the shukla - eliasson potential has been derived from linearized qhd , below we reproduce its most recent variant @xcite . it involves the electron plasma frequency @xmath25 , [ @xmath26 is the effective electron mass ] and the electron fermi speed @xmath27 , where @xmath28 is proportional to the wigner - seitz radius . the term @xmath29^{1/2}$ ] can be regarded as ( optional ) correction that is supposed to account for exchange - correlation effects [ it was proposed in ref . @xcite based on the exchange - correlation potential of density functional theory ] . since the other potentials do not include such a correction and it does not qualitatively alter the potential , in the comparisons below we will use @xmath30 . the se potential , for @xmath31 , is given by @xcite : @xmath32 e^{-k'_+ r } , \label{eq : sep}\end{aligned}\ ] ] with the parameter @xmath33 . the wave vectors are given by @xcite @xmath34 , where @xmath35 is similar ( but not identical ) to the inverse thomas - fermi screening length @xmath19 , eq . ( [ ksc ] ) . the se potential for the second parameter range , @xmath36 , is given by : @xmath37.\end{aligned}\ ] ] with the definitions @xmath38 and @xmath39 . note that the shukla - eliasson potential is a zero - temperature approximation , and the potential depends only on density . numerical results are included in fig . [ fig : all_vs_rpa0 ] . here we reproduce the electrostatic potential derived by akbari - moghanjoughi ( we use the most recent corrected version , ref . this potential has the same mathematical form as the se potential @xmath40 , for @xmath41 : @xmath42 \ , e^{-k'_+ r}. \label{eq : amp>}\end{aligned}\ ] ] the difference lies in the definition of the parameter @xmath43 and of the wave numbers @xcite @xmath44 where @xmath45 is the inverse thomas - fermi screening length . these formulas are connected to the se versions by @xmath46 and an additional factor of @xmath1 in the definition of @xmath47 . the potential for @xmath48 is not given in ref . @xcite . assuming the same relation to the se result it has the form @xmath49 , \label{eq : amp<}\end{aligned}\ ] ] with @xmath50 and @xmath51 . as the se potential , this is a zero - temperature result which , therefore , depends only on the density . numerical results are included in fig . [ fig : all_vs_rpa0 ] . here we reproduce the screened potential of stanton and murillo following the latest corrected version @xcite . again , there are two cases . for @xmath52 the potential is given by @xcite @xmath53 \ , e^{-k'_+ r},\end{aligned}\ ] ] with @xmath54 , @xmath55 , @xmath56 , and @xmath57 . @xmath58 denotes the fermi integral and @xmath59 its derivative with respect to @xmath60 . @xmath61 is determined by the normalization , @xmath62 with the inverse temperature @xmath63 . the inverse thomas - fermi screening length for finite temperatures is given as @xmath64 . a second version of the potential follows for the case @xmath65 : @xmath66,\end{aligned}\ ] ] where @xmath67 and @xmath68 . we note that the sm expressions are identical to the am version in the limit @xmath9 , if @xmath69 is rescaled by a factor of @xmath70 , @xmath71 . numerical results are included in figs . [ fig : all_vs_rpa0 ] and [ fig : ms_rpa_t ] . ] potentials to the zero - temperature full rpa potential ( full black line ) for three densities . top : @xmath72 , center : @xmath73 , bottom : @xmath74 . the sm potential shows increasing deviations from the rpa potential @xmath0 for decreasing density , cf . top and middle figures . at high density ( bottom figure ) the sm potential exhibits very good agreement with the rpa potential . the se potential exhibits large deviations from the rpa at all densities . insets show the behavior for large distances and confirm that all models ( except for the rpa ) are unable to describe the friedel oscillations correctly.,title="fig : " ] + ] potentials to the zero - temperature full rpa potential ( full black line ) for three densities . top : @xmath72 , center : @xmath73 , bottom : @xmath74 . the sm potential shows increasing deviations from the rpa potential @xmath0 for decreasing density , cf . top and middle figures . at high density ( bottom figure ) the sm potential exhibits very good agreement with the rpa potential . the se potential exhibits large deviations from the rpa at all densities . insets show the behavior for large distances and confirm that all models ( except for the rpa ) are unable to describe the friedel oscillations correctly.,title="fig : " ] + ] potentials to the zero - temperature full rpa potential ( full black line ) for three densities . top : @xmath72 , center : @xmath73 , bottom : @xmath74 . the sm potential shows increasing deviations from the rpa potential @xmath0 for decreasing density , cf . top and middle figures . at high density ( bottom figure ) the sm potential exhibits very good agreement with the rpa potential . the se potential exhibits large deviations from the rpa at all densities . insets show the behavior for large distances and confirm that all models ( except for the rpa ) are unable to describe the friedel oscillations correctly.,title="fig : " ] we now present a numerical comparison of the three lqhd potentials to the rpa potential ( [ pot_stat ] ) . in the following , all results are given in hartree atomic units . since the am potential conincides with the sm potential taken at @xmath9 we will , in the following , not distinguish both potentials . we use the shortcut `` sm '' for both cases . to begin the analysis , we recall that the se , sm [ and am ] ion potentials are all derived by applying the long - wavelength limit to the full potential . correspondingly , in fourier space one should expect accurate results only for small wave numbers , @xmath75 , where @xmath76 is the fermi wavenumber . for a comprehensive comparison we consider a broad density range , covering 6 orders of magnitude with the three values of the brueckner parameter , @xmath77 ( similar values were studied in ref . @xcite ) . the first observation , cf . [ fig : all_vs_rpa0 ] , is that all four potentials show the same overall behavior with the deviations growing when the density is reduced . the next observation is that the sm potential conincides with the rpa potential at large distances : for @xmath78 , for the @xmath79 and @xmath22 , respectively . however , a closer inspection of the large distance behavior shows that substantial deviations remain : none of the three approximate potentials reproduces the friedel oscillations of the rpa potential , see also figs . [ fig : rpa_friedel ] and [ fig : rpa_friedel2 ] . nevertheless , the sm potential provides a good fit to the rpa potential if one discards the ( small ) fourier component with @xmath80 . consider now the case @xmath73 , cf . middle part of fig . [ fig : all_vs_rpa0 ] . here one clearly recognizes the difference between the sm ( and am ) potentials , compared to the se potential . while the se potential exhibits strong deviations from the rpa potential , the sm potential is very close to the rpa result . at the highest density , @xmath74 , bottom figure , the deviations between sm and se are very small . but this is not surprising as , in this case , the bohm term in the qhd equations is negligible , so the influence of the different prefactors of this term ( @xmath81 , for se , versus @xmath1 , for sm ) is insignificant . finally , at the lowest density , @xmath72 , top figure , all three potentials show substantial deviations from the rpa for small distances below @xmath82 , although the deviations of the se potential are significantly larger , in particular for larger distances . however , at these low densities , the plasma is strongly correlated and the rpa potential itself is not applicable . thus , from the zero - temperature behavior we conclude that the sm ( and am ) potential shows a substantial improvement over the se potential , confirming the correctnesss and importance of the prefactor @xmath1 in front of the bohm term @xcite . in fact , we will show in sec . [ s : dis ] that this coefficient is not a free parameter but , for the description of the long - wavelength properties of the plasma ( such as the statically screened potential ) the value @xmath1 follows rigorously . quantum hydrodynamics , as used in the plasma physics community , is a zero - temperature theory , so the results of se @xcite for the effective ion potential and the recent correction by akbari - moghanjoughi @xcite are restricted to the ground state . at the same time , for relevant applications to dense plasmas , warm dense matter or laser plasmas , finite temperature effects are usually non - negligible . it was shown in refs . @xcite that the qhd equations are directly linked to the thomas - fermi theory which allows for a straightforward incorporation of finite - temperature effects . this link was exploited in ref . @xcite to derive a statically screened potential for arbitrary temperatures . it is now interesting to compare the corresponding finite - temperature sm potential to the rpa potential at the same temperature . this is done in fig . [ fig : ms_rpa_t ] for different values of the degeneracy parameter @xmath83 at constant density . the figure shows that , for high density , @xmath84 , the agreement is excellent , for all temperatures . this is not surprising since , at these densities gradient corrections ( the bohm term of qhd ) are small , as noted above . here the temperature dependence is governed by the one of the ideal fermi gas ( fermi pressure ) . more interesting is the comparison at lower density , @xmath73 , bottom part of fig . [ fig : ms_rpa_t ] where correlation effects ( and the bohm contribution ) are significant . here the agreement between sm and rpa potentials is slightly worse and improves with increasing temperature . this is due to smoothening of the friedel oscillations with increased temperature . overall , we conclude that the agreement between the sm potential and the rpa , even at finite temperature , is very impressive . this confirms again that the sm potential correctly captures the long - wavelength properties of the static rpa potential ( [ pot_stat ] ) . , bottom : @xmath73 . as it is seen the sm potential accurately describes the screening effect at high temperatures and densities.,title="fig : " ] , bottom : @xmath73 . as it is seen the sm potential accurately describes the screening effect at high temperatures and densities.,title="fig : " ] and two collision frequencies ( in units of the plasma frequency ) . the behavior is representative for other densities and temperatures and indicates that the electron collisions do not affect the static screened potential , within the relaxation time approximation ( mermin model ) . ] finally , let us analyze the effect of collisions ( correlations ) on the dielectric function . it is well known that these effects become increasingly important outside the weak - coupling limit , i.e. for @xmath85 . yet it is not clear , a priori , how important these effects are for the static ion potential . we , therefore , computed the static potential ( [ pot_stat ] ) with the mermin dielectric function for a broad range of temperatures , densities and collision frequencies . a typical result is shown in fig . [ fig : ms_mermin_t ] where the rpa potential is compared to the mermin dielectric function with the collision frequency @xmath86 , at zero temperature . the differences are extremely small , and the same behavior is observed at other densities and elevated temperatures . the explanation is , of course , that correlation effects are particularly relevant at short distances where the coulomb potential is large anyway . we note that the situation is very different in streaming plasmas that are out of equilibrium . here scattering effects have an important impact on the effective ion potential @xcite . we have presented an analysis of three recently proposed static ion potentials in quantum plasmas and compared them to the static limit of the rpa potential ( [ pot_stat ] ) . the first of these potentials ( se ) was the one derived by shukla and eliasson @xcite who predicted an attractive minimum of the potential [ cf . middle part of fig . [ fig : all_vs_rpa0 ] ] that would be responsible for ordering of ions in dense plasmas . our analysis reveals that no such minimum exists in the rpa potential , so this minimum should be regarded an artefact of the used lqhd model @xcite . the comparison with the two other potentials , the ones of stanton and murillo ( sm ) @xcite and of akbari - moghanjoughi ( am ) @xcite which do not exhibit a comparable minimum reveals the origin of this discrepancy ( see below ) . the sm ( and am ) potential exhibits very good agreement with the rpa potential at @xmath9 indicating that it correctly captures the long - wavelength properties of the rpa , in contrast to the se potential . since the only difference between the se potential ( [ eq : sep ] ) and the sm ( and am ) potential ( [ eq : sm > ] ) is the different prefactor of the bohm term in the underlying qhd equations , the good accuracy of the latter indicates that the choice of the prefactor @xmath1 is the correct one . in fact , this question is easily answered by recalling the one - to - one correspondence between qhd and thomas - fermi theory with gradient corrections @xcite . in thomas - fermi theory and , more generally , in density functional theory , it is well known that an expansion of the ( non - interacting ) kinetic energy ( together with exchange contributions ) in a power series in terms of density gradients converges to the full hartree - fock energy @xcite . in the present case , we start from the statically screened potential ( [ pot_stat ] ) in rpa , so we should expect agreement from a gradient expansion of the thomas - fermi kinetic energy @xmath87 $ ] without exchange and correlation corrections . since the bohm term of qhd corresponds just to the first non - vanishing gradient correction , @xmath88 , ( quadratic in density gradients , see appendix ) , the pre - factors of both are linked inseparably via the equation @xcite @xmath89}{\delta n } = \gamma\ , \frac{\hbar^2}{8m}\left ( \left| \frac{\nabla n}{n } \right|^2 - 2 \frac{\nabla^2 n}{n } \right ) = \gamma \,v^{\rm mh}_b[n ] , \label{eq : grad = vb1}\ ] ] where @xmath90 is the bohm potential for quantum systems with bose statistics . for the case of quantum plasmas ( fermions ) the same potential ( i.e. @xmath91 ) was postulated by manfredi and haas @xcite . however , the correct value ( at @xmath9 ) is @xmath92 which is clearly confirmed by the comparison of the se and sm potentials to the rpa in the present work . let us briefly discuss the origin and validity of the prefactor @xmath1 . this , coefficient was first obtained by kompaneets @xcite and kirzhnitz @xcite , almost 60 years ago , who used a matsubara green functions method to obtain the gradient correction to the thomas - fermi energy functional . in the same year golden @xcite derived this factor by expanding the density matrix of a many - electron system . later hohenberg and kohn @xcite presented a systematic derivation in terms of a gradient expansion to all orders , on the basis of the electron polarization function for zero temperature . mermin @xcite extended the hohenberg and kohn method to finite temperature but did not give a derivation of the gradient correction at finite temperature . this was done by krizhnitz _ _ @xcite and by perrot @xcite who minimized the grand canonical potential using the known form of the free energy of the electrons . at last , the feynman path - integral method was used by yang @xcite to obtain the single - particle green function and to prove the factor @xmath1 . over the recent decades occasionally additional derivations have appeared . in particular , the recent analysis of akbari - moghanjoughi @xcite essentially reproduces earlier derivations based on the calculation of the long - wavelength limit of the inverse of the lindhard dielectric function , @xmath93 at @xmath9 , whereas the analysis of stanton and murillo @xcite directly starts from perrot s result @xcite . we also note that there exists a different approach to the gradient corrections in thomas - fermi theory where the prefactor of the respective term is used as a fit parameter to reproduce improved simulations ( that go beyond the rpa ) , e.g. @xcite . similarly , for the case of electrons bound in atoms , the prefactor can be optimized to better match the hartree - fock ground state energy , yielding @xmath94 @xcite . however , in the context of the screened potential of free charged particles in a plasma one has to use the static long - wavelength limit of the inverse dielectric function which is exactly given by the rpa , so there exists no freedom of choice , see sec . [ ss : p - expansion ] . a key input for the gradient correction of the kinetic energy is the wavenumber expansion of the inverse of the static density response function , i.e. of the longitudinal polarization function @xmath95 . the result is ( for the derivation , see the appendix ) @xmath96 this is a completely general result where the whole microphysics is contained in the value of the expansion coefficients . from this we can directly obtain systematic approximations for the statically screened ion potential ( [ pot_stat ] ) without recourse to the connection between thomas - fermi theory and qhd , eq . ( [ eq : grad = vb1 ] ) . indeed , what is needed to evaluate ( [ pot_stat ] ) is the inverse of the dielectric function which is related to the polarization function by @xcite @xmath97 from this and eq . ( [ eq : pinv_exp ] ) we obtain the desired wavenumber expansion of the inverse static dielectric function [ and of the screened potential ( [ pot_stat ] ) ] @xmath98 using , as the lowest order approximation , the rpa polarization , we obtain for the first five non - vanishing coeffcients ( see appendix ) @xmath99 where the @xmath100 involve fermi integrals of different orders [ see eq . ( [ eq : bi ] ) of the appendix ) ] . note that the finite temperature results for @xmath101 and @xmath102 are given here for the first time . this result allows one to systematically derive approximations for the screened potential by considering the long - wavelength limit , @xmath103 of the inverse dielectric function ( [ eq : eps_inv_exp ] ) . * lowest order result : * the lowest order is given by neglecting all @xmath104-dependent terms in the parantheses of eq . ( [ eq : eps_inv_exp ] ) , @xmath105 , and we obtain @xmath106 inserting this into eq . ( [ pot_stat ] ) , immediately yields the yukawa potential ( [ yukawa ] ) . * second order result : * the next order is obtained by retaining the terms quadratic in @xmath104 , @xmath107 the result for @xmath108 is ( see appendix ) @xmath109 in the ground state , @xmath9 , this coefficient becomes @xmath110 . equation ( [ eq : eps_2 ] ) is the approximation used by stanton and murillo @xcite and ( for @xmath9 ) by akbari moghanjoughi ( am ) @xcite . the approximation of shukla and eliasson is obtained for @xmath9 by using @xmath111 without the prefactor @xmath112 . this coefficient yields the ratio of the corresponding results for fermions and bosons ( or spinless particles ) @xcite . from the temperature dependence of @xmath108 we can extract the temperature dependence of the coefficient @xmath113 which is displayed in fig . [ fig : alpha_3 ] . @xmath113 increases monotonically from @xmath1 to @xmath114 @xcite , for temperatures large compared to @xmath115 . evidently this factor is crucial for the correct treatment of fermions within thomas - fermi theory or in quantum hydrodynamics . in front of the bohm potential @xmath116 , cf . ( [ eq : grad = vb1 ] ) , for fermions as a function of temperature . in the zero - temperature limit , @xmath113 approaches @xmath1 , and it increases monotonically to @xmath114 @xcite . in contrast , for bosons or spinless particles this factor equals @xmath81 ( dashed line ) , as in the case of a single quantum particle @xcite . ] to compare the first and second approximations to the static inverse dielectric function , eqs . ( [ eq : eps_0 ] ) and ( [ eq : eps_2 ] ) , we plot again the associated static potentials computed according to eq . ( [ pot_stat ] ) . in the top two parts of fig . [ fig : all_vs_rpa0_tf ] we show again the zero temperature results of se and sm and , in addition the thomas - fermi screended yukawa result , eq . ( [ yukawa ] ) . thus , we are able to assess the quality of the different orders of the wavenumber expansion of @xmath117 , in comparison to the full rpa result . for @xmath73 we see that the yukawa potential ( zeroth order of the expansion ) provides a good approximation , but the next correction , proportional to @xmath118 , that is contained in eq . ( [ eq : eps_2 ] ) and , thus , in the sm potential constitutes an improvement . in contrast , the se potential which contains the @xmath118 correction with a nine times larger prefactor exhibits a poor performance and is even significantly less accurate than the zeroth order correction alone . the same trend is observed at higher density ( top figure ) where the zeroth and first order are both indistinguishable from the full result whereas the se potential is substantially less accurate . the bottom part of fig . [ fig : all_vs_rpa0_tf ] shows the performance of the zeroth and second order potentials [ tf and sm , eqs . ( [ eq : eps_0 ] ) and ( [ eq : eps_2 ] ) , respectively ] at finite temperature . while at low temperature , @xmath119 , the second order provides a significant impovement , for higher temperature , @xmath120 , the expansion exhibits sign alternating convergence and both approximations are of comparable quality with a slightly better performance of the sm potential . in other words , for @xmath120 , the standard static yukawa potential provides a fairly accurate description of the static limit of the rpa . but with the yukawa potential , eq . ( [ yukawa ] ) , `` tf '' , added for comparison of the ground state behavior . bottom : temperature dependence of the potentials . same as bottom part of fig . [ fig : ms_rpa_t ] with the yukawa potential ( [ yukawa ] ) added.,title="fig : " ] + but with the yukawa potential , eq . ( [ yukawa ] ) , `` tf '' , added for comparison of the ground state behavior . bottom : temperature dependence of the potentials . same as bottom part of fig . [ fig : ms_rpa_t ] with the yukawa potential ( [ yukawa ] ) added.,title="fig : " ] + but with the yukawa potential , eq . ( [ yukawa ] ) , `` tf '' , added for comparison of the ground state behavior . bottom : temperature dependence of the potentials . same as bottom part of fig . [ fig : ms_rpa_t ] with the yukawa potential ( [ yukawa ] ) added.,title="fig : " ] * higher order results . correlations : * it poses no principal problem to extend the above results for the inverse dielectric function to higher powers of @xmath104 . with the equations ( [ eq : a4][eq : a8 ] ) the next three higher orders are available , for arbitrary temperature . this should improve the behavior of the static potential also at shorter distances . however , for the short - distance behavior also other effects will be important , most noticeably exchanges and correlations . one way to include correlations is to use static local field corrections , as proposed in ref . a more systematic and powerful approach is to use many - body theory which provides dynamic correlation corrections , e.g. refs . @xcite and @xcite and the cited references . the conclusion from the above comparison of approximate static ion potentials to the rpa result is that , in fact , there is no freedom in the choice of the prefactor of the bohm term for quantum hydrodynamics for fermions . the qhd equations follow from the thomas - fermi theory by performing a systematic gradient expansion of the kinetic energy or , equivalently , a long - wavelength expansion of the inverse polarization function @xmath121 , eq . ( [ eq : pinv_exp ] ) . this expansion directly leads to the coefficient @xmath1 , at @xmath9 , in front of the bohm term compared to the cases of bosons or of a single - particle , whereas for higher temperatures the coefficient increases monotonically to @xmath122 ( in 1d and 2d different coefficients apply @xcite ) . this coefficient is a consequence of the pauli principle and has important implications for a hydrodynamic modeling of dense quantum plasmas . for example , we have shown that a neglect of this coefficient gives rise to a statically screened ion potential that is far less accurate than the standard thomas - fermi ( yukawa ) potential , rather than an improvement . it also indicates that earlier derivations of qhd equations are invalid for quantum plasmas and earlier qhd results that were using the prefactor @xmath81 do not describe fermions , in general , nor dense plasmas , in particular . this work has been supported by the deutsche forschungsgemeinschaft via grant sfb - tr 24 , project a9 and the ministry of education and science of kazakhstan . here we present a systematic derivation of the gradient correction to the thomas fermi energy using mermin s approach for finite temperatures . this appears to be the most systematic and general approach and allows one to derive higher order corrections as well . following hohenberg and kohn @xcite , and mermin @xcite , the electron energy is a functional of the density : @xmath123=\int \ ! v({\mathbf{r}})n({\mathbf{r}})\ , \mathrm{d}{\mathbf{r}}+\frac{1}{2}\int \ ! \frac{n({\mathbf{r}})n({\mathbf{r}}^\prime)}{\mid { \mathbf{r}}-{\mathbf{r}}^\prime\mid } \ , \mathrm{d}{\mathbf{r}}\mathrm{d}{\mathbf{r}}^\prime+ t[n],\ ] ] where the first term is the energy due to external field and the second is the interaction energy of the electrons in mean field approximation . $ ] is the kinetic energy ( we neglect additional exchange - correlation terms ) which is well known in the local approximation ( thomas - fermi theory ) , @xmath124 = \int d{\bf r } \ , t[n({\bf r})]\ , n({\bf r}),\ ] ] where the kinetic energy density is related to the ( local ) fermi energy , @xmath125=\frac{3}{5}e_f[n]$ ] . to obtain corrections to this local approximation , we expand the kinetic energy in terms of the density perturbation , @xmath126 , @xmath127=t_0[n_0]+\int \ ! k({\mathbf{r}}-{\mathbf{r}}^\prime)\tilde{n}({\mathbf{r}})\tilde{n}({\mathbf{r^{\prime}}})\ , \mathrm{d}{\mathbf{r}}\mathrm{d}{\mathbf{r}}^\prime+\\ + \int \ ! l({\mathbf{r}},{\mathbf{r}}^\prime,{\mathbf{r}}^{\prime \prime})\tilde{n}({\mathbf{r}})\tilde{n}({\mathbf{r^{\prime}}})\tilde{n}({\mathbf{r^{\prime \prime}}})\ , \mathrm{d}{\mathbf{r}}\mathrm{d}{\mathbf{r}}^\prime\mathrm{d}{\mathbf{r}}^{\prime \prime}+ .... \end{gathered}\ ] ] we now transform the second term to an expansion in powers of the density gradient ( gradient corrections ) , by fourier expanding the kernel ( @xmath128 is the volume ) , @xmath129 from eqs . ( [ gn ] ) and ( [ k ] ) we have , using the convolution theorem , @xmath130=t_0[n_0]+\frac{1}{\omega } \sum_{{\mathbf{k } } } { \tilde k}({\mathbf{k}})\tilde{n}({\mathbf{k}})^2 + \dots.\ ] ] we are now looking for the long - wavelength limit and expand @xmath131 since @xmath132 is real , eq . ( [ k ] ) leads to @xmath133 , therefore , the expansion ( [ k3 ] ) contains only even powers of @xmath134 . from this , we obtain for @xmath135 , using ( [ k ] ) , @xmath136 substituting ( [ k4 ] ) into ( [ gn ] ) one obtains @xmath137 & = t_0[n_0 ] + \sum_{l=1 } t_{2l}[n ] , \\ t_{2l}[n ] & = \int \mathrm{d}{\mathbf{r}}\,a_{2l}\left[n\right ] \mid \nabla^l n({\mathbf{r}})\mid^{2 } , \label{term_2l } \ ] ] where we took into account that @xmath138 and identifiy @xmath139 . thus we have rewritten the kinetic energy in terms of a local term plus gradient corrections . it was shown in ref . @xcite that this energy functional leads directly , via functional differentiation with respect to the density profile , to quantum hydrodynamic equations with the local term yielding the fermi energy and the first gradient term giving rise to a bohm - type potential , @xmath89}{\delta n } = \gamma\ , \frac{\hbar^2}{8m}\left ( \left| \frac{\nabla n}{n } \right|^2 - 2 \frac{\nabla^2 n}{n } \right ) = \gamma \,v^{\rm mh}_b[n ] , \label{eq : grad = vb}\ ] ] where @xmath90 is the potential postulated by manfredi and haas @xcite . what remains is to establish the relation between the coefficient @xmath113 and the coeffient @xmath140 of the expansion ( [ term_2l ] ) . in the following we prove that , for zero temperature , to lowest order in the correlation energy , @xmath141 which means that @xmath142 . to evaluate the coefficient @xmath139 we recall that the energy associated with a charged particle density fluctuation can be expressed , in linear response , in terms of an effective potential ( external plus induced potential ) @xmath143=t_0[n_0]+\frac{1}{2\omega } \sum_{{\mathbf{k}}}\tilde{n}({\mathbf{k } } ) \tilde{u}_{\rm eff}({\mathbf{k}}).\ ] ] recalling the definition of the longitudinal polarization function , @xmath144 , eqs . ( [ gn2 ] ) and ( [ gn3 ] ) allow us to identify @xmath145 and to make use of known results for @xmath146 below . the lowest order many - body approximation for @xmath146 is the random phase approximation ( rpa ) which reads , for an arbitrary temperature @xcite @xmath147,\ ] ] where @xmath148 , @xmath149 , @xmath150 , @xmath151 , @xmath152 , and @xmath153 where @xmath154 , and @xmath155 is the chemical potential . to obtain the long - wavelength limit of @xmath121 , eq . ( [ k3 ] ) , we now expand the inverse of the static ( i.e. @xmath156 ) polarization function in the limit @xmath157 : @xmath158 h_1(\eta ) & = \frac{\sqrt{\theta}}{2 } \ , i_{-1/2}(\eta ) , \quad h_2(\eta)=\frac{1}{2\sqrt{\theta } } \,i_{-3/2}(\eta).\end{aligned}\ ] ] the first term in ( [ k5 ] ) yields the coefficient @xmath159 in eq . ( [ k3 ] ) which does not contribute to the kinetic energy since the integral of @xmath160 vanishes whereas the second term in ( [ k5 ] ) yields the coefficient @xmath140 : @xmath161,\theta\right ) & = -\frac{\pi e^2}{2k_f^2\chi_0 ^ 2 h_1(\eta ) } , \\ a_2\left(\left[n\right],\theta\right ) & = -\frac{\hbar^2 h_2(\eta)}{72m_e n h_1 ^ 2(\eta)}.\end{aligned}\ ] ] in the ground state , @xmath162 , we have the limits @xmath163 and @xmath164 , so @xmath165 , and ( [ a2 ] ) gives @xmath166 meaning that the coefficient in front of the bohm potential in eq . ( [ eq : grad = vb ] ) becomes @xmath142 . the temperature dependence of this coefficient was displayed in fig . [ fig : alpha_3 ] . finally , substituting @xmath140 from ( [ a2 ] ) to the formula ( [ gn4 ] ) , for the kinetic energy at finite temperature we have @xcite : @xmath167=t_0[n_0]+ \frac{1}{9 } \frac{\hbar^2}{8m_e } 4\beta^{3/2 } \int \ ! e_{f}^{3/2}(n)\frac{i_{-1/2}^{\prime}(\eta)}{i_{-1/2}^2(\eta ) } \frac{\mid \nabla n\mid^{2}}{n } \ , \mathrm{d}{\mathbf{r}},\ ] ] were we eliminated the fermi integral of order @xmath168 using the formula @xmath169 . now consider the higher order ( @xmath171 ) terms of the gradient expansion ( [ gn4 ] ) . in the long - wavelength limit , @xmath172 the kernel @xmath121 can be written as : @xmath173 where @xmath174,\theta\right)=\frac{\theta^{-i}}{2i+1 } \frac{i_{-i-1/2}(\eta)}{i_{-1/2}(\eta)}. \ ] ] expanding ( [ k6 ] ) it is straightforward to obtain the higher ( even ) order terms of the gradient correction : @xmath175=\frac{\hbar^2 \pi^2}{16 m_e } \int \ ! \frac{b_1 ^ 2-b_2}{\theta\left[n\right]^{1/2}k_f^5i_{-1/2}(\eta)}\mid \nabla^2 n\mid^{2 } \ , \mathrm{d}{\mathbf{r } } , \nonumber\ ] ] @xmath176=\frac{\hbar^2 \pi^2}{64 m_e } \int \ ! \frac{b_1 ^ 3 - 2b_1b_2+b_3}{\theta\left[n\right]^{1/2}k_f^7i_{-1/2}(\eta)}\mid \nabla^3 n\mid^{2 } \ , \mathrm{d}{\mathbf{r } } , \nonumber\ ] ] @xmath177=\frac{\hbar^2 \pi^2}{256 m_e}\times \\ \int \ ! \frac{b_1 ^ 4 - 3b_1 ^ 2b_2+b_2 ^ 2 + 2b_1b_3-b_4}{\theta\left[n\right]^{1/2}k_f^9i_{-1/2}(\eta)}\mid \nabla^4 n\mid^{2 } \ , \mathrm{d}{\mathbf{r}}. \nonumber \end{gathered}\ ] ] in the zero - temperature limit , @xmath178 and @xmath179 were obtained by hodges @xcite and murphy @xcite , respectively . for finite temperature these terms and , additionally , @xmath180 are given here for the first time . here we do not discuss the convergence of this gradient series . obviously , before using higher order terms one should verify that all terms of the same order arising from three - particle correlations [ third term in ( [ gn ] ) ] are included as well . further , it is well possible that there exist comparable terms arising from correlation effects beyond the rpa , e.g. @xcite . finally we note that these results apply only to three - dimensional plasmas . results for 1d and 2d systems were discussed in ref . extensions of these results to finite temperatures are given in ref . 10 t. kluge , c. gutt , l. huang , j. metzkes , u. schramm , m. bussmann , and t.e cowan , _ using xfels for probing of complex interaction dynamics of ultra - intense lasers with solid matter_. phys . plasmas * 21 * , 033110 ( 2014 ) . m.r . _ , _ experimental demonstration of fusion - relevant conditions in magnetized liner inertial fusion _ , phys . rev . lett . * 113 * , 155003 ( 2014 ) ; p.f . et al . _ , . lett . * 113 * , 155004 ( 2014 ) . h. khlert , g. kalman , and m. bonitz , _ dynamics of strongly correlated confined plasmas _ , e * 90 * , 011101(r ) ( 2014 ) ; _ linear fluid theory for weakly inhomogeneous plasmas with strong correlations _ plasma phys . * 55 * , 352 ( 2015 ) . p. k. shukla , and b. eliasson , _ novel attractive force between ions in quantum plasmas _ , 108 * , 165007 ( 2012 ) ; _ erratum _ , _ ibid _ : * 108 * , 219902 ( 2012 ) ; _ erratum _ , _ ibid _ : * 109 * , 019901 ( 2012 ) . m. bonitz , e. pehlke , and t. schoof,_comment on discussion on novel attractive force between ions in quantum plasmas failure of simulations based on a density functional approach " _ , phys . scr . * 88 * , 057001 ( 2013 ) govind s. krishnaswami , rajaram nityananda , abhijit sen , anantanarayanan thyagaraja , _ a critique of recent semi - classical spin - half quantum plasma theories _ , contrib . plasma phys . ( 2014 ) , doi : 10.1002/ctpp.201400043 , arxiv:1407.6865 sergey v. vladimirov , _ plasma and potential distributions of moving objects in classical and quantum plasmas _ , _ plasma phys . control . fusion _ * 53 * , 074005 ( 2011 ) . n. crouseilles , p .- a . hervieux , g. manfredi , _ quantum hydrodynamic models for nonlinear electron dynamics in thin metal films _ , b * 78 * , 155412 ( 2008 ) . in the original formulas @xmath181 are denoted by @xmath182 and @xmath183 , respectively . a. moldabekov , p. ludwig , j.p . joost , m. bonitz , and t. ramazanov , _ dynamical screening and wake effects in classical , quantum , and ultrarelativistic plasmas , _ , contrib . plasma phys . * 55 * , 186 ( 2015 ) . a. s. kompaneets , e. s. pavlovskii , zh . 31 * , 427 ( 1956 ) . d. a. kirzhnitz , _ quantum corrections to the thomas - fermi equation _ , zh . eksp . . fiz . * 32 * , 115 ( 1957)(sov jetp , * 5 , 64 * ) . sidney golden , _ statistical theory of many - electron systems . general considerations pertaining to the thomas - fermi theory _ , phys . rev . * 105 * , 604 ( 1957 ) . p. hohenberg , w. kohn , _ inhomogeneous electron gas _ , phys . rev . * 136 * , b864 ( 1964 ) . n. david mermin , _ thermal properties of the inhomogeneous electron gas _ , phys . rev . * 137 * , a1441 ( 1965 ) . kirzhnitz , yu.e . lozovik , and g. v. shpatakovskaya , _ statistical model of matter _ , usp . nauk * 117 * , 3 - 47 ( 1975 ) [ sov . , * 18 * , no . 9 , ( 1976 ) ] f. perrot , _ gradient correction to the statistical electronic free energy at nonzero temperatures : application to equation - of - state calculations _ , phys . rev . a , * 20 * , 586 ( 1979 ) . weito yang , _ gradient correction in thomas - fermi theory _ a , * 34 * , 4575 ( 1986 ) . g. manfredi , and f. haas , _ self - consistent fluid model for a quantum electron gas _ , phys . b * 64 * , 075316 ( 2001 ) . nestor r. arista and werner brandt , _ dielectric response of quantum plasmas in thermal equilibrium _ , a. , * 29 * , 1471 ( 1984 ) . h. xu , and j .- hansen , _ plasma - insulator transition of spin - polarized hydrogen _ , e * 60 * , r9(r ) ( 1999 ) . k. yonei , and y. tomishima , j. phys . japan * 20 * , 1051 ( 1965 ) . gross , and r.m . dreizler , phys . a * 20 * , 1789 ( 1979 ) . c. h. hodges , _ quantum corrections to the thomas - fermi approximation the kirzhnits method _ , can . j. phys . * 51 * , 1428 ( 1973 ) . d. r. murphy , _ sixth - order term of the gradient expansion of the kinetic - energy density functional _ , phys . a. , * 24 * , 1682 ( 1984 ) . kwong , and m. bonitz , _ real - time kadanoff - baym approach to plasma oscillations in a correlated electron gas _ , _ phys . lett . _ * 84 * , 1768 ( 2000 ) . a. moldabekov , t. s. ramazanov , and m. bonitz , to be published .

the tohoku - pacific ocean earthquake happened in japan on march 11 , 2011 , and many people failed to evacuate and lost their lives due to severe attack by tsunamis . from the viewpoint of disaster prevention from city planning and evacuation planning , it has now become extremely important to establish effective evacuation planning systems against large scale disasters in japan . in particular , arrangements of tsunami evacuation buildings in large japanese cities near the coast has become an urgent issue . to determine appropriate tsunami evacuation buildings , we need to consider where evacuation buildings are located and how to partition a large area into small regions so that one evacuation building is designated in each region . this produces several theoretical issues to be considered . among them , this paper focuses on the location problem of the evacuation building assuming that we fix the region such that all evacuees in the region are planned to evacuate to this building . in this paper , we consider the simplest case for which the region consists of a single road . + in order to represent the evacuation , we consider the _ dynamic _ setting in graph networks , which was first introduced by ford et al . @xcite . in a graph network under the dynamic setting , each vertex is given supply and each edge is given length and capacity which limits the rate of the flow into the edge per unit time . we call such networks under the dynamic setting _ dynamic networks_. unlike in static networks , the time required to move supply from one vertex to a sink can be increased due to congestion caused by the capacity constraints , which require supplies to wait at vertices until supplies preceding them have left . in this paper , we consider the flow on dynamic networks as continuous , that is , each input value is given as a real number , and supply , flow and time are defined continuously . then each supply can be regarded as fluid , and edge capacity is defined as the maximum amount of supply which can enter an edge per unit time . the _ 1-sink location problem in dynamic networks _ is defined as the problem which requires to find the optimal location of a sink in a given dynamic network so that all supplies are sent to the sink as quickly as possible . + in order to evaluate an evacuation , we can naturally consider two types of criteria : _ completion time criterion _ and _ total time criterion_. in this paper we adopt the latter one ( for the former one , refer to @xcite ) . we here define a _ unit _ as an infinitesimally small portion of supply . given a sink location @xmath1 in a dynamic network , let us consider an evacuation to @xmath1 starting at time @xmath4 and define the _ evacuation time _ of a unit to @xmath1 as the time at which the unit reaches @xmath1 in the evacuation . the total time for the evacuation to @xmath1 is defined as the sum of evacuation times over all infinitesimal units to @xmath1 . then , the minimum total time for all possible evacuations to @xmath1 could be the criterion for the optimality of sink location , which we adopt . given a dynamic network , we define the _ 1-median problem _ as the problem which requires to find a sink location minimizing the minimum total time , and the optimal solution is called the _ median_. + although the above criterion is reasonable for the sink location , it may not be practical since the number of evacuees in an area may vary depending on the time ( e.g. , in an office area in a big city , there are many people during the daytime on weekdays while there are much less people on weekends or during the night time ) . so , in order to take into account the uncertainty of population distribution , we consider the _ maximum regret _ for a sink location as another evaluation criterion assuming that for each vertex , we only know an interval of vertex supply . then , the _ minimax regret 1-median problem in dynamic path networks _ is formulated as follows . a particular assignment of supply to each vertex is called a _ scenario_. here , for a sink location @xmath1 and a scenario @xmath0 , we denote the minimum total time by @xmath5 . also let @xmath6 denote the median under @xmath0 . the problem can be understood as a 2-person stackelberg game as follows . the first player picks a sink location @xmath1 and the second player chooses a scenario @xmath0 that maximizes the _ regret _ defined as @xmath7 . the objective of the first player is to choose @xmath1 that minimizes the maximum regret . + related to the minimax regret facility location in graph networks , especially for trees , some efficient algorithms have been presented by @xcite . for dynamic networks , cheng et al . @xcite first studied the _ minimax regret 1-center problem _ in path networks , which requires to find a sink location in a path that minimizes the maximum regret where the completion time criterion is adopted instead of the total time one . they presented an @xmath8 time algorithm . higashikawa et al . @xcite improved the time bound by @xcite to @xmath9 , and also wang @xcite independently achieved the same time bound of @xmath9 with better space complexity . very recently , bhattacharya et al . @xcite have improved the time bound to @xmath10 . the above problem was extended to the multiple sink location version by arumugam et al . @xcite . for the minimax regret @xmath11-center problem in dynamic path networks with uniform capacity , they presented an @xmath12 time algorithm , and this time bound was improved to @xmath13 recently @xcite . on the other hand , for dynamic tree networks , only the minimax regret 1-center problem was solved in @xmath14 time @xcite . this paper first considers the minimax regret median problem in dynamic networks while all the above works for dynamic networks treated center problems . in this paper , we address the minimax regret 1-median problem in dynamic path networks with uniform capacity and present an @xmath2 time algorithm . let @xmath15 be an undirected path with ordered vertices @xmath16 and edges @xmath17 where @xmath18 for @xmath19 . let @xmath20 be a dynamic network with the underlying path graph @xmath21 ; @xmath22 is a function that associates each edge @xmath23 with positive length @xmath24 , @xmath25 is a function that associates each vertex @xmath26 with positive weight @xmath27 , amount of supply at @xmath26 ; @xmath28 is the capacity , a positive constant representing the amount of supply which can enter an edge per unit time ; @xmath29 is a positive constant representing the time required for a flow to travel a unit distance . in our model , instead of the weight function @xmath25 on vertices , we are given the weight interval function @xmath30 that associates each vertex @xmath31 with an interval of supply @xmath32 $ ] with @xmath33 . we call such a network @xmath34 with path structures _ a dynamic path network under uncertain supplies_. in the following , we write @xmath35 to indicate that a point is a vertex of @xmath21 or lies on one of the edges of @xmath21 . for any point @xmath35 , we abuse this notation by also letting @xmath36 denote the distance from @xmath37 to @xmath36 . informally we can regard @xmath21 as being embedded on a real line with @xmath38 . for two points @xmath39 satisfying @xmath40 , let @xmath41 $ ] ( resp . @xmath42 , @xmath43 $ ] and @xmath44 ) denote an interval in @xmath21 consisting of all points @xmath45 such that @xmath46 ( resp . @xmath47 , @xmath48 and @xmath49 ) . let @xmath50 denote the cartesian product of all @xmath51 for @xmath52 : @xmath53 an element of @xmath50 , i.e. , a particular assignment of weight to each vertex , is called a _ scenario_. given a scenario @xmath54 , we denote by @xmath55 the weight of a vertex @xmath26 under @xmath0 . in our model , the supply is defined continuously . we define a unit as an infinitesimally small portion of supply . given a sink location @xmath45 and a scenario @xmath54 , without loss of generality , an evacuation to @xmath1 under @xmath0 is assumed to satisfy the following assumptions . when a unit arrives at a vertex @xmath56 on its way to @xmath1 , it has to wait for the departure if there are already some units waiting for leaving @xmath56 . all units waiting at @xmath56 for leaving @xmath56 are processed in the first - come first - served manner . for a given @xmath45 and @xmath54 , let us consider an evacuation to @xmath1 under @xmath0 starting at time @xmath4 and define the evacuation time of a unit to @xmath1 under @xmath0 as the time at which the unit reaches @xmath1 . let @xmath5 denote the sum of evacuation times over all infinitesimal units to @xmath1 under @xmath0 . also let @xmath57 ( resp . @xmath58 ) denote the sum of evacuation times to @xmath1 under @xmath0 for all units on @xmath59 ( resp . then , @xmath5 is obviously the sum of @xmath57 and @xmath58 , i.e. , @xmath61 without loss of generality , we assume @xmath62 and @xmath63 . we will show the formula of @xmath5 that has been proved in @xcite . suppose that @xmath1 is located in an open interval @xmath64 with @xmath65 , i.e. , @xmath66 . we here show only the formula of @xmath57 ( the case of @xmath58 is symmetric ) . first , let us define the vertex indices @xmath67 inductively as @xmath68 where @xmath69 . obviously @xmath70 holds . we then call a set of all units on @xmath71 $ ] _ i - th cluster _ , and a vertex @xmath72 the _ head _ of @xmath73-th cluster ( see fig . [ fig : cl ] ) . where @xmath73-th cluster is headed by a vertex @xmath72,width=340 ] also , for @xmath74 , we define @xmath75 as @xmath76 , which is called the weight of @xmath73-th cluster . the interpretation can be derived from @xcite as follows . the first unit on each @xmath72 does not encounter any congestion on its way to @xmath1 . here , although @xmath77 may not be uniquely determined by ( [ ms2.5 ] ) , we choose the maximum index as @xmath77 . by this assumption and ( [ ms2.5 ] ) , the following inequality holds for @xmath78 : @xmath79 in other words , even if we transform the input so that all units on @xmath80 are moved to @xmath72 for @xmath74 , the sum of evacuation times over @xmath81 never changes . in the following , we call such a transformation the _ left - clustering _ for @xmath1 and clusters obtained by the left - clustering for @xmath1 _ left clusters _ for @xmath1 ( the right - clustering and right clusters are symmetric ) . then as in @xcite , @xmath57 is written as @xmath82 for a scenario @xmath54 , let @xmath6 be a point in @xmath21 that minimizes @xmath5 over @xmath45 , called the _ median _ under @xmath0 . we now define the _ regret _ for @xmath1 under @xmath0 as @xmath83 moreover , we also define the _ maximum regret _ for @xmath1 as @xmath84 if @xmath85 , we call @xmath86 the _ worst case scenario _ for @xmath1 . the goal is to find a point @xmath87 , called the _ minimax regret median _ , that minimizes @xmath88 over @xmath45 , i.e. , the objective is to @xmath89 we here show some properties on the 1-median problem in a dynamic path network @xmath90 when a scenario @xmath54 is given , which were basically presented in @xcite . we first introduce the following two lemmas . for a scenario @xmath54 , @xmath6 is at a vertex in @xmath91 . [ lem : ms ] for a scenario @xmath54 , all @xmath92 over @xmath52 can be computed in @xmath10 time in total . [ lem : pht ] we then can see a corollary of these lemmas . for a scenario @xmath54 , @xmath6 and @xmath93 can be computed in @xmath10 time . [ coro : mst ] now let us look at the formula of ( [ eq : ms3 ] ) . even if @xmath1 is moving on an edge @xmath23 ( not including endpoints @xmath26 and @xmath94 ) , the formation of left clusters for @xmath1 does not change over @xmath95 . therefore , @xmath57 is a linear function of @xmath95 , and symmetrically , @xmath58 is also a linear function . for @xmath19 , letting @xmath96 and @xmath97 be the values such that for @xmath95 , @xmath98 we can derive the following lemma from @xcite . for a scenario @xmath54 , all @xmath96 and @xmath97 over @xmath19 can be computed in @xmath10 time in total . [ lem : abt ] in this section , we show the important properties which worst case scenarios have . in our problem , a main difficulty lies in evaluating @xmath99 over @xmath54 to compute @xmath88 even for a fixed @xmath1 since the size of @xmath50 is infinite . we thus aim to find a scenario set with a finite size ( in particular , a polynomial size ) which includes a worst case scenario for any @xmath45 . in order to do this , we introduce a new concept , the _ gap _ between two points @xmath100 under a scenario @xmath54 , defined by @xmath101 by lemma [ lem : ms ] and the definition of ( [ eq : rg ] ) , we have @xmath102 and by ( [ eq : rmax ] ) and ( [ eq : rg2 ] ) , @xmath103 from ( [ eq : rmax2 ] ) , if we can compute @xmath104 for a fixed pair @xmath105 , @xmath88 can also be computed by repeating the same maximization over @xmath106 . we call a scenario that maximizes @xmath107 for a fixed @xmath108 a worst case scenario for @xmath108 . in the following , we show a scenario set of size @xmath10 that includes a worst case scenario for a fixed @xmath108 , which implies a scenario set of size @xmath109 that includes a worst case scenario for a fixed @xmath1 . we first introduce the concept of the _ bipartite scenario _ , which was originally introduced as the _ dominant scenario _ in @xcite . let us consider a scenario @xmath54 . a scenario @xmath0 is said to be _ left - bipartite _ ( resp . _ right - bipartite _ ) if @xmath110 ( resp . @xmath111 ) over @xmath112 and @xmath113 ( resp . @xmath114 ) over @xmath115 for some @xmath19 . obviously the number of such scenarios is @xmath10 . the authors of @xcite treated the minimax regret 1-center problem in dynamic path networks , which requires to find a sink location in a path that minimizes the maximum regret similarly defined as ( [ eq : rmax ] ) where the completion time criterion is adopted instead of the total time one . they proved that for any point in an input path , at least one worst case scenario is left - bipartite or right - bipartite . we here introduce the concept of the _ pseudo - bipartite scenario_. a scenario @xmath0 is said to be _ left - pseudo - bipartite _ ( resp . _ right - pseudo - bipartite _ ) if @xmath110 ( resp . @xmath111 ) over @xmath116 and @xmath113 ( resp . @xmath114 ) over @xmath115 for some @xmath117 . in this definition , we do not care about the weight of a vertex @xmath26 , called the _ intermediate vertex_. given a pseudo - bipartite scenario with the intermediate vertex @xmath26 , we call intervals @xmath118 and @xmath119 $ ] the _ left part _ and the _ right part _ , respectively . let @xmath120 ( resp . @xmath121 ) denote a set of all left - pseudo - bipartite scenarios ( resp . right - pseudo - bipartite scenarios ) . we then prove the following lemma ( the proof is given in appendix a ) . given a pair @xmath105 satisfying @xmath122 ( resp . @xmath123 ) , there exists a worst case scenario for @xmath108 belonging to @xmath120 ( resp . @xmath121 ) such that @xmath124 ( resp . @xmath1 ) is in the left part and @xmath1 ( resp . @xmath124 ) is in the right part . [ lem : wcsxy ] by lemma [ lem : wcsxy ] , we studied the property of a worst case scenario for a fixed @xmath105 , however the sizes of @xmath120 and @xmath121 are still infinite since the weight of the intermediate vertex in a pseudo - bipartite scenario is not fixed . in the rest of this section , we focus on the weight of the intermediate vertex in a pseudo - bipartite scenario which is worst for @xmath108 . given a pair @xmath105 satisfying @xmath122 , let us consider a scenario @xmath125 such that the intermediate vertex is @xmath26 and @xmath126 . suppose that the weight of @xmath26 is set as the minimum , i.e. , @xmath127 . performing the right - clustering for @xmath124 under @xmath0 ( mentioned in section [ subsec : to ] ) , we will get right clusters for @xmath124 such that for @xmath128 , the head of @xmath22-th cluster is @xmath129 and the weight of @xmath22-th cluster is @xmath130 . then , suppose that the intermediate vertex @xmath26 belongs to @xmath131-th cluster . now let us increase the weight of @xmath26 , little by little , without changing the weight of any other vertex . let @xmath132 be a scenario in @xmath120 such that the intermediate vertex is @xmath26 whose weight is @xmath133 $ ] . suppose that when the weight of @xmath26 reaches some value @xmath134 , the following equality holds : @xmath135 note that @xmath136 corresponds to the weight of @xmath131-th cluster under @xmath137 . at that moment , referring to ( [ eq : ms2.6 ] ) , @xmath138-th cluster is merged to @xmath131-th cluster . we then call @xmath137 a _ critical left - pseudo - bipartite scenario _ for @xmath124 . also , @xmath139 and @xmath140 are assumed to be critical left - pseudo - bipartite scenarios for @xmath124 even if any merge does not occur at those moments . _ critical right - pseudo - bipartite scenarios _ for @xmath124 are symmetrically defined . let @xmath141 denote a set of all critical left - pseudo - bipartite scenarios and critical right - pseudo - bipartite scenarios for @xmath124 , and @xmath142 . we will show two lemmas ( the proof of lemma [ lem : wcsxy2 ] is given in appendix b ) . given a pair @xmath105 , there exists a worst case scenario for @xmath108 belonging to @xmath141 . [ lem : wcsxy2 ] for a vertex @xmath106 , the size of @xmath141 is @xmath10 , and all scenarios in @xmath141 can be computed in @xmath10 time . [ lem : csy ] we first prove that the number of critical left - pseudo - bipartite scenarios for @xmath124 is @xmath10 ( the case of critical right - pseudo - bipartite scenarios is symmetric ) . suppose that @xmath143 . for @xmath144 and @xmath133 $ ] , let @xmath145 be a scenario in @xmath120 such that the intermediate vertex is @xmath26 whose weight is @xmath25 . here , let us define the order between two scenarios @xmath145 and @xmath146 : @xmath147 holds if and only if ( a ) @xmath148 or ( b ) @xmath149 and @xmath150 . for @xmath144 , we also define @xmath151 and @xmath152 as follows . let @xmath151 be the number of critical left - pseudo - bipartite scenarios for @xmath124 such that the intermediate vertex is @xmath26 ( including @xmath153 and @xmath154 ) . let @xmath152 be , under a scenario @xmath154 , the number of right clusters for @xmath124 that follow a cluster including @xmath26 . let us consider computing all critical left - pseudo - bipartite scenarios for @xmath124 in ascending order , and suppose that the weight of @xmath26 now increases from @xmath155 to @xmath156 . while it increases , since @xmath157 critical left - pseudo - bipartite scenarios for @xmath124 occur ( except @xmath153 and @xmath154 ) , and at each such scenario , one or more clusters are merged into one of @xmath26 , at least @xmath157 clusters are merged into one of @xmath26 in total . we thus have @xmath158 for @xmath144 , i.e. , @xmath159 note that the total number of critical left - pseudo - bipartite scenarios for @xmath124 is exactly @xmath160 . by ( [ eq : lem1.21 ] ) , we have @xmath161 which is @xmath10 since @xmath162 and @xmath163 . in the rest of the proof , we show that all critical left - pseudo - bipartite scenarios for @xmath143 can be computed in @xmath10 time . recall that all critical left - pseudo - bipartite scenarios for @xmath124 are computed in ascending order . the algorithm first gets @xmath164 , and performs the right clustering for @xmath124 under @xmath164 . as claimed in @xcite , it is easy to see that the right clustering for a fixed @xmath124 can be done in @xmath10 time . suppose that for particular @xmath144 and @xmath165 $ ] , @xmath166 is critical for @xmath124 , and the algorithm has already obtained @xmath166 and the right clusters for @xmath124 . we then show how to compute the subsequent critical left - pseudo - bipartite scenario . let @xmath167 be a right cluster for @xmath124 including @xmath26 and @xmath168 be a right cluster for @xmath124 immediately following @xmath167 . also , let @xmath169 ( resp . @xmath170 ) be the index of a vertex that corresponds to the head of @xmath167 ( resp . @xmath168 ) , and @xmath171 ( resp . @xmath172 ) be the weight of @xmath167 ( resp . @xmath168 ) . there are two cases : [ case 1 ] @xmath173 ; [ case 2 ] @xmath174 . for case 2 , we notice that @xmath175 is equivalent to @xmath154 . therefore , this case immediately results in case 1 by letting @xmath73 be @xmath176 and @xmath134 be @xmath177 ( although a right cluster for @xmath124 including @xmath94 may be @xmath168 , not @xmath167 ) . we thus consider only case 1 in the following . the algorithm will compute the subsequent critical left - pseudo - bipartite scenario @xmath178 where @xmath179 satisfies @xmath180 . in order to compute @xmath179 , the algorithm test if there exists @xmath181 $ ] such that @xmath182 which is similar to ( [ eq : cpbs1 ] ) . if yes , for such @xmath25 , the algorithm returns @xmath183 and updates the right clusters for @xmath124 by merging @xmath168 into @xmath167 . otherwise , @xmath184 is just returned . such testing and updating are done in @xmath185 time . since the number of critical left - pseudo - bipartite scenarios for @xmath124 is @xmath10 and each of those is computed in @xmath185 time , we completes the proof . by ( [ eq : rmax2 ] ) , we have a corollary of lemma [ lem : wcsxy2 ] . given a point @xmath45 , there exists a worst case scenario for @xmath1 belonging to @xmath186 . [ coro : wcsx ] also , a corollary of lemma [ lem : csy ] immediately follows . the size of @xmath186 is @xmath109 , and all scenarios in @xmath186 can be computed in @xmath109 time . [ coro : cusy ] in this section , we show an algorithm that computes the minimax regret median , which minimizes @xmath88 over @xmath45 . the algorithm basically consists of two phases : + * [ phase 1 ] * compute @xmath187 over @xmath52 , and + * [ phase 2 ] * compute @xmath188 over @xmath19 . + after these , the algorithm evaluates all the @xmath189 values obtained and finds the minimax regret median in @xmath10 time . by corollary [ coro : wcsx ] , we only have to consider scenarios in @xmath186 to compute @xmath88 for any @xmath45 . therefore , the algorithm computes all scenarios in @xmath186 in advance , which can be done in @xmath109 time by corollary [ coro : cusy ] . subsequently , it computes all the values @xmath93 over @xmath190 for phase 1 and phase 2 . by corollaries [ coro : mst ] and [ coro : cusy ] , this can be done in @xmath2 time in total . first let us see details in phase 1 . for a fixed scenario @xmath190 , since all @xmath92 over @xmath52 can be computed in @xmath10 time by lemma [ lem : pht ] and @xmath93 has already been computed before phase 1 , all @xmath191 over @xmath52 can also be computed in @xmath10 time ( refer to ( [ eq : rg ] ) ) . after the algorithm obtains @xmath192 over @xmath190 in @xmath2 time , for each @xmath52 , @xmath191 over @xmath190 are evaluated to obtain @xmath187 . thus , it is easy to see that phase 1 can be done in @xmath2 time in total . we next focus on phase 2 . as mentioned at the end of section [ subsec : kp ] , for a fixed scenario @xmath190 , @xmath5 is a linear function of @xmath95 for each @xmath19 ( not including @xmath26 and @xmath94 ) . therefore , @xmath99 is also linear for @xmath95 for each @xmath73 . referring to ( [ eq : aibi ] ) , a function @xmath99 on an edge @xmath23 is written as @xmath193 recall that @xmath93 has already been computed . then , by lemma [ lem : abt ] , @xmath99 on @xmath23 over @xmath19 can be computed in @xmath10 time . after the algorithm does the same computation over @xmath190 in @xmath2 time , on each edge @xmath23 , we have @xmath109 linear functions @xmath99 over @xmath190 . by the definition of ( [ eq : rmax ] ) , @xmath194 can be obtained by solving a linear programming problem in two dimensions with @xmath109 constraints , i.e. , @xmath195 this problem can be solved in @xmath109 time by @xcite . repeating the same operations over @xmath19 , phase 2 is completed in @xmath2 time . the minimax regret 1-median problem in dynamic path networks with uniform capacity can be solved in @xmath2 time . in this paper , we address the minimax regret 1-median problem in dynamic path networks with uniform capacity and present an @xmath2 time algorithm . additionally , this is the first work that treats the minimax regret facility location problem in dynamic networks where the total time criterion is adopted . two natural questions immediately follow . the first one is whether we can reduce the number of scenarios to be considered . the other one is whether we can extend the problem to the @xmath11-median version with @xmath196 , or the problem in more general networks . 99 g. p. arumugam , j. augustine , m. j. golin , p. srikanthan , a polynomial time algorithm for minimax - regret evacuation on a dynamic path " , , arxiv:1404.5448 . i. averbakh and o. berman , `` algorithms for the robust @xmath197-center problem on a tree '' , , 123(2 ) , pp . 292 - 302 , 2000 . b. bhattacharya and t. kameda , a linear time algorithm for computing minmax regret @xmath197-median on a tree " , ( cocoon 2012 ) , lncs 7434 , pp . 1 - 12 , 2012 . b. bhattacharya , t. kameda , improved algorithms for computing minmax regret 1-sink and 2-sink on path network " , ( cocoa 2014 ) , lncs 8881 , pp . 146 - 160 , 2014 . b. bhattacharya , t. kameda and z. song , a linear time algorithm for computing minmax regret @xmath197-median on a tree network " , , pp . 1 - 20 , 2013 . g. s. brodal , l. georgiadis and i. katriel , an @xmath9 version of the averbakh - berman algorithm for the robust median of a tree " , , 36(1 ) , pp . 14 - 18 , 2008 . b. chen and c. lin , `` minmax - regret robust 1-median location on a tree '' , , 31(2 ) , pp . 93 - 103 , 1998 . s. w. cheng , y. higashikawa , n. katoh , g. ni , b. su and y. xu , minimax regret 1-sink location problems in dynamic path networks " , ( tamc 2013 ) , lncs 7876 , pp . 121 - 132 , 2013 . e. conde , `` a note on the minmax regret centdian location on trees '' , , 36(2 ) , pp . 271 - 275 , 2008 . m. e. dyer , `` linear time algorithms for two- and three - variable linear programs '' , , 13(1 ) , pp . 31 - 45 , 1984 . l. r. ford jr . , d. r. fulkerson , constructing maximal dynamic flows from static flows " , , 6 , pp . 419 - 433 , 1958 . y. higashikawa , studies on the space exploration and the sink location under incomplete information towards applications to evacuation planning " , , kyoto university , 2014 . y. higashikawa , j. augustine , s. w. cheng , n. katoh , g. ni , b. su and y. xu , minimax regret 1-sink location problem in dynamic path networks " , , doi : 10.1016/j.tcs.2014.02.010 , 2014 . y. higashikawa , m. j. golin , n. katoh , minimax regret sink location problem in dynamic tree networks with uniform capacity " , ( walcom 2014 ) , lncs 8344 , pp . 125 - 137 , 2014 . y. higashikawa , m. j. golin , n. katoh , multiple sink location problems in dynamic path networks " , ( aaim 2014 ) , lncs 8546 , pp . 149 - 161 , 2014 . y. higashikawa , m. j. golin , n. katoh , minimax regret sink location problem in dynamic tree networks with uniform capacity " , , 18(4 ) , pp . 539 - 555 , 2014 . y. higashikawa , m. j. golin , n. katoh , multiple sink location problems in dynamic path networks " , , doi:10.1016/j.tcs.2015.05.053 . s. mamada , t. uno , k. makino and s. fujishige , an @xmath8 algorithm for the optimal sink location problem in dynamic tree networks " , , 154(16 ) , pp . 2387 - 2401 , 2006 . h. wang , minmax regret 1-facility location on uncertain path networks " , , lncs 8283 , pp . 733 - 743 , 2013 . _ proof of claim [ clm:2 ] . _ we here assume @xmath199 : if @xmath200 , the proof is straightforward . let @xmath201 be a worst case scenario for @xmath108 . if there are more than one worst case scenario , we choose the one such that all weights are lexicographically maximized in the order of ascending indices among all worst case scenarios . we prove by contradiction : suppose @xmath202 for some vertex @xmath198 $ ] . under a scenario @xmath201 , let us perform the left - clustering for @xmath1 and @xmath124 , respectively ( refer to section [ subsec : to ] ) . performing the left - clustering for @xmath1 , let @xmath203 be a left cluster for @xmath1 including @xmath26 , @xmath204 be the index of a vertex that corresponds to the head of @xmath203 , and @xmath205 be the weight of @xmath203 . also , performing the left - clustering for @xmath124 , @xmath167 , @xmath169 and @xmath171 are similarly defined . when @xmath167 evacuates to @xmath1 , by the definition of a cluster , the first unit of @xmath167 does not encounter any congestion in the interval @xmath206 , however , it may encounter congestions at some vertices in @xmath207 . if it does not encounter any congestion in @xmath208 , @xmath167 is never merged to any other cluster for @xmath1 , which implies @xmath209 and @xmath210 . otherwise , @xmath167 is eventually merged into a left cluster for @xmath1 headed by a vertex @xmath211 , i.e. , @xmath203 . we thus consider two cases : [ case 1 ] @xmath212 and @xmath210 ; [ case 2 ] @xmath213 and @xmath214 . now let @xmath215 be a scenario obtained from @xmath201 by increasing the weight of @xmath26 by infinitesimally small @xmath216 , i.e. , @xmath217 and @xmath218 for @xmath219 . we then show the following claim for the proof of claim [ clm:2 ] . _ proof of claim [ clm:1 ] . _ performing the left - clustering for @xmath1 under @xmath201 , let @xmath220 be a left cluster for @xmath1 immediately following @xmath203 , and @xmath221 be the index of a vertex that corresponds to the head of @xmath220 . referring to ( [ eq : ms2.6 ] ) , the following inequality holds : @xmath222 then , for a sufficiently small @xmath223 , we have @xmath224 the inequality of ( [ eq : clm1.1 ] ) means that after performing the left - clustering for @xmath1 under @xmath201 , the first unit of @xmath220 does not catch up with the last unit of @xmath203 at @xmath225 , and by ( [ eq : clm1.2 ] ) , this remark also holds even for @xmath215 . under @xmath215 , if @xmath220 is not merged to @xmath203 , any other merge never occurs . thus , the formation of left clusters for @xmath1 does not change , and similarly , it does not change for @xmath124 . by claim [ clm:1 ] and the definitions of ( [ eq : ms1 ] ) and ( [ eq : ms3 ] ) , we have @xmath226 and similarly , @xmath227 also by the definition of ( [ eq : gap ] ) , we have @xmath228 from ( [ eq : lem1.1 ] ) , ( [ eq : lem1.2 ] ) and ( [ eq : lem1.3 ] ) , we can derive @xmath229 if case 2 occurs , performing the left - clustering for @xmath1 merges @xmath167 into @xmath203 as mentioned above , and then , all units on @xmath231 $ ] are also merged into @xmath203 . therefore , if we consider an input such that all units on @xmath232 are moved to @xmath225 , the first unit of @xmath167 must catch up with the last unit of supply at @xmath225 , i.e. , @xmath233 since @xmath203 includes @xmath167 and all units on @xmath231 $ ] , we have @xmath234 note that in ( [ eq : lem1.6 ] ) , the left side is less than the right side when @xmath203 also includes some clusters for @xmath124 following @xmath167 . from ( [ eq : lem1.4 ] ) , ( [ eq : lem1.5 ] ) and ( [ eq : lem1.6 ] ) , we can derive @xmath235 which contradicts that @xmath201 is a worst case scenario for @xmath108 . if we consider a worst case scenario for @xmath108 such that the weight of every vertex @xmath198 $ ] is @xmath156 and weights of all other vertices in @xmath236 $ ] are lexicographically minimized in the order of descending indices , the following claim is also proved in a similar manner as in the proof of claim [ clm:2 ] . given a pair @xmath105 satisfying @xmath122 , there exists a worst case scenario for @xmath108 such that the weight of every vertex @xmath198 $ ] is @xmath156 and the weight of every vertex @xmath237 $ ] is @xmath155 . [ clm:3 ] now , let @xmath238 be a worst case scenario for @xmath108 such that the weight of every vertex @xmath198 $ ] is @xmath156 , the weight of every vertex @xmath237 $ ] is @xmath155 , and weights of all other vertices in the open interval @xmath239 are lexicographically maximized in the order of ascending indices . then , @xmath240 can be proved . we prove by contradiction : there exist two vertices @xmath241 satisfying @xmath242 such that @xmath243 and @xmath244 . let @xmath245 be a scenario obtained from @xmath238 by increasing the weight of @xmath26 by infinitesimally small @xmath216 and decreasing the weight of @xmath246 by the same @xmath247 , i.e. , @xmath248 , @xmath249 and @xmath250 for @xmath251 . then , we immediately see @xmath252 and @xmath253 , therefore @xmath254 the inequality of ( [ eq : lem1.8 ] ) implies that @xmath245 is also a worst case scenario for @xmath108 , which contradicts the lexicographical maximality of weights on the open interval @xmath239 under @xmath238 . for a fixed pair @xmath105 satisfying @xmath122 , let us consider a worst case scenario in @xmath120 such that the intermediate vertex is @xmath26 ( @xmath255 ) . we now consider the weight of @xmath26 as a variable @xmath133 $ ] , and let @xmath132 be a scenario in @xmath120 such that the intermediate vertex is @xmath26 whose weight is @xmath25 . then , let @xmath256 denote the gap between @xmath1 and @xmath124 under @xmath132 , i.e. , @xmath257 suppose that @xmath132 is critically left - pseudo - bipartite for @xmath124 when @xmath258 , where @xmath36 is a positive integer and @xmath259 . in the following , we prove that a function @xmath256 is convex and piecewise - linear for @xmath260 $ ] for every @xmath261 . under a scenario @xmath132 , let us perform the left - clustering for @xmath1 and the right - clustering for @xmath124 , respectively . performing the left - clustering for @xmath1 , let @xmath262 be a left cluster for @xmath1 including @xmath26 , @xmath263 be the index of a vertex that corresponds to the head of @xmath262 , and @xmath264 be the weight of @xmath203 . also , performing the right - clustering for @xmath124 , @xmath265 , @xmath266 and @xmath267 are similarly defined . _ proof of claim [ clm:4 ] . _ the statement is equivalent to @xmath268 . \label{eq : lem1.11}\end{aligned}\ ] ] we here prove ( [ eq : lem1.10 ] ) ( the case of ( [ eq : lem1.11 ] ) is similarly treated ) . by ( [ eq : lem1.9 ] ) , we only have to prove @xmath269 note that , similarly as in claim [ clm:1 ] , while @xmath132 changes to @xmath270 , the formation of left clusters for @xmath1 remains the same . therefore , by the definitions of ( [ eq : ms1 ] ) and ( [ eq : ms3 ] ) , we have @xmath271 which leads ( [ eq : lem1.12 ] ) by letting @xmath247 go to positive zero . for an integer @xmath261 , we consider the right - derivative of @xmath256 for @xmath272 , i.e. , @xmath273 similarly to ( [ eq : lem1.13 ] ) , we have @xmath274 from ( [ eq : lem1.9 ] ) , ( [ eq : lem1.13 ] ) and ( [ eq : lem1.15 ] ) , we derive @xmath275 and by ( [ eq : lem1.14 ] ) and ( [ eq : lem1.16 ] ) , @xmath276 we here notice that as @xmath25 increases , @xmath263 and @xmath266 never change ( even if left clusters for @xmath1 following @xmath262 and right clusters for @xmath124 following @xmath265 are merged to @xmath262 and @xmath265 , respectively ) . also , since both of @xmath262 and @xmath265 include @xmath26 , @xmath264 and @xmath267 can be represented as follows : @xmath277 where @xmath278 and @xmath279 are functions of @xmath25 . in addition , @xmath278 increases only if a left cluster for @xmath1 following @xmath262 is merged to @xmath262 , i.e. , @xmath278 is an increasing step function , and @xmath279 is a constant function of @xmath272 since the formation of right clusters for @xmath124 does not change over @xmath272 ( recall the definition of critical left - pseudo - bipartite scenarios for @xmath124 in section [ subsec : cpbs ] ) . from the above observations , and ( [ eq : lem1.17 ] ) , ( [ eq : lem1.18 ] ) and ( [ eq : lem1.19 ] ) , we derive that for @xmath272 , @xmath280 which is an increasing step function . by this fact and the continuity of @xmath256 by claim [ clm:4 ] , we have the following claim .

rare event physics at the low energy frontier is playing a significant role in particle physics and cosmology . examples of such rare phenomena could be the detection of non - baryonic particle dark matter ( axions or wimps ) , supposedly filling a substantial part of the galactic haloes , or neutrinoless double beta decay . these rare signals , if detected , would be important evidences of a new physics beyond the standard model of particle physics , and would have far - reaching consequences in cosmology . the experimental achievements accomplished during the last decade in the field of ultra - low background detectors have led to sensitivities capable of searching for such rare events . dark matter detection experiments have largely benefited from the techniques developed for double beta decay searches . both types of investigation , which have very close relation from an experimental point of view , are the two main scientific objectives of cuoricino and cuore . due to the low probability of both types of events , the essential requirement of these experiments is to achieve an extremely low radioactive background . for that purpose the use of radiopure detector components and shieldings , the instrumentation of mechanisms of background identification , the operation in an ultra - low background environment , in summary , the use of the state - of - the - art of low background techniques is mandatory . in some of these phenomena , like the case of the interaction of a particle dark matter with ordinary matter , a very small amount of energy is deposited , and the sensitivity needed to detect events within such a range of energies , relies on how low the energy threshold of detection is . in addition , to increase the chances of observing such rare events a large amount of detector mass is in general advisable , and so most of the experiments devoted to this type of searches are planned to have large detector mass , while keeping the other experimental parameters ( background , energy thresholds and resolutions ) optimized . on the other hand , the recent development of cryogenic particle detection @xcite has led to the extended use of thermal detectors @xcite to take advantage of the low energy threshold and good energy resolution that are theoretically expected for the thermal signals . this detection technology has also the advantage of enlarging the choice of materials which can be used , either as dm targets or @xmath2 decay emitters . after a long period of r&d to master the techniques used in cryogenic particle detectors , low temperature devices of various types are now applied to the detection of double beta decay or particle dark matter @xcite . a major exponent of this development is the midbd ( _ milano double beta _ ) experiment @xcite , which successfully operated a 20 teo@xmath3 crystal array of thermal detectors of a total mass of 6.8 kg , the largest cryogenic mass operated up to now . with the objective of going to larger detector masses and of improving the sensitivity achieved in the smaller arrays , the cuore ( _ cryogenic underground observatory for rare events _ ) project @xcite was born some years ago as a substantial extension of midbd . the objective of cuore is to construct an array of 1000 bolometric detectors with cubic crystal absorbers of tellurite of 5 cm side and of about 760 g of mass each . the crystals will be arranged in a cubic compact structure and the experiment will be installed in the gran sasso underground laboratory . the material to be employed first is teo@xmath3 , because one of the main goals of cuore is to investigate the double beta decay of @xmath1te , although other absorbers could also be used to selectively study several types of rare event phenomenology . apart from the midbd experiment , already completed , a wide r&d program is under way in the framework of cuoricino , a smaller and intermediate stage of cuore , which consists of 44 of the above crystals ( 790 g in this case ) and 18 smaller crystals ( 330 g ) for a total mass of @xmath440 kg . cuoricino has been recently installed in the gran sasso laboratory ( lngs ) and is , by far , the largest cryogenic detector on the stage . the preliminary running tests are encouraging @xcite . in the present work the prospects are presented for the cuore and cuoricino experiments with respect to their double beta decay discovery potential and their detection capability of wimp and axions . in section [ bc ] considerations of the expected background levels in the energy regions of interest are made taking into account all relevant sources . the theoretical motivation of neutrinoless double beta decay experiments as well as the experimental prospects of both cuoricino and cuore are shown in section [ dbd ] . in the same way , the physics potential of the experiments is discussed for direct detection of wimps in section [ wimps ] and for solar axion searches in section [ axions ] . from the radioactive background point of view , thermal detectors are very different than , say , conventional ionization detectors ; the former are sensitive over the whole volume , implying that surface impurities may play an important role . in principle , bolometric detectors like those used in midbd , cuoricino and cuore experiments are expected to have radioactive backgrounds larger than those of conventional germanium ionization detectors , also because the more complex technology of cryodetectors had not yet been fully optimized from the point of view of the radiopurity of the components near the detector . on the other hand , the production of radiopure absorbers for thermal detectors is less mastered than the production of radiopure ge crystal . however , after a considerable r&d effort a significant improvement in the radiopurity of bolometers has been accomplished during recent years , and now these devices have achieved very competitive backgrounds @xcite . the midbd background @xcite at low energy ( @xmath52 c / kev / kg / d at threshold -10 kev- and at 3500 m.w.e . ) is similar to the measured event rate anticoincident with the veto in the cdms experiment @xcite ( @xmath5 2 c / kev / kg / d at 10 kev and at 20 m.w.e . ) and in the edelweiss experiment @xcite(@xmath5 1.8 c / kev / kg / d at 30 kev and 4000 m.w.e . ) , without a veto , ( in both cases , obviously , prior to charge - heat discrimination ) and is roughly equal to that of dama @xcite and anais @xcite ( @xmath5 1.5 c / kev / kg / d , at 2 kev and 3500 m.w.e . and at 4 kev and 2450 m.w.e . respectively ) but still one order of magnitude worse than that of igex @xcite ( @xmath5 0.2 and 0.05 c / kev / kg / d at 4 kev and 10 kev respectively , and at 2450 m.w.e . ) . in the double beta decay region ( @xmath5 2.5 mev ) the midbd background values ( @xmath5 0.3 and 0.6 c / kev / kg / y in the old and new set - up respectively ) are competitive but still higher than those of the igex and heidelberg - moscow experiments @xcite ( @xmath5 0.05 c / kev / kg / y at 2 mev ) ( which uses pulse shape discrimination ) . the challenge of cuore is to significantly reduce the midbd background values ( both in the low and high energy regions ) by , say , two orders of magnitude without using background discrimination mechanisms like the simultaneous measurement of charge ( or light ) and heat . this reduction will be accomplished by working on both the selection of low contamination materials and on the improvement of the detector geometry and the shielding efficiency . as one of the main ingredients in the prospective physics potential of the cuoricino and cuore experiments is the level of background achievable , an evaluation of the expected background must be done , including the contribution of all the possible background sources . a preliminary estimation is contained in ref . @xcite , a more complete estimate is underway . in the following we will give a briefing of that reference and will add further considerations . there are several background sources to be considered : the environmental backgrounds of the underground site ( neutrons from the rocks , environmental @xmath6 flux ... ) , the intrinsic radioactivities of the detectors components and shielding ( bulk and surface ) , the cosmic muon induced backgrounds ( neutrons , muon direct interactions ) , as well as the possible cosmogenic induced activities produced when the detector and components were outside the underground laboratory . even the small leakage of two neutrino double beta decay counts emitted by the absorber s nuclei in the relevant region of analysis might be a potential source of background . the depth at which the experiment will be performed , plus the addition of an efficient cosmic veto and a suitable passive shielding ( lead , polyethylene , ) will effectively reduce the external background . so we will refer mainly to the intrinsic background , although we will make some remarks concerning the other sources . on the other hand , the experience gained in the midbd experiment , where the measured background is used as a test - bench for checking the mc estimates , has been essential in predicting the expectations for cuoricino and cuore . using the midbd results and the cuoricino tests a monte carlo study of the expected intrinsic background has been carried out for the cuoricino and cuore set - ups . radioactive impurities in the bulk of the tellurite crystals , as well as in the dilution refrigerator and surrounding shielding were assumed . background suppression due to the anticoincidence between crystals ( which will be significant in the cuore set up ) was calculated . the code is based on the geant-4 @xcite package ; it models in detail the shields , the cryostat , the detector structure and the detector array for the midbd , cuoricino and cuore experiments . it includes the propagation of photons , electrons , @xmath7 particles and heavy ions ( nuclear recoils from @xmath7 emission ) as well as neutrons and muons . for the simulated radioactive chains ( @xmath8 and @xmath9 ) or radioactive isotopes ( @xmath10 , @xmath11 ) alpha , beta and @xmath6 /x ray emmissions are considered according to their branching ratios . the time structure of the decay chains is taken into account and the transport of nuclear recoils from alpha emissions is included . the 20 crystal midbd array was operated in two different configurations @xcite , the difference between the two was that in the latter configuration the roman lead shield surrounding the detectors was improved by increasing its thickness , and that the crystals as well as their copper mounting structures underwent a new surface treatment . the crystals were re - polished with superpure powders while the copper was cleaned with a chemical process . the result was an improvement in the background measured in the second configuration ( midbd - ii ) with respect to that measured in the previous one ( midbd - i ) either in the gamma region ( below 3 mev ) and in the @xmath7 region ( between 3 and 10 mev ) . the @xmath7 peaks identified in midbd - i as due to a surface u and th contamination of the crystals were reduced by a factor 2 . the reduction of the @xmath7 contaminations , together with the reduction of the @xmath12tl line at 2615 kev is probably responsible of the improvement obtained in the 2448 - 2556 kev ( 0@xmath132@xmath14 decay ) region where the counting rate changed from b=0.59@xmath150.06 c / kev / kg / y in midbd - i to b=0.33@xmath150.11 c / kev / kg / y in midbd - ii @xcite . during midbd - ii run , a 10 cm thick borated polyethylene shield was mounted outside the external lead shield but no improvement in the counting rate was obtained proving that the backgroud was still dominated by other radioactive sources . with the above background value the lower limit of the neutrinoless half - life of @xmath16 obtained is @xmath17 y , or equivalently , an upper bound for the majorana neutrino mass @xmath18 of 0.9 - 5.2 ev ( depending on the nuclear matrix element calculation ) , which is the second - best published result @xcite . as far as the low energy region is concerned , the background of midbd stands around 2.3 c / kev / kg / day between 10 and 50 kev and 0.5 c / kev / kg / day between 50 and 80 kev , region where the dark matter signal is expected @xcite . with these background values , the exclusion plot of wimps interacting coherently with te and o is depicted as the ( dashed ) contour of figure [ cuoricino_exclusion ] . to understand what are the main sources responsible of the midbd background the mc code developed to study cuore background was used to reproduce the results obtained in the two configurations in which the midbd 20 crystal array was operated @xcite . the radioactive impurities used as inputs were all the contaminants identified in the midbd background spectrum , through an analysis of the @xmath7 and @xmath6 lines : @xmath19u chain in secular equilibrium , @xmath19u chain broken at the level of @xmath20ra , @xmath21bi or @xmath22pb , @xmath23th chain in secular equilibrium , @xmath24k and @xmath25co . either bulk contaminations ( with a uniform and isotropic distribution ) and surface contaminations ( with an exponential density profile and various depths ) were considered . the details of this analysis are reported in ref . @xcite , the results obtained can be summarized as follows : the mc simulation is able to reproduce the measured background spectrum , allowing the identification and localization of the most probable sources of background . it proves that a large contribution to the counting rate in both the low energy region and the 0@xmath132@xmath14 region comes from the surface contaminations of the crystals and of the copper structure facing the bare crystals . in the case of the crystals , the origin of this contamination is known : it is due to the use of contaminated powders during the optical polishing of the crystals produced by the shanghai quinhua material company ( sqm ) in china . indeed the re - treatment of their surfaces by means of low contamination powders reduced the contamination to one half . for this reason it was agreed that the cuoricino crystals would undergo only a minor surface treatment in china ( the one indispensable to check crystal quality ) and were subsequently polished in italy using low contamination powders . on the contrary , the origin of the copper surface contamination is still uncertain : it could either have been produced during the machining of the metal or during the surface acid treatment . as the cuoricino copper structure was produced with the same technique of the midbd - ii a similar contamination may have been produced . for cuore a dedicated r@xmath26d is foreseen to reduce or cancel this kind of contamination . finally , other important contributions come from bulk contamination of the structures outside the roman lead shield that surrounds the detector and from outside the cryostat . evaluating each impurity for each simulated element , the maximum contribution to the measured background by bulk contamination of teo@xmath3 , copper and roman lead were obtained and have been used in cuore simulation . cuoricino will be operated in the same set - up as that of the midbd experiment : the same dilution refrigerator , the same cryostat , the same external lead and borated - polyethylene shield , at the same underground location ( a description of the set - up can be found in ref . the crystals will be different but the measured upper limits of their u and th contamination are again of the order of 10@xmath27 g / g . therefore bulk contribution to cuoricino background should be not larger than that measured in midbd - ii . on the other hand , for surface contamination some improvement with respect to midbd - ii result is foreseen thanks to the different surface treatment of the crystals and to a reduced exposure of detectors to dust and rn . a reasonable guess for cuoricino is therefore a background of @xmath5 0.1 c / kev / kg / y in the @xmath28 region , slightly better than the one measured in midbd - ii ; while in the dark matter region a background of the order of 1 - 0.1 c / kev / kg / d is anticipated . on the contrary no extrapolation of the background achievable in cuore is possible from the midbd data , nor from the cuoricino results . this is true for several reasons : the different structure of the detector ( in the case of cuore an efficient reduction of background events will be provided by operating the detectors of the array in anticoincidence , without loosing efficiency in detecting wimps or 0@xmath132@xmath14 decay events ) , the different cryostat and shielding systems , and finally because a much more severe selection of the constructing materials and a much better control on their mechanical working and chemical cleaning will be applied . also predictions based on mc simulations are not straightforward since they depend critically on the assumed inputs : contamination levels , detector design , etc . here we present a very conservative evaluation of the background reachable with cuore , mainly based on the state of the art of detector design and on the knowledge of radioactive contamination . the cuore construction will require about five years and in the meantime an @xmath29 dedicated to background reduction will be realised . this approach for intrinsic background evaluation could be rather pessimistic but it is the only one that presently guarantees a reliable but possibly conservative prediction . in cuore ( a description of the set - up can be found in ref . @xcite ) the main contribution to background due to bulk contaminations comes from the heavy structures near the detectors ( the copper mounting structure of the array , the roman - lead box and the two lead disks on the top of the array ) and from the detectors themselves ( the @xmath30 crystals ) . in the simulation , the @xmath9 , @xmath8 , @xmath10 , @xmath31 , @xmath32 , @xmath33 and @xmath34 impurities of @xmath30 , copper and lead as well as the @xmath11 cosmogenic contamination of copper are conservatively assumed equal to the 90% c.l . upper limits obtained for the contamination levels of those materials measured in midbd experiment or in low hp@xmath35 spectrometry @xcite . the @xmath11 cosmogenic contamination of @xmath30 is on the other hand evaluated according to the time required to produce @xmath30 powder from the ore , grow the crystals and store them underground @xcite . with these contaminations , assuming a threshold of 10 kev and using the results obtained after the reduction by anti - coincidence between detectors , the background due to bulk contaminations is @xmath36 counts/(kev kg y ) in the @xmath37 decay region and @xmath38 counts/(kev kg d ) in the low energy region ( 10 - 50 kev ) . on the other hand , surface contaminations contribute to background only when they are localized on the crystals or on the copper mounting structure directely facing them . in the simulation , u and th surface contaminations are assumed to be : for crystals @xmath5 100 times lower and for copper @xmath5 50 times lower than the corresponding contamination in midbd experiment . the midbd crystals were contaminated during the polishing procedure due to the use of highly contaminated powders . polishing powders with a radioactive content 1000 times lower are however commercially available and have already been used for cuoricino , an improvement of the surface contamination of a factor 100 is therefore fully justified . a similar situation holds for copper . in the midbd experiment the copper surfaces were treated with an etching procedure studied to reduce impurities on surfaces before the sputtering process . this procedure resulted in an improvement of the surface quality of copper , however it is not optimized from the point of view of background . the use -for the surface treatment- of low contamination liquids in a low background environment , and special care in the mechanical working and handling will allow an improvement between one and two orders of magnitude in the surface contamination of copper . with such a contamination the surface contribution is estimated to be 2.8@xmath39 counts/(kev kg y ) in the 0@xmath132@xmath14 region and @xmath40 counts/(kev kg d ) in the dark matter region . finally , the unavoidable background produced by the @xmath41 decay region is lower than 10@xmath42 counts/(kev kg y ) in the 0@xmath132@xmath14 region and @xmath43 counts/(kev kg d ) in the dark matter region ( assuming 3.8@xmath44 y for the @xmath1te @xmath45 half - life @xcite ) . concerning the cosmogenic activation produced by cosmic rays when the crystals are above ground ( during fabrication and shipping of the crystals from the factory to the underground laboratory ) we used a code based on computed cross sections to estimate the amount and type of radionuclides produced by cosmic rays on teo@xmath3 @xcite . the comparison between our prediction and the midbd - i and ii run data is discussed in ref . the radionuclei produced by the activation of tellurium are mostly tellurium isotopes ( a=121,123,125,127 ) as well as @xmath46sb , @xmath47sb , @xmath25co and tritium , these last three being of more concern because of their long half - life ( @xmath47sb : beta decay of 2.7 years , end - point energy of 767 kev , @xmath25co : beta decay of 5.27 years end - point energy 2823 kev and @xmath48h : beta decay of 12.3 years , end - point energy of 18 kev ) . in midbd - i no clear evidence of cosmogenic activation was present ( the crystals were stored underground for several months ) , while in midbd - ii the @xmath6 lines of the tellurium isotopes were clearly seen , with an intensity that is in good agreement with our prediction ( based on the period they were reexposed because of the repolishing process ) . the history of the cuoricino crystals is rather similar to those of midbd , therefore their cosmogenic activation is already included in the extrapolation of cuoricino background from midbd data . in the case of cuore the control on crystal production will be more severe . the sqm that produces the crystals , oxidizing the tellurium ore and growing a 760 g crystal in about two months , will require two years to grow the 1000 cuore detectors . once grown , the crystals will be shipped to italy and stored underground , therefore their total exposition to cosmic rays will be limited to about 4 months . the total induced activities remaining after 2 years underground have been estimated @xcite and the consequent contribution to the detector counting rate was deduced by a mc simulation . the radionuclei that contribute to background through their @xmath49 decay are : * in the 0@xmath132@xmath14 region , the long lived @xmath25co isotope with an activity of @xmath50.2 @xmath50bq / kg , while a minor contribution is due to the isotopes @xmath51ag and @xmath46sb whose activity is 4 times lower and fast decreasing with time ; * in the dark matter region , the long lived nuclei of @xmath48h and @xmath47sb with an activity of @xmath57 @xmath50bq / kg for the former and of @xmath515 @xmath50bq / kg for the latter . the influence of @xmath25co in the cuore background was already considereded in the evaluation of the contributions due to bulk contaminations of the crystals , while the contribution of @xmath48h and @xmath47sb to the dark matter region is completely negligible ( @xmath52 counts/(kev kg d ) ) if compared to the intrinsic background . as noted before , neither contributions from underground cosmic muons nor neutrons have been taken into account in detail in the estimation of the background . however , the following simplified arguments will serve to have an approximate idea of their contribution . the depth of the lngs ( 3500 m.w.e ) reduces the muon flux to @xmath52@xmath53 @xmath54s@xmath55 , but a further effective reduction could be obtained with the use of an efficient ( 99.9% ) active veto to detect muons traversing the detectors and to tag possible events associated with them . consequently , the muon - induced events contributing to the background are expected to be only a small component of it . on the other hand , the shielding will substantially reduce the event rate due to particles external to the detector from various sources ( neutrons and photons ) , from radioactivity in the environment ( natural decay chains u / th , @xmath22pb , @xmath24k , ) , as well as muon - induced in the surroundings or in the shielding itself . the passive shielding either for midbd , for cuoricino and for cuore consists of a neutron screen ( blocks of -borated- polyethylene of 10 cm thickness ) to moderate and attenuate neutrons and a lead shell to attenuate the incoming external photons ; the innermost part ( 10 cm thick ) is made of special lead with a small content ( @xmath5616 bq / kg ) of @xmath22pb ( half - life 22 years ) while the outer part ( 10 cm thick ) is made of modern lead ( @xmath22pb @xmath56150bq / kg ) both with a low level of th and u. neutrons may constitute a worrisome background in dark matter experiments because for appropriate neutron energies ( few mev ) they can produce nuclear recoils in the detector target nuclei which would mimic wimp interactions . simple kinematics tells that in the case of tellurium , neutrons of 1 to 5 mev could elastically scatter off tellurium nuclei producing recoils of energies up to 31 to 154 kev . in general , one considers neutrons of two origins : from radioactivity in the surroundings or muon - induced . depending on the overburden of the underground site ( i.e. , depending on the muon flux ) , muon - induced neutrons are produced , at lesser or greater rate , both inside and outside the shielding . they are moderated according to their energies by the polyethylene and lead shield ( when produced outside ) or tagged by the muon veto coincidence ( when produced within the passive shielding ) . a certain fraction of the neutrons of a few mevs produced outside can pass the veto reaching the detector and producing `` dangerous '' nuclear recoils and @xmath6 background . however , a significant fraction of the associated events can be vetoed because of their interaction with the veto and also because of the hadronic showers initiated by the muons ( see ref . @xcite ) . in the case of external neutrons ( from the rocks , from fission processes or from ( n,@xmath7 ) reactions , as well as neutrons originated by muons in the walls of the underground site ) , the environmental neutron flux has been measured in lngs . the resulting flux is of @xmath57 @xmath54s@xmath55 for the thermal component , @xmath58 @xmath54s@xmath55 for the epithermal and @xmath59 @xmath54s@xmath55 for energies over 2.5 mev @xcite . they are fairly well moderated by the polyethylene and eventually absorbed or captured . we have carried out a monte carlo simulation of the propagation of neutrons through the 10 cm thick borated polyethylene shield of cuoricino and cuore . the result is that the neutron induced event rate on the entire energy range ( from threhsold to 10 mev ) is much lower than the contribution due the bulk contamination of crystals . other neutrons , produced by muon interactions inside the shielding materials , are very scarce and tagged as events coincident with the muon veto . as it is well - known , muon - induced neutrons are originated in a variety of processes . the reduction of muon flux in underground sites results in a substantial depletion of the associated neutrons and below @xmath5100 m.w.e . the dominant sources of neutrons are nuclear fission processes and ( n,@xmath7 ) reactions in rocks and other environmental material with sizeable content of u / th . the energy spectrum of muon - induced neutrons is approximated by an inverse energy power law ( @xmath60 for 1 - 50 kev and @xmath61 above 50 kev ) but other neutron energy spectra have been proposed . the neutron yield per muon can be approximately evaluated for an organic compound through the simple expression @xmath62 neutrons/(g @xmath54 ) per muon @xcite which fits the value of 1.5@xmath63 neutrons/(g @xmath54 ) of the muon - induced neutron flux measured by the lvd experiment at gran sasso @xcite . for the lngs muon flux ( 2.5@xmath53 @xmath50/(@xmath64s ) ) , muons would produce in the cuore shielding of polyethylene ( 10 cm ) and lead ( 20 cm ) ) about @xmath50.04 neutrons/(m@xmath65 day ) in the polyethylene shield and @xmath525 neutrons/(m@xmath65 day ) in the lead shell . so , independently of the mechanism set up to reject or tag the events associated to neutrons , their rather small number is expected to play a secondary role in the total background compared with other , intrinsic , sources of background . a preliminary evaluation of the influence of the environmental @xmath6 background in gran sasso @xcite resulted in a negligible contribution for the @xmath28 region and a contribution similar to that of bulk contaminations for the dark matter region . a more complete and detailed study of the external , non intrinsic background rates for cuore is underway . summarizing , for cuoricino a background similar to the one measured in midbd - ii is the most conservative prediction either in the dark matter region or in the @xmath28 region . for cuore on the other hand , the previous sections have shed light on how the background contamination dominating the two regions of interest is mostly the intrinsic background due to bulk and surface contaminations of the contructing materials . with the presently achieved quality of low contamination materials and considering the worst possible condition for bulk contaminations ( i.e. all the contamination equal to the present 90%/,c.l . measured upper limits ) we have proved that cuore background will be @xmath66 counts/(kev kg y ) at the @xmath28 transition energy and @xmath67 counts/(kev kg d ) near threshold . even at these conservative values cuore is , as we will prove in the following , a powerful instrument to search for double beta decay and dark matter . however , the goal of the cuore technical @xmath68 of the next few years will be the reduction of background to the level of 0.001 counts/(kev kg y ) at the @xmath28 transition energy and to 0.01 counts/(kev kg d ) at threshold . the potential of the experiment in this background configuration will also be discussed . regarding the expected threshold and resolution , in the cuoricino test experiment @xcite energy thresholds of @xmath69 kev have been obtained and a resolution of 1 kev at the 46 kev line of @xmath22pb was achieved in some of the detectors . details can be found in ref . we will assume a conservative value of 10 kev for the energy threshold of both cuoricino and cuore , and energy resolutions of 1 kev at threshold . as far as the energy resolutions obtained in the double beta decay region , values of 3 kev at 2615 kev were achieved in some crystals but they are worse ( by a factor two ) in others . an energy resolution of 5 kev will be assumed both for cuore and for cuoricino . taking into account these expectations , we discuss in the following the prospects of cuoricino and cuore for double beta decay searches ( section 3 ) , for wimp detection ( section 4 ) and for solar axion exploration ( section 5 ) . one of the main scientific objectives of the cuore detector is the search for the neutrinoless double - beta decay of the @xmath1te isotope contained in the ( natural ) teo@xmath3 crystals . the importance of nuclear double - beta decay as an invaluable tool to explore particle physics beyond the standard model has been repeatedly emphasized and widely reported @xcite . in the standard model of particle physics neutrinos are strictly massless , although there is no theoretical reason for such a prejudice . on the experimental side , moreover , there exist strong evidences from atmospheric neutrino data ( from superkamiokande @xcite ) , from experiments with solar neutrinos ( from homestake , gran sasso and kamioka @xcite ) and from reactor experiments ( from kamland @xcite ) which suggest that neutrinos have indeed masses and oscillate among the three species . the results of the solar @xmath13 experiment from sno with both cc ( charged current ) and nc ( neutral current ) interactions @xcite , also combined with superk have definitively provided a strong evidence that neutrinos change of flavour and , consequently , the existence of non - zero mass neutrinos . the recent results of kamland @xcite exclude all oscillation solutions but the large mixing angle solution to the solar neutrino problem using reactor antineutrino sources . however , neutrino oscillation experiments provide the squared mass difference between the neutrino species but not their absolute value and scale . the neutrinoless double beta decay would help to solve this question and to disentangle the hierarchy scheme of the neutrino flavours . most of the models ( see @xcite ) indicate that the majorana neutrino mass parameter could be around ( or slightly below ) @xmath70 ev , value within reach of the future double beta decay experiments , like cuore . majorana massive neutrinos are common predictions in most theoretical models , and the value of a few tens of millielectronvolts predicted for its effective mass , if reached experimentally -as expected- will test its majorana nature . double - beta decay experiments with even better sensitivities ( of the order of a few mev ) will be essential to fix the absolute neutrino mass scale and possibly to provide information on cp violation @xcite . on the other hand , galaxy formation requires a small amount of hot non - baryonic dark matter likely in form of neutrinos to match properly the observed spectral power at all scales of the universe . the question of the neutrino mass is one of the main issues in particle physics . in the standard model , neutrinos and antineutrinos are assumed to be different particles , but no experimental proof has been provided so far . nuclear double - beta decay addresses both questions : whether the neutrinos are self - conjugated and whether they have majorana mass . in fact , the lepton number violating neutrinoless double beta decay @xmath71 ( @xmath72 ) is the most direct way to determine if neutrinos are majorana particles . another form of neutrinoless decay , @xmath73 may reveal also the existence of the majoron ( @xmath74 ) , the goldstone boson emerging from the spontaneous symmetry breaking of the @xmath75 symmetry which is of most relevance in the generation of majorana neutrino mass and has far - reaching implications in astrophysics and cosmology . moreover , the observation of a @xmath72 decay would imply a lower bound for the neutrino mass , i. e. at least one neutrino eigenstate has a non - zero mass . it is quite straightforward to demonstrate how the determination of the effective majorana mass of the electron neutrino constrains the lightest neutrino mass eigenvalue . this was done recently by barger , glashow , marfatia and whistnant @xcite . however it is very convenient to use the approximation that @xmath76 and @xmath77 . with the further assumption @xmath78 in the normal hierarchy case we obtain : @xmath79 where @xmath80 . similarly in the case of inverted hierarchy : @xmath81 a related quantity , @xmath82 has great importance in cosmology . it is related to the quantity @xmath83 , the fraction of the critical density , @xmath84 ( that would close the universe ) that is in the form of neutrinos , where : @xmath85 and @xmath86 is the dimensionless hubble constant . it was shown in reference @xcite that @xmath87 obeys the following inequality with respect to @xmath88 : @xmath89 where the @xmath90 and @xmath91 signes refer to normal and inverted hierarchies respectively , and @xmath92 . if this relation is written as two equalities , and each side is solved for @xmath93 in terms of @xmath87 , quadratic equations result with the constant terms @xmath94 . values of @xmath95 of cosmological interest are much larger than @xmath96 ev ( 99.73% c.l . ) . accordingly the above inequality reduces to the trivial form : @xmath97 this clearly demonstrates the direct connection between neutrinoless double beta decay and neutrino dark matter . cosmic microwave background and galaxy cluster surveys have set the bound @xmath98 @xcite . it is expected that the map satellite will produce data that will result in a sensitivity @xmath99 . one possible interesting scenario results in the case the map satellite , if for example it observes a clear singal well above their limiting sensitivity , and cuore ( as well as other next generation experiments ) have a negative results at sensitivities of @xmath87 far below that of the satellite data . this would be a clear indication that the phenomenon causing the density fluctuations implied by the cmb data was not caused by majorana neutrino dark matter . in addition a positive measurement of the mass @xmath100 by katrin @xmath101 beta end - point spectrum measurement would have to yield @xmath102 . the cuore experiment will have a far greater sensitivity , and if it found a negative result , the mistery of dirac or majorana character of neutrinos will be solved . these and other issues , make the search for the neutrinoless double beta decay an invaluable tool for the exploration of non - standard model physics , probing energy scales well above those reachable with accelerators . that is the motivation of why there are dozens of experiments underway looking for the double beta decay of various nuclei @xcite like @xmath103 ( igex @xcite , heidelberg - moscow @xcite ) , @xmath104 and others ( nemo @xcite , elegants @xcite ) and @xmath16 ( midbd , cuoricino ) and a few big experimental projects , like cuore , majorana @xcite ( @xmath103 ) , moon @xcite ( @xmath104 ) and exo @xcite ( @xmath105 ) . the cryogenic thermal detectors provide new double - beta emitter nuclei to be explored in `` active '' source / detector calorimeters . some of them have been tested and others are already in running detectors , like @xmath106ca in caf@xmath3 , @xmath1te in teo@xmath3 , and @xmath107cd in cdwo@xmath108 . as far as the tellurium oxide is concerned , the @xmath1te isotope is a good candidate for double beta decay searches : its isotopic content in natural tellurium is 33.87% , and its 2@xmath14 q - value ( @xmath109 kev ) is reasonably high to escape from the main radioimpurity lines when looking for a neutrinoless signal . moreover , this q - value happens to be between the peak and the compton edge of the 2615 kev line of @xmath12tl , which leaves a clean window to look for the signal . finally , it has a fairly good neutrinoless nuclear factor - of - merit @xmath110 , where @xmath111 is an integrated kinematical factor qualifying the goodness of the @xmath112 value ( large phase space ) and @xmath113 is the neutrinoless double - beta decay nuclear matrix elements characterizing the likeliness of the transition . in table [ tab_doblebeta ] we quote the values of @xmath114 in the case of @xmath1te calculated in various nuclear models @xcite , together with those of other emitters used in source / detector calorimeters . it can be seen that no matter the nuclear model is used to compute the neutrinoless double - beta decay matrix elements , the merits of @xmath1te are a factor 510 more favorable than those of @xmath115ge ( the emitter where the best neutrinoless double beta decay half - life limits have been achieved so far ) , which translates into a factor 2 to 3 better as far as the @xmath116 ( majorana neutrino mass parameter ) bounds are concerned . the detector factor - of - merit @xmath117 , or detection sensitivity , introduced earlier by fiorini , provides an approximate estimate of the neutrinoless half - life limit achievable with a given detector . for source / detector devices , it reads : @xmath118 where @xmath119 is the atomic mass , @xmath120 is the isotopic abundance , @xmath121 the detector mass in kg , @xmath122 the background in c / kev / kg / y in the 2@xmath14 neutrinoless decay region , @xmath123 the running time in years , @xmath124 the fwhm energy resolution in kev and @xmath125 the detector efficiency ( which is 0.86 in the case of the cubic 5x5x5 @xmath126 crystals ) . in the case of a teo@xmath3 crystal detector , @xmath127 , with @xmath121 the crystal mass in kg and @xmath122 the background in counts per kev and year per kg of detector mass . a simple projection for cuoricino sensitivity can be obtained assuming a background of b=0.1 c / kev / kg / y ( as discussed in section 2 ) and a resolution of @xmath128 5 kev in the 2.5 mev region . in that case , one would have @xmath129 years , with @xmath121 ( kg ) the mass of the teo@xmath0 crystal array . for the mass of cuoricino ( m=40.7 kg ) , one will have a sensitivity of @xmath130 years . using a typical average value of @xmath131 y@xmath55 , as obtained in qrpa @xcite , cuoricino will have a mass sensitivity of @xmath132 ev in one year , in the least favourable case . notice that the bound on the effective mass of the electron neutrino , @xmath133 , is given by @xmath134 . using for comparison the value of @xmath135 y@xmath55 of ref . @xcite which is usually employed in the @xmath116 bound ( @xmath50.33 ev ) derived from the ge experiments , with these assumptions cuoricino would provide @xmath1360.24 ev . to go further , one needs to increase the mass of teo@xmath0 and to reach even a lower background . in our very conservative projection for the cuore array ( 760 kg of teo@xmath0 ) a background of b=0.01 c / kev / kg / y can be easly achieved , assuming as in the case of cuoricino a resolution @xmath124(2.5 mev)=5 kev , one would get @xmath137 years , which in one year of statistics would provide @xmath116 bounds ranging from 0.07 ev @xcite , 0.08 ev @xcite , 0.15 ev @xcite or 0.04 ev @xcite just to mention results using a few nuclear matrix element estimates . however , the r&d to be carried out in cuore , if succesful , would provide a value of @xmath138 c / kev / kg / y i.e. , a detection sensitivity of @xmath139 years , or @xmath116 bounds ranging from @xmath140 ev ( in @xcite ) to @xmath141 ev ( in @xcite ) . teo@xmath0 crystals made with @xmath142te enriched material have been already tested during the midbd experiment , making an enriched cuore a feasible option . assuming a 95% enrichment in @xmath142te and a background level of b=0.001 c / kev / kg / y , the sensitivity becomes @xmath143 years . for an exposure of 5 years , the corresponding @xmath144 bounds range from 9 mev to 56 mev depending on the nuclear matrix element calculations . recent cosmological observations @xcite provide compelling evidence for the existence of an important component of non - baryonic cold dark matter in the universe . among the candidates to compose this matter , weakly interacting massive particles ( wimps ) and axions are the front runners . the lightest stable particles of supersymmetric theories , like the neutralino @xcite , constitute a particular class of wimps . under the hypothesis that wimps are the main component of the dark matter , these particles should fill the galactic haloes and explain the flat rotation curves which are observed in many galaxies . the detection of such particles could be attempted both by means of direct and indirect methods . the direct detection of wimps relies on the measurement of their elastic scattering off the target nuclei of a suitable detector@xcite . the non relativistic and heavy ( gev tev ) wimps could hit a detector nucleus producing a nuclear recoil of a few kev . because of the small wimp - matter interaction cross sections the rate is extremely low . in the case of susy wimps , most of the cross section predictions @xcite ( derived using mssm as the basic frame implemented with different unification hypothesis ) encompass a range of values several orders of magnitude ( the so - called scatter plots ) providing rates ranging from 1 c / kg / day down to 10@xmath145 c / kg / day according to the particular susy model . it is well known that the predicted signal for the wimp elastic scattering has an exponentially decaying energy dependence , hardly distinguishable from the background recorded in the detector . the simple comparison of the theoretical wimp spectrum with the one experimentally obtained , provides an exclusion curve ( at a given confidence level ) , as dark matter component of the halo , of those wimps with masses ( @xmath146 ) and cross sections on nucleons ( @xmath147 ) which yield spectra above the measured experimental rate . to claim a positive identification of the wimp , however , a distinctive signature is needed . the only identification signals of the wimp explored up to now are provided by the features of the earth s motion with respect to the dark matter halo . in particular , the annual modulation @xcite is originated by the combination of the motion of the solar system in the galactic rest frame and the rotation of the earth around the sun . due to this effect , the incoming wimp s velocities in the detector rest frame change continuously during the year , having a maximum in summer and a minimum in winter . therefore the total wimp rate changes in time with an oscillating frequency which corresponds to an annual period and a maximum around the beginning of june . the relative annual variation of the signal is small ( a few percent ) so in order to detect it one needs large detector masses to increase statistics and several periods of exposure to minimize systematics . several experiments have already searched for this effect @xcite and since 1997 one group has reported a positive signal @xcite which has been appearing throughout four yearly periods . the present situation is no doubt exciting : on one hand that result has triggered an intense activity in the field ; on the other , the experimental sensitivities of various types of underground detectors are entering the supersymmetric parameter space @xcite and in particular several of them @xcite are excluding , to a larger or lesser degree , the region of mass and cross - section where the reported wimp is supposed to exist . new data from one of them @xcite have excluded totally the dama region . we will discuss in the following the capabilities of cuoricino and cuore to exclude or detect wimps using the total time - integrated experimental rate and comparing it with the predicted nuclear recoil rate . to look for the annual modulation signal in cuoricino / cuore experiments , which in principle have enough mass to be sensitive to it , one needs to know their stability performances . the analysis of the cuore / cuoricino potential for annual modulation searches will be performed -following statistical consideration- ( see ref . @xcite ) , with the proviso that systematic uncertainties are under control . data on the stability of cuoricino will be crucial to assess such hypotheses . to calculate the theoretical wimp rate , standard hypothesis and astrophysical parameters are assumed , i.e. , that the wimps form an isotropic , isothermal , non - rotating halo ( the isothermal sphere model ) of density @xmath148 gev/@xmath126 , which has a maxwellian velocity distribution with @xmath149 km / s ( with an upper cut corresponding to an escape velocity of 650 km / s ) , and a relative earth - halo velocity of @xmath150 km / s ) . other , more elaborated halo models , which have been considered recently @xcite would lead to different results . the same applies when other astrophysical parameters are employed or when uncertainties in the halo wimps velocity distribution are included @xcite . the theoretical predicted rate is expressed in terms of the mass and cross - section of the wimp - matter interaction . the cross sections are normalized per nucleon assuming a dominant scalar interaction , as is expected , for instance , for one of the most popular dark matter candidates , the neutralino : @xmath151 where @xmath119 is the target ( oxygen and tellurium ) mass number , @xmath152 , is the wimp - nucleus reduced mass , and @xmath153 the wimp - nucleon reduced mass . the helm parameterization @xcite is used for the scalar nucleon form factor . the @xmath154 exclusion plot is then derived by requiring the theoretically predicted signal for each @xmath146 and @xmath147 in each energy bin to be less than or equal to the ( 90% c.l . ) upper limit of the ( poisson ) recorded counts . the bin width is assumed to be equal to the detector resolution . in figure [ cuoricino_exclusion ] , the exclusion plots for coherent spin - independent wimp - matter interaction are shown for two possible values of the background of cuoricino , 1 and 0.1 c / kev / kg / day . the first value is of the order of the background already achieved from threshold onwards ( 10 - 50 kev ) in the midbd latest results ( see @xcite ) . the value 0.1 c / kev / kg / day is a one - order - of - magnitude extrapolation from that currently achieved in midbd ( see discussion on section 2 ) and is close to the one obtained above 50 kev . in the case of cuore , background values of 0.05 and 0.01 c / kev / kg / day will be assumed . notice , moreover , that values of a few 0.01 c / kev / kg / day have been obtained above 10 kev in the raw spectra of germanium experiments ( like igex @xcite ) without using mechanisms of background rejection , and so it does not seem impossible to achieve such equivalent small values in crystal thermal detectors of tellurium ( only phonons , and no discrimination mechanism ) . to draw the two exclusion contours of fig . [ cuoricino_exclusion ] , a low energy resolution of 1 kev and an energy threshold of 10 kev have been assumed as well as an exposure of 2 years of cuoricino ( 81 kg@xmath155year ) . the projected exclusion contours are compared with the one currently obtained from midbd ( dashed line ) . in figure [ cuore_exclusion ] , the exclusions for the two quoted values of the background of cuore , 0.05 and 0.01 c / kev / kg / day , are similarly presented for an exposure of 1 year ( 760 kg@xmath155year ) . as previously noted , cuore and to some extent cuoricino have detector masses large enough to search for the annual modulation signal . as it is well known , an essential requirement to estimate the prospects of any detector to search for annual modulation is to have a superb control of systematic errors and to assure that the stability of the various experimental parameters , which might mimic periodic variations of the signals , are kept within a small fraction of the ( already tiny ) expected signal . the various changes of the set - up , crystals and shielding of the midbd experiment have not provided a definitive estimation of the long - term stability parameters of midbd . possible instabilities are that of the electronic gain and the ensuing time fluctuation of the energy scale ( both in energy thresholds and energy resolutions ) , the temperature variations , the possible fluctuation in time of the efficiency with which the triggered noise is rejected and others . they must be kept well below the small expected seasonal modulation of the wimp signal the fact that we are dealing with a very small signal depending on time , which typically amounts to a fraction between 1% and 7% of the average count rates , reinforces the need for a control of the stability of the experiment well below that range over long periods of time . when assuming that all these fluctuations are controlled well below the levels needed ( @xmath561% ) , then one can proceed to analyze the sensitivity of cuoricino / cuore to the annual modulation signal on purely statistical grounds . this has been first attempted in @xcite and @xcite , but a more extensive and rigorous approach is followed in ref . @xcite where sensitivity plots for several types of detectors ( and experimental parameters ) are presented , and in particular , for cuore and cuoricino . the sensitivity of a given experimental device to the annual modulation signal ( according to the detector material employed and the experimental parameters of the detectors ) has been extensively studied in ref . @xcite on purely statistical grounds . following the guidelines of that reference , it can be precisely quantified by means of the @xmath156 parameter , defined from the likelihood function or , equivalently , from the @xmath157 function of the cosine projections of the data ( for further details see ref . @xcite ) : @xmath158 this parameter measures the statistical significance of the modulation signal detected in an experimental set of data . however , for a given ( @xmath159 ) and a given experiment the expected value @xmath160 can be estimated using the expression derived in ref . @xcite : @xmath161 where @xmath162 and @xmath163 are the modulated and non - modulated parts of the wimp signal in the @xmath164th energy bin of @xmath165 width , @xmath166 is the background in that energy bin and @xmath167 the effective exposure , being @xmath7 a coefficient accounting for the temporal distribution of the exposure time around modulation maxima and minima ( @xmath168 for @xmath169 temporal bins ) . using this equation we have estimated the region that could be within reach for cuore and cuoricino with the above mentioned assumptions on the background levels . we have fixed a value of 5.6 for @xmath160 that corresponds to 50% probability of obtaining a positive result at 90% c.l .. in figure [ cuoricino_mod ] curves are shown obtained for a threshold of 10 kev , two years of exposure with cuoricino ( 81 kg year ) and two assumed flat backgrounds of 1 and 0.1 c / kev / kg / day . one can see that cuoricino could already explore most of the dama region looking for a positive annual modulation signal . in figure [ cuore_mod ] similar curves are presented , assuming flat backgrounds of 0.05 and 0.01 c / kev / kg / day , two years of exposure of cuore ( 1500 kg year ) and a threshold of 10 kev ( solid lines ) . the possibility of a lower thresholds of 5 kev with a background of 0.01 c / kev / kg / day is also shown ( dashed line ) . in conclusion , cuore and cuoricino will be able to explore and/or exclude wimps lying in large regions of their parameter space . the capability of cuoricino / cuore to investigate the dama region through the exclusion plot ( time integrated method ) relies in getting a background of 0.1 c / kev / kg / day from 10 kev onwards , independently of more elaborated time modulation methods which require an exhaustive control of the stability of the experiment . however , cuoricino and cuore could also attempt to look for annual modulation of wimp signals provided that the stability of the experiment is sufficient . axions are light pseudoscalar particles which arise in theories in which the peccei - quinn u(1 ) symmetry has been introduced to solve the strong cp problem @xcite . they could have been produced in early stages of the universe being attractive candidates for the cold dark matter ( and in some particular scenarios for the hot dark matter ) responsible to 1/3 of the ingredients of a flat universe . dark matter axions can exist in the mass window @xmath170 ev @xmath171 ev , but hadronic axions could exist with masses around the ev . axions could also be copiously produced in the core of the stars by means of the primakoff conversion of the plasma photons . in particular , a nearby and powerful source of stellar axions would be the sun . the solar axion flux can be easily estimated @xcite within the standard solar model , resulting in an axion flux of an average energy of about 4 kev that can produce detectable x - rays when reconverted again in an electromagnetic field . moreover , it has been pointed out recently that the dimming of supernovae snia might be due to the conversion of photons into axions in the extra - galactic magnetic field @xcite . a photon - axion ( @xmath6-a ) oscillation could make unobservable about 1/3 of the sn emitted light and so , they would appear fainter than implied by the luminosity - distance versus redshift relation , without need to invoke an accelerated expansion of the universe . the sn result would be matched by axions of mass @xmath172 ev and coupling to photons @xmath173 gev@xmath55 . so stellar axions may play an important role in cosmology . we would like to stress that , although we focus on the axion because its special theoretical motivations , all this scenario is also valid for any generic pseudoscalar ( or scalar ) particle coupled to photons @xcite . needless to say that the discovery of any type of pseudoscalar or scalar particle would be extremely interesting in particle physics . we will keep our discussion , however , restricted to the case of solar axions . crystal detectors provide a simple mechanism for solar axion detection @xcite . axions can pass in the proximity of the atomic nuclei of the crystal where the intense electric field can trigger their conversion into photons . the detection rate is enhanced if axions from the sun coherently convert into photons when their incident angle with a given crystalline plane fulfills the bragg condition . this induces a correlation of the signal with the position of the sun which can be searched for in the data and allows for background subtraction . the potentiality of primakoff conversion in crystals relies in the fact that it can explore a range of axion masses ( @xmath174 kev ) not accessible to other direct searches . moreover it is a relatively simple technique that can be directly applied to detectors searching for wimps . primakoff conversion using a crystal lattice has already been employed in two germanium experiments : solax @xcite and cosme - ii @xcite with the ensuing limits for axion - photon coupling @xmath175 gev@xmath55 and @xmath176 gev@xmath55 respectively . also the dama collaboration has analyzed 53437 kg - day of data of their nai set up @xcite , in a search for solar axions , following the techniques developed in ref . @xcite , where a calculation of the perspectives of various crystals detectors ( including nai ) for solar axion searches has been made . the dama result @xmath177 gev@xmath55 improves slightly the limits obtained with other crystal detectors @xcite and agrees with the result predicted in ref . these `` crystal helioscopes '' constraints are stronger than that of the tokyo axion helioscope @xcite for @xmath178 ev and do not rely on astrophysical considerations ( i.e. on red giants or hb stars dynamics @xcite ) . the orientation of the crystal was not known so that the data were analyzed taking the angle corresponding to the most conservative limit . it has been noted that the model that yields the solar axion fluxes used to calculate the expected signals is not compatible with the constraints coming from helioseismology if @xmath179 gev@xmath55 @xcite . this would imply a possible inconsistency for solar axion limits above that value , and sets a minimal goal for the sensitivity of future experiments . the use of cuore to search for solar axions via bragg scattering should have a priori some advantages with respect to germanium detectors , because of the larger mass and the known orientation of the crystals . on the other hand , as the cross - section for primakoff conversion depends on the square of the atomic number , teo@xmath0 will be a priori a better candidate than germanium . needless to say that a low energy threshold is mandatory because the expected signal lies in the energy region 2 kev @xmath180 10 kev and is peaked at @xmath181 kev . a detailed analysis has been performed @xcite for a teo@xmath3 crystal ( which has a tetragonal structure @xcite ) assuming different values for the experimental parameters . as it is shown in ref . @xcite , the bound on axion - photon coupling which a given experiment can achieve can be estimated through the expression : @xmath182 where @xmath164 depends on the crystal structure and material , as well as on the experimental threshold and resolution . for the case of teo@xmath3 and a threshold of 5 kev , @xmath164 has been calculated to be @xmath183 assuming an energy resolution of 1 kev . the computation of this expression for some assumed values of the experimental parameters is shown in table [ tab_cuore ] for cuoricino and cuore . in all cases flat backgrounds and 2 years of exposure are assumed . it is worth noticing the faible dependence of the ultimate achievable axion - photon coupling bound on the experimental parameters , background and exposure mt : the 1/8 power dependence of @xmath184 on such parameters softens their impact in the final result . the best limit shown in table [ tab_cuore ] is in fact only one order of magnitude better than the present limits of solax and cosme - ii . the @xmath185 bound that cuore could provide is depicted comparatively to other limits in figure [ fig ] . the limit which can be expected from the cuoricino experiment is comparable to the helioseismological bound mentioned before ( see table [ tab_cuore ] ) . cuore could go even further ( see figure [ fig ] and table [ tab_cuore ] ) . notice that in both cases an energy threshold of @xmath186 kev ( and resolution of @xmath51 kev ) has been assumed . that value will be confirmed only after knowing the performances and preliminary runs of cuoricino . as was described at length in ref . @xcite , the crucial parameters for estimating the perspectives on solar axion detectors with crystals rely on the energy threshold and resolution ( appearing in @xmath164 ) , and the level of background achieved ( although the influence of this parameter is damped by a factor 1/8 ) . in particular , a threshold of 5 - 8 kev would loose most of the axion signal . other crystal detectors with , say , ge or nai ( genius , majorana , gedeon , dama , libra , anais , ) could surpass cuore as axion detectors because the energy thresholds of these projects are supposed to be significantly lower . also the background is expected to be better . nevertheless , it should be stressed that the bounds on @xmath187 obtained with this technique in the various proposed crystal detector arrays stagnate at a few @xmath188 gev@xmath55 , not too far from the goal expected for cuore , as has been demonstrated in @xcite . there are no realistic chances to challenge the limit inferred from hb stars counting in globular clusters @xcite and a discovery of the axion by cuore would presumably imply either a systematic effect in the stellar - count observations in globular clusters or a substantial change in the theoretical models that describe the late - stage evolution of low - metallicity stars . to obtain lower values of @xmath185 one should go to the magnet helioscopes like that of tokio @xcite and that of cern ( cast experiment @xcite currently being mounted ) . in particular , the best current experimental bound of @xmath185 published comes from the tokyo helioscope : @xmath189 gev@xmath55 for @xmath190 ev and @xmath191 gev@xmath55 for @xmath192 ev . the sensitivity of cast is supposed to provide a bound @xmath193 gev@xmath55 or even lower . a recent , preliminary run of cast , with only @xmath55 hours of `` axion light '' has improved already the tokyo limit . we have reported the perspectives of cuore , a projected massive 760 kg array of 1000 teo@xmath3 bolometers , and of its first stage cuoricino , with 40 kg of the same crystals , as far as their physics potential to detect various types of rare events is concerned . the estimated background and resolution , based on monte carlo studies , together with the results obtained in recent improvements in the performances of the midbd experiment and the information obtained from the preliminary tests of cuoricino , have allowed us to assess the potentialities of these experiments for double beta decay searches , solar axion detection and wimp exclusion or identification . in these three types of searches , cuore and to some extent cuoricino will be powerful tools to explore , with higher sensitivity , such rare phenomena . this work has been partially supported by the spanish cicyt ( contract aen99 - 1033 ) , italian infn , the us national science foundation and the eu network ( contract erb - fmrx - ct-98 - 0167 ) . thanks are due to our student f. capozzi for the work done on the mc simulations , and to g. luzon for discussions on the cosmogenic activation of the crystals . booth , b. cabrera and e. fiorini , ann . nucl . part . * 46 * ( 1996 ) 471 . k. pretzl , nucl . intrum . and meth . a * 454 * ( 2000 ) 114 . e. fiorini , nucl . * 91 * ( 2001 ) 262 . l. mosca , review talk given at the impf2002 , feb . 2002 , jaca , huesca ( spain ) , to appear in nucl . b proc.suppl . 2003 . for a recent survey of wimp detection , see for instance a. morales , nucl . * 87 * ( 2000 ) 477 [ astro - ph/9912554 ] and review talk given at taup 2001 , sept . 2001 , lngs , [ astro - ph/0112550 ] , nucl . * 110 * ( 2002 ) 39 . r. abusaidi _ et al . _ [ cdms collaboration ] , phys . * 84 * ( 2000 ) 5699 [ astro - ph/0002471 ] . d. abrams _ et al . _ [ cdms collaboration ] , [ astro - ph/0203500 ] . a. benoit _ et al . _ [ edelweiss collaboration ] , phys . * b513 * ( 2001 ) 15 [ astro - ph/0106094 ] . g. gerbier , talk given at neutrino 2002 , munich 30 may 2002 . m. altmann _ et al _ , proceedings of the x international symposium on lepton and photon interactions at high energies , 2001 , [ astro - ph/0106314 ] . j. jochum _ et al _ , nucl . b. ( proc . suppl . ) 87 ( 2000 ) 70 . r. bernabei _ et al . _ [ dama collaboration ] , phys . b * 480 * ( 2000 ) 23 ; r. bernabei _ et al . _ [ dama collaboration ] , phys . b * 450 * ( 1999 ) 448 ; r. bernabei _ et al . _ , phys . b * 424 * ( 1998 ) 195 . s. cebrin _ et al . _ , . suppl . * 110 * ( 2002 ) 94 , [ hep - ex/0111075 ] . a. morales _ et al . _ [ igex collaboration ] , phys . b * 489 * ( 2000 ) 268 [ hep - ex/0002053 ] ; a. morales _ et al . _ [ igex collaboration ] , phys . b * 532 * ( 2002 ) 8 [ hep - ex/0110061 ] . d. gonzalez _ et al . _ , nucl . * 87 * ( 2000 ) 278 . h. v. klapdor - kleingrothaus _ et al . _ , eur . phys . j. a * 12 * ( 2001 ) 147 . a. alessandrello _ et al . _ , submitted to nim a , `` cuore : a cryogenic underground observatory for rare events '' . the evaluation of cosmogenic activation is obtained with a code based on the measured cosmic ray flux and on the interaction cross sections contained in the following articles : r. silberberg and r. tsao , astrophys . * 220 * ( 1973 ) 315 for an updated review , see for instance a. morales , nucl . * 77 * ( 1999 ) 335 [ hep - ph/9809540 ] ; h. ejiri , nucl . phys . * 91 * ( 2001 ) 255 . e. fiorini , nucl . * 91 * ( 2001 ) 262 and o. cremonesi , review talk given at neutrino 2002 , munich , 30 may 2002 . w. hampel _ _ , the gallex collaboration , phys . lett . * b 388 * , 384 ( 1996 ) ; g. n. abdurashitov _ _ , the sage collaboration , phys . * 77 * , 4708 ( 1996 ) ; b. t. cleveland _ et al . _ , astrophysics j. * 496 * , 505 ( 1998 ) ; r. davis , prog . part . phys . * 32 * , 13 ( 1994 ) ; y.fukuda _ et al . lett . * 86 * , 5651 ( 2001 ) ; * 86 * , 5656 ( 2001 ) . q. r. ahmad _ et al . _ [ sno collaboration ] , phys . rev . lett . * 87 * ( 2001 ) 071301 , [ nucl - ex/0106015 ] . s. pascoli and s. t. petcov , phys . b 544 ( 2002 ) 239 , [ hep - ph/0205022 ] . b. kayser , talk given at neutrino 2002 , 30 may 2002 . j. g. hirsh _ et al . _ , a589 * ( 1995 ) 445 ; c. r. ching _ et al . _ , phys . rev . * c40 * ( 1984 ) 304 ; x. r. wu _ et al . _ , phys . lett . * b272 * ( 1991 ) 169 ; x. r. wu _ et al . lett * b276 * ( 1992 ) 274 e. caurier _ et al . _ , rev . lett . * 77 * ( 1996 ) 54 ; retamosa _ et al . _ , phys . rev . * 51 * ( 1995 ) 371 ; a. poves _ et al . _ , phys . lett . * b361 * ( 1995 ) 1 ; e. caurier _ et al . _ , nucl . phys . * a654 * ( 1999 ) 973c . a. k. drukier , k. freese and d. n. spergel , phys . d * 33 * ( 1986 ) 3495 . sarsa , a. morales , j. morales , e. garca , a. ortiz de solrzano , j. puimedn , c. senz , a. salinas , j.a . villar , _ phys . lett . _ * b386 * ( 1996 ) 458 . d. abriola , f.t . avignone , iii , r.l . brodzinski , j.i . collar , d.e . di gregorio , h.a . farach , e. garca , a.o . gattone , c.k . gurard , f. hasenbalg , h. huck , h.s . miley , a. morales , j. morales , a. ortiz de solrzano , j. puimedn , j.h . reeves , a. salinas , m.l . sarsa , j.a . villar , _ astrop . * 10 * ( 1999 ) 133 . s. cebrian _ et al . _ , astropart . * 14 * ( 2001 ) 339 [ hep - ph/9912394 ] . f. donato , n. fornengo and s. scopel , astropart . * 9 * ( 1998 ) 247 [ hep - ph/9803295 ] ; m. kamionkowski and a. kinkhabwala , phys . rev . * d 57 * ( 1998 ) 3256 [ hep - ph/9710337 ] ; j. d. vergados , phys . rev . * 83 * ( 1999 ) 3597 ; j. d. vergados , phys . rev . * d62 * ( 2000 ) 023519 [ astro - ph/0001190 ] ; p. ullio and m. kamionkowski , jhep 0103 ( 2001 ) 049 [ hep - ph/0006183 ] ; a. m. green , phys.rev . * d63 * ( 2001 ) 043005 [ astro - ph/0008318 ] ; c. j. copi , j. heo and l. m. krauss , phys.lett . * b461 * ( 1999 ) 43 ; p. belli _ et al . _ , [ hep - ph/0203242 ] . m. brhlik and l. roszkowski , phys . lett . * b464 * ( 1999 ) ; p. belli , r. bernabei , a. bottino , f. donato , n. fornengo , d. prosperi y s. scopel , phys . * d61 * ( 2000 ) 023512 [ hep - ph/9903501 ] ; a. bottino , f. donato , n. fornengo and s. scopel , phys . rev . * d62 * ( 2000 ) 056006 [ hep - ph/0001309 ] . r. d. peccei and h. r. quinn , phys . * 38 * ( 1977 ) 1440 . k. van bibber , p. m. mcintyre , d. e. morris and g. g. raffelt , phys . d * 39 * ( 1989 ) 2089 . r. j. creswick , f. t. avignone , h. a. farach , j. i. collar , a. o. gattone , s. nussinov and k. zioutas , phys . * b427 * ( 1998 ) 235 [ hep - ph/9708210 ] . c. csaki , n. kaloper and j. terning , phys . * 88 * ( 2002 ) 161302 . f. t. avignone , d. abriola , r. l. brodzinski , j. i. collar , r. j. creswick , d. e. digregorio , h. a. farach , a. o. gattone , c. k. gurard , f. hasenbalg , h. huck , h. s. miley , a. morales , j. morales , s. nussinov , a. ortiz de solrzano , j. h. reeves , j. a. villar , and k. zioutas , phys . rev . * 81 * ( 1998 ) 5068 [ astro - ph/9708008 ] . i. g. irastorza , s. cebrian , e. garca , d. gonzalez , a. morales , j. morales , a. ortiz de solrzano , a. peruzzi , j. puimedn , m.l . sarsa , s. scopel and j.a . villar , nucl . * 87 * ( 2000 ) 102 [ astro - ph/9912491 ] ; a. morales , f.t . avignone iii , r.l . brodzinski , s. cebrin , e. garca , d. gonzlez , i.g . irastorza , h.s . miley , j. morales , a. ortiz de solrzano , j. puimedn , j.h . reeves , m.l . sarsa , s. scopel , j.a . villar , astropart . * 16 * ( 2002 ) 325 [ hep - ex/0101037 ] f. t. avignone iii _ et al . _ , suppl . ) 72 ( 1999 ) 176 . s. cebrian , e. garca , d. gonzalez , i.g . irastorza , a. morales , j. morales , a. ortiz de solrzano , j. puimedn , a. salinas , m.l . sarsa , s. scopel , j.a . villar , astropart . * 10 * ( 1999 ) 397 [ astro - ph/9811359 ] . y. inoue , t. namba , s. moriyama , m. minowa , y. takasu , t. horiuchi and a. yamamoto , [ astro - ph/0204338 ] . for an updated review see , for instance , p. sikivie , nucl . * 87 * ( 2000 ) 41 and g. raffelt , nucl . * 77 * ( 1999 ) 459 . [ hep - ph/9806506 ] ) plane for cuoricino , assuming a threshold of 10 kev , flat background @xmath195 1 ( solid line ) and 0.1 c / kev / kg / day ( dashed line ) and two years of exposure ( 81 kg year ) . it has been calculated for @xmath196 5.6 ( see the text ) . the closed contour represents the 3@xmath147 cl region singled out by the modulation analysis performed by the dama experiment @xcite.,scaledwidth=100.0% ] ) plane for cuore , assuming a threshold of 10 kev , two years of exposure ( 1500 kg year ) and flat backgrounds of 0.05 and 0.01 c / kev / kg / day . it has been calculated for @xmath196 5.6 ( see the text ) . the sensitivity curve has been also calculated for a possible threshold of 5 kev with a background of 0.01 c / kev / kg / day ( dashed line ) . the closed contour represents the 3@xmath147 cl region singled out by the modulation analysis performed by the dama experiment @xcite.,scaledwidth=100.0% ]

the construction of several new detectors of gravitational radiation is currently approaching completion . these instruments are of a different design and have significantly better sensitivity and broader bandwidth than previous detectors . they include the ligo detector being built in the united states by a joint caltech / mit collaboration @xcite , the virgo detector being built near pisa by an italian / french collaboration @xcite , the geo-600 detector being built in hannover by an anglo / german collaboration @xcite , and the tama-300 detector near tokyo @xcite . there are also several resonant bar detectors currently in operation @xcite , and several more refined bar and interferometric detectors presently in the planning and proposal stages @xcite . these instruments search for very weak signals . for the most likely sources , the signals will be buried in the noise of the detectors , and need to be extracted with sophisticated optimal signal - processing strategies @xcite . the standard assumption made in the literature is that the detector noise has multivariate gaussian statistics . this assumption is certainly incorrect : every sensitive gravitational wave detector operated to date has been characterized by noise that is both non - stationary and non - gaussian . some experimentation has shown that this is a serious matter @xcite : existing detection strategies for both deterministic and stochastic signals do not perform nearly as well when non - gaussian noise is present . roughly speaking , if the non - gaussian fluctuations are large , they bias the statistics and make it more difficult to achieve a given level of statistical confidence . in this paper , we develop a new set of statistical signal - processing techniques to search for deterministic and stochastic gravitational waves . these techniques are _ robust _ , meaning that they will work well even if the detector noise is not gaussian but falls into a broader statistical class that we expect includes realistic detectors . in large part , these new methods are similar to the older ones : one constructs matched filters to search for known waveforms or cross - correlates the instrument outputs at the different detector sites to search for a stochastic background . the essential difference is that by using locally optimal methods @xcite these statistical measures are modified . the effect is to _ truncate _ the statistics : detector samples that fall outside the central gaussian - like part of the sample distribution ( i.e. , the outliers ) are excluded from ( or saturated when constructing ) the measurement statistic . for both cases , a robust statistic is found which performs better than the optimal linear filter in the case where the detector noise is non - gaussian , and almost as well in the gaussian - noise case . the paper is organized as follows . in sec . [ s : deterministic ] we derive the locally most powerful signal - processing tests for deterministic signals . we begin in sec . [ ss : sdwnkw ] with a derivation of the neyman - pearson criteria for optimality , in the case where a known waveform is hidden in white noise . we define the power function of a test and derive a criteria for the locally optimal test in the weak - signal regime . the locally optimal test is analyzed for a number of different types of non - gaussian noise , and we show that the locally optimal decision statistic is a matched filter where the non - gaussian sample values are truncated or excluded . in sec . [ ss : sdcnkw ] the results are generalized to the case where a known waveform is hidden in colored noise , and we introduce models for non - gaussian colored noise . in sec . [ s : stochastic ] , we turn to the detection of a stochastic background . section [ ss:2dwnss ] considers the case of a stochastic signal ( i.e. , where the waveform is not known ) and derives the locally optimal statistic which can be used to correlate two identical detectors , where we assume that each detector has independent white noise and is co - aligned and coincident . in sec . [ ss:2dcnss ] , these results are generalized to the case where the noise is colored , and the detectors are in different locations , and not aligned in parallel . in sec . [ s : impsim ] , we discuss an implementation of these statistics , and we illustrate how one can compare the performance of different statistics using monte carlo simulations . section [ s : conclude ] contains a short conclusion and summary . in order to describe the idea in a simple way we first discuss the case where we are searching for a known signal in the data stream of a single detector , where the time - domain detector noise samples are independent in the time - domain . denote the data stream of the first detector by @xmath0 for @xmath1 . in this section , since we are going to only consider this single detector , we will drop the subscript `` 1 . '' imagine that we are looking for a signal of known waveform but unknown amplitude @xmath2 , which we will denote by @xmath3 . our primary interest is in the case where the amplitude @xmath2 is either small , or zero . for convenience , imagine for the moment that this parameter can have only two possible values , either @xmath4 or @xmath5 . the detection problem that we need to solve is to partition the space of possible observations @xmath6 into two disjoint subsets . when the observation @xmath7 falls into one of these , we conclude that @xmath4 and that the null hypothesis is true . when the observation falls into the other set we conclude that the signal has been observed with @xmath8 . to describe the partition of @xmath6 into two regions , define a function @xmath9 which is zero in the null hypothesis region and unity elsewhere . this function is called a _ test_. our goal is to find the `` best '' choice of a test @xmath10 . to help characterize tests @xmath10 , it is helpful to define the _ power _ function of a test : @xmath11 here @xmath12 is the probability distribution of the measurement @xmath7 given signal amplitude @xmath2 . for example , for additive white , stationary gaussian noise of unit variance and vanishing mean , @xmath13 the quality of the test can be expressed in terms of the power function . we characterize the quality of the test by the false alarm and the false dismissal probabilities . the false alarm probability is the probability with which we conclude that @xmath8 when in fact @xmath2 vanishes . this is given by @xmath14 . the false dismissal probability is the probability with which we conclude that @xmath4 when in fact it is @xmath15 . this is given by @xmath16 . one standard definition of the `` best '' test @xmath10 is that it minimizes the false dismissal probability for a given false alarm probability . this is called the neyman - pearson test . one can find this test using calculus of variations , with a lagrange multiplier @xmath17 to enforce the constraint that the false alarm probability is fixed . the best test is obtained by partitioning @xmath6 as follows . choose a constant @xmath18 . then , set @xmath19 in regions where the likelihood ratio @xmath20 is greater than @xmath21 . set @xmath22 elsewhere . ( we assume that the boundary between these two regions is a set of probability measure zero . ) the value of the constant @xmath21 determines the false alarm probability . thus , the likelihood ratio is a `` decision statistic '' : a number that can be calculated from the observed data . if the statistic is less than some value , we conclude that the null hypothesis holds . if the statistic is greater than this value , we conclude the opposite . the decision statistic provides a partition of the space of observations into two disjoint regions . in the case where the noise is gaussian ( [ e : gaussnoise ] ) this criteria is easily understood . the optimal neyman - pearson test divides the space of observation along an @xmath23-dimensional plane . on one side of this plane @xmath19 and on the other side @xmath10 vanishes . the plane is defined by setting the likelihood ratio ( [ e : like ] ) to a constant . for the gaussian probability distribution ( [ e : gaussnoise ] ) the plane is defined by @xmath24 this plane is perpendicular to the vector @xmath25 . different choices of this plane correspond to different false alarm rates . in the case where the noise is not gaussian , the problem becomes more challenging . in the gaussian test , the decision statistic is independent of the signal amplitude @xmath2 . however when the noise in not gaussian , the choice of decision statistic depends upon @xmath2 . consider the graph in fig . [ f : powerfn ] showing the power function @xmath26 as a function of @xmath2 for several different tests . all the tests have the same false alarm rate , but the optimal test depends upon the value of @xmath2 . for the case of weak signals in non - gaussian noise , there is a useful test called the `` locally optimal '' test . for a given noise probability distribution , the locally optimal test is easy to describe , and leads to a simple decision statistic which can be calculated from the observed data @xcite . to define this test , it is useful to again consider the set of all tests with a given false alarm rate , as shown for example in fig . [ f : powerfn ] . the locally optimal test is the one that maximizes @xmath27 at @xmath4 for a fixed false alarm probability . as above , one can show that the locally optimal test sets @xmath19 inside the region where @xmath28_{\epsilon=0}>\text{constant}\ ] ] for some constant , and @xmath22 elsewhere ( see fig . [ f : powerfn ] ) . the value of the constant determines ( or is determined by ) the false alarm probability . more generally , if the first derivative vanishes , the locally optimal test is determined by the first non - zero @xmath29 to understand the implications of this , it is helpful to consider several examples . the examples here are for the case where the ( additive ) detector noise is independent for each sample value ( so the noise spectrum is white ) but has an arbitrary probability distribution . for convenience , we write @xmath30 where the function @xmath31 is a quadratic function of its argument for the case where the probability distribution of the noise is gaussian . [ note : any probability distribution for stationary additive noise where the sample values are independent can be written in this way . if the noise is not stationary but is still additive and independent , then each function @xmath31 appearing in ( [ e : nongaussnoise ] ) may be different @xmath32 . ] the first derivative of the pdf ( [ e : nongaussnoise ] ) with respect to @xmath2 is @xmath33 where @xmath34 denotes the derivative of @xmath31 with respect to its argument . setting @xmath4 in this expression one can easily find the locally optimal test ( [ e : lotest ] ) . this is defined by setting @xmath19 in the region @xmath35 and setting @xmath22 elsewhere . [ note : if @xmath2 can take either sign @xmath36 then an absolute value sign should enclose the lhs of the inequality in eq . ( [ e : forexamples ] ) . ] as before , the value of the constant determines the false alarm probability . here are several examples : * * gaussian noise * : @xmath37 , so @xmath38 . for this case the locally optimal test ( [ e : forexamples ] ) and the optimal test ( [ e : simplegauss ] ) both give the same statistic : @xmath39 . this is the standard optimal linear filter . * * exponential noise * : @xmath40 , so @xmath41 . here the locally optimal statistic is given by ( [ e : forexamples ] ) as @xmath42 where the @xmath43 function is @xmath44 for @xmath45 and @xmath46 for @xmath47 . * * sum of distinct gaussian processes * : this is a white - noise version of the model given in @xcite : @xmath48 where @xmath49 and @xmath50 . usually one also has @xmath51 . this noise model is discussed in more detail later in this paper . it often arises when the most common source of noise is gaussian , but there is also a `` tail '' of `` outlier '' events which dominates the wings of the distribution . here the locally optimal statistic is defined by ( [ e : forexamples ] ) where @xmath52.\ ] ] this function is shown ( for the case @xmath53 , @xmath54 , @xmath55 ) in fig . [ f : weight2 ] . roughly speaking , for @xmath56 small compared to @xmath57 one has @xmath58 . for large @xmath56 one has @xmath59 . * * gaussian noise plus uniform background * : here , we have a ( small ) uniform background superposed on gaussian noise of zero mean and unit variance . this is defined for ( small ) @xmath60 by @xmath61 here we assume that @xmath62 is the scale size of the uniform background ( the probability distribution is correctly normalized only in the limit @xmath63 ) . in this case one finds that @xmath64 for @xmath65 and @xmath66 for @xmath67 . while the results for the different probability distributions are technically different , they all carry same message , which is the central result of this paper : _ if the distribution of sample values has a central gaussian region , then sample values falling in this region should be correlated exactly as they would be in the gaussian case . if a sample value falls outside this region , its value should be truncated ( or clipped ) to the largest allowed value in the central region , or even dropped from any correlation statistic , depending upon the shape of the probability distribution_. let us repeat this central point one more time . the results show that when the noise is not gaussian , the normal optimal filter used to construct a decision statistic is replaced by a somewhat different sum . the values of the expected signal @xmath68 are multiplied , not by the observed data @xmath69 but by some non - linear function of that data , then summed . in the event that the probability distribution of the noise has a non - gaussian tail , the effect of this non - linear function is to `` clip '' or truncate sample values which fall outside the central bulge of the probability distribution function . if the detector s noise spectrum is colored rather than white , then the previous analysis does not apply : the assumption that the different sample values are uncorrelated no longer holds . however the analysis can be generalized to the colored case if we make assumptions that are motivated by the properties of stationary detector noise . in explaining this , it helps to begin by describing the stationary gaussian case . for a colored gaussian process , the probability density function ( pdf ) of the detector samples may be expressed as @xmath70 where the @xmath71 correlation matrix @xmath72 is a positive - definite real symmetric matrix with @xmath73 real degrees of freedom and @xmath74 . we have assumed that the process has zero mean . the volume element associated with this pdf is @xmath75 . in the time - domain , @xmath7 is a vector of real numbers so @xmath76 . in the case where the random process is stationary , the matrix @xmath77 is a toeplitz matrix , which depends only upon @xmath78 . such a process is defined by the first row or first column of the matrix and has only @xmath79 real degrees of freedom . thus stationary gaussian processes are a tiny subset of _ all _ gaussian processes . now consider the pdf of new random variables that are linear combinations of the old ones : @xmath80 . take @xmath81 to be an arbitrary unitary matrix . clearly the pdf of these new variables @xmath82 is still gaussian . the matrix @xmath81 can be chosen to diagonalize the correlation matrix : this is called a karhunen - loeve transformation . in the limit where the time interval occupied by the @xmath79 samples is much larger than the correlation time of the noise , the linear combinations of random variables that diagonalize the correlation matrix asymptotically approach the discrete fourier transform ( dft ) . this is given by @xmath83 thus , if @xmath79 is sufficiently large , to a good approximation the pdf of the new variables in the gaussian case may be written as @xmath84 } 2\pi^{-1}p_k^{-2 } \exp(-2|\tilde x_k|^2/p_k)\ ] ] where @xmath85 is the ( real , positive ) mean spectral amplitude in the @xmath86th frequency bin : @xmath87 for @xmath88 $ ] ; thus @xmath89 $ ] . in other words , it is a good approximation to express the pdf of a stationary colored gaussian process as a diagonal process in frequency space . the limits of the product in eq . ( [ e : fspacegauss ] ) appear strange because @xmath82 can not take on arbitrary values since @xmath7 is real . the consequences include : * @xmath90 . hence the amplitudes of @xmath91 for @xmath92 + 1,\ldots , n-1 $ ] are completely determined by @xmath91 for @xmath93 $ ] . * @xmath94 and , for even @xmath79 , @xmath95 are real . however , we assume that the data set has had the mean value ( dc term ) removed : @xmath96 . since gravitational wave detectors are ac - coupled and have no useful low - frequency response , this is a valid assumption . it implies that @xmath94 is identically zero . second , when @xmath79 is even , we assume that there is no energy in the nyquist frequency bin : @xmath97 also vanishes identically . this is a very reasonable assumption , since an experiment will include an anti - aliasing filter whose response ( as a function of frequency ) falls off rapidly as the nyquist frequency is approached . * the volume element associated with this pdf is therefore @xmath98}d(\re\tilde x_k)d(\im\tilde x_k)$ ] . the likelihood ratio in the case of colored , stationary , gaussian noise is @xmath99 or , in the frequency domain , @xmath100}\tilde s_k^\ast \tilde x_k / p_k.\ ] ] thus , the matched filter statistic , with a weighting equal to the inverse of the noise spectrum , is the optimal detection statistic . this motivates a more general model for the statistical distribution of colored non - gaussian detector noise , assuming that it is still stationary . in this case , to good approximation , the two - point correlation matrix @xmath101 is diagonal . there may be higher - order correlations present between the fourier amplitudes at different frequencies , but we will assume that this additional correlation is negligible , and that to a reasonable approximation the probability distribution of the noise in the non - gaussian case is described by a pdf in which the different frequency components are independent : @xmath102 } 2 \pi^{-1 } p_k^{-2 } \exp[-2g_k(|\tilde x_k|^2/p_k)],\ ] ] with volume element @xmath98}d(\re\tilde x_k)d(\im\tilde x_k)$ ] . the functions @xmath103 depend upon the frequency bin index @xmath86 , so that the statistical distribution can depend upon the frequency . for the colored gaussian case the functions are @xmath104 . in order that the pdf be properly normalized , and that @xmath105 , the functions @xmath103 must obey @xmath106 respectively , these constrain the additive constant in the definition of @xmath107 , and the multiplicative scale of the argument of @xmath107 . this is _ not _ the most general possible form of the probability distribution of a stationary random process , but in many situations it should be a reasonable approximation , particularly if the quantities of interest are dominated by the second moments . the locally optimal statistic may now be easily derived . letting @xmath108 denote the dft of the expected waveform , and as before zeroing its dc and nyquist components , the conditional probability distribution of the detector output is given by @xmath109 } 2 \pi^{-1 } p_k^{-2 } \exp[-2g_k(|\tilde x_k-\epsilon\tilde s_k|^2/p_k)]\ ] ] the locally optimal test can then be obtained from the first derivative : @xmath110}\re(\tilde s_k^\ast \tilde x_k / p_k ) g_k'(|\tilde x_k|^2/p_k).\ ] ] in the colored gaussian case @xmath111 this is the ordinary optimal linear matched filter . the contributions of the different frequency bins are weighted by the inverse noise power spectrum in that bin . in the non - gaussian case , just as for the case of uncolored white noise , the correlation in frequency space is clipped or truncated for ( frequency - bin ) samples that lie outside the central gaussian part of the probability distribution , where @xmath112 . an example of this may be seen in fig . [ f : weight2 ] : for the illustrated case @xmath113 . let us consider another form of non - gaussian noise that describes a process in which there is an ambient gaussian noise background interrupted occasionally by a large noise burst , which we will model as a second component of gaussian noise with a much larger variance . the probability distribution we adopt is @xcite : @xmath114 where @xmath77 is the autocorrelation matrix for the normal ambient detector noise and @xmath115 is the composite autocorrelation matrix for the detector noise when a noise burst is present . the noise bursts occur with probability @xmath116 in this model . also , @xmath74 and @xmath117 . we assume that @xmath118 is much smaller than @xmath119 , meaning that @xmath120 for all vectors @xmath7 . the locally optimal statistic is @xmath121 where @xmath122\ ] ] is a detector of possible bursts . when a burst is absent , @xmath123 is typically small and the locally optimal statistic reduces to the matched filter . however , when a burst is present , @xmath123 is typically large and the matched filter is suppressed . thus the locally optimal statistic is nearly equivalent to the matched filter statistic with a veto if a segment of data has a large amount of excess power as measured by @xmath124 or , in the frequency domain , @xmath125 } |\tilde x_k|^2/p_k.\ ] ] the lengths ( in time ) of the data chunks used to estimate the autocorrelation matrices should be choosen to be significantly longer than the characteristic time of the signals being searched for , but still short enough that the detector behavior is quasi - stationary . for inspiral signals , typical signals are in the detector band for tens of seconds , so the matrix estimation time should be at least of order tens of minutes . for stochastic background detection , the correlation time between the two instruments is tens of milliseconds , so that the matrix estimation time should be at least a few seconds . based on the two forms of non - gaussian noise considered in this section , it seems reasonable to adopt the following detection rules : ( i ) veto immediately any segment of data that has an excess of power as measured by the excess power statistic ; ( ii ) for segments of data without an excess of power , construct the matched filter in the frequency domain , but exclude those frequency bins in which the detector power is too large . the resulting ( truncated ) matched filter is a good approximation of the locally optimal statistic for a wide variety of possible non - gaussian noise distribution . in this sense , it is a robust , nearly optimal detection statistic . observational limits from nucleosynthesis demonstrate that the stochastic background of gravitational radiation has such small amplitude that it would not be detectable with a single instrument @xcite . in a single instrument , there would be no practical way to discriminate between intrinsic detector noise and the small additional noise - like output arising from a stochastic background . however , one can correlate the outputs of two different instruments and search for a common signal in this way . if the instrumental noise is not gaussian , then the previous single - detector analysis can be easily generalized . we begin by considering the simple case in which the two detectors are coincident and co - aligned , so that they have identical output contributions from the stochastic background but independent intrinsic noise . we also assume that the intrinsic noise samples in each detector are independent , and hence white . if the signal were deterministic ( known ) then the joint probability distribution for the samples in the two detectors could be written as @xmath126 this system can be analyzed in exactly the same way as in sec . [ ss : sdwnkw ] . however the stochastic background does not produce a known ( deterministic ) signal , so that the probability distribution needs to be averaged over its expected distribution @xmath127 ( which , by reason of the central limit theorem , is almost certainly a multivariate gaussian ) . this leads to a joint probability distribution which is given by @xmath128 here @xmath2 may be thought of as the coupling of the detector . the case of small @xmath2 corresponds to a detector that is only weakly coupled to the signal . for this non - deterministic signal , it is still straightforward to construct a locally optimal test , and a corresponding decision statistic or threshold criterion . the locally optimal statistic is obtained from the derivative of the probability distribution with respect to @xmath2 . this is given by @xmath129 ) \right ) \nonumber\\ & & \times\prod_{i=0}^{n-1 } e^{-f_1(x_{1,i}-\epsilon s_i)-f_2(x_{2,i}-\epsilon s_i)}.\end{aligned}\ ] ] setting @xmath4 and dividing by @xmath130 yields the locally optimal statistic : @xmath131 \times\int s_j p_{\text{sb}}({{\mathbf{s } } } ) d{{\mathbf{s}}}.\ ] ] unfortunately this vanishes if the random process described by @xmath132 has vanishing mean , since in this case @xmath133 . this is indeed the case for the gravitational - wave stochastic background . when the _ first _ derivative vanishes , the locally optimal statistic is defined by having the largest _ second _ derivative at @xmath4 . see fig . [ f : powerfn ] for example . taking another derivative of ( [ e : firstder ] ) and setting @xmath4 yields @xmath134{l } \biggl\{\bigl(\sum_{j=0}^{n-1 } s_j [ f_1'(x_{1,j})+f_2'(x_{2,j } ) ] \bigr)^2\\ \quad-\sum_{j=0}^{n-1}s_j^2[f_1''(x_{1,j})+f_2''(x_{2,j } ) ] \biggr\ } \end{array}\ ] ] the terms that appear in this statistic have different character , and before moving on , some discussion is required . the locally optimal statistic depends upon the statistical character of the stochastic background radiation through the second - order moments . we will assume that the stochastic background is a stationary process , so that the second order correlation @xmath135 is a function of the lag @xmath78 only : @xmath136 in a stochastic background search , the `` signal model '' only requires an assumption about the form of the spectrum . this is ( roughly ) the fourier transform of @xmath137 . without loss of generality we normalize @xmath137 so that @xmath138 ( this simply scales the value of @xmath2 ) . expressing the locally optimal statistic in terms of the correlation function @xmath139 then gives @xmath140 \nonumber\\ & & + \sum_{j , k=0}^{n-1 } c(|j - k| ) [ f_1'(x_{1,j } ) f_1'(x_{1,k } ) \\ & & \qquad + f_2'(x_{2,j } ) f_2'(x_{2,k } ) + 2 f_1'(x_{1,j } ) f_2'(x_{2,k } ) ] . \nonumber\end{aligned}\ ] ] each of the five terms that appears in ( [ e : lostochbg ] ) has a specific interpretation . the first four terms that appear in the locally optimal estimator @xmath141 are generalized `` single - detector '' statistics which do not cross - correlate the two detectors . they are generalized measures of the `` energy '' received by each individual detector , and provide useful information only if the stochastic background contributes substantially more to the measured signal than the detector output does , or if the detector s intrinsic noise contributions can somehow be separated from the noise contribution arising from the stochastic background . ( this will not be the case for the first few generations of gravitational wave detectors ) . the last term in ( [ e : lostochbg ] ) is a generalized cross - correlation ( gcc ) statistic , that provides useful information even if the detector noise dominates the signal : the expected case for gravitational - wave stochastic background . to quote from kassam ( following eq . ( 7 - 24 ) in ref . @xcite ) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ it is important to note that the increase in power level occurs whenever random signals are present at the individual receivers of the array regardless of whether the signals _ across _ the array are one common signal or are completely uncorrelated . the gcc part of the locally optimal ( lo ) statistic responds only to a _ common _ signal or at least to signals which are spatially correlated across the array elements . this is a major reason why it is useful to employ only the gcc part of the lo statistic in applications involving detection as well as location of signal sources . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for this reason ( and others is positive definite ) and that for the expected `` gaussian + tail '' detector probability distribution , term 1 is also positive definite . baysian analysis ( in part ii of this series of papers ) also shows that the single - detector terms must be dropped . this may also be seen in the analysis at the end of sec . [ ss : estimators ] . ] ) we drop the single - detector terms from the statistic , and define the gcc statistic as : @xmath142 this generalized cross - correlation statistic reduces to the ordinary cross - correlation statistic in the case where the detector noise is gaussian : @xmath143 . it can be easily generalized to the case of three or more detectors @xcite . in practical work @xmath139 will vanish for lags greater than the light travel time between the two detectors ( i.e. , 10 ms for the ligo detectors ) . this means that even if @xmath79 is chosen to be very large , @xmath144 only correlates samples from the two detectors taken within this time window . ( note : if the detector noise is colored , then the time window may be larger , as will be seen shortly . ) in this section , we generalize the work of sec . [ ss:2dwnss ] to the case where the two detectors are not coincident or co - aligned , and their noise power spectrum is not white . we assume that the intrinsic detector noise of the two detectors is independent . if the two detectors are widely separated and subject to different environmental influences , this assumption should hold . let us start by assuming that the two detectors each have internal ( instrumental ) colored gaussian noise with known autocorrelation matrices @xmath145 and @xmath146 , and that the instrumental noise of the two detectors is independent . the stochastic background produces an additional source of colored gaussian noise that is correlated between the two detectors . the stochastic background noise is measured by the autocorrelation matrices @xmath147 , @xmath148 , and the cross - correlation matrices @xmath149 , @xmath150 . since the stochastic background is isotropic , @xmath151 ( the stochastic background contribution to the detector s autocorrelation matrices ) and @xmath152 ( the cross - correlated noise between the detectors due to the stochastic background ) . the total autocorrelation noise of the two detectors are @xmath153 and @xmath154 . in the presence of the stochastic background , the likelihood ratio is @xmath155 where @xmath156 \quad\text{and}\quad { { \mathbf{\sigma } } } = \left [ \begin{array}{cc } { { \mathbf{r}}}_1 & \epsilon^2{{\mathbf{s } } } \\ \epsilon^2{{\mathbf{s } } } & { { \mathbf{r}}}_2 \end{array } \right].\ ] ] in the weak signal approximation , @xmath157 \nonumber\\ & & - \epsilon^2 \left [ \begin{array}{cc } { { \mathbf{q}}}_{\text{in},1}\cdot{{\mathbf{r}}}_{\text{sb}}\cdot{{\mathbf{q}}}_{\text{in},1 } & { { \mathbf{q}}}_{\text{in},1}\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_{\text{in},2 } \\ { { \mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_{\text{in},1 } & { { \mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{r}}}_{\text{sb}}\cdot{{\mathbf{q}}}_{\text{in},2 } \end{array } \right ] \nonumber\\ & & + \epsilon^4 \biggl\ { \left [ \begin{array}{c } { { \mathbf{q}}}_{\text{in},1}\cdot{{\mathbf{r}}}_{\text{sb}}\cdot{{\mathbf{q}}}_{\text{in},1 } \cdot{{\mathbf{r}}}_{\text{sb}}\cdot{{\mathbf{q}}}_{\text{in},1 } , \quad { { \mathbf{0 } } } \\ { { \mathbf{0 } } } , \quad { { \mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{r}}}_{\text{sb}}\cdot { { \mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{r}}}_{\text{sb}}\cdot{{\mathbf{q}}}_{\text{in},2 } \end{array}\right ] \nonumber\\ & & \quad + \left [ \begin{array}{c } { { \mathbf{q}}}_{\text{in},1}\cdot{{\mathbf{s}}}\cdot { { \mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_{\text{in},1 } , \quad { { \mathbf{0 } } } \\ { { \mathbf{0 } } } , \quad { { \mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{s}}}\cdot { { \mathbf{q}}}_{\text{in},1}\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_{\text{in},2 } \end{array } \right ] \biggr\ } \nonumber\\ & & + o(\epsilon^6),\end{aligned}\ ] ] @xmath158 \nonumber\\ & & - \epsilon^4 [ \frac{1}{2}{{\mathrm{tr}}}({{\mathbf{q}}}_{\text{in},1}\cdot{{\mathbf{r}}}_{\text{sb } } \cdot{{\mathbf{q}}}_{\text{in},1}\cdot{{\mathbf{r}}}_{\text{sb } } ) \nonumber\\ & & \qquad + \frac{1}{2}{{\mathrm{tr}}}({{\mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{r}}}_{\text{sb } } \cdot{{\mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{r}}}_{\text{sb } } ) \nonumber\\ & & \qquad + { { \mathrm{tr}}}({{\mathbf{q}}}_{\text{in},2}\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_{\text{in},1 } \cdot{{\mathbf{s } } } ) ] \nonumber\\ & & + o(\epsilon^6),\end{aligned}\ ] ] and @xmath159 where @xmath160 and @xmath161 . the last two terms represent the autocorrelation `` energy '' detectors . the following question now becomes important : how does one obtain the quantities @xmath145 and @xmath146 . there are two possible methods : ( i ) by a theoretical understanding of the detector , or ( ii ) by shielding the instrument from the stochastic background and measuring the noise autocorrelation . for gravitational wave searches , method ( ii ) is not available as there is no way to shield the detector from a stochastic background of gravitational waves . method ( i ) holds more promise , but if the stochastic background is expected to be weak , it is unlikely that our understanding of the detector will be sufficient to distinguish between the noise autocorrelations @xmath162 and @xmath163 . we expect that the noise matrices that should be used are the _ measured _ noise matrices @xmath164 and @xmath165 , which contain both the internal , instrumental noise as well as the stochastic background `` noise . '' since it is these quantities rather than @xmath145 and @xmath146 that are known , the previous analysis must be modified . we now have @xmath166 - \epsilon^2 \left [ \begin{array}{cc } { { \mathbf{0 } } } & { { \mathbf{q}}}_1\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_2 \\ { { \mathbf{q}}}_2\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_1 & { { \mathbf{0 } } } \end{array } \right ] \nonumber\\ & & + \epsilon^4 \left [ \begin{array}{c } { { \mathbf{q}}}_1\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_2\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_1,\quad { { \mathbf{0 } } } \\ { { \mathbf{0}}},\quad { { \mathbf{q}}}_2\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_1\cdot{{\mathbf{s}}}\cdot{{\mathbf{q}}}_2 \end{array } \right ] \nonumber\\ & & + o(\epsilon^6),\end{aligned}\ ] ] @xmath167 and @xmath168 where @xmath169 and @xmath170 . the locally optimal detection statistic ( which is appropriate for weak signals ) is the cross - correlation statistic . to generalize to non - gaussian noise , it is helpful to use moment generating functions . suppose the vector @xmath171 represents the internal ( instrumental ) noise in the first detector . the moment generating function for @xmath171 is @xmath172 and the probability distribution for @xmath171 is the fourier transform of the moment generating function : @xmath173 the moment generating function @xmath174 for the internal noise in detector 2 is defined similarly . we assume that the stochastic background is a multivariate gaussian with a moment generating function @xmath175 with @xmath176 \quad\text{and}\quad { { \mathbf{\sigma}}}_{\text{sb}}=\left [ \begin{array}{cc } { { \mathbf{r}}}_{\text{sb } } & { { \mathbf{s } } } \\ { { \mathbf{s } } } & { { \mathbf{r}}}_{\text{sb } } \end{array } \right].\ ] ] then the moment generating function for the detectors output is @xmath177 and the joint probability distribution is @xmath178 thus , if we ignore the autocorrelation terms , the locally optimal statistic is @xmath179 this equation for the locally optimal statistic is good for the time domain , in which the detectors output vectors are real and so the derivative is meaningful . to extend the result to complex vectors , and thus to a frequency - domain representation , we use the following formal replacement : replace every complex number @xmath180 and derivative @xmath181 with the matrices @xmath182 \quad\text{and}\quad \nabla \to \underline\nabla = \frac{1}{2 } \left [ \begin{array}{rr } \partial/\partial a & -\partial/\partial b \\ \partial/\partial b & \partial/\partial a \end{array } \right].\ ] ] note that this means @xmath183 is represented by @xmath184 and @xmath185 by @xmath186 . also , the meaning of @xmath187 is @xmath188 . the locally optimal statistic is @xmath189 for example , for the noise model in which @xmath190}g_k(|\tilde x_k|^2/2p_k)$ ] and @xmath191 $ ] , the locally optimal statistic is @xmath192 } \frac{\gamma_k\sigma^2_k\tilde x_{1,k}^\ast\tilde x_{2,k}}{p_{1,k}p_{2,k } } \nonumber\\ & & \quad\times g_{1,k}'(|\tilde x_{1,k}|^2/p_{1,k } ) g_{2,k}'(|\tilde x_{2,k}|^2/p_{2,k}).\end{aligned}\ ] ] before we examine specific non - gaussian noise models , we will describe the form of the matrices @xmath193 and @xmath194 . a stochastic background , if present , contributes to the signal amplitude at each detector . to simplify the analysis , in sec . [ ss:2dwnss ] , we assumed that the detectors were coincident and co - aligned , so that the amplitude contribution in each individual detector are identical . here , we drop that assumption . because the detectors are not co - aligned , the axes of the two interferometer arms point in different directions , and are sensitive to different linear combinations of the two possible gravitational wave polarizations . this reduces the correlation between the amplitudes in the two detectors , since we will assume that the stochastic background is unpolarized . an additional loss of correlation occurs because the two detectors are separated . this loss of correlation becomes increasingly greater for shorter wavelengths . roughly speaking , there is no significant loss of correlation for wavelengths much longer than the inter - detector distance , and there is a complete loss of correlations for wavelengths much shorter than this@xcite . the loss of amplitude correlation due to the separation and non - alignment of the two detectors may be described ( for an unpolarized and isotropic stochastic background ) in terms of the overlap reduction function @xmath195 defined by flanagan @xcite . this quantity is the average value of the product of the detector outputs , for a stochastic background of a given frequency @xmath31 , averaged over the possible directions of arrival and phases . it is given by : @xmath196 here @xmath197 is a unit - length vector on the two - sphere , @xmath198 is the separation between the two detector sites , and @xmath199 is the response of detector @xmath200 to the @xmath201 or @xmath202 polarization . for the @xmath200th detector @xmath203 one has @xmath204 where @xmath205 are the gravitational wave polarization tensors for a wave propagating in direction @xmath197 . the normalization of @xmath195 is chosen so that for coincident and co - aligned detectors , @xmath206 . for co - aligned but not coincident detectors , @xmath207 . for coincident but unaligned detectors , @xmath195 is a frequency - independent constant that depends only upon the relative orientation of the two detectors , and vanishes if the two detectors are sensitive to orthogonal polarizations . general expressions for @xmath195 for arbitrary detectors may be found in refs . @xcite . for the pair of ligo detectors @xmath195 is shown in fig . [ f : gammaligo ] , and is given by @xmath208 where @xmath209 is a frequency variable , @xmath210 is the detector separation , @xmath211 is the speed of light , and @xmath212 is a spherical bessel function . it is helpful to introduce notation for the overlap reduction function s values in the frequency bins of interest . let @xmath213 denote the frequency of the @xmath86th bin , with @xmath214 $ ] . here @xmath215 is the sample interval and @xmath216 is the total observation time . then @xmath217 are the values of the overlap reduction function in the @xmath86th bin . the stochastic background is characterized by its dimensionless energy density @xmath218 where @xmath219 is the energy density of the gravitational radiation contained in the frequency range @xmath31 to @xmath220 , and @xmath221 is the critical energy density required ( today ) to close the universe : @xmath222 @xmath223 is the hubble expansion rate ( today ) : @xmath224 and @xmath225 is a dimensionless factor that we have included to account for the different values of @xmath223 that are quoted in the literature . is probably very close to 0.65 . ] the pdf of the stochastic background strain can usually be expressed in closed form . the central limit theorem shows that if the stochastic background has been produced ( as it is in many scenarios ) by an incoherent sum of many small processes , then its statistics will be stationary@xcite and gaussian . this means that it is characterized by the single - site second moments @xmath226 with @xmath227 as before , we have assumed that @xmath79 is chosen so that @xmath228 is much larger than the correlation time of the stochastic background ( filtered by the instrument response function ) , so that the rhs of ( [ e : sbg2nd ] ) is proportional to @xmath229 . the expectation value of the product of the strain at the two different sites is reduced by the overlap reduction function : @xmath230 this follows from eqn . ( 3.56 ) of reference @xcite . in practice , since the shape of the stochastic background spectrum is not know , the dependence of the @xmath231 on @xmath86 should be assumed to fit some simple parameterized model , such as a power law @xmath232 for a reasonable range of @xmath123 . we can now express the locally optimal detection statistic for a stochastic background in colored gaussian noise . it is : @xmath233 } \gamma_k \sigma^2_k x_{1,k}^\ast x_{2,k}/(p_{1,k}p_{2,k})\ ] ] where @xmath234 and @xmath235 are the measured noise spectra in the two detectors . let us now turn to our first non - gaussian noise model . our starting point is a pdf for the noise in the two detectors in the absence of any stochastic background signal . we make the same assumptions about the detector noise as in sec . [ ss : sdcnkw ] . the pdf is given in frequency space by a product of two terms identical to ( [ e : assumedform ] ) , @xmath236 } 2 \pi^{-1 } p_{1,k}^{-2 } e^{-2g_{1,k}(|\tilde x_{1,k}|^2/p_{1,k } ) } \\ & & \times \prod_{k'=1}^{[(n-1)/2 ] } 2 \pi^{-1 } p_{2,k'}^{-2 } e^{-2g_{2,k'}(|\tilde x_{2,k'}|^2/p_{2,k ' } ) } \\ & = & \prod_{k=1}^{[(n-1)/2 ] } 4 \pi^{-2 } p_{1,k}^{-2 } p_{2,k}^{-2 } \\ & & \times e^{-2g_{1,k}(|\tilde x_{1,k}|^2/p_{1,k } ) -2g_{2,k}(|\tilde x_{2,k}|^2/p_{2,k})}.\end{aligned}\ ] ] the statistical distribution of the stochastic background is @xmath237 } ( \pi\sigma_k^2)^{-2}(1-\gamma_k^2)^{-1 } \nonumber\\ & & \quad \times \exp\left ( - \frac{|\tilde s_{1,k}|^2 + |\tilde s_{2,k}|^2 - 2\gamma_k \re(\tilde s_{1,k}^\ast \tilde s_{2,k } ) } { \sigma_k^2 ( 1-\gamma_k^2 ) } \right).\end{aligned}\ ] ] we can now find the locally optimal statistic . since the detector is linear , as before , one has a joint probability distribution for the observed fourier amplitudes : @xmath238 this corresponds to a stochastic background with a characteristic energy - density function @xmath239 . should really be interpreted as `` coupling to the detector '' . the locally optimal statistic is the best choice in the `` weakly coupled detector '' limit . ] the locally optimal statistic is @xmath240 } & & \biggl\ { \frac{\langle\tilde s_{1,k}^\ast\rangle \tilde x_{1,k } g_{1,k}'(|\tilde x_{1,k}|^2/p_{1,k})}{p_{1,k } } \nonumber\\ & & \quad + \frac{\langle\tilde s_{2,k}^\ast\rangle \tilde x_{2,k } g_{2,k}'(|\tilde x_{2,k}|^2/p_{2,k})}{p_{2,k } } \biggr\}\end{aligned}\ ] ] where the quantities @xmath241 and @xmath242 are mean values of the stochastic background s fourier amplitudes at each of the two detector sites . these both vanish : @xmath243 since the mean values of the fourier amplitudes are zero . hence , as in sec . [ ss:2dwnss ] one must look for the locally optimal statistic at the next order in @xmath2 . taking an additional derivative , one can easily compute @xmath141 . as in sec . [ ss:2dwnss ] this consists of two types of terms . for the same reasons as before , we discard from this decision statistic all the _ single detector _ terms . this leaves us with the following generalized cross - correlation statistic : @xmath244 } \frac{\re(\tilde s_{1,k}^\ast\tilde x_{1,k } ) g_{1,k}'(|\tilde x_{1,k}|^2/p_{1,k})}{p_{1,k } } \nonumber\\ & & \qquad \times \frac{\re(\tilde s_{2,k'}^\ast\tilde x_{2,k ' } ) g_{2,k'}'(|\tilde x_{2,k'}|^2/p_{2,k'})}{p_{2,k'}}.\end{aligned}\ ] ] since the expectation value of the product of the stochastic background at the two sites is given by @xmath245 one obtains the generalized cross - correlation statistic @xmath246 } \frac{\gamma_k\sigma^2_k\tilde x_{1,k}^\ast\tilde x_{2,k}}{p_{1,k}p_{2,k } } \nonumber\\ & & \quad\times g_{1,k}'(|\tilde x_{1,k}|^2/p_{1,k } ) g_{2,k}'(|\tilde x_{2,k}|^2/p_{2,k}).\end{aligned}\ ] ] if the functions @xmath247 are replaced by unity , this reduces to the standard result for the optimal filter for the case where the detector noise is assumed to be stationary and gaussian . for typical non - gaussian noise models , the effect of the @xmath247 functions is to exclude those frequency bins in which @xmath248 is large in either detector . our second non - gaussian noise model is similar to the noise burst model used in sec . [ ss : sdcnkw ] , generalized to the two detector case . the composite pdf for this model is @xmath249 where @xmath250 and @xmath251 are the probabilities of bursts in detectors 1 and 2 . the matrices @xmath252 , @xmath253 , and @xmath254 represent the correlation matrices when a noise burst is present . as in sec . [ ss : sdcnkw ] , a burst effectively changes the noise level for the detector experiencing the burst . thus , if there is a burst in detector 1 , simply replace @xmath255 with @xmath256 in @xmath257 to obtain @xmath252 . then we find @xmath258 + o(\epsilon^4)\ ] ] and @xmath259 to first order in @xmath260 and similarly for @xmath253 . we also have @xmath261\ ] ] and @xmath262 to first order in @xmath260 and @xmath263 . we can now compute the locally optimal statistic : @xmath264 where @xmath265 and @xmath266 is given by a similar expression . here we have neglected all @xmath118 terms . the terms @xmath267 and @xmath266 detect bursts , and their role is to suppress @xmath141 when a burst is present in either detector . in analyzing experimental data , there are different possible goals . one goal might be to set an upper limit ( with a certain statistical confidence ) on the stochastic background energy density in a particular frequency band . another goal might be to estimate this energy density in a particular frequency band . for this latter purpose , there are different possible estimators that might be used . one standard estimator is the maximum likelihood estimators ( mle ) . in this section , we show how this estimator is related to the cross - correlation statistic . recall that the probability distribution for the joint detector output is @xmath268 suppose we wish to estimate the strength @xmath269 of the stochastic background . the maximum likelihood estimator is the value @xmath270 for which this probability is maximized : @xmath271_{\epsilon^2_{\text{mle } } } = 0 $ ] . the result is @xmath272 where @xmath273 is a measure of how how sensitive the detectors were to the stochastic background . normally @xmath274 will be on the order of unity so @xmath270 is approximately just the cross - correlation statistic . however , if the detector were abnormally noisy , then @xmath274 would be less that unity and the estimate of the stochastic background strength would be smaller than the cross - correlation statistic would indicate : this is a correction that compensates for artificially large values of the cross - correlation statistic due to noise fluctuations . another possible estimator is the bayesian estimator . in the long measurement , weak signal case this again yields the same result as the mle estimator . a nice feature of these techniques is that in practice , they should be easy to implement . work by scott and whiting @xcite has shown that the pdfs of the fourier amplitudes in different frequency bins can be easily obtained . since the characteristic time - scale for stochastic background correlation is @xmath275 , these can be computed using data - segments with lengths seconds or tens of seconds . these pdfs can then be used to determine where to truncate or clip the correlation , frequency - bin by frequency - bin . provided that the instrument s characteristics are stable over periods of minutes or hours , it should be simple to accumulate sufficient statistics to determine the pdfs and therefore the truncation or weighting functions with reasonable accuracy . in practice , it may also be desirable to `` discard '' a small part of the `` attainable - in - principle '' correlation in exchange for obtaining more robust statistics . for example , on can arbitrarily zero the 1% of frequency bins that are the largest number of standard deviations away from the mean value ( for that bin ) . since the dominant contribution in any bin always comes from the detector noise , this is only very weakly correlated with the actual stochastic background signal , and the net effect is to discard just a bit more than 1% of the `` in principle '' attainable signal - to - noise ration . but in exchange , the detection statistic becomes far less sensitive to non - gaussian detector fluctuations . the precise effects of such treatment , and the appropriate truncation thresholds , can be easily determined with monte carlo simulations using simulated signals added into real detector noise . in searching for a known waveform ( e.g. , binary inspiral ) the methods are again easily implementable . here , since the signal timescale is less than a minute , the frequency - bin by frequency - bin statistics take a bit more time to accumulate , and the detector s statistical properties have got to be stable over a slightly longer time - scale ( an hour , perhaps ) . this appears likely . since certain non - gaussian noise features are more likely to appear as outlier points in the time - domain , and others in the frequency - domain , a combination of the time- and frequency - domain methods may be desirable . unfortunately , if the detector noise is not white , this may require the removal ( vetoing ) of entire small sections of time - series data . this is easy in the stochastic background case , where only tens of milliseconds around a glitch need excision . it may be more problematic for signals like binary inspiral chirps that have longer duration . in secs . [ s : deterministic ] and [ s : stochastic ] , we derived locally optimal statistics to search for deterministic and stochastic gravitational wave signals in the presence of non - gaussian noise . these statistics reduce to the standard matched - filtering and cross - correlation statistics when the detector noise is gaussian . but they are more _ robust _ ( i.e. , less sensitive to outliers ) when the detector noise has non - gaussian components . for both cases , the standard and robust statistics take ( as input ) the output of one or more detectors , and return ( as output ) a single real number . but the statistics also depend on the gravitational wave signal and detector noise models , which are not directly observable . different choices for the signal and noise models correspond to different statistics , and these different statistics will in general perform differently given the same detector output . in order to compare and evaluate the statistics , we need a way to quantify their performance . as mentioned in sec . [ s : deterministic ] , the quality of a test ( i.e. , a decision rule based on a particular statistic ) is characterized by its _ false alarm _ and _ false dismissal _ probabilities for a given source . these are , respectively , the probability that the test leads us to conclude that a signal is present , when in fact it is absent ( @xmath4 ) , and the probability that the test leads us to conclude that a signal is absent , when in fact it is present ( @xmath276 ) . these two probabilities ( denoted @xmath123 and @xmath277 ) completely specify the long - term performance of a statistic . but to rank different tests , we need to reduce these multi - dimensional error measures to a single figure of merit . how we do this depends on the problem we are trying to solve ( see , e.g. @xcite ) , but in the context of gravitational wave detection , it is common to look for a test that minimizes the false dismissal probability , keeping the false alarm probability less than or equal to some maximum tolerable value . this criterion is known as the _ neyman - pearson _ criterion , and it was used in sec . [ s : deterministic ] to define the locally optimal statistics . thus , to compare the performance of different statistics , we should plot false dismissal versus false alarm curves for different values of the signal amplitude @xmath2 . the best test ( or best statistic ) is the one that has the smallest false dismissal probability @xmath278 , for fixed false alarm probability @xmath123 and fixed signal amplitude @xmath2 . note that since the false dismissal probability depends on both @xmath123 and @xmath2 , it is possible that the best test for one choice of @xmath279 is not the best test for a different choice of @xmath279 . note also that this method of comparing statistics is different than simply comparing expected signal - to - noise ratios . what is important when determining error rates ( and hence the performance of a particular test ) is not the expected value of the statistic , but rather its probability _ distribution_. for sufficiently simple statistics with sufficiently simple signal and noise models , it may be possible to analytically calculate the corresponding false dismissal versus false alarm curves . but for most cases of interest , we must resort to _ monte carlo simulations _ to generate the curves . this consists of adding simulated signals to simulated ( or real ) detector noise , and then processing the resulting data with a statistic . for each stretch of data , the statistic outputs a single number which is then compared to a threshold to determine if we should claim detection . since we know if a signal is present in the data , we can easily determine the fraction of times that the decision rule was in error . in the absence of a signal , this procedure yields the false alarm probability @xmath123 as a function of the threshold @xmath21 . in the presence of a signal having fixed amplitude @xmath2 , we obtain the false dismissal probability @xmath277 , again as a function of the threshold . if we invert @xmath280 for @xmath281 , and substitute this expression back into @xmath282 , we obtain the false dismissal versus false alarm curve @xmath283 . we can then repeat these steps for a different signal amplitude @xmath284 to produce a new curve @xmath285 . the final result will be a set of curves similar to those shown in fig . [ f : typical1 ] . alternatively , we can plot @xmath286 or @xmath287 versus @xmath123 , as shown in figs . [ f : typical2 ] and [ f : typical3 ] . note that the quantity @xmath288 is the difference of two probabilities : @xmath289 is the probability that the statistic exceeds some threshold in the presence of a signal ( @xmath290 , while @xmath123 is the probability that the statistic exceeds the same threshold in the absence of a signal ( i.e. , @xmath4 ) . although , figs . [ f : typical2 ] and [ f : typical3 ] contain the same information as the false dismissal versus false alarm curves ( fig . [ f : typical1 ] ) , plotting @xmath287 versus @xmath123 has the nice property that , for stochastic signals , the curves have a well - defined @xmath291 limit . to illustrate how we can compare different statistics using monte carlo simulations , consider the simple case of a search for a white , gaussian stochastic background signal using two independent , identical , coincident and coaligned detectors . statistic 1 will be the standard cross - correlation statistic defined by a white , gaussian stochastic background signal and white , gaussian detector noise . statistic 2 will be a locally optimal statistic , also defined by a white , gaussian stochastic background signal , but with white , 2-component , mixture gaussian noise with an arbitrary knee . we will assume that we know ( a priori ) that the two detectors are identical and have uncorrelated white noise . we will not assume , however , that we know ( a priori ) the parameters describing the statistical properties of the detector noise or the overall amplitude of the stochastic background signal . each statistic will have to internally estimate the parameters from the detector output , without any other prior knowledge . we perform monte carlo simulations of the two statistics for the following three cases : * uncorrelated , white , gaussian detector noise with zero mean and unit variance . * uncorrelated , white , 2-component , mixture gaussian detector noise with zero mean , unit variance , @xmath292 , and @xmath55 [ see eq . ( [ sgs ] ) ] . * uncorrelated , white , exponential detector noise with zero mean and unit variance [ see eq . ( [ e : exp ] ) ] . the first two simulations test the optimal behavior of the statistics . statistic 1 is designed for the data of case ( i ) , and statistic 2 is designed for the data of case ( ii ) . the third simulation tests the two statistics in a sub - optimal situation , representative of a real search where we do not know in advance the exact statistical character of the detector noise . details of the monte carlo simulations are summarized below : \(i ) a single stretch of data consists of @xmath293 discrete - time samples . this @xmath79 is sufficiently large that the large observation time approximation is valid . since we are considering white noise ( which has zero correlation length ) , any @xmath294 would do . \(ii ) the simulated stochastic gravitational - wave signal strengths are @xmath295 , .005 , .010 , .020 , and .040 , where @xmath2 is the ratio of the rms amplitude of the stochastic background signal to the rms amplitude of the detector noise . these signal strengths correspond to signal - to - noise ratios ( @xmath296 ) ranging from approximately 0.1 to 1 for a single stretch of data . note : since a real stochastic background is expected to have a smaller value of @xmath269 ( @xmath297 ) , we would need a much longer observation time to build - up similar signal - to - noise ratios in a real search . the purpose of this example , however , is to illustrate how one can compare two different statistics ; it is not meant to simulate a real ( @xmath298 month ) stochastic background search . \(iii ) for all three types of simulated detector noise , the standard cross - correlation statistic estimates the variance of the noise by calculating the sample variance of a stretch of detector output equal to @xmath299 . since the detector output consists in general of signal plus noise , the estimate of the noise variance gets worse as the signal amplitude increases . the sample variance is needed to specify the white , gaussian noise model that enters the definition of the standard cross - correlation statistic [ c.f . [ e : gcc ] ] : @xmath300 where @xmath301 and @xmath302 are the estimated variances of the noise in detectors 1 and 2 , respectively . \(iv ) in addition to estimating the variance of the detector noise , the locally optimal statistic also estimates the variances @xmath57 , @xmath303 , and breakpoint @xmath304 of the 2-component , mixture gaussian model that define this statistic . it does this by fitting two straight lines to a @xmath305 vs. @xmath306 plot obtained from a histogram of a stretch of detector output , again equal to @xmath307 . best - fit lines at small @xmath308 and large @xmath308 , respectively , yield estimates of @xmath57 and @xmath303 , while the intersection of the lines yields an estimate of @xmath304 . actually , only the breakpoints for the detector noise are needed to define the following locally optimal statistic : @xmath309 which is a truncated version of @xmath310 . ( see the discussion of truncation in the previous subsection . ) here @xmath311 is the usual step function , which equals 0 if @xmath47 , and equals 1 if @xmath312 . note : in order to handle pure gaussian noise ( which is a pathological case when one tries to model it as a 2-component , mixture gaussian distribution ) , the locally optimal statistic sets the breakpoint @xmath304 to @xmath313 whenever the estimated slopes at small and large values of @xmath308 have a percent difference less than @xmath314 . by doing this , the locally optimal statistic @xmath315 effectively reduces to the standard cross - correlation statistic @xmath316 when the noise is pure gaussian . \(v ) we use @xmath317 trials to generate each false dismissal versus false alarm curve . \(vi ) the simulations were written in matlab @xcite . the results of the simulation are shown in figs . [ f : noise1_beta]-[f : noise3_gamma ] . as noted in ( iv ) above , our implementation of the locally optimal statistic reduces to the standard cross - correlation statistic when the detector noise is pure gaussian . that is why the false dismissal versus false alarm curves for the two statistics are effectively identical in fig . [ f : noise1_beta ] . from fig . [ f : noise2_beta ] we see that the locally optimal statistic performs better than the standard cross - correlation statistic when the simulated detector noise is mixture gaussian . for each value of the stochastic signal strength @xmath269 and for each false alarm probability @xmath123 , the false dismissal probability @xmath283 for the locally optimal statistic is less than that for the standard cross - correlation statistic . this is as expected , since the locally optimal statistic was constructed precisely to handle mixture gaussian noise . finally , from figs . [ f : noise3_beta]-[f : noise3_gamma ] we see that the locally optimal statistic also performs better than the standard cross - correlation statistic when the simulated detector noise has an exponential distribution . the difference in performance between the two statistics for this case is less than that for mixture gaussian noise , but it is still noticeable . ( figure [ f : noise3_blowup ] focuses attention on the false dismissal versus false alarm curves for small values of the false alarm probability , while fig . [ f : noise3_gamma ] is a plot of @xmath318 versus @xmath123 , which highlights the difference between the two statistics in the small signal limit . ) this behavior is again as expected , since a locally optimal statistic is constructed to be less sensitive to the tails of a non - gaussian distribution . in this paper , we have constructed a replacement for the standard linear matched filter estimators used for gravitational wave detection . the replacements are more robust because they are less susceptible to corruption by non - gaussian detector noise . we have explicitly illustrated the locally optimal detection strategies for a variety of different noise pdfs , and for two different detection problems ( single detector known waveform , and two - detector stochastic background ) . in all cases , the optimal strategy is similar to the one for gaussian noise except that data samples that lie outside the central part of the distribution ( the outliers ) are excluded from the sums which form the estimators . this research was supported in part by nsf grants phy-9728704 , phy-9722189 , phy-9981795 , phy-0071028 , nasa grant nasa - jpl 961298 , the sloan foundation , and by the max planck society ( albert einstein institute , potsdam ) . a. abramovici , w. e. althouse , r. w. p. drever , y. grsel , s. kawamura , f. j. raab , d. shoemaker , l. sievers , r. e. spero , k. s. thorne , r. e. vogt , r. weiss , s. e. whitcomb , and m. e. zucker , science * 256 * , 325 ( 1992 ) . c. bradaschia et al . methods * a289 * , 518 ( 1990 ) ; also in _ gravitation 1990 _ , proceedings of the banff summer institute , banff , alberta , 1990 , edited by r. mann and p. wesson ( world scientific , singapore , 1991 ) . i. s. heng , d. g. blair , e. n. ivanov , m. e. tobar , phys . a * 218 * , 190 ( 1996 ) ; e. mauceli et al . , 1996 phys . d * 54 * , 1264 ( 1996 ) ; p. astone et al . , _ ibid_. * 47 * , 362 ( 1993 ) ; m. cerdonio et al . , in _ gravitational wave experiments _ , proceedings of the edoardo amaldi conference , edited by e. coccia , g. pizzella , and f. ronga ( world scientific , singapore 1995 ) , p. 176 ; e. coccia , _ ibid_. , p. 161 . simple experiments carried out by allen and romano using the grasp data analysis package ` http://www.lsc-group.phys.uwm.edu ` indicate that the large non - gaussian fluctuations present in the data from the ligo 40-m prototype created large biases in the cross - correlation estimators . the 1994 data stream was split into two parts , and each was assumed to come from an independent detector . b. allen , in _ proceedings of the les houches school on astrophysical sources of gravitational waves _ , edited by jean - alain marck and jean - pierre lasota ( cambridge university press , cambridge , england , 1997 ) . . flanagan , phys . d * 48 * , 2389 ( 1993 ) . note that the second term on the rhs of eq . ( b6 ) should read @xmath319 rather than @xmath320 , that the sliding delay function shown in fig . 2 is incorrect , and that the right hand side of eq . ( 2.8 ) should be divided by @xmath321 . matlab : ` http://www.mathworks.com ` . the matlab scripts that we used to generate the monte carlo simulations are available at : ` http://feynman.utb.edu/\simjoe/research/ stochastic / algorithms / robust / simulations / code/ ` .

with the successful launch in 1999 of the _ chandra x - ray observatory _ @xcite came the opportunity to explore the x - ray universe with greater angular resolution and sensitivity in the energy range than ever before . one of the primary goals of the mission is to perform surveys of the extragalactic sky at these higher energies . the longest _ chandra _ pointings thus far are the _ chandra _ deep field - north @xcite and -south @xcite , with 2 ms and 1 ms exposure times , respectively . these fields have explored faint sources , with many objects reaching below the luminosity range of active galactic nuclei ( agn ) to include normal " galaxies . though the deep fields are pushing to lower fluxes than previously attainable , they cover only a small area ( @xmath4 deg@xmath1 each ) , and do not have good statistics for the flux range in which the dominant contribution to the x - ray background arises . to get such statistics , a larger area of the sky must be covered by combining multiple _ _ pointings . the _ chandra _ archive provides the opportunity to explore multiple , mid - depth ( @xmath5 ks ) exposures . in we presented the serendipitous extragalactic x - ray source identification ( sexsi ) program design , our x - ray analysis techniques and a catalog of over one thousand sources , as well as some initial x - ray results and comparisons with previous work . with data from 27 archival _ fields , sexsi is designed to focus on a large - area ( @xmath6 deg@xmath1 ) sample of x - ray sources in the intermediate flux range ( @xmath7 ) . our goal is to determine spectroscopically the redshift distributions and nature of hard ( ) x - ray source populations from which the bulk of the x - ray background arises . the campaign of single - color photometry presented here attempts a minimum depth for every field of @xmath8 , consistent with identifying counterparts for which classifiable spectra can be obtained with a 10-m telescope in a @xmath9 hour integration . this strategy maximizes the number of hard sources identified spectroscopically in our fields for our fixed observing campaign . other programs are also aiming to understand the x - ray source population at medium depth with degree - scale _ chandra _ and _ xmm - newton _ surveys . examples of such surveys include hellas2xmm ( e.g. , * ? ? ? * ; * ? ? ? * ) and the _ chandra _ multiwavelength project ( champ ; e.g. , * ? ? ? @xcite presents a comprehensive list of the many ongoing x - ray background surveys . the sexsi program is unique among the _ chandra _ serendipitous surveys in concentrating on obtaining a uniform spectroscopic survey restricted to the hard source population . the champ , in contrast , is a comprehensive program of imaging and spectroscopy to follow - up both soft and hard sources in a large number of publicly - accessible extragalactic _ the sexsi spectroscopic catalog will therefore contain a larger fraction of obscured agn than champ . eight of the fields are common to both surveys , and for one of these champ has published some spectroscopic followup @xcite . even for the common fields , sexsi plans deeper spectroscopy and should identify a greater number of hard x - ray sources . between the large number of distinct fields and the different focus for ground - based followup , the sexsi survey provides an essential complement to current surveys of the sources responsible for the x - ray background . in this paper , we present optical @xmath3-band imaging of the sexsi _ chandra _ fields , as well as the methodology used to derive x - ray source identifications and a catalog of @xmath3-band counterpart magnitudes . in addition , we explore relationships between the x - ray flux , hardness ratio , and @xmath3-band magnitude . a third paper in this series ( eckart et al . , in preparation ; paper iii ) will include our optical spectroscopic results , with redshifts and classifications for @xmath10 of these objects , and a discussion of the luminosity distribution , redshift distribution , and composition of the sample . we have imaged the sexsi fields using the palomar 60-inch ( p60 ) and 200-inch ( p200 ) telescopes , the mdm 2.4-meter ( mdm-2.4 ) and 1.3-meter ( mdm-1.3 ) telescopes , and the keck i telescope . much of the early optical imaging for this program was done with small field - of - view , single - ccd array cameras . for example , the p60 ccd13 camera has a 12.6@xmath1112.6 field of view . this requires several pointings to map a typical @xmath12 _ chandra _ field , making multi - color imaging very time consuming . since spring 2000 , both the p200 and mdm-2.4 have had large - format ( @xmath13 ) ccd cameras available , allowing us to cover nearly all of a _ field in a single pointing . the p200 large format camera ( lfc ; * ? ? ? * ) uses the sloan digital sky survey ( sdss ) filter set ( @xmath14 , @xmath15 , @xmath16 , and @xmath17 ; * ? ? ? * ) , while , for our mdm-2.4 8k camera @xcite imaging , we used johnson - cousins filters ( @xmath18 , @xmath19 , @xmath3 , and @xmath20 ) . in each filter band , our goal has been to obtain images with a limiting vega magnitude of @xmath21 , the practical limit for obtaining counterpart spectra at the keck telescope in reasonable integration times . because the large - format cameras were not available for the first 1.5 years of the sexsi project , some fields are imaged only in @xmath3 band ( due to observing time constraints ) , while the fields observed later in the program generally have multicolor data . for the purposes of this paper , we use only the @xmath3-band data . table [ tbl : cameras ] summarizes the characteristics of the telescopes and imaging instruments used in this program . we reduced the optical images using standard techniques , relying upon iraf tasks for the preliminary steps . median - combined bias frames , taken the same night as the science data , were subtracted from the images . for flatfielding , we used either dome flats or sky flats generated by combining all the images from a night , including twilight flats , with min / max rejection to remove stars . we used standard techniques to mask out cosmic rays and bad pixels prior to combining data . for astrometric calibration of most of the data , we used the dophot software package to locate all nonsaturated , point - like objects in individual exposures . comparing with the usno - a2.0 catalog @xcite , we then used the object - matching program starmatch ( by doug reynolds ) to both align the images before stacking and to obtain an astrometric solution for the final , combined image . the standard deviations of the astrometric solutions are generally @xmath22 in each axis . we reduced images using the same methodology for both the lfc data from the p200 telescope and the 8k data from the mdm-2.4 telescope , although slightly different suites of software were required . we relied upon the mscred package within iraf , designed for reduction of large - format , multi - array optical data from the kitt peak / cerra tololo mosaic cameras , following the informative reduction pipeline outlined by b. jannuzi . for the lfc images this necessitated first changing the data into the requisite format using the lfcred package created by m. huntmph / lfcred . ] . we relied on a variety of methods to provide photometric calibration of the data , depending upon the conditions under which data for a sexsi field were taken . throughout , unless otherwise noted , we refer our magnitudes to the vega - based system . if conditions were photometric , we determined magnitude zeropoints from @xcite fields observed the same night , at an airmass close to that of the field in question . in cases where a non - photometric image overlaps a photometric image from another night , we determined the zeropoint magnitude by matching the photometry of overlapping regions . in cases where we obtained no photometric images of a field , we relied on shallow sky surveys to provide magnitude zeropoints . where available , we used @xmath15 photometry from the sdss first data release @xcite , converting from ab magnitudes to vega magnitudes with @xmath23 alternatively , we used the usno - b1.0 catalog @xcite to establish the zeropoint . typical photometric calibration errors for this technique were 0.2 mag . for the purposes of this paper , we make the assumption that the sdss @xmath15 and johnson - cousins @xmath3 filters are similar . convolving the respective filter transmission curves with the @xcite composite sdss quasar spectrum redshifted from @xmath24 to @xmath25 , we find that this assumption provides a @xmath26 mag systematic uncertainty in our photometry . performing the same exercise on the @xcite composite galaxy templates , we find systematic offsets @xmath27 mag for @xmath24 , with larger systematic offsets at higher redshifts . for example , the s0 composite template at @xmath28 has a 0.39 mag difference between the @xmath15 and @xmath3 vega magnitudes due to the 4000 break redshifting out of the @xmath15 filter . though imprecise , our assumption that the @xmath15 and @xmath3 filters are similar is adequate to establish the faint source limit of each image , and to plan spectroscopic observing runs . for extraction of sources from our combined , calibrated optical images , we used the sextractor code of @xcite to generate catalogs for both the full optical images and for the smaller , central portions of each image . because they avoid the ccd edges where the increased noise leads to spurious sources , these latter catalogs are used to derive photometric depths . quoted magnitudes refer to the sextractor magbest , which usually reports photometry within elliptical apertures @xcite , but reverts to a corrected isophotal magnitude if a source is strongly contaminated by nearby sources . to determine limiting magnitudes for each image , we made histograms of the @xmath3-band source density in 0.5 magnitude bins . figure [ fig : magcutoff ] shows an example histogram from sexsi field q2345 . for each sexsi field image , we compare the measured number counts @xmath29 to that derived by @xcite from deep subaru imaging of the _ hubble _ deep field - north : @xmath30 where @xmath29 has units of number degree@xmath31 0.5 mag@xmath32 . for @xmath3-band magnitudes in the range @xmath33 , @xcite finds @xmath34 and @xmath35 . converting our @xmath3-band photometry into the ab system with equation 1 , we plot our measured number counts for each image with the published , deep - field fit ( see figure [ fig : magcutoff ] ) . in order to derive a crude depth for our images , we compare our number counts in each 0.5 mag bin to equation 2 ( at bin center ) . limiting magnitudes are defined by the inter - bin flux where our number counts go from greater than 80% to less than 80% of the @xcite value : this provides an @xmath36 completeness depth for each image . by the nature of our algorithm , then , these limits are good to @xmath37 mag . table [ tbl : images ] summarizes the @xmath3-band optical imaging we have obtained for the sexsi fields . the average x - ray to optical offsets , @xmath38 and @xmath39 , are discussed in [ sec : matching ] below . we derived @xmath3-band galactic extinction for the central position of each x - ray image using the ned galactic extinction calculator . checking the variance of the ned extinction values across the _ chandra _ field of view , we find that the typical error on the extinction is @xmath40 mag . three of our fields , however , have significantly higher extinctions than the other fields , and thus have an uncertainty of @xmath41 mag in their extinction values for individual sources . these fields are flagged in table [ tbl : images ] . approximately half of the sexsi fields were observed with a single , uniform depth image ; the other half rely on imaging from multiple cameras and telescopes over several observing runs . of the fifteen fields with a single optical image , six reach @xmath3-band depths @xmath42 , and ten reach @xmath3-band depths @xmath43 . the shallowest field , cl 0442 + 0202 , only reaches @xmath44 , while the deepest field , cl 0848 + 4454 , reaches @xmath45 . of the twelve fields with imaging from multiple pointings , eight have sections reaching depths of at least @xmath46 . a histogram of imaging depth for each source is shown in figure [ fig : depth_hist ] . we match x - ray and optical sources in several steps . the procedure is slightly different depending on whether there is one optical image or multiple optical images per _ chandra _ field . for fields with a single optical image , we first identify the closest optical source within 4 of each _ chandra _ source position . excluding sources displaced by more than 1.5@xmath47 from the mean offset , we use these identifications to calculate a first estimate of the average offset between the _ chandra _ and usno astrometry . for the following iterations , the search radius used to identify optical counterparts depends upon the off - axis angle ( oaa ) of the x - ray source : we use the larger of 1.5 and psf/3 , where psf is the full - width , half - maximum of the _ chandra _ point spread function at each oaa . again , we exclude sources displaced by more than 1.5@xmath47 from the mean as we iteratively correct the x - ray source positions . we continue iterating until the corrections to the offsets are less than the corresponding standard deviation . two to three iterations are typically required , with the biggest shift applied after the initial matching . as shown in table 2 , the typical standard deviation on this astrometric correction is @xmath48 0.3 , comparable to the accuracy of the optical astrometry . for fields with multiple optical images we modify this algorithm slightly . some images cover an area with few ( or one ) x - ray sources , which , if we were to follow the procedure described above , could lead to incorrect offsets and matches . instead , for a given _ field , we first take each optical image and find the closest match within a 4 radius of all _ chandra _ source positions that fall within the image . we record these astrometric differences . for x - ray sources falling on multiple optical images , we use the data from the image with the deepest limiting magnitude . as before , we use this list of multi - image astrometric differences to calculate an average _ chandra _ to usno offset , again eliminating sources greater than 1.5@xmath47 from the mean offset . this astrometric correction is then applied to all _ chandra _ positions for the sexsi field considered . a second pass at optical identifications is then made with the oaa - dependent matching radius , using the larger of 1.5 and psf/3 . of the 998 sexsi sources with optical coverage , 655 used a 1.5 search radius , while 343 used the psf/3 search radius . we expect few false matches due to an optical source randomly overlapping the matching search area . the number of false matches per field depends upon both the number of x - ray sources detected in the _ chandra _ field , and the depth of the optical image . we tested several fields from table [ tbl : images ] for false matches by finding the total area covered by the 1.5 radius search circles and multiplying that area by the optical source density . we find that for medium - depth x - ray and optical images , the number of false sources detected per field is generally less than one . for example , we predict 0.6 false matches for q2345 , 0.7 false matches for cl 0442 + 0202 , and 1.3 false matches for 1156 + 295 , where these numbers are for the 1.5 search radius sources only . for the 34% of the sources with larger oaas and thus larger search areas we expect a larger fraction of false matches , though still fewer than 3 per field for most fields . for example , we predict 1.3 false matches from the @xmath49 match radius sources in q2345 , 1.4 false matches in cl 0442 + 0202 , and 2.4 false matches in 1156 + 295 . for several of the fields , including the three example fields just mentioned , we shifted all x - ray source positions by 1 , producing a `` fake '' x - ray catalog , and ran the matching routine again with the average x - ray - to - optical offset forced to zero . this process , repeated with several different shifts of the x - ray source positions , gives a check on the number of false detections predicted above . for all of the fields tested the number of false detections are consistent with the values we calculate . our spectroscopic results are also consistent with this matching scheme and will be discussed further in paper iii . once matching is complete , we produce first - draft optical postage stamp cutouts for our entire hard x - ray catalog ( see figure [ fig : postage ] ) . each stamp is centered on the astrometrically - corrected x - ray source position , and the size of the _ chandra _ psf - dependent search radius is shown by the ( centered ) solid circle . a dashed circle , located at the un - corrected _ chandra _ position , illustrates the need for this offset correction . arrows point to identified optical counterparts , while the absence of an arrow signals a non - detection . if the magnitude found by sextractor is fainter than the limiting optical image , we annotate the photometry with an asterisk . as in paper cl " flag is added if a source is potentially within 1 mpc of a target cluster center . next we visually inspect postage stamps for each x - ray source , flagging sources with potentially inaccurate photometry as required . these flags , detailed in [ sec : catalog ] , identify sources which either ( 1 ) have multiple optical identifications within the psf - dependent search area , or ( 2 ) are , or are near , a saturated source in our optical image . a final matching iteration is then done , excluding the newly flagged sources . we produce new optical catalogs and postage stamp cutouts , omitting stamps for the handful of saturated sources where counterpart magnitudes are drawn from the literature and the 65 sexsi sources that either lack optical coverage in our data or have an unknown limiting magnitude due to nearby bright - source contamination ( optical flag = 6 , see [ sec : catalog ] ) . an example of six postage stamp cutouts is included as figure [ fig : postage ] ; the entire catalog of postage stamps is provided in the online version of the manuscript . for the 262 sources with @xmath50 , 160 ( 61% ) have identified counterparts , while for the 434 sources with @xmath51 , 291 ( 67% ) have identified counterparts , and for the 167 sources with @xmath52 , 124 ( 74% ) have identified counterparts . our total sample of 947 sources with unambiguous photometry ( no contamination ) identifies 603 counterparts ( 64% ) . in table [ tbl : catalog ] we present the catalog of 1034 hard - band sexsi sources the table is published in its entirety in the electronic version of the _ astrophysical journal_. columns 1 7 present x - ray source information for easy reference , while the optical photometric data are presented in columns 8 15 . complete x - ray source information is presented in paper i , table 4 . the x - ray source positions in table [ tbl : catalog ] are corrected for mean optical to x - ray offsets . note that since the source names ( column 1 ) , identical to the source names in paper i , are derived from the hard - band x - ray images , the refined positions of columns 2 3 will not exactly match those of column 1 ( though mean offsets are typically less than 1 ) . column 4 lists the off - axis angle ( oaa , _ i.e. _ , the angular distance , in arcmin , of the source position from the telescope aim point ) . the flux ( in units of @xmath53 ) and detection snr are shown in columns 5 6 , while column 7 gives the hardness ratio , @xmath54 , where @xmath55 and @xmath56 are the counts cm@xmath31 in the and bands , respectively . here , as distinct from paper i , we record the hardness ratio derived from the net soft x - ray counts recorded at the hard - band source position when there was not a significant soft - band source detected ( in paper i these cases are reported as @xmath58 ) . in addition , for a subset of these cases , when the soft - band counts recorded at the hard - band source position were less than twice the soft - band background counts , the @xmath59 is considered a lower - limit , flagged as such in the catalog , and set to @xmath60 , where @xmath61 soft - band background counts . column 8 contains an optical flag code essential to interpretation of the optical data : 0 = no optical coverage , 1 = a solid optical i d , 2 = no optical counterpart ( the magnitude listed is then a lower limit ) , 3 = saturated in the sexsi optical image ( @xmath3-band magnitude taken from the guide star catalog ii @xcite or a secondary source in vizier database ) , 4 = a solid i d but @xmath3-band magnitude affected by nearby bright source , 5 = more than one optical source in x - ray error circle ( the source with the smallest positional offset is recorded in the table ) , and 6 = lower limit ( no optical counterpart ) but area contaminated by nearby bright source so limiting magnitude is unknown . column 9 is the @xmath3-band magnitude of the optical counterpart , with its error shown in column 10 . column 11 gives the limiting magnitude for the image from which the optical counterpart , or a lower limit thereto , was derived . column 12 lists @xmath62 in arcsec , where @xmath63 is the astrometrically - corrected x - ray source ra ( from column 2 ) and @xmath64 is the optical counterpart ra ; column 13 is the corresponding value for declination . column 14 , @xmath65 ( @xmath66 ) , is the x - ray to optical position difference , also in arcsec . following @xcite and @xcite , the logarithmic x - ray - to - optical flux ratio ( column 15 ) is given by the relation @xmath67 derived from the kron - cousins @xmath3-band filter transmission function . column 16 presents the sextractor class_star parameter @xcite for basic star - galaxy separation , with values ranging from 0.0 for significantly extended sources to 1.0 for sources consistent with perfectly stellar psfs . the stellarity value of the 8 bright sources with optflag = 3 ( from the literature ) is set to 1.00 . we present the stellarity data with the caveat that the values should only be used for broad separation , for example , of sources near 0 verses 1 , but not for detailed quantitative analysis . note that for all of the optical counterpart data columns , special attention must be paid to the optical flag . for example , if the code is a 2 ( optical counterpart is not detected ) , column 9 describes a lower limit , columns 10 , 12 14 , and 16 have no data , and the x - ray - to - optical flux ratio in column 15 is a lower limit . we now discuss the results of the optical identifications of the hard x - ray source counterparts and how the counterpart properties relate to the x - ray properties . throughout this section , only sources with a solid optical counterpart or upper limit thereto are plotted and analyzed ; we exclude sources with @xmath3-band magnitudes or limits contaminated by a nearby bright source and sources with more than one optical source within the identification radius . this provides a sample of 947 hard x - ray sources with unambiguous counterpart photometry , of which 603 are identifications and 344 have limits to the counterpart optical magnitude . figure [ fig : r_hflux_scatter ] presents the @xmath3-band magnitudes , or limits thereto , plotted as a function of x - ray flux . dashed lines show constant x - ray - to - optical flux ratios . the bulk of the sexsi sources have @xmath68 and are fainter than @xmath69 . shallow , wide - area x - ray surveys , such as the _ asca _ hard x - ray survey reported by @xcite , find that the majority of extremely bright x - ray sources ( @xmath70 ) are agn and have @xmath71 . many of the sexsi sources are likely fainter and/or more distant analogs . as has been found in other surveys ( e.g. , * ? ? ? * ; * ? ? ? * ) , sexsi detects a population of sources overluminous in the x - ray for their optical magnitudes . this population is found over the entire x - ray flux range sampled . two explanations have been commonly discussed to explain large values of @xmath72 : ( i ) extremely high - redshift agns that might have bright x - ray fluxes but faint @xmath3-band fluxes due to absorption from the lyman transitions of hydrogen along our line of sight , or ( ii ) heavy obscuration by material in the host galaxy . the latter interpretation for the majority of sources is supported by near - ir studies , which find a large fraction to be very red early - type galaxies at @xmath73 @xcite . in addition , a number of _ type ii quasars _ , which would have similar properties , have been identified in the deepest _ chandra _ surveys ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . by virtue of the large area surveyed , sexsi contains numerous sources with @xmath74 at moderate x - ray fluxes , @xmath75 . many of these sources have @xmath76 , well within the reach of 8- to 10-meter class telescopes for spectroscopic follow - up . these are brighter examples of the new x - ray overluminous population . deep survey versions of these sources are typically extremely faint as an example , @xcite report on seven `` extreme x - ray / optical '' ( exo ) sources in the 2 msec _ chandra _ deep field - north ( cdf - n ) which , despite having extremely robust detections in the _ chandra _ data ( @xmath77 counts ) , remain undetected ( @xmath78 ) in the great observatories origins deep survey imaging of the cdf - n @xcite . @xcite report on the spectra of 13 sources with @xmath79 from the hellas2xmm survey . they find 8 narrow - lined sources , with @xmath80 erg s@xmath32 . our survey contains 109 such sources with @xmath79 of which 64 have an optical limiting magnitude and @xmath81 have spectral data . these high @xmath82 sources provide a useful catalog for future infrared surveys . we also find a number of sources with very bright ( @xmath83 ) optical counterparts . the bulk of these sources ( 7 out of 8) are identified as stars in the literature . figure [ fig : hfluxhist ] presents a histogram of the number of sources in sexsi and in the cdf - n @xcite as a function of hard x - ray flux and split by @xmath3-band magnitude at @xmath84 . this magnitude was chosen to separate our sources approximately in half . the small number of sources from images with @xmath85 were not included in this plot or in figures [ fig : hrhist ] , [ fig : meanhr ] , and [ fig : fxfohr ] . the majority of the sexsi sources have @xmath86 , a flux level that lies between the _ asca _ and _ bepposax _ sensitivity limits and the _ chandra _ deep survey capability . this important flux range corresponds to the regime in which the relation changes slope and from which the bulk of the x - ray background arises @xcite . compared to the deep surveys , sexsi has nearly an order of magnitude more sources at this x - ray depth , approximately half of which have @xmath87 , making them easy spectroscopic targets for 8- to 10-meter class telescopes . figure [ fig : hrhist ] shows the hardness ratio histogram of the sexsi sources , again split at @xmath84 . the optically - brighter sources are peaked at a low hardness ratio , while the fainter sources have a much harder , and broader , distribution , showing that the optically - fainter portion of the sample has a higher fraction of sources with a flat x - ray spectral slope . the large peak of the @xmath88 sources is near @xmath89 , which corresponds to a power - law photon index ( @xmath90 ) between 1.8 and 1.9 , a typical value for unobscured broad - line agn . this suggests that the majority of the optically - brighter sources are broad - lined agn . spectroscopic followup ( paper iii ) will be able to specifically address this hypothesis . the harder and broader distribution of the @xmath91 sources indicates that obscuration at the source is likely involved . @xcite shows a similar trend in the 1 ms cdf - n data . figure [ fig : meanhr ] , which displays the mean hardness ratio as a function of hard - band flux separately for bright and faint counterpart magnitudes , again emphasizes that distinct populations are contributing to the overall source counts . the mean values for this figure are calculated using source @xmath59s regardless of if they are flagged as lower limits . setting all such flagged sources to the upper extreme , @xmath92 , shifts the points only slightly , if at all . for reference , the @xmath59 of a @xmath93 spectrum with varying intrinsic absorbing column density and redshift is presented in the adjacent panel . this shows that modest - redshift , heavily - obscured sources produce the highest @xmath59s . for sources fainter than @xmath84 , the mean hardness ratio is essentially constant at @xmath94 over two orders of magnitude in x - ray flux . for those sources with brighter magnitudes , however , there is a significant decline in mean hardness ratio , from @xmath95 at @xmath96 to @xmath97 at @xmath98 . in a simple - minded two - component model with median hardness ratios of @xmath99 and @xmath100 for the two source classes , the faint sources are dominated by the hard population with perhaps an admixture of 10% soft sources at all flux levels , while for the brighter sources , the ratio of hard to soft sources changes from 20:80 at @xmath96 to 80:20 at @xmath98 . this is qualitatively consistent with the relations derived in paper i , where we showed that hard sources follow a constant power - law over three orders of magnitude in hard - band flux , while softer sources show a distinct break at @xmath101 . in this picture , the soft sources are the standard , largely unobscured agn at all redshifts , while the harder component represents the lower mean - redshift , lower - luminosity populations revealed in the deep surveys ( e.g. , * ? ? ? * ; * ? ? ? figure [ fig : fxfohr ] shows the hardness ratio as a function of x - ray to optical flux ratio for the sexsi sample , again with the @xmath59 of a @xmath93 spectrum with varying intrinsic absorbing column density as a function of redshift presented in the adjacent panel . the main plot shows a trend toward larger hardness ratios as @xmath82 increases . this is in general consistent with the notion that the increasing @xmath82 results from larger absorbing columns which attenuate the @xmath3-band flux much more severely than the hard x - ray flux ( e.g. , * ? ? ? it is interesting , however , that the most extreme sources with @xmath79 are not all hard , but exhibit a wide range of @xmath59s , implying that obscuration may not be the sole explanation for the dim optical counterparts . @xcite find that for non - broad - lined sources ( e.g. , sources with optical spectra showing narrow agn emission lines or early - type galaxy absorption lines ) there is a linear correlation between @xmath72 and @xmath102 . this relationship implies that some of the high @xmath82 sources with low @xmath59s are high luminosity , high redshift , narrow - lined agn . we also explored the optical source densities in the vicinity of the x - ray sources . for each optical counterpart , we counted the number of sources in the surrounding region and compared the source density to the overall field density to calculate overdensities . we employed 20 , 30 , 40 , and 60 radius circles around each optical counterpart and counted sources with @xmath3-magnitude of @xmath103 and @xmath104 for each of the sources . we found no significant overdensities around the x - ray sources . we present @xmath3-band imaging of @xmath105% of the x - ray sources in the sexsi survey . we describe our optical data reduction and the x - ray to optical source matching algorithms employed , and present a catalog of @xmath3-band counterpart photometry and astrometry . while the power of the sexsi sample is fully realized only with the addition of our collection of optical counterpart spectra , the photometric identification of the sources is an important step that provides clues to the composition of the source population . we find that by splitting the sources at @xmath84 and analyzing each group s x - ray properties , we begin to see the emergence of what appear to be physically distinct populations . for example we find that the hardness ratio distribution of the optically brighter sources is sharply peaked near @xmath106 , typical of unobscured agn , while the fainter , @xmath91 sources have a much broader and harder spectral distribution . the findings of this paper suggest that the analysis of optical spectral data from counterparts fainter than @xmath84 ( attainable with 10-meter - class telescopes ) , will be an essential part of exploring the x - ray background composition . this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration ( nasa ) . some of the data presented herein were obtained at the w.m . keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and nasa . the observatory was made possible by the generous financial support of the w.m . keck foundation . the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community . we are most fortunate to have the opportunity to conduct observations from this mountain . this work has made use of the usnofs image and catalog archive operated by the united states naval observatory , flagstaff station . the guide star catalog was produced at the space telescope science institute under u.s . government grant with data based on photographic data obtained using the oschin schmidt telescope on palomar mountain and the uk schmidt telescope . this research has made use of nasa s astrophysics data system . this work has been supported by nasa nag5 - 6035 ( djh ) , as well as by a small _ chandra _ archival grant . the work of ds was carried out at the jet propulsion laboratory , california institute of technology , under a contract with nasa . llccl 8k & mdm 2.4 m & 0.18 & @xmath107 & @xcite + ccd13 & palomar 60 & 0.37 & @xmath108 & + cosmic & palomar 200 & 0.29 & @xmath109 & @xcite + echelle & mdm 1.3 m & 0.50 & @xmath110 & + echelle & mdm 2.4 m & 0.28 & @xmath111 & + lfc & palomar 200 & 0.18 & @xmath112 & @xcite + lris & keck i & 0.22 & @xmath113 & @xcite + templeton & mdm 1.3 m & 0.50 & @xmath114 & + [ tbl : cameras ] lcccclllcc ngc 891 & 51 & -0.9 @xmath115 0.3 & 0.6 @xmath115 0.2 & 0.17 & 02 22 40 & + 42 26 16 & 8k & 24.3 & 1.6 + awm 7 & 48 & 0.5 @xmath115 0.3 & 0.2 @xmath115 0.1 & 0.30 & 02 54 45 & + 41 40 10 & 8k & 24.2 & 1.7 + xrf 011130 & 30 & 0.4 @xmath115 0.2 & -0.5 @xmath115 0.2 & 0.26 & 03 05 28 & + 03 48 52 & lfc & 23.2 & 1.2 + ngc 1569 & 97 & -0.3 @xmath115 0.2 & 0.6 @xmath115 0.3 & 2.00 & 04 29 08 & + 64 45 42 & echelle ( 1.3 ) & 21.5 & 1.7 + & & & & & 04 29 42 & + 64 42 35 & lris & 21.5 & 0.9 + & & & & & 04 29 53 & + 64 39 47 & ccd13 & 21.0 & 1.4 + & & & & & 04 30 36 & + 64 51 06 & lris & 23.0 & 1.2 + & & & & & 04 31 15 & + 64 51 06 & lris & 22.5 & 0.9 + 3c 123 & 47 & -0.1 @xmath115 0.1 & 0.0 @xmath115 0.0 & 2.61 & 04 36 52 & + 29 40 08 & 8k & 21.9 & 1.4 + cl 0442 + 0202 & 44 & -0.2 @xmath115 0.2 & -1.2 @xmath115 0.3 & 0.41 & 04 42 17 & + 02 03 35 & lfc & 21.1 & 1.2 + cl 0848 + 4454 & 186 & -0.3 @xmath115 0.1 & -1.2 @xmath115 0.1 & 0.08 & 04 48 49 & + 44 54 09 & lfc & 24.4 & 1.1 + rx j0910 & 171 & -0.3 @xmath115 0.1 & -2.3 @xmath115 0.2 & 0.05 & 09 10 33 & + 54 19 38 & 8k & 24.0 & 1.6 + 1156 + 295 & 49 & -0.5 @xmath115 0.2 & -1.1 @xmath115 0.3 & 0.05 & 11 59 32 & + 29 16 19 & 8k & 24.5 & 1.5 + ngc 4244 & 49 & -0.2 @xmath115 0.2 & 0.0 @xmath115 0.1 & 0.06 & 12 16 46 & + 37 51 46 & ccd13 & 22.9 & 1.8 + & & & & & 12 17 34 & + 37 46 54 & echelle ( 1.3 ) & 23.4 & 1.8 + & & & & & 12 17 54 & + 37 55 15 & ccd13 & 23.4 & 1.5 + ngc 4631 & 59 & 0.4 @xmath115 0.3 & -0.3 @xmath115 0.2 & 0.05 & 12 41 29 & + 32 36 46 & ccd13 & 23.0 & 1.6 + & & & & & 12 41 39 & + 32 30 44 & ccd13 & 23.0 & 2.5 + & & & & & 12 42 03 & + 32 40 53 & echelle ( 2.4 ) & 23.0 & 1.2 + & & & & & 12 42 21 & + 32 37 09 & ccd13 & 23.0 & 1.7 + & & & & & 12 42 59 & + 32 32 49 & templeton & 23.0 & 1.4 + hcg 62 & 49 & 0.1 @xmath115 0.1 & -0.1 @xmath115 0.1 & 0.13 & 12 53 04 & @xmath11609 14 20 & echelle ( 2.4 ) & 22.4 & 1.2 + & & & & & 12 53 06 & @xmath11609 05 48 & echelle ( 2.4 ) & 23.4 & 1.0 + & & & & & 12 53 04 & @xmath11609 14 20 & ccd13 & 22.4 & 1.6 + rx j1317 & 111 & -1.6 @xmath115 0.2 & -1.6 @xmath115 0.2 & 0.03 & 13 17 13 & + 29 11 32 & lfc & 23.5 & 1.5 + bd 1338 & 38 & -1.3 @xmath115 0.2 & -1.1 @xmath115 0.1 & 0.04 & 13 37 44 & + 29 25 24 & ccd13 & 22.5 & 1.3 + & & & & & 13 37 53 & + 29 29 44 & ccd13 & 22.5 & 1.7 + & & & & & 13 38 15 & + 29 36 51 & ccd13 & 22.5 & 2.0 + & & & & & 13 38 26 & + 29 24 53 & ccd13 & 22.5 & 1.5 + rx j1350 & 58 & 0.7 @xmath115 0.3 & -2.5 @xmath115 0.5 & 0.03 & 13 50 39 & + 60 04 11 & lfc & 22.5 & 0.9 + 3c 295 & 23 & 0.2 @xmath115 0.2 & -0.3 @xmath115 0.3 & 0.05 & 14 11 01 & + 52 21 18 & ccd13 & 22.5 & 2.2 + & & & & & 14 11 24 & + 52 13 36 & ccd13 & 21.5 & 2.8 + & & & & & 14 11 46 & + 52 05 42 & ccd13 & 22.0 & 2.3 + grb 010222 & 18 & -0.6 @xmath115 0.2 & 0.1 @xmath115 0.1 & 0.06 & 14 51 59 & + 43 08 54 & ccd13 & 21.9 & 2.1 + & & & & & 14 52 13 & + 43 01 06 & ccd13 & 22.9 & 1.4 + & & & & & 14 52 43 & + 42 55 12 & ccd13 & 22.4 & 1.6 + & & & & & 14 52 59 & + 43 06 07 & ccd13 & 22.9 & 1.3 + qso 1508 & 89 & -0.6 @xmath115 0.2 & -0.2 @xmath115 0.1 & 0.03 & 15 09 50 & + 57 04 16 & lfc & 24.0 & 1.2 + mkw 3s & 57 & -0.8 @xmath115 0.1 & 0.1 @xmath115 0.1 & 0.09 & 15 21 30 & + 07 47 43 & ccd13 & 23.4 & 1.2 + & & & & & 15 21 37 & + 07 39 08 & ccd13 & 22.9 & 1.1 + & & & & & 15 22 12 & + 07 48 36 & ccd13 & 22.9 & 1.3 + & & & & & 15 22 13 & + 07 47 21 & ccd13 & 22.9 & 1.6 + ms 1621 & 30 & -1.2 @xmath115 0.1 & 0.3 @xmath115 0.2 & 0.09 & 16 23 06 & + 26 36 33 & ccd13 & 22.4 & 2.3 + & & & & & 16 23 13 & + 26 30 22 & cosmic & 24.4 & 1.4 + & & & & & 16 23 26 & + 26 38 28 & cosmic & 23.9 & 1.6 + & & & & & 16 23 42 & + 26 43 49 & ccd13 & 22.9 & 2.3 + & & & & & 16 24 03 & + 26 35 32 & cosmic & 23.9 & 1.2 + grb 000926 & 32 & -0.7 @xmath115 0.2 & -0.6 @xmath115 0.3 & 0.08 & 17 03 19 & + 51 47 58 & 8k & 22.9 & 1.2 + rx j1716 & 52 & -0.5 @xmath115 0.1 & 1.2 @xmath115 0.1 & 0.09 & 17 15 36 & + 67 19 45 & ccd13 & 22.4 & 1.7 + & & & & & 17 15 55 & + 67 05 59 & cosmic & 24.4 & 1.3 + & & & & & 17 16 39 & + 67 13 27 & cosmic & 23.9 & 1.4 + & & & & & 17 17 11 & + 67 01 40 & cosmic & 24.4 & 1.2 + & & & & & 17 17 42 & + 67 12 36 & ccd13 & 22.9 & 1.3 + & & & & & 17 17 56 & + 67 09 08 & cosmic & 23.9 & 1.6 + & & & & & 17 18 04 & + 67 12 01 & ccd13 & 22.4 & 2.2 + ngc 6543 & 46 & -0.5 @xmath115 0.2 & -0.2 @xmath115 0.1 & 0.12 & 17 57 44 & + 66 44 40 & ccd13 & 22.9 & 1.3 + & & & & & 17 58 31 & + 66 34 02 & ccd13 & 22.9 & 1.4 + & & & & & 18 00 25 & + 66 30 52 & ccd13 & 22.9 & 1.2 + xrf 011030 & 47 & -0.2 @xmath115 0.2 & -0.4 @xmath115 0.2 & 1.09 & 20 44 02 & + 77 21 06 & lfc & 21.4 & 2.2 + ms 2053 & 44 & -1.1 @xmath115 0.1 & 0.4 @xmath115 0.1 & 0.22 & 20 55 48 & @xmath11604 31 10 & ccd13 & 22.8 & 1.7 + & & & & & 20 55 59 & @xmath11604 35 47 & cosmic & 23.8 & 1.4 + & & & & & 20 56 09 & @xmath11604 44 01 & cosmic & 23.8 & 1.4 + & & & & & 20 56 30 & @xmath11604 38 17 & lris & 23.8 & 1.0 + & & & & & 20 56 33 & @xmath11604 30 53 & ccd13 & 22.3 & 2.1 + & & & & & 20 56 42 & @xmath11604 41 28 & cosmic & 23.8 & 1.2 + rx j2247 & 49 & -0.8 @xmath115 0.1 & 1.0 @xmath115 0.1 & 0.15 & 22 47 21 & + 03 39 25 & 8k & 23.4 & 1.7 + q2345 & 74 & 0.0 @xmath115 0.1 & -0.3 @xmath115 0.2 & 0.07 & 23 48 25 & + 01 01 12 & 8k & 23.4 & 1.5 + [ tbl : images ] lccrrrrclccrrcrrc j022142.6 + 422654 & 02 21 42.75 & + 42 26 53.5 & 9.49 & 23.10 & 2.83 & -0.33 & 1 & 18.73 & 0.16 & 24.3 & 1.0 & 1.8 & 2.1 & -0.64 & 0.66 + j022143.6 + 421631 & 02 21 43.72 & + 42 16 31.0 & 8.33 & 74.80 & 8.81 & -0.28 & 6 & & & & & & & & + j022151.6 + 422319 & 02 21 51.76 & + 42 23 18.7 & 6.17 & 6.55 & 2.06 & -0.16 & 2 & 24.3 & & 24.3 & & & & 1.04 & + j022204.9 + 422338 & 02 22 05.08 & + 42 23 37.7 & 4.24 & 8.73 & 2.74 & 0.84 & 1 & 20.37 & 0.16 & 24.3 & -0.2 & 0.5 & 0.5 & -0.41 & 0.02 + j022205.1 + 422213 & 02 22 05.21 & + 42 22 12.7 & 3.45 & 6.82 & 2.38 & 0.07 & 2 & 24.3 & & 24.3 & & & & 1.07 & + j022207.1 + 422918 & 02 22 07.19 & + 42 29 18.2 & 8.93 & 18.90 & 3.84 & -0.33 & 1 & 21.56 & 0.16 & 24.3 & 2.3 & -0.8 & 2.4 & 0.40 & 0.96 + j022210.0 + 422956 & 02 22 10.08 & + 42 29 55.7 & 9.38 & 17.30 & 3.52 & -0.62 & 1 & 18.88 & 0.16 & 24.3 & 0.2 & -0.6 & 0.6 & -0.71 & 0.03 + j022210.8 + 422016 & 02 22 10.93 & + 42 20 16.1 & 2.17 & 3.63 & 1.53 & -0.46 & 1 & 21.57 & 0.17 & 24.3 & 0.6 & -0.4 & 0.8 & -0.31 & 0.95 + j022211.7 + 421910 & 02 22 11.79 & + 42 19 10.1 & 2.55 & 6.87 & 2.36 & 0.14 & 6 & & & & & & & & + j022215.0 + 422341 & 02 22 15.12 & + 42 23 41.0 & 3.15 & 15.40 & 3.88 & -0.41 & 1 & 22.52 & 0.18 & 24.3 & 0.1 & 0.0 & 0.1 & 0.70 & 0.64 + j022215.1 + 422045 & 02 22 15.19 & + 42 20 44.5 & 1.32 & 42.90 & 7.49 & -0.42 & 1 & 17.33 & 0.16 & 24.3 & -0.3 & 0.1 & 0.3 & -0.94 & 0.03 + j022215.5 + 421842 & 02 22 15.63 & + 42 18 41.7 & 2.46 & 6.79 & 2.34 & 0.28 & 1 & 20.33 & 0.16 & 24.3 & -0.4 & 0.1 & 0.4 & -0.54 & 0.03 + j022219.3 + 422052 & 02 22 19.40 & + 42 20 51.6 & 0.54 & 6.38 & 2.31 & 0.37 & 1 & 23.82 & 0.22 & 24.3 & -0.9 & 0.6 & 1.1 & 0.83 & 0.38 + j022224.3 + 422138 & 02 22 24.45 & + 42 21 38.4 & 0.91 & 64.60 & 8.92 & -0.44 & 1 & 17.41 & 0.16 & 24.3 & 0.0 & 0.0 & 0.0 & -0.73 & 0.98 + j022225.2 + 422451 & 02 22 25.33 & + 42 24 50.9 & 4.07 & 128.00 & 12.99 & -0.34 & 1 & 19.31 & 0.16 & 24.3 & 0.1 & 0.2 & 0.2 & 0.33 & 0.98 + j022226.5 + 422154 & 02 22 26.63 & + 42 21 54.4 & 1.35 & 6.48 & 2.32 & -0.35 & 2 & 24.3 & & 24.3 & & & & 1.04 & + j022232.5 + 423015 & 02 22 32.61 & + 42 30 14.6 & 9.61 & 53.60 & 6.50 & 0.12 & 1 & 20.88 & 0.16 & 24.3 & 0.4 & 0.4 & 0.5 & 0.58 & 0.77 + j022236.3 + 421730 & 02 22 36.45 & + 42 17 30.2 & 4.23 & 12.00 & 3.30 & 0.01 & 1 & 22.38 & 0.18 & 24.3 & 0.3 & -0.9 & 1.0 & 0.53 & 0.42 + j022236.7 + 422858 & 02 22 36.88 & + 42 28 57.7 & 8.57 & 12.90 & 3.00 & -0.68 & 1 & 22.05 & 0.17 & 24.3 & -1.5 & 2.1 & 2.6 & 0.43 & 0.81 + j022259.1 + 422434 & 02 22 59.18 & + 42 24 33.6 & 7.77 & 32.50 & 5.43 & 0.49 & 2 & 24.3 & & 24.3 & & & & 1.74 & + j022334.0 + 422212 & 02 23 34.13 & + 42 22 11.6 & 13.34 & 37.50 & 4.47 & 0.15 & 2 & 24.3 & & 24.3 & & & & 1.81 & + [ tbl : catalog ]

the study of the electronic structure of materials usually begins with noninteracting electrons due to the vast number of particles involved . many - body methods then provide a hierarchy of corrections to account for the coulomb interaction to various order . many - body approaches such as gw@xcite and the bethe - salpeter equation ( bse)@xcite are frequently and successfully employed in the calculation of the electronic structure and excitations of materials . though accurate and physically sound , these many - body methods can become cumbersome and impractical for large systems due to the steep scaling of the numerical cost versus the system size . alternatively , density - functional theory ( dft ) and time - dependent density - functional theory ( tddft)@xcite are popular methods for calculating electronic ground states and excitations , respectively , and are widely used in chemistry , physics , materials science , and other areas . density - functional methods solve the many - body problem by constructing a noninteracting system which reproduces the electronic density of the interacting , physical system . the favorable balance between accuracy and efficiency makes the resulting dft and tddft schemes unrivaled for large but finite system sizes.@xcite considerable effort has been spent to replicate this success of tddft for periodic solids.@xcite generally speaking , tddft works very well for simple metallic systems , where the excitation spectrum is dominated by collective plasmon modes . the reason is that common local and semilocal exchange - correlation ( xc ) functionals are based on the homogeneous electron liquid as reference system , which is an ideal starting point to describe electrons in metals . the situation is more complicated in insulators and semiconductors . the first problem that comes to mind is that of the band gap , which is typically strongly underestimated by most popular xc functionals of dft . in principle , tddft provides a mechanism to obtain the correct band gap,@xcite but this puts very strong demands on the xc kernel @xmath0 ( it has to simulate a discontinuity , which requires a strong frequency dependence ) . the second difficulty are excitonic effects . it is a well - known fact that standard local and semilocal xc functionals do not produce any excitonic binding;@xcite again , the proper choice of @xmath0 is crucial . there are many examples in the literature of successful tddft calculations of excitonic effects , using exact exchange , @xcite an effective xc kernel engineered from the bse,@xcite a meta - gga kernel,@xcite and a recent ` bootstrap ' xc kernel.@xcite these kernels all have in common that they have a long spatial range ; however , it has also been shown that certain excitonic features can be equally well reproduced by simple short - range kernels . @xcite this calls for further explanation . due to the complexity of real solids , the question of the general requirements for excitonic binding in tddft has been difficult to analyze . as a first step towards a simplified tddft approach for excitons , a two - band model was recently developed , which was used to test the performance of simple xc kernels for calculating excitonic binding energies in several iii - v and ii - vi semiconductors.@xcite in this paper we will push this reductionist approach further and propose a minimal tddft model for excitons . our model is one - dimensional ( 1d ) and uses two simple kronig - penney - type bands as input . we show that the minimal model reproduces and reveals many aspects of excitons . the model is accessible and relatively easy to implement , and it can be used to identify important aspects of the xc functional for excitonic effects . it clearly shows that excitons are collective excitations of the many - body system : the appropriate phase - coherent mixing of the single - particle excitations is accomplished via a coupling matrix featuring @xmath0 . the properties of this coupling matrix are analyzed and compared with its bse counterpart . in textbooks , excitons are usually introduced through the two - body wannier equation , which describes an electron and a hole which interact via a screened coulomb potential . while this arguably constitutes the simplest model for excitons , it is based on several drastic assumptions which are not fulfilled in general.@xcite we discuss how and under what circumstances a wannier - like equation emerges from our minimal tddft model . the paper is structured as follows . we give an introduction to wannier excitons in sec . [ sect : bkgnd ] and to tddft for solids in sec . [ sect : tddft ] , followed by a description of the minimal model in sec . [ sect : model ] . section [ sect : results ] then presents results for the minimal exciton model comparing tddft with the bse , and discusses various implications . in sec . [ sect : wannier ] we show how the wannier equation emerges from tddft . conclusions are given in in sec . [ sect : app : details ] . details of the bse method are provided in appendix [ sect : app : bse ] . atomic units ( @xmath1 ) are used throughout unless otherwise stated , and we will only consider spin - unpolarized systems . the electronic structure of crystalline solids is described by the bloch theory , where the electrons move in a periodic effective single - particle potential which reflects the crystal symmetry . as a result , the electronic states form energy bands . in insulators and semiconductors , electronic excitations take place between the occupied ( valence ) and unoccupied ( conduction ) bands . these interband transitions can be described within a simple independent - particle approach based on fermi s golden rule ; one thus obtains a reasonable qualitative account of the optical properties in these materials.@xcite however , experiments reveal that there are important modifications to this picture , as illustrated schematically in fig . [ fig : bkgnd : illustrationexcitons ] . above the band gap , the spectrum appears strongly enhanced , and below the band gap one may find discrete absorption peaks known as bound excitons . the origin of these modifications are coulomb interactions : the simple picture of independent single - particle excitations is replaced by a more complex scenario where these excitations are dynamically coupled . excitonic effects are ubiquitous in nature , and occur in 3d , 2d and 1d systems alike.@xcite the details of the excitonic modifications to the noninteracting spectrum depend on the dimensionality of the system.@xcite the standard textbook explanation of excitons is based on the simple picture of an electron - hole pair held together by coulomb interactions , see fig . [ fig : bkgnd : wannierexciton ] . one thus arrives at a two - body problem similar to the positronium atom , with a center - of - mass momentum @xmath2 and relative motion described by the wannier equation:@xcite @xmath3\psi_\nu({{\bf r}})=e_\nu\psi_\nu({{\bf r } } ) , \label{eqn : bkgnd : wannier}\ ] ] where @xmath4 is the reduced mass , defined as @xmath5 . @xmath6 and @xmath7 are the effective masses of the conduction band electrons and the valence band holes . @xmath8 is the coulomb interaction between the electron and the hole , divided by the static dielectric constant of the system . @xmath9 and @xmath10 are the excitonic wave function and binding energy , respectively . in 1d systems , the coulomb interaction is ill - defined and requires , in general , some parametrized form;@xcite we will use the following soft - coulomb interaction : @xmath11 where @xmath12 and @xmath13 are parameters . in the following , we will use @xmath14 throughout . the wannier equation ( [ eqn : bkgnd : wannier ] ) has the form of a hydrogenic schrdinger equation , so it possesses a rydberg series ( even for 1d cases with soft - coulomb interaction ) with infinitely many eigenvalues below the band gap , and a continuum above the gap.@xcite the excitons in the wannier model for 3d and 2d cases both enhance the optical absorption near the band gap , while the presence of excitons in 1d systems suppresses the optical absorption just above the band gap . sham and rice@xcite showed how the wannier equation can be derived from first principles starting from the bse , under the assumption that the bohr radius of the exciton is much bigger than the lattice constant . the resulting wannier picture of excitons appears clear and intuitive , but this simplicity is somewhat deceptive . in reality , excitons are a dynamical many - body phenomenon and require a subtle coordination and cooperation of many single - particle transitions between two bands . in the following , we will develop a model based on tddft which will illustrate the true physical nature of excitons , but which will remain sufficiently transparent to allow a simple interpretation of the collective many - body effects that are responsible for the excitonic binding . tddft@xcite is an in principle exact approach for electron dynamics , based on the uniqueness of the mapping between the time - dependent electronic density and the external potential.@xcite the key equation of tddft is the time - dependent kohn - sham ( tdks ) equation : @xmath15\phi_i({{\bf r}},t ) , \label{eqn : bkgnd : tddft}\end{aligned}\ ] ] where @xmath16 are the tdks orbitals of the noninteracting kohn - sham system which reproduces the density of the real interacting system . @xmath17 and @xmath18 are the external potential of the physical system and the hartree potential , respectively , and the xc potential @xmath19 is the only piece that needs to be approximated in practice . the excitation spectrum of a system can be calculated via time propagation of eq . ( [ eqn : bkgnd : tddft ] ) following a suitably chosen initial perturbation . alternatively , one can obtain excitation energies and optical spectra directly from linear - response tddft , using the so - called casida equation:@xcite @xmath20 equation ( [ eqn : bkgnd : casida ] ) is a generalized eigenvalue equation . one obtains the optical transition frequencies from the eigenvalues @xmath21 , and the corresponding eigenvectors tell us how the kohn - sham single - particle transitions are mixed to form the transitions of the interacting system . the matrices @xmath22 and @xmath23 in eq . ( [ eqn : bkgnd : casida ] ) are defined as @xmath24 where @xmath25,@xmath26,@xmath27,@xmath28 are labels for ground - state kohn - sham orbitals , and the @xmath29 s are the associated kohn - sham orbital energies . @xmath30 in eq . ( [ eqn : bkgnd : ab ] ) is defined as @xmath31 where the @xmath32 s are the ground - state kohn - sham orbitals , and @xmath33 is the hartree - exchange - correlation ( hxc ) kernel , defined as a fourier transform of @xmath34 the xc kernel @xmath35 has to be approximated in practice . @xmath36 and @xmath37 make up the eigenvector in eq . ( [ eqn : bkgnd : casida ] ) , and they describe excitations and de - excitations , respectively . a commonly used approximation to eq . ( [ eqn : bkgnd : casida ] ) , known as tamm - dancoff approximation ( tda),@xcite is to set @xmath38 , so that the casida equation reduces to @xmath39\rho_{mn}(\omega)=\omega\rho_{ij}(\omega ) . \label{eqn : bkgnd : tddftworking}\ ] ] this decouples excitations and de - excitations , and the computational cost is reduced . there are situations , for instance for molecular excitations of open - shell systems,@xcite in which the tda is preferred in practice over the full casida equation ( [ eqn : bkgnd : casida ] ) . we find that the tda can also be advantageous for excitons ( see sec . [ sect : results ] ) . by considering only a single kohn - sham transition in eq . ( [ eqn : bkgnd : casida ] ) one arrives at the small - matrix approximation ( sma):@xcite @xmath40 where @xmath41 is the kohn - sham transition frequency to be corrected . one can further simplify this by making the tda , which yields the single - pole approximation ( spa ) : @xmath42 the sma and spa are valid when the considered excitation is far away from other transitions in the system . though they are not usually accurate enough for real calculations , their simplicity makes them very useful for theoretical analysis and development . in periodic solids , the kohn - sham orbitals are labeled with the band index @xmath25 and the wavevector @xmath43 and have the bloch form : @xmath44 it is in principle possible to adapt the casida equation ( [ eqn : bkgnd : casida ] ) for the case of orbitals of the form ( [ eqn : bkgnd : blochform]),@xcite and use this to calculate the excitation spectrum . however , to obtain the optical spectrum in solids it is more convenient to calculate the macroscopic dielectric function:@xcite @xmath45 where the @xmath46 s are reciprocal lattice vectors , and @xmath47 is the linear response function . @xmath48 can be expressed as@xcite @xmath49 where @xmath50 labels the solutions of eq . ( [ eqn : bkgnd : tddftworking ] ) for an extended system . beyond linear response , few real - time tddft calculations exist for periodic solids.@xcite instead of directly solving the tdks equation ( [ eqn : bkgnd : tddft ] ) , the tdks orbitals can be expanded in terms of ground - state kohn - sham bloch functions as @xmath51 the time - dependent density matrix is defined as @xmath52 the equation of motion of the density matrix is then @xmath53 , \label{eqn : bkgnd : tddmeom}\ ] ] with the tdks hamiltonian matrix @xmath54 defined by @xmath55 where @xmath56 is the volume of the unit cell , and @xmath57 is the tdks hamiltonian of eq . ( [ eqn : bkgnd : tddft ] ) . this density - matrix approach has been used to derive the tddft version of the semiconductor bloch equations.@xcite in this formalism we only consider vertical transitions , where the bloch wavevector @xmath2 does not change during the dynamics . nonvertical excitations are not considered , since they involve indirect ( i.e. , phonon - assisted ) processes which we ignore here . solids are formally described by the many - body schrdinger equation . since exact solutions are not possible , it is instructive to resort to model systems to eliminate undesired details of the many - body system and provide clear illustrations of the specific features one is interested in . the wannier equation for excitons presented in sec . [ sect : bkgnd ] is such a model . while intuitive , the wannier model assumes the electron - hole interaction as given , so the excitonic effects are already built in by default however , this does not explain under what conditions one expects to see the formation of excitons , and the many - body nature of excitonic effects remains hidden . we therefore propose a minimal model for excitons which lowers the abstraction level of the wannier model , and where excitonic effects show up without any _ ad hoc _ assumptions . for excitations near the band gap , a reasonable approximation is to use a two - band model , i.e. , only to consider the highest valence band @xmath58 and the lowest conduction band @xmath59 . this means that we only need to consider those elements of the time - dependent density matrix @xmath60 [ eq . ( [ tddm ] ) ] for which @xmath61 and @xmath62 ; the latter index will be dropped in the following . for the case when a small perturbative electric field is applied to the system , it is sufficient , to lowest order in the perturbation , to consider only the time evolution of the off - diagonal part of the density matrix,@xcite @xmath63 . one then obtains from eq . ( [ eqn : bkgnd : tddmeom ] ) @xmath64 where @xmath65 , and @xmath66 . here , @xmath67 denotes the matrix elements of @xmath68 , defined similarly to eq . ( [ hmat ] ) . there is no external perturbation in eq . ( [ eqn : bkgnd : rhocvintime ] ) ; we assume free propagation of the system . our reasoning is that individual excitations can be viewed as the eigenmodes of the system and do not depend on the specific form of the perturbative field . fourier transformation of eq . ( [ eqn : bkgnd : rhocvintime ] ) gives @xmath69 where the factor 2 accounts for the spin , and @xmath70 equation ( [ eqn : bkgnd : rhocv ] ) is equivalent to the sma for finite systems . while the sma only refers to the transition between one individual occupied and one individual unoccupied orbital , eq . ( [ eqn : bkgnd : rhocv ] ) considers the transitions between the valence and the conduction band as a whole . ignoring the coupling between excitations and de - excitations by setting @xmath71 ( i.e. , making the tda ) , one arrives at the solid - state analog of the spa : @xmath72\rho_{{\mathbf{k}}'}^{cv}(\omega ) = \omega\rho_{\mathbf{k}}^{cv}(\omega ) . \label{eqn : bkgnd : rhocvspa}\ ] ] this is the central equation which we will use to describe excitonic effects . it requires as input the kohn - sham bloch functions for the valence and the conduction band of an insulator or semiconductor . to keep things as simple as possible , we will consider the kronig - penney ( kp ) model@xcite rather than a real material . the kp model is a 1d noninteracting system with a periodic potential of square wells . within the unit cell @xmath73 $ ] , the potential is @xmath74 a typical example for the band structure of the kp model is plotted in fig . [ fig : bkgnd : kppot ] . despite its simple appearance , the kp model is very versatile : by varying the values of the lattice constant @xmath75 , barrier width @xmath76 , and barrier height @xmath77 , a wide range of band gaps and band curvatures can be achieved . a square - well potential of finite depth does not support an infinite number of bound states as the coulomb potential does ; but this , in fact , closely reflects the reality of the effective potential felt by the valence electrons in a solid , which is relatively shallow due to the screening of the bare nuclear charges by the core electrons . in practice , many solid - state calculations account for this screening by using muffin - tin potentials or pseudopotentials . @xcite the kp model can be viewed as an elementary version of this approach . in the following , we always choose the first two bands to be fully occupied and the higher bands to be empty . we then make the two - band approximation for band 2 and 3 , since the shapes of these bands resemble the highest valence band and lowest conduction band in direct - gap materials such as gaas . the bands in the kp model are sufficiently well separated , so the two - band approximation is justified . to establish a connection with tddft , we assume the solution to the noninteracting kp model as our ground - state kohn - sham system . in other words , the potential in eq . ( [ eqn : bkgnd : kppot ] ) represents the _ exact _ kohn - sham potential @xmath78 which corresponds to a physical system whose external potential @xmath79 is uniquely determined thanks to the hohenberg - kohn theorem of dft . for our purpose , it is not necessary to know what this external potential looks like . the kohn - sham bloch functions can then be determined in an elementary fashion.@xcite for comparison , we will also carry out bse calculations in our model system ( see appendix [ sect : app : bse ] for technical details ) . bse calculations are typically based on ground - state quasiparticle states obtained from the gw method.@xcite this is because the single - particle gap in gw is usually closer to experiment than the approximate kohn - sham gap . however , in our case this distinction is not important because we use the given kp band structure as input for both bse and tddft . in summary , our minimal tddft model for excitons consists of the following two ingredients : \(1 ) a two - band model for the vertical transitions between the highest valence band and the lowest conduction band , see eq . ( [ eqn : bkgnd : rhocvspa ] ) ; \(2 ) the band structure from a 1d kp model . of course , the model is not complete without a choice for the xc kernel @xmath35 . this will be discussed below . the exact xc kernel @xmath80 is unknown and must be approximated ; we restrict ourselves to adiabatic kernels that have no frequency dependence . the adiabatic local - density approximation ( alda ) , as well as all semilocal , gradient - corrected xc kernels , are known to be unable to describe excitonic effects.@xcite the exact xc kernel has a long - range decay of @xmath81 , which is absent in all ( semi)local xc kernels derived from the uniform electron gas . this long - range part is thought to be essential for excitons.@xcite the long - range behavior of @xmath35 depends on the dimensionality , and in our 1d model system we define the following long - ranged ( or ` soft - coulomb ' ) xc kernel : @xmath82 we also consider an extremely short - ranged contact xc kernel : @xmath83 these model xc kernels depend on the constants @xmath84 and @xmath85 , which we will treat as fitting parameters in the following . the idea is to tune the parameters in the model xc kernels so that bound excitons are produced , and to align the lowest bound exciton in the tddft spectrum with the lowest bound exciton in the bse spectrum . results for the imaginary part of the dielectric function are presented in figs . [ fig : results : tddftoneexciton ] and [ fig : results : tddfttwoexcitons ] . we find that both the long - ranged @xmath86 and the short - ranged @xmath87 produce bound excitons , and thus , strictly speaking , the long - range behavior of the xc kernel is not really required for excitonic effects . the bse results in figs . [ fig : results : tddftoneexciton ] and [ fig : results : tddfttwoexcitons ] show several identifiable bound excitons , in agreement with the rydberg series predicted by the wannier model ; the number of visible bound excitons somewhat depends on the numerical resolution in momentum space . for the kp model parameters of fig . [ fig : results : tddftoneexciton ] , we find that the adiabatic tddft can only bind a single excitonic state . for other kp parameters ( specifically those in which the lowest conduction band is above the barrier ) , tddft produces two excitons , see fig . [ fig : results : tddfttwoexcitons ] , which agree well with the lowest two excitons in bse . there are additional , higher - lying bound excitons in bse which are very faint and difficult to resolve numerically . for all the kp systems we tested , we never found more than two bound excitons with tddft . this indicates the limitations of the adiabatic xc kernels used here . + as mentioned in sec . [ sect : model ] , the wannier model does not clearly demonstrate that excitons are collective excitations . since the wannier model assumes a single electron - hole pair picture , one can not immediately see that excitonic excitations are composed of a coherent superposition of many single - particle excitations . in our minimal model , we solve the eigenvalue equation ( [ eqn : bkgnd : rhocvspa ] ) , and the eigenvectors @xmath88 ( which depends on @xmath21 parametrically ) describe how the transitions between noninteracting orbitals form the transitions in the interacting system . @xmath89 is the percentage of a noninteracting transition in the transition of the interacting system . two typical cases are plotted in figs . [ fig : results : excitoniceigenvect ] and [ fig : results : nonexcitoniceigenvect ] . figure [ fig : results : excitoniceigenvect ] clearly shows that excitons are collective excitations which are formed by mixing a wide distribution of single - particle transitions . as expected , the lowest exciton eigenfunction is nodeless and the second excitonic eigenfunction has a single node . with purely parabolic bands , the results from the wannier model would be recovered , as we will show below . in contrast , the transitions in the continuum shown in fig . [ fig : results : nonexcitoniceigenvect ] have a strong single - particle character ( the two peaks arise from the @xmath90 degeneracy in the kp model ) . equation ( [ eqn : bkgnd : rhocvspa ] ) is equivalent to the spa for finite systems , which ignores the coupling between excitations and de - excitations ( tda ) . we also investigated what happens when we do not make the tda , i.e. , when we work instead of eq . ( [ eqn : bkgnd : rhocvspa ] ) with the full equation for the two - band model , eq . ( [ eqn : bkgnd : rhocv ] ) . as long as we describe relatively weakly bound excitons that are not too far below the band gap , we find that the difference between the two methods is very minor . however , we also discovered that , under rare circumstances , eq . ( [ eqn : bkgnd : rhocv ] ) can lead to tddft excitonic binding energies that are purely imaginary . in our minimal model , such instabilities arise when the interaction strength @xmath12 in eq . ( [ eqn : results : fxcsc ] ) and ( [ eqn : results : fxccont ] ) increases so that the excitonic binding energy becomes greater than the band gap . this situation is comparable to the well - known triplet instability in tddft , for which the tda generally leads to an overall better behavior;@xcite for excitons binding energies in our minimal model , we draw similar conclusions . the bse scheme is commonly implemented within an adiabatic scheme ( see appendix [ sect : app : bse ] ) ; as we have seen , it produces a series of bound excitons . since we assumed that the kp model is the kohn - sham ground state in tddft and the gw quasiparticle ground state in bse , the difference between tddft and bse becomes easily comparable , since the central equation to be solved have the same form , eq . ( [ eqn : bkgnd : rhocvspa ] ) . the @xmath91 coupling matrices for tddft and bse are @xmath92 where @xmath93 is the hartree kernel ( the 1d soft - coulomb interaction ) , and @xmath94 is the screened interaction . aside from the change from @xmath35 to @xmath94 , the most prominent difference between bse and tddft is the order of the indices for @xmath94 in eq . ( [ eqn : results : fbse ] ) . since the noninteracting ground - state wave functions have the bloch form ( [ eqn : bkgnd : blochform ] ) , we can see from eq . ( [ eqn : bkgnd : fijmn ] ) that @xmath95 has a strong @xmath96 dependence ; in fig . [ fig : results : bsecoupling ] this shows up as a dominance along the diagonal . by contrast , this @xmath96 dependence is clearly absent in @xmath97 , as demonstrated in fig . [ fig : results : tddftcoupling ] . @xmath98 and @xmath99 with only vertical transitions can be expressed in momentum space as @xmath100 @xmath101 the xc matrix ( [ eqn : results : fxctrans ] ) only depends on the long - range ( @xmath102 ) behavior of its momentum space representation @xmath103 , while the @xmath94 matrix ( [ eqn : results : wtrans ] ) also depends on other @xmath104 values in its momentum space representation @xmath105 . it is impossible to find an adiabatic @xmath35 that reproduces the bse coupling matrix as in fig . [ fig : results : bsecoupling ] , since @xmath105 has an extra degree of freedom over @xmath106 . one can only hope to reproduce a portion of the bse coupling matrix with adiabatic tddft ( as pointed out in ref . ) , or make the xc kernel frequency dependent so that the information from the @xmath104-dependence in @xmath105 is mapped into the frequency dependence in @xmath107 . considering the nature of the objects involved in this mapping , a highly nontrivial frequency dependence in @xmath35 is required to reproduce a series of bound excitons . for example , one can easily construct an @xmath35 which reproduces a given series of bound excitons by using a different contact kernel in the region @xmath108 surrounding each exciton at frequency @xmath109 : @xmath110\theta[(\omega_i+\eta_i^+)-\omega].\ ] ] here , the @xmath111 s are parameters which are adjusted so that a tddft calculation with the frequency - independent kernel @xmath112 would produce @xmath109 as the lowest excitonic binding energy . such an @xmath35 is of course completely ad hoc , but the fact that the excitonic series can be reproduced in this way demonstrates that the inclusion of the frequency dependence would greatly improve the flexibility of the tddft scheme . on the other hand , within the adiabatic approximation the characteristics of the @xmath113 coupling matrix are important for excitonic effects . to emphasize this , we now show that in very special cases the number of discrete excitonic eigenvalues can be derived . consider a real matrix @xmath114 , where @xmath115 is a diagonal real matrix with @xmath116 , and @xmath117 has the symmetry @xmath118 . within second order perturbation theory , there is at most one discrete eigenvalue of @xmath114 in the limit where @xmath119 become continuous.@xcite though this case does not correspond to the matrices that would occur in real calculations , it indicates the close relationship between the properties of the coupling matrix and excitonic effects . it is also possible to derive properties of the discrete eigenvalues if @xmath120 and @xmath121 are completely decoupled in the xc kernel . owing to the symmetry @xmath122 implied by eq . ( [ eqn : bkgnd : fijmn ] ) , such separable kernels can only have the form @xmath123 for an excitation below the band gap with frequency @xmath21 , we can show@xcite that it must satisfy @xmath124 where the sum is carried out over the first brillouin zone ( fbz ) . equation ( [ eqn : results : eqnsepkernel ] ) shows that eq . ( [ eqn : results : sepkernel ] ) must have the negative sign in order to have bound excitons . the left - hand side of eq . ( [ eqn : results : eqnsepkernel ] ) is monotonically increasing with @xmath21 , so for separable kernels of the form of eq . ( [ eqn : results : sepkernel ] ) , there is only one bound excitonic solution . as shown in fig . [ fig : results : tddftcoupling ] , tddft coupling matrices lack the strong dependence of @xmath125 as in bse coupling matrices . expanding the tddft coupling matrices into a power series of separable matrices and truncating at the first order would be a reasonable approximation , explaining why tddft produces fewer bound excitons ( if any at all ) than many - body methods such as bse . the contact xc kernel and the soft - coulomb xc kernel in eqs . ( [ eqn : results : fxcsc ] ) and ( [ eqn : results : fxccont ] ) have the following simple form in momentum space : @xmath126 where @xmath127 , @xmath128 and @xmath128 are reciprocal lattice vectors , and @xmath129 is the modified bessel function of the second kind . it is customary to refer to the matrix elements where @xmath130 as ` head ' , @xmath131 or @xmath132 as ` wings ' , and @xmath133 as ` body ' . the 3d coulomb potential has the form @xmath134 in momentum space . however , in 1d systems there is no real coulomb interaction which behaves as @xmath135 for @xmath136 , and one has to use the soft - coulomb interaction instead . though there are many flavors of the soft - coulomb interaction , they all have the same @xmath137 behavior for @xmath136 . however , the linear response function @xmath138 always behaves as @xmath135 for @xmath136 and does not depend on the dimensionality . this renders quantities like the macroscopic dielectric function ( [ eqn : bkgnd : macrodielec ] ) ill - defined for strictly 1d systems . furthermore , the bootstrap xc kernel@xcite and other xc kernels that depend on the cancellation of the 3d coulomb @xmath135 singularity will not work as designed in strictly 1d and 2d systems . therefore @xmath139 shown in fig . [ fig : results : tddftoneexciton ] and [ fig : results : tddfttwoexcitons ] are calculated at a small but finite @xmath140 . the coupling matrix @xmath141 can be written in momentum space as @xmath142 \nonumber\\ & & \times{\left\langle c , k \left| e^{igx } \right| v , k \right\rangle}{\left\langle v , k ' \left| e^{-ig'x } \right| c , k ' \right\rangle}. \label{eqn : app : detail : fvcvc}\end{aligned}\ ] ] for any xc kernel that behaves as @xmath135 for @xmath136 , one can further simplify the calculation by ignoring the so - called local field effects,@xcite i.e. instead of summing over @xmath128 and @xmath143 in eq . ( [ eqn : app : detail : fvcvc ] ) , only the head is considered . in 3d systems , a prominent example is the long - range kernel @xmath144 , which is obtained as an effective xc kernel with only head matrix elements from inverting the bse of contact excitons.@xcite on the other hand , any xc kernel that diverges more slowly than @xmath135 for @xmath136 changes the spectrum only through the local field effects , i.e. all @xmath128 and @xmath143 must be summed in eq . ( [ eqn : app : detail : fvcvc ] ) . in other words , effective xc kernels with only the head are not feasible in strictly 1d systems due to the asymptotic behavior of the soft - coulomb potential discussed above . for 1d systems with @xmath131 , we have @xmath145 and @xmath35 s with 1d long - range behavior such as the soft - coulomb kernel also behave as @xmath146 . considering eq . ( [ eqn : app : detail : fvcvc ] ) , these asymptotic properties imply that the head and wing contributions to @xmath91 always vanish in strictly 1d systems for physically meaningful xc kernels . due to these dimensionality restrictions , the xc kernel changes the strictly 1d system only through the local field effects . in 3d the head contribution to the coupling matrix @xmath147 is much more important than the local field effects , which is the reason that long - range kernels ( with nonzero head ) work much better than local xc kernels ( with vanishing head ) such as alda . in our strictly 1d model system , the head contribution is zero even for the bse , and thus the long - range kernel does not outperform local kernels such as the contact kernel . these peculiarities only occur when one considers strictly 1d and 2d systems . in a more realistic picture , one encounters quasi-2d systems@xcite and quasi-1d systems ( such as quantum wires with finite radius or nanotubes@xcite ) , in which the movement of electrons is confined in certain directions such that the transverse motion can be averaged in comparison with the longitudinal motion . though these systems show low - dimensional characteristics in various properties due to confinement , in the limit of @xmath136 they eventually differ from strictly low - dimensional systems . our minimal model and the wannier exciton picture can be connected by considering the fourier transform of eq . ( [ eqn : bkgnd : rhocvspa ] ) . we define an effective two - body potential @xmath148 via the fourier transform of @xmath149 : @xmath150 where @xmath151 is a direct lattice vector . the fourier transform of the density matrix is @xmath152 since wannier excitons extends over many lattice constants , we approximate @xmath151 as a continuous variable @xmath153 . assuming the effective mass approximation , eq . ( [ eqn : bkgnd : rhocvspa ] ) becomes @xmath154 where @xmath155 is the excitonic binding energy , and the integration is carried out over all space . we call eq . ( [ eqn : bkgnd : tddftwannier ] ) the tddft wannier equation , since it has the same form as eq . ( [ eqn : bkgnd : wannier ] ) . with proper choice of the approximated xc kernel , the nonlocal effective electron - hole interaction potential @xmath148 supports bound excitonic states . since the bse and the tddft are formally similar within the minimal model , eq . ( [ eqn : bkgnd : rhocvspa ] ) can also be applied to the bse results . [ fig : results : veh ] shows the effective interaction potential @xmath148 for tddft and bse . \(a ) + ( b ) the tddft wannier equation provides an intuitive way of describing the effective nonlocal electron - hole interaction , and of explaining why adiabatic tddft usually has fewer excitons than bse and the wannier model . however , in most cases the tddft wannier equation is not suitable for quantitative use due to the approximations involved . the approximation where we take the lattice vector @xmath151 as a continuous variable assumes that the exciton radius is much larger than the lattice constant ; this works fine in most cases we tested . but the effective mass approximation where @xmath156 is approximated by @xmath157 is only good for transitions near the band gap , thus requiring these transitions of the noninteracting system to dominate the exciton , which is equivalent to the exciton extending over many lattice constants . one obtains the @xmath158 term in eq . ( [ eqn : bkgnd : tddftwannier ] ) from @xmath157 in the limit where the lattice constant @xmath159 , and this approximation is not valid for most systems . although @xmath148 is a nonlocal potential , in most cases we find that the @xmath148 s for both tddft and bse are dominated by the diagonal part , so the exciton problem is in analogy to one - body systems . [ fig : results : vehdiag ] shows the diagonal part of @xmath148 , which can be taken as the effective one - body potential . the wannier model in 1d has the soft - coulomb potential , which supports an infinite number of bound excitons ( the soft - coulomb interaction is fitted so that the binding energy of the first exciton matches that of the bse ) . we find in general that the diagonal parts of @xmath148 for both bse and tddft are much more shallow than the soft - coulomb potential , and the tddft one is more narrow than the bse one . thus , bse and tddft are not able to produce a complete excitonic rydberg series , and tddft in general produces fewer bound excitons than the bse . the tddft wannier equation is not suitable for quantitative use for most of our model systems , despite the success of the wannier model in describing real semiconductors.@xcite since the approximations involved in eq . ( [ eqn : bkgnd : tddftwannier ] ) require that the exciton radius is large compared to the lattice constant , this suggests that this discrepancy is due to the special nature of 1d systems : namely , for similar effective masses the exciton radius in 1d is much smaller than in 3d and 2d.@xcite the purpose of this paper was to construct a transparent and accessible minimalist model system that produces excitonic effects in a non-_ad hoc _ fashion using tddft . the model , as presented here , is not intended to be a testing ground for xc kernels . thus despite the dimensionality restrictions for the strictly 1d system , our results in sec . [ sect : results ] carry over to 3d systems . with our minimal model , we show that adiabatic tddft is capable of producing bound excitons through the local field effect even when the xc kernel is local in space , provided the strength of the kernel is strong enough . this statement is still true in 3d ; however , due to the non - vanishing head contribution of the exact @xmath35 , we expect that the deviation of the effective interaction strength of a local @xmath35 from the real , nonlocal @xmath35 becomes larger than the in our strictly 1d model . in this sense the long - range kernel , though very favorable , is not a necessary condition for excitonic effects . we show the connection between tddft and the wannier model for excitons by deriving the tddft wannier equation , which describes a real - space system featuring a nonlocal effective electron - hole interaction . such a connection intuitively demonstrates how adiabatic tddft generally produces fewer bound excitons than bse , and does not have a complete rydberg series . the eigenvectors of the excitonic excitations in the minimal model clearly show their collective nature , which is not obvious from the wannier model alone . excitonic instabilities may show up in tddft with approximate xc kernels , and this suggests that the tda tends to be more reliable for excitons than the formally exact method . the frequency dependence of the exact xc kernel , @xmath160 , is usually ignored . despite the fact that adiabatic xc kernels have met with considerable recent success in producing optical spectra of insulators and semiconductors ( see the discussion in the introduction ) , they are incapable of producing excitonic rydberg series . our model system gives an explanation for why this is the case . this failure of the adiabatic approximation for @xmath35 is quite different from that which is responsible for the inability of adiabatic tddft to produce double excitations in finite systems or certain classes of charge - transfer excitations.@xcite this calls for continuing efforts in the search for nonadiabatic xc kernels for excitons . this work was supported by nsf grant no . dmr-1005651 . electrons and holes near the fermi surface are well described in the quasiparticle picture . the quasiparticle green s function @xmath128 is related to that of the noninteracting system , @xmath161 , through the use of the self - energy @xmath162 : @xmath163 where the arguments denote sets of space and time variables . a widely used approximation for the self - energy is the gw approximation : @xcite @xmath164 where @xmath94 is the screened interaction , @xmath165 @xmath166 is the bare coulomb interaction , and the inverse dielectric function @xmath167 is obtained as @xmath168 the linear response function @xmath138 can be calculated by its lehmann representation : @xmath169 where @xmath170 are quasiparticle states , @xmath171 are quasiparticle energies , and @xmath172 are occupation numbers . in practice , evaluating @xmath138 through eq . ( [ eqn : app : bse : chi ] ) can be quite time - consuming , and @xmath138 is often calculated with the plasmon - pole model.@xcite however , our minimal model is simple enough to allow us to use eq . ( [ eqn : app : bse : chi ] ) directly . the gw quasiparticle green s function obtained from eq . ( [ eqn : app : bse : gwsigma ] ) misses important dynamical many - body effects , such as the electron - hole ( excitonic ) interaction . the two - particle green s function includes these effects . the bse@xcite describes the relation between the four - point polarization function @xmath173 of an interacting system and the corresponding object of the quasiparticle system : @xmath174 in which @xmath175 is @xmath176 and assuming the gw approximation , the kernel @xmath129 is @xmath177 here , @xmath178 denotes the coulomb interaction with the long - range part removed.@xcite in practice the bse is often solved in the transition space , which is spanned by single - particle excitations . a four - point function such as @xmath179 then becomes @xmath180 where the @xmath32 s can be any complete basis set . ( [ eqn : app : bse : l ] ) in the transition space becomes @xmath181^{-1}_{(ij)(mn)}(f_m - f_n),\ ] ] where the excitonic hamiltonian matrix is @xmath182 we make the adiabatic approximation for @xmath183 and arrive at the following eigenvalue problem : @xmath184 and @xmath185 can be expressed in terms of these eigenvectors by @xmath186 only the transitions between the valence and conduction bands contribute . within our two - band model , the excitonic hamiltonian has the following block matrix form : @xmath187 ignoring the off - diagonal part in eq . ( [ eqn : app : bse : hexc ] ) is equivalent to the tda . as shown in sec . [ sect : results ] , it is possible that instabilities show up in the full bse results when the underlying ground - state calculation is not exact . such instabilities in the minimal model are an artifact originating from the assumption that the solution of the kp model constitutes the ground state of the many - body system . however , this is not a matter of great concern in practice . in principle , the transition space spans all possible combinations of valence and conduction orbitals , including nonvertical transitions connecting different bloch wavevectors . the kernel @xmath188 of the bse in momentum space , eq . ( [ bsekernel ] ) , has the following ingredients : @xmath189 and @xmath190 only the excitations with the same momentum transfer @xmath191 are coupled due to the @xmath192 functions in eq . ( [ eqn : app : bse : v ] ) and ( [ eqn : app : bse : w ] ) , so we only need to include vertical transitions in the calculations for optical properties .

the study of the formation and evolution of protoplanetary disks is an important topic for our understanding of the formation of extrasolar planets and the planets in our solar system . the gas content of the disk that represents most of its mass and has an imprint of earlier evolutionary history is best studied in molecular lines . since 1997 , a number of rotational transitions from simple molecules like co , cs , cn , hcn , hco@xmath1 , n@xmath2h@xmath1 , h@xmath2co , c@xmath2h , and some of their isotopes have been firmly detected in several nearby protoplanetary disks with single - dish telescopes and millimeter interferometers ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? the wealth of information about disk physical structure and chemical composition can only be acquired with multi - molecule , multi - line interferometric observations at ( sub- ) millimeter wavelengths and at sub - arcsecond resolution ( e.g. , * ? ? ? * ; * ? ? ? in 2005 , we started a comprehensive observational program at the iram 30-m antenna and the plateau de bure interferometer , aimed at studying protoplanetary disks with several key molecular tracers of disk temperature , density , kinematics , and ionization degree ( chemistry in disks ( cid ) project ; for the first results see * ? ? ? * ) . however , in order to extract all the relevant information from molecular lines and thermal dust continuum one has to cope with the inverse ( line ) radiative transfer ( lrt ) problem , which can hardly be solved for complex geometrical configurations . instead , one solves the direct lrt problem : models with varied parameters are compared to the observational data until the best - fit is found by using some iterative minimization technique ( e.g. * ? ? ? * ; * ? ? ? naturally , this task requires enormous amounts of computing power that severely restricts the choice of applicable computational methods . this implies that often fast but only approximate algorithms can be used . thus it is crucial to investigate the applicability of such approximate methods for the lrt modeling of protoplanetary disks . various one- and multi - dimensional exact and approximate lrt methods have been developed for modeling of stellar atmospheres , stellar winds , interstellar clouds , and star - forming regions ( for a good overview see , e.g. , * ? ? ? * ) . among the simplest lrt concepts are the lte ( local thermodynamical equilibrium ) and lvg ( large velocity gradients ) approaches in which the lrt problem is treated locally ( see * ? ? ? * ; * ? ? ? the lvg method is just one among the sub - class of `` escape probability '' approaches that allow the straightforward calculation of the escape probability . these simplified lrt methods have been applied to constrain the basic physical parameters of dense molecular cores using single - dish molecular line data ( see , e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? later more sophisticated , non - local lrt approaches have been developed , starting from the 1d methods ( e.g. , * ? ? ? * ; * ? ? ? * ) to the multi - dimensional ( accelerated ) monte carlo , local linearization , and multi - zone escape probability methods , e.g. @xcite ( for a recent review see * ? ? ? being more accurate , modern lrt methods are routinely applied to study the kinematics and chemical structure of prestellar cores in detail ( e.g. * ? ? ? * ; * ? ? ? * ) , but their use for protoplanetary disks is still the exception ( e.g. * ? ? ? * ; * ? ? ? the global appearance of the physical structure of protoplanetary disks is commonly treated in the framework of steady - state passive or active flaring disk models with or without an inner rim , which seem to be successful in explaining many observational facts ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? the modeled structure can further be used to unravel the chemical evolution of the disk with a sophisticated gas - grain chemical model including surface processes , and even dynamical mixing and advective motions ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? finally , the disk physical structure and chemical composition together with appropriate collision rates provide the necessary inputs for the lrt simulations . since the iterative procedure to search for the best fit to the observed spectra can easily become computationally prohibitive when the full lrt has to be considered , it is of great importance to identify those fast approximate lrt methods that give sufficiently accurate results ( with respect to the observational limitations ) and to investigate the limits of their applicability . the analysis is further complicated when a realistic disk chemical structure is taken into account . to the best of our knowledge , a detailed benchmarking of various approximate lrt approaches for protoplanetary disks has not yet been attempted in the literature apart from the work by @xcite and therefore constitutes a major goal of the present paper . in this paper , we compare a full 2d non - lte monte carlo code with several approximate line radiative transfer methods by considering a few representative radiative transfer cases . these problems are thought to cover all possible situations encountered in disks , namely , the excitation of optically thin / thick emission lines in low / high density regions . consequently , we consider low and high transitions of key molecules in disks tracing temperature ( co and isotopologues ) , column density ( @xmath4co or c@xmath0o ) , density ( cs and h@xmath2co ) , deuterium fractionation ( dco@xmath1 ) , ionization degree ( hco@xmath1 ) , and photochemistry ( hcn ) . for this exploratory study we adopt an irradiated flaring model of a young , low - mass t tauri protoplanetary disk , and use both uniform molecular abundances and those from an advanced chemical model ( section [ model ] ) . our 2d monte carlo algorithm is based on the 1d code `` ratran '' @xcite with two acceleration modifications for disks ( see section [ method ] ) . the adopted approximate methods include the classical lte approach , the full escape probability ( fep ) and vertical escape probability ( vep ) methods , lvg , and a non - local 1d - method vor ( vertical one ray method ) . the approximate methods are all described in section [ approxi ] . using the input disk model and the above lrt methods , in section [ results ] we calculate the corresponding excitation temperatures , synthetic spectra , and channel maps and compare these data to the results obtained with the monte carlo code . the discussion and final conclusions follow . we consider a typical protoplanetary disk surrounding a t tauri - like star in the nearby taurus - auriga molecular complex at 140 pc . a flaring steady - state model representative of a class ii disk with vertical temperature gradient is adopted @xcite . the disk has inner and outer radii of 0.037 and 800 au , respectively , a mass of @xmath5 , and a mass accretion rate of @xmath6yr@xmath7 . the central star has a radius of @xmath8 , a mass of @xmath9 , an effective temperature of 4000 k , and an age of about 5 myr ( see table 2 in * ? ? ? we assume that the disk has a constant microturbulent velocity of 0.1 kms@xmath7 and a purely keplerian rotation profile , @xmath10 @xcite . the disk structure is shown in fig . [ struct ] and the parameters are summarized in table [ disk_param ] . in the present study we focus on the molecules that have been detected in protoplanetary disks and are readily used as probes of the disk physical and chemical structure , high - energy radiation , and deuterium fractionation : co , @xmath4co / c@xmath0o , hco@xmath1 , dco@xmath1 , hcn , cs , and h@xmath2co . corresponding to the different excitation conditions , the various rotational lines of these molecules are thought to be generated in distinct disk regions ranging from the cold midplane and warm molecular layer to the hot disk surface layer . normally one has to use a sophisticated gas - grain chemical network with surface reactions in order to obtain realistic abundance distributions of these observable species in the disk ( see * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? for many molecules their highest concentrations are reached in the disk intermediate layer ( at @xmath11 pressure scale heights ) that is sufficiently opaque to destructing high - energy radiation and yet too warm for effective freeze - out of gas - phase molecules ( e.g. * ? ? ? * ; * ? ? ? as a result , in these models , from a qualitative point of view the abundance distributions of many molecules can be represented by a single layer with particular concentration , thickness , and location in the disk ( see , however , * ? ? ? * ; * ? ? ? * ) . however , for the first simple analysis of the observational data it is often sufficient to assume that molecular abundances are constant ( with respect to hydrogen ) . moreover , the use of uniform molecular abundances can help to disentangle most important line radiative transfer effects from strong chemical gradients in the disks . we adopt both approaches . the values of the uniform molecular concentrations and parameters of the molecular layers are obtained by using the gas - grain chemical model with surface reactions of @xcite . in essence , this model is based on the umist95 database of gas - phase reactions @xcite , a set of surface reactions from @xcite and @xcite on @xmath12 m silicate grains , and includes a small deuterium network from @xcite . the location of each molecular layer in the disk is constrained by the upper and lower boundaries that can be approximated by a power law : @xmath13 ( where @xmath14 is the radius in au and @xmath15 is the pressure scale height as defined in * ? ? ? we assume that the molecular abundances are constant across a molecular layer and utilize the maximum values in the uniform disk model ( fig . [ struct ] ) . note that the cs abundances peak toward the inner disk region and can not be accurately represented by such a simple one - zone model . the parameters of the uniform and layered chemical structures , i.e. parameters @xmath16 and @xmath17 describing the upper and lower boundaries as defined above , are presented in table [ chem_param ] . finally , the collision rates for the line radiative transfer modeling are taken from the publicly available `` leiden atomic and molecular database , '' see @xcite . in this section we describe our implementation of the accelerated monte carlo technique in the 2d non - lte line radiative transfer code art , which is part of the software package `` uran(ia ) '' aimed at modeling of various objects with different exact / approximate lrt methods . the general idea of all monte carlo radiative transfer methods is to follow the propagation of randomly ejected photons or photon packages through the medium . these codes are based on a straightforward formalism that is easy to implement numerically and are capable of modeling radiative transfer in various complex multidimensional objects with arbitrarily distributed sources of radiation ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? . however , in order to lower random noise and thus retain reliable accuracy of the results , monte carlo simulations typically require enormous amounts of computing time , especially for highly optically thick objects . in certain cases the numerically most demanding part of monte carlo codes ( ray tracing ) can be significantly optimized by special iterative schemes for the acceleration of convergence . the algorithm of our 2d non - lte line radiative transfer code art is fully described in @xcite and only briefly summarized here . in essence , this monte carlo code is based on the accelerated scheme originally proposed and implemented in the lrt code `` ratran '' by @xcite and in the modified monte carlo method of @xcite . the main equations of the line radiative transfer can be found elsewhere ( e.g. , * ? ? ? * ; * ? ? ? * ) and are not repeated here . first , the computational domain is split into cells . all physical quantities are assumed to be constant within each disk cell @xmath18 , except for the velocity that can vary . in the first iteration , initial level populations and a set of photon random directions @xmath19 through the grid are defined ( this is the monte carlo part of the calculations ) . the same set of photon random directions is used in all subsequent iterations for accurate ray tracing , producing a solution free of random noise but with possible undersampling of directions and velocities . using these values , the line intensities @xmath20 are computed for each disk cell @xmath18 by explicit direct integration of the radiative transfer equation with the long characteristic method along the predefined set of photon rays @xmath19 , starting with the cmb field at the edge of the disk ( or interstellar radiation field , or anisotropic stellar radiation ) . the integration path is split into sub - intervals with a step that is smaller than the cell size . ideally , one would need to define this integration step based on local gradients of the physical conditions in the disk and the optical depth of the line . however , this is not easily achievable in practice , so an empirical value has to be used . we adopt a step size of 1/10 of the cell size in our computations . next , the resulting mean line intensity @xmath21 is calculated in each disk cell by spatial averaging of the derived intensities @xmath20 over directions @xmath19 and by averaging over the line profile . the computed mean intensities are utilized in the next iteration step to refine the level populations by solving the balance equations in all disk cells . the convergence of the integration procedure for optically thick lines can be substantially accelerated by additional internal sub - iterations similar to the accelerated lambda iterations ( ali ) scheme @xcite . the ali method employs the fact that the calculated mean line intensity at each particular grid cell can be separated into an internal component generated in the cell itself and an external contribution coming from other grid cells . as a result , additional sub - iterations are applied to bring into an agreement the internal mean line intensity and the corresponding level populations in every grid cell . when these local sub - iterations converge , the global iterations over the entire grid are continued until the level populations in each cell do not differ by more than @xmath22% . note that this scheme may not be fully appropriate in cases of particularly thick lines with @xmath23 when such a simple convergence criterion fails , e.g. for water lines @xcite . after the final molecular level populations are obtained , the maximum random error associated with the monte carlo computations is estimated . for that all iterations are repeated again with another set of preselected random photon rays . we choose the number of random photon paths such that the relative error in the computed level populations is smaller than 1% . although the code is currently restricted to axially symmetric ( 2d ) radiative transfer problems in polar coordinates , it can be extended to problems with any number of dimensions and for any coordinate system . finally , these resulting level populations are used in the 3d ray tracing part of the `` uran(ia ) '' package to produce synthetic molecular line spectra , channel maps , integrated line intensity maps , velocity maps etc . , with an arbitrary number of velocity channels , and for any disk viewing geometry . beam convolution can be applied if needed . the `` uran(ia ) '' code was carefully benchmarked against other lrt codes by solving several one- and two - dimensional problems . in the one - dimensional regime , the computed molecular level populations and excitation temperatures are fully consistent with all tests proposed during the conference on molecular radiative transfer ( leiden university , 1999 , see * ? ? ? several additional tests of our code were performed in 2d mode , using a few simple problems with analytically predictable solutions . in addition , together with m. hogerheijde the developer of the two - dimensional lrt code `` ratran '' we successfully compared the results for both codes for keplerian disks . the method of long characteristics is one of the most reliable algorithms to calculate the mean intensity and thus to obtain an accurate solution to the radiative transfer problem ( e.g. , * ? ? ? however , a direct realization of this method is computationally expensive for multi - dimensional radiative transfer simulations . for example , to compute molecular level populations for a 2d disk model having 150 vertical and 50 radial grid points and 1000 photon packages per cell , one needs @xmath24 days of computational time on a modern pc ( pentium xeon 2.4 ghz , 1 gb ram ) . however , for certain cases the computational time needed for the ray tracing part can be efficiently reduced . to illustrate this for a rotating keplerian disk with radial velocity gradient , let us consider a cell at a given radius @xmath14 . the local line width that strongly depends on the temperature and microturbulent velocity in the cell is typically @xmath25 kms@xmath7 , which is significantly smaller than the total velocity dispersion across the disk , @xmath26 kms@xmath7 . therefore , due to the doppler shift the photons emitted at a peak frequency in a cell rotating with a given velocity can only be absorbed locally by some of the neighboring cells having similar velocities ( the so - called `` interaction area '' ) . an example of such an interaction area is shown in fig . [ interac ] , where regions of the keplerian disk that are radiatively coupled to the cell located at @xmath27 au ( white spot in the figure ) are shown in black . we modified the ray - tracing algorithm used in the rt solver of `` uran(ia ) '' such that the iterative integration of the rt equation is only performed over the radiatively coupled regions of the disk . during the first iteration the radiatively coupled regions for each cell are identified using the full ray tracing and stored in a file . this information is utilized during the next iterations to integrate the rt equation over the radiatively coupled regions only . note that the radiatively coupled region will be different for each line of each of the considered molecules . however , for those molecules for which the local lines widths are dominated by non - thermal motions ( e.g. , microturbulence ) this difference will be small . this modification preserves the flexibility of the accelerated monte carlo approach and allows to decrease the computation time for protoplanetary disks by about one order of magnitude and more , retaining the same high level of accuracy . one may use an analytical prescription instead of the preliminary ray tracing , but such a prescription will be model - dependent and thus can not be easily generalized . though our concept of the interaction areas in rotating configurations somewhat resembles the idea of the approximate lvg method ( see sect . [ lvg ] ) , it leads to an _ exact _ solution of the lrt problem . the densities in the disk interior are so high that the low level populations of many molecules are thermalized . naturally , the thermalized disk cells can be effectively excluded from the global iteration calculations since the radiative effects do not play a role in populating the levels . we use the following approach to determine whether a disk cell is thermalized or not . first , the level populations in the disk are calculated using the simple lte and full escape probability ( fep ) approaches ( see next section ) . these two approximate methods represent two extreme cases to the solution of the lrt problem . the lte approach tends to overestimate while fep method tends to underestimate the populations of high levels in the case of radiative coupling . the cell is supposed to be thermalized if the relative difference between the corresponding lte and fep level populations is smaller than an adopted threshold of 10@xmath2810@xmath29 . this modification leads to an additional computational acceleration of about 10% for hco@xmath1 and 30% for co and co isotopes . in this section , we summarize the approximate lrt methods that will be compared with the results of the monte carlo simulations from the art code . a schematic overview and comparison of the utilized lrt approaches is given in table [ lrt_methods_table ] . in the case of local thermodynamical equilibrium ( lte ) , the molecular level populations are thermalized . they follow a boltzmann distribution with kinetic gas temperature @xmath30 : @xmath31 where @xmath32 and @xmath33 are the statistical weights and @xmath34 is the frequency of the @xmath35 transition . in terms of radiative transfer , thermalization means that the excitation and kinetic temperatures are equal for a given transition : @xmath36^{-1}.\ ] ] when the excitation temperature is higher or lower than the kinetic temperature , the level populations are super- or subthermally excited . there are two general situations when lte holds : 1 ) the hydrogen density is substantially higher than the critical density of a transition , or 2 ) the optical depth of a transition is so high that a perfect equilibrium between the collisional and radiative ( de-)excitations is reached . the lte approach is appropriate for the lower rotational transitions of co and its isotopologues ( see @xcite ) . it has also been used in the analysis of molecular line data even if the conditions for its applicability are not fully justified because it is the easiest lrt method to implement ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? another extreme is to assume that a transition has a very low optical depth , so that the newly generated photons escape the disk freely , without any further interaction with the medium . or , by other words , in the fep approach the escape probability @xmath37 is not calculated and always set to 1 ( see also * ? ? ? * ) . under such conditions the lrt problem becomes local and can be easily solved . the full escape probability ( fep ) level populations are obtained as a solution of the statistical equilibrium equations with the mean intensity @xmath38 substituted by the mean external radiation field , e.g. cmb : @xmath39 . as in the lte case , the fep approach requires only very little computational time . fep can be an adequate method for molecules with very low abundances . in some objects such as expanding stellar envelopes or galaxies , radial velocity gradients can be much larger than the local thermal and microturbulent velocities ( e.g. , * ? ? ? * ; * ? ? ? in this case the photons that are emitted by a certain region can only be absorbed locally . the size of this region is determined by the local velocity field and the line width . if one assumes that the physical conditions in this local environment are homogeneous , the radiative transfer problem can be substantially simplified ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? the keplerian rotation of a circumstellar disk causes strong velocity variations of @xmath405 kms@xmath7 , which are larger than the local thermal ( @xmath41 kms@xmath7 ) and microturbulent ( @xmath41 kms@xmath7 ) velocities . as shown in fig . [ interac ] , the size of the radiatively coupled zone is significantly smaller than the entire disk plane . therefore , one may expect the lvg approach to be efficient for the line radiative transfer modeling . in the lvg method , the mean intensity @xmath38 in a disk cell is approximated by a combination of the source function @xmath42 for a given transition and the external radiation field @xmath43 : @xmath44 where @xmath37 is the probability for the photons to freely escape the medium , @xmath45 . the escape probability @xmath37 is usually expressed in the form : @xmath46 where @xmath47 is the effective line optical depth because of the local velocity gradient . note that for optically thin lines the escape probability @xmath37 reaches unity and the lvg method turns into the fep approach . the effective optical depth @xmath47 is evaluated as : @xmath48 where @xmath49 is the local absorption coefficient and @xmath50 is the coherence length i.e. the mean distance for a photon to escape the medium . the coherence length @xmath50 in the equatorial plane of the keplerian disk can be roughly estimated as follows ( see fig . [ scheme_lvg ] ) . the photons emitted in a disk cell at a radius @xmath51 can not escape in the radial direction @xmath52 since there is no velocity shift in this direction . instead , we assume that the photons can only escape in the tangential direction . ideally , one would need to apply spatial averaging to get a correct value of the coherence length , but this would significantly complicate the realization of the lvg method . the approximate coherence length can be expressed via the local line width and the keplerian velocity . the velocity shift along the coherence path is , by definition , equal to the local line width @xmath53 : @xmath54 where @xmath55 and @xmath56 are the keplerian velocities at the distances of @xmath51 and @xmath57 , respectively , and @xmath58 is the corresponding angle between the velocity vectors ( fig . [ scheme_lvg ] ) . expanding the keplerian velocity into taylor series , we find that @xmath59 , while @xmath60 . this allows to derive the value of @xmath61 from eq . ( [ dv ] ) : @xmath62 note that @xmath61 is the width of the ring - like radiatively coupled region depicted in fig . [ interac ] . from fig . [ scheme_lvg ] we also find that the coherence length @xmath50 can be related to the radial shift @xmath61 and radius @xmath51 as @xmath63 , and thus @xmath64 for a constant line width @xmath53 . this approach is adopted in our lvg method . it typically takes only several seconds of cpu time to calculate the molecular level populations with the lvg approach . one has to keep in mind that in our disk models there is no velocity gradient in the vertical direction and the disk cells are radiatively coupled in vertical direction , while in the lvg approach , the disk is effectively decoupled in independent parallel planes . let us compare the coherence length @xmath50 with the scale height of the disk @xmath65 . for the geometrically thin , isothermal disk @xmath66 ( for our definition of @xmath65 , see * ? ? ? * ) , where @xmath67 is the sound speed and @xmath68 is the keplerian angular velocity at radius @xmath51 . for grey dust opacities the radial temperature profile is @xmath69 @xcite and thus @xmath70 , similar to the scaling of the coherence length . for our disk model @xmath71 au at @xmath51=100 au , while @xmath72 au . therefore , the radiative effects in vertical and radial directions are comparable over the entire keplerian disk ( at least around the disk mid - plane ) and , by design , the lvg method can only partly account for the radiative coupling . to better illustrate the role of radiative coupling in vertical and radial directions , we computed with art the ratio between tangential and vertical optical depths for the uniform disk model ( see fig . [ tau_ratio ] ) . the typical value of this ratio is only about a few over the entire disk , so both directions are important for the line radiative transfer . using art , we also calculated the mean coherence lengths in the equatorial plane of the disk . this is done by spatial averaging of individual coherence lengths . the calculated mean and individual coherence lengths are compared to the values from eq . ( [ coherent ] ) in fig . [ av_coh_length ] . as can be clearly seen , our analytical expression reproduces well the radial dependence of the mean coherence length , but a number of the individual lengths have larger values because of the photon trapping along directions with zero doppler shifts . as a result , the mean value of the coherence length is typically about 10 times longer than predicted by eq . ( [ coherent ] ) . however , the excitation conditions vary so greatly over the mean coherence length that it can not be directly put into eqs . ( [ eqlvg1 ] ) and ( [ eqlvg2 ] ) . in order to properly evaluate the total escape probability @xmath37 one has to average individual escape probabilities rather than the coherence lengths . this requires information about exact physical parameters along all coupled directions , which makes this approach non - local and thus far too complicated . in the lvg approximation the coherence length is determined by the velocity gradients resulting from the keplerian rotation of the disk . however , as we have shown above the radial radiative coupling is comparable to the coupling in vertical direction . one can assume that the vertical radiative coupling is more important than the radial coupling , which leads to significant simplification of the radiative transfer . this idea is utilized in the vep method where photons are assumed to escape only in the vertical direction perpendicular to the disk plane . a simple correction is made to take into account the large asymmetry between the two opposite escape directions : the escape probability is computed by taking the sum of escape probability upwards and downwards : @xmath73 where @xmath74 is the optical depth towards the disk surface ( upward opacity ) , and @xmath75 is the optical depth through the disk midplane ( downward opacity ) . the optical depth in upward and downward directions is calculated using the local excitation conditions and the molecular surface densities along the corresponding directions . note that this assumption is rather coarse in the presence of strong vertical gradients of physical conditions . similar to lvg , in the vep approach the escape probability , source function , and the local optical depth are obtained iteratively . however , the direct implementation of the iterative vep approach can be dangerous for disk cells with negative excitation temperature . in vep these local conditions are extrapolated over the total vertical extent of the disk and the resulting optical depth can be negative and large . this leads to the physically unrealistic situation that the escape probability can become large , especially during the iterations required to solve the statistical equilibrium equations and the vep method may fail to converge . to avoid this problem , we control the downward line optical depth @xmath75 such that it can not become negative . the upward optical depth is allowed to remain negative , in order to represent weak inversions at the surface layers . the method was initially developed in the diskfit software package ( see section 4.6 ) . like fep , the vep method takes only a few seconds of cpu time to solve the lrt problem ( see table [ lrt_methods_table ] ) . in the lte and fep methods radiative coupling effects are completely ignored , while in the lvg and vep approach radiative trapping is only partially treated . since there is no vertical velocity gradient in the adopted disk model , the cells are generally radiatively coupled in vertical direction . furthermore , fig.[interac ] shows that , with the exception of a relatively narrow cone in the radial direction , a given cell in the disk is only coupled with cells at the same radius , i.e. with regions with the same 1d structure . using these ideas , we construct a non - local approximate lrt method in which the radiative coupling in only the vertical direction is taken into account ( `` vor '' vertical one ray method ) . this simplifying assumption allows to use a 1d approach for calculating the mean line intensity @xmath38 in a disk cell : @xmath76 where @xmath77 and @xmath78 are the intensities of the radiation coming along vertical direction from the top and bottom of the disk , respectively . to further accelerate this method , only one central frequency for each emission line is taken into account . this non - local rt problem is finally solved using only 2 photon paths along vertical direction for the ray tracing . computational demands of the 1d vor method are much less than those of the exact 2d code art . it usually takes about a few minutes of cpu time to obtain the level populations . note that our vor approach is another implementation of the feautrier method ( see * ? ? ? * ) and resembles the 1d scheme of @xcite . the 1d methods have been used to represent spheres as well as infinite planes . in the latter case , an eddington factor of 3 is used to take into account the mean opacity over all possible angles . in our case , since the interaction only occurs in a ring centered around the star , the eddington factor should be smaller . we have simply used the eddington factor of 1 , like for a sphere . the described approximate methods are included in the `` uran(ia ) '' package . in our study we also used another software package `` diskfit '' . the program diskfit allows to model line radiative transfer in disks using various approximate methods as well as to reconstruct basic disk parameters using an advanced @xmath79-minimization scheme in the @xmath80-plane @xcite . the program offers three possibilities to solve the coupled equations of statistical equilibrium and rt : lte , vep , and a non - local 1d approach . the latter is based on the feautrier scheme @xcite and is an adaptation to the disk geometry of the code developed by @xcite for the line radiative transfer modeling of interstellar clouds . this 1d method is thus similar to vor , but uses several frequency points across the line profile . the eddington factor may also be adjusted if needed . once the level populations are derived , the interferometric visibilities as well as beam - convolved channel maps can be computed using the 3d ray tracing part @xcite . any arbitrary orientations of the disk defined by the inclination and position angles can be used . for that , a three dimensional regular grid is used , with additional refinement for the central disk region where the surface brightness gradient is very strong . the diskfit radiative transfer methods and the ray tracing part were thoroughly compared against those implemented in the `` uran(ia ) '' package and good overall agreement was achieved for the disk level populations and spectral maps for both the optically thin and thick molecular transitions . therefore , for the rest of the paper we compare approximate and exact lrt methods from the `` uran(ia ) '' package only . prior to the detailed comparison of various radiative transfer methods described above , we discuss the problem of molecular line formation and excitation conditions in disks in general . since the thermal and density structure of protoplanetary disks is in relatively well understood ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , we can discriminate distinct regimes of line formation . a particular rotational transition is excited when the molecular hydrogen density exceeds a critical density @xmath81 : @xmath82 here @xmath83 is the einstein coefficient and @xmath84 are the collisional rates from the upper level @xmath85 to all other lower levels @xmath18 . the density , @xmath86 , at which the molecular levels become predominantly populated by collisions and thus are thermalized is typically about one order of magnitude higher than the critical density . we will further refer to @xmath86 as to the thermalization density . consequently , the disk can be roughly divided in two regions . in the so - called `` super - critical '' region near the disk midplane , where the density is larger than the thermalization density , molecular level populations follow the boltzmann distribution and the excitation temperature is equal to the kinetic temperature , @xmath87 . in this region a simple lte approach can be safely utilized . in the upper , more diffuse `` sub - critical '' region of the disk the line excitation is partly controlled by collisions and partly by internal / external radiation . the crucial parameter that determines the excitation conditions in this region is the molecular column density . if this column density is so high that the sub - critical region is opaque to the internal radiation , then the collisional and radiative ( de-)excitations are in perfect balance and the corresponding molecular levels are thermalized , @xmath87 . in contrast , when the column density is so low that the sub - critical region is fully transparent to the internal radiation , the corresponding molecular levels are mostly radiatively populated and therefore subthermally excited : @xmath88 is the temperature of the background radiation , e.g. 2.73 k for the cmb field ] . in this case a simple full escape probability approximation can be used ( see section [ results ] ) . the characteristics of the emergent molecular line spectrum will depend on the relative contribution from each of the super - critical and sub - critical regions . this simple scheme becomes more complicated when the effects of chemical stratification are taken into account . in this situation three distinct regimes of line excitation can be distinguished . in the first , and most simple case the molecular layer is located deeply inside the super - critical density region and the molecular level populations are in lte ( see left panel in fig . [ scheme_lrt ] ) . good examples of that kind are the lower rotational transitions of co and its isotopologues @xcite . in the second case , the molecular layer is entirely confined within the sub - critical density region . the distribution of the level populations will then depend on the molecular column density as discussed above ( middle panel , fig . [ scheme_lrt ] ) . when this column density is low and radiative coupling is not important , the fep approximation works . in the opposite case the radiative transfer problem has to be tackled with a more sophisticated non - local method . the line excitation proceeds in this way in gas - deficient transition disks @xcite . finally , in the most general case the molecular layer is located between the sub - critical and super - critical disk regions ( right panel in fig . [ scheme_lrt ] ) . in this situation the lte approach does not work , whereas the fep approximation is valid only if the optical depth of the sub - critical layer is low and the molecular emission is dominated by the super - critical ( thermalized ) layer . the lower rotational lines of e.g. hco@xmath1 , hcn , and cs are likely excited in such a way . if the molecular column densities are so large that the sub - critical region becomes optically thick for internal radiation , the simple fep approach fails and non - local radiative transfer modeling is required . the non - lte effects in disks are likely important for the rotational lines of complex species , e.g. h@xmath2co , but may also be important for higher transitions of many other simple molecules ( such as hcn , cs , etc . ) . in the next section we verify whether non - lte effects are important in protoplanetary disks and for which molecular transitions this is the case . there we use a simple preliminary analysis of the excitation conditions . it is often useful to understand under what conditions a given transition is excited before any lrt modeling is performed . one example is the iterative @xmath79-minimization analysis of interferometric data where an approximate lrt method has to be utilized due to limited computational resources ( e.g. , * ? ? ? * ; * ? ? ? usually the basic properties of the physical structure and sometimes the chemical structure of the disks are at least partly known . when the chemical structure is not known , one can use uniform abundance distributions with the abundance values taken from theoretical models . using the density and chemical structure of the disk and molecular data as input , one can determine the line optical depth and the amounts of emitting molecules in the sub- and super - critical regions . as discussed in section [ vor ] , radiative coupling is strong in vertical direction due to the absence of velocity gradients , and thus radiative effects on the line excitation are most profound for face - on oriented disks . plotted as a function of radial distance , the molecular column densities in the sub- and super - critical disk regions as well as the molecular column density needed to reach an optical depth of unity may serve as an indicator whether and where the considered transition is non - lte excited and which approximate or exact method is appropriate to model the radiative transfer in this case ( see previous section ) . in fig . [ analysis_lrt ] we present such critical column density ( ccd ) diagrams for some transitions of the molecules under investigation : co , cs , hcn , h@xmath2co , and hco@xmath1 . we focus on low rotational transitions and adopt approximate values of their thermalization densities , viz . , @xmath89 cm@xmath28 for the co lines and @xmath91 cm@xmath28 for all other molecules . the column densities to attain @xmath92 are computed for the central velocity channel of the ( 1 - 0 ) transition ( @xmath93 for para - h@xmath2co ) assuming a constant excitation temperature of 10 k. not surprisingly , the column density of co is so high , @xmath94cm@xmath29 , that the lines are very optically thick , @xmath9610@xmath97 . since the co lines are easily excited already at densities of about 10@xmath97cm@xmath28 , essentially all co molecules are located within the super - critical region ( case `` a ) '' in fig . [ scheme_lrt ] ) . molecules tracing the disk interior are excited in similar manner , e.g. h@xmath2d@xmath1 , n@xmath2h@xmath1 , dco@xmath1 @xcite . note that in our model , the co abundances are high ( see table 2 ) . when comparing with real disks , where co can be significantly depleted , our test case for c@xmath0o can be applied to @xmath4co observations , while for @xmath98co , the lte approximation will work even better ( since @xmath92 will be reached at higher densities ) . for other , less abundant species that have higher critical densities and are concentrated in the disk intermediate layer , the ccd diagrams show a bimodal behavior . in the inner , dense disk region at @xmath99600au almost all molecules are located in the super - critical region , and their levels are lte - populated ( similar to co ) . in the outer , less dense disk region the molecules are mainly concentrated in the sub - critical layer ( case `` b ) '' in fig . [ scheme_lrt ] ) . the optical thickness of the sub - critical region for the disk model with chemical stratification is lower or close to one . consequently , the molecular sub - critical layers are expected to be optically thin or moderately optically thick in the outer disk region and the fep approximation holds ( see hcn and hco@xmath1 in fig . [ analysis_lrt ] ) . thus , non - lte effects should not play a major role for low rotational lines of most molecules in protoplanetary disks . a notable exception is water that has a complex level structure with a number of maser transitions and some highly optically thick lines @xcite . in contrast to the layered chemical structure , for the uniform model the amount of molecules is so high in the sub - critical region that this region is optically thick and radiative trapping becomes efficient . as a representative example , we present ccd diagrams for the uniform hco@xmath1 model ( lower right panel in fig . [ analysis_lrt ] ) . in what follows , we will focus on this uniform hco@xmath1 model to analyze in detail the importance of non - lte excitation for both low and high transitions as modeled by various lrt methods . in fig . [ texc ] we show the spatial distributions of the excitation temperature in the uniform disk model as computed by the different lrt methods for the hco@xmath1 ( 1 - 0 ) , ( 4 - 3 ) , and ( 7 - 6 ) transitions . by definition , the lte excitation temperature for any transition is equal to the gas kinetic temperature . the vertical temperature gradient and cold isothermal midplane are clearly visible ( fig . [ texc ] , top row ) . in contrast , the excitation temperature distributions calculated with other lrt methods show a departure from lte conditions , in particular for high transitions and for the upper , less dense disk regions . the most extreme deviation from lte is the negative excitation temperature caused by the inversion in the populations of the first two rotational levels of hco@xmath1 ( maser effect ) , which is reached in the warm upper disk region , at @xmath100 au ( white zone in fig . [ texc ] , left panels ) . in this part the ( self-)radiation is not intense , and the levels are excited and de - excited by collisions but only radiatively de - excited . consequently , lte is not longer valid and collisional pumping of the rotational level populations appears as a result of a specific ratio between radiative and collisional ( de)-excitation probabilities . such inversion is characteristic of some linear molecules like cs , hco@xmath1 , and sio and requires high temperatures and a specific density range @xcite . however , this inversion does not result in maser amplification of the hco@xmath1(1 - 0 ) line intensity even for highly inclined disks . as soon as the line opacity and thus the self - radiation becomes significant , the stimulated radiative transitions depopulate the upper levels and destroy the inversion . note that in the fep approximation the effect of self - radiation is completely neglected and therefore the negative excitation temperature zone is strongly overestimated ( fig . [ texc ] , left bottom panel ) . another non - lte zone corresponds to the subthermally excited sub - critical disk region , which occupies a significant disk volume for high hco@xmath1 transitions . to quantify the difference between the excitation temperature @xmath101 computed with an approximate and the exact art methods , we introduce the local error @xmath102 in each disk cell : @xmath103 a relative difference of @xmath104 in the excitation temperature corresponds to @xmath105 , and a @xmath106 difference leads to @xmath107 . consequently , the global accuracy @xmath108 of an approximate method over the entire disk can be evaluated with the following expression : @xmath109 where the absolute values of the local errors @xmath110 are summed over the optically thin disk cells and weighted by the number of molecules @xmath111 in each cell . @xmath112 is a total number of molecules in the optically thin region . the optical depth at the line central frequency , @xmath113 , is calculated in the vertical direction using the art level populations . this global accuracy criterion is designed to directly relate the difference in the excitation temperatures to the difference that will appear in the spectra . we consider disk cells with @xmath114 ( in vertical direction ) because this is the region where most of the line emission comes from ( for face - on disks ) . in essence , the errors in the @xmath115 regions are not important , since these regions are not visible . certainly , this criterion is not a unique one and has to be used with care . we state that an approximate method gives `` good '' accuracy ( `` a '' ) if the global criterion @xmath116 , `` reasonable '' accuracy ( `` b '' ) when @xmath117 , and `` bad '' accuracy ( `` c '' ) when @xmath118 . the distributions of the local error @xmath102 and global accuracy @xmath108 of the applied approximate lrt methods for the uniform hco@xmath1 model are shown in fig . [ texc_rel ] . all lrt methods are accurate in the super - critical disk region , @xmath119 cm@xmath28 . however , the thermalized disk region extends toward lower densities because the hco@xmath1 lines are highly optically thick and the perfect balance between radiative and collisional ( de-)populations is reached ( see sect . [ line_form_uni ] ) . the thermalized region is particularly large for low rotational transitions that are excited at lower densities . the lte approximation tends to significantly overestimate the excitation temperatures for the high transitions with high critical densities but is still appropriate for the @xmath120(1 - 0 ) line ( @xmath121 , fig . [ texc_rel ] ) . in contrast , fep fails for the ( 1 - 0 ) transition since it greatly overestimates the region of negative excitation temperature . it is rather accurate ( @xmath122 ) for the ( 4 - 3 ) line because it adequately describes the excitation conditions in the small , optically thin zone above @xmath92 in the outer disk region , at @xmath123 au , which occupies the largest volume in the disk . fep is also accurate for ( 7 - 6 ) transition ( @xmath124 ) , since it works in the large optically thin region that is decoupled from the internal radiation . the lvg , vep and vor methods are in general much more accurate than lte and fep because they account for the radiative trapping . in particular , the maser zone in the disk is well reproduced by lvg and vor . the lvg approach shows the same tendency for the excitation temperature deviations as fep because it underestimates the radiative coupling . the non - local vor approach gives the most accurate results for all considered transitions since it treats the radiative transfer in vertical direction , where velocity gradients are absent in our model . a minor problem of vor is the overestimation of the photon trapping in the very upper disk regions where the radiation can escape also in other directions . in general , for the uniform hco@xmath1 model and all considered transitions fep and lte are not accurate lrt methods ( `` c '' grade ) , while lvg and vep are acceptable ( `` b '' ) , and vor is good ( `` a '' ) . this conclusion is valid in general for all other key molecules apart from co isotopologues where all methods are accurate because their lines arise from the super - critical region ( see table [ applic_methods_uni ] ) . for the disk model with chemical stratification the overall accuracy of all lrt methods is better because the lines are more optically thin and radiative effects are not that important ( table [ applic_methods_chem ] ) . for many molecular transitions lte and fep are reliable approximate methods for the line radiative transfer modeling of chemically stratified disks . however , in this case both lvg and vep provide an even better approximation , at the expense of very little extra computational time . using the excitation temperatures calculated for the uniform chemical structure , we plot the corresponding synthetic hco@xmath1 spectra for the @xmath125 , ( 4 - 3 ) , and ( 7 - 6 ) transitions in fig . [ single - dish_uniform ] . all spectra are convolved with the same @xmath126 beam and are computed for two disk inclinations : @xmath127 ( face - on ) and @xmath128 ) . the lvg , vep and vor methods result in line profiles with intensities deviating by no more than 30% from art , whereas the fep and lte intensities and line profiles differ more strongly from those of art , in agreement with the difference in excitation temperature maps ( see fig . [ texc_rel ] and tables [ applic_methods_uni],[applic_methods_chem ] ) . the line intensities computed with fep and lte are either overestimated or underestimated , depending on transition . the overall disagreement between these approximate methods and art does not depend strongly on the disk orientation . for the hco@xmath1(1 - 0 ) transition the lte spectrum is close to the art line profile , but the fep intensity is 3 times higher for the face - on orientation and even higher for the inclined disk . as discussed in the previous section , this is due to the maser effect in the first two hco@xmath1 level populations . the same effect applies to the lower transitions of cs ( not shown here ) . however , for the higher hco@xmath1 transitions the fep spectra are in better agreement with the full monte carlo simulations , whereas lte tends to significantly overestimate the line intensities . this is because the size of the radiatively coupled part is decreasing for higher transitions , having higher thermalization densities and thus a more extended sub - critical region where @xmath129 ( compare top panels in fig . [ texc_rel ] ) . at face - on disk orientation a self - absorption dip arises in the ( 4 - 3 ) and ( 7 - 6 ) art spectra at the central velocity channels . this dip is caused by absorption of emitted radiation in the upper disk region , at @xmath130cm@xmath28 , due to the vertical excitation temperature gradient ( fig . [ texc ] ) . both lte and fep fail to accurately reproduce this effect . in contrast , the double - peaked hco@xmath1 spectra of the inclined disk are not due to self - absorption but appear as a result of the beam convolution over disk cells with different projected velocity . in this case all approximate methods correctly predict the shape of the line profiles . note that while integrated line intensities do not strongly depend on the disk orientation , the peak intensities do differ strongly , and are lower for higher transitions . the effects of chemical stratification on the integrated spectra are shown in fig . [ single - dish_chem ] . since hco@xmath1 occupies only a fraction of the entire disk , the computed line intensities and their optical depths are much lower compared to the uniform model . the self - absorption dip is absent in the face - on line profiles since there is no more hco@xmath1 in the upper disk region . the overall agreement between the different lrt methods becomes better ( see also table [ applic_methods_chem ] ) . still , lte fails for higher transitions . the detailed comparison for the other species considered here yields similar results . the criterion @xmath108 that gives a global accuracy estimate for the excitation temperature calculated with an approximate lrt method is also valid for the single - dish spectra . to investigate how good this criterion is for the synthetic spectral maps , we plot the hco@xmath1(4 - 3 ) map at the 0.68kms@xmath7 velocity channel for the same @xmath131-inclined disk , but without beam convolution ( fig . [ spectral_map ] ) . the model with the uniform hco@xmath1 abundances leads to a fan - like structure with two low - intensity lanes and a high - intensity interior zone ( see fig . [ spectral_map ] , upper row , central panel ) . the low - intensity lanes correspond to the cold disk midplane with low excitation temperatures . similar to the single - dish spectrum ( fig . [ single - dish_uniform ] ) , lte ( fep ) tends to overestimate ( underestimate ) the total map intensity . the lte significantly overestimates the excitation temperature in the upper disk region , leading to the formation of two high - intensity features near the vertical edges of the image . the contrast of the fep spectral image is much smaller , yet the cold midplane and high - temperature inner zone are clearly visible in the map . the lte , fep , and art hco@xmath1(4 - 3 ) maps simulated with the layered chemical structure do not differ so much as for the uniform model ( see also fig . [ single - dish_chem ] ) . however , the spectral images show a more complicated pattern that resemble two thick elliptical rings overlaid on each other with an angular shift of about @xmath132 and fixed at the center . the three low - intensity lanes correspond to the disk region close to the midplane that is devoid of hco@xmath1 emission at the chosen velocity channel of 0.68kms@xmath7 . furthermore , the hco@xmath1 ions are also absent in the disk upper layer and the overall extension of the map is smaller in vertical direction compared to the uniform model . our analysis of the line radiative transfer in protoplanetary disks supports and extends the analysis performed by @xcite . in particular , they argued that se(@xmath133=0 ) calculations ( in our notation fep ) provide better agreement with se calculations ( in our notation art ) compared to lte . according to our results , the fep method is indeed more correct than lte for the upper molecular transitions which were considered by @xcite . however , the fep approach may give completely wrong intensities for the lowest molecular transitions such as @xmath134 and @xmath135 since it overestimates the maser effect , resulting from the specific ratio between collisional ( de)excitation rates . hence , the lte approach provides more reliable intensities of the lowest transitions than the fep method . however , lte fails to reproduce important features of the line profiles like the self - absorption dips . despite the fact that either lte or fep can be sufficient for the simulations of the particular molecular lines , we recommend to use lvg or vep since they provide in general better agreement at the same cost of complexity and cpu requirement . the more exact , but more time - consuming methods such as vor or art can be used to check the validity of the above approximate methods . an interesting result of our study is the fact that the synthetic images calculated with vep are quite close to the exact art spectra , for both chemically uniform and layered models . this may look surprising since the idea of the vep method is to use local excitation conditions and to extrapolate them along the vertical direction to calculate escape probabilities . thus , one may expect that this `` homogeneous '' approach is inappropriate for the disk models with their strong gradients of physical conditions and molecular abundances . the `` good behavior '' of vep is a consequence of the very strong ( vertical ) density gradient compared to the temperature gradient . in the layered disk model molecules are concentrated in thin layers where physical conditions do not change much in vertical direction . in the uniform disk model the strong vertical density gradients result in a super - critical region around the midplane where molecular emission is thermalized , while emission from a low - density surface layer is negligible . in between lies a narrow sub - critical disk region , where molecular opacities can be high enough to affect the emergent spectra ( see sect . [ line_form ] ) . for optically thin lines , the contribution of this narrow layer to the emission is small . for optically thick lines , the likelihood that the line opacity reaches about unity in this region is very low . the exact location of the @xmath92 layer determines the brightness temperature , but is not very critical since vertical temperature gradients are not very large . recently @xcite have proposed a new approximate multi - zone escape probability ( mep ) method to model water line emission from a pdr region @xcite . in contrast to our vep approach , in mep the probability for photons to escape a cell is computed in 3d , by spatial averaging of the optical depths over several ( 6 ) directions . due to its 3d nature , mep may treat more consistently the photon escape from complicated radiatively coupled regions in keplerian disks ( see fig . [ interac ] ) than our vertical escape probability method . since the 1d vep approach already reasonably accurate in most cases , the use of 3d mep for the lrt modeling of protoplanetary disks is recommended in situations when even higher accuracy is required at the expense of extra cpu costs for the spatial averaging . another recent 1d escape probability method has been derived by @xcite . their coupled escape probability ( cep ) approach is based on mathematical transformation of the coupled system of the integro - differential rt equation and linear balance equations into one system of implicit non - linear equations . due to fast convergence of this implicit scheme , cep gains about 12 orders of magnitude in computational speed compared to conventional ali schemes . however , this method can not be easily adapted to multidimensional lrt problems with multi - level molecules because of the mathematical `` trick '' that lies behind its algorithm . taken at face value , in 1d geometry , it would be conceptually similar to , but more sophisticated than the 1d feautrier approach implemented in our vor method . in the present study we did not consider molecules with complex level structure , e.g. molecules with hyperfine components since they require more detailed investigations . for instance , @xcite used new collisional rates to simulate the n@xmath2h@xmath1 lines from prestellar cores and showed that the usual assumption that n@xmath2h@xmath1 sub - level populations follow the boltzmann distribution may lead to significant errors in the intensities of the hyperfine lines . @xcite mentioned severe problems in lrt simulations of h@xmath2o lines . apparently , the formation of such complex molecular lines in protoplanetary disks requires a more extended analysis which has to be based on the appropriate collisional rates . in the present paper , we also did not treat the effects of dust emission on the line formation . for the protoplanetary disks the dust emission at millimeter wavelengths is negligible compared to the molecular transitions . however , at shorter wavelengths the dust emission may be sufficiently strong to excite upper molecular transitions and , therefore , should be considered . nor did we include the effects of vertical mixing and turbulence in our model . from a chemical point of view , turbulent diffusion acts as an efficient non - thermal desorption mechanism and smoothes molecular abundance gradients in disks @xcite . consequently , the results for such a disk model would be in between the results for our uniform and layered chemical models . we analyzed the line radiative transfer and the formation of rotational lines of several key molecules ( co , cs , hco@xmath1 , h@xmath2co , hcn ) in protoplanetary disks . a detailed model of the disk physical structure with vertical temperature gradient and uniform / layered chemical structure is used . we used the exact accelerated monte carlo code art and compare the results with several approximate lrt approaches , including lte , the full escape probability ( fep ) , lvg , and the vertical escape probability method ( vep ) and a 1d non - local code vor . our main conclusions are as follows : 1 . prior to any modeling of the radiative transfer , one can roughly estimate the importance of non - lte effects for the line excitation and thus choose an appropriate approximate or exact method by using critical column density ( ccd ) diagrams . the basic idea of the ccd diagrams is to calculate and compare the amount of emitting molecules in the sub- and super - critical regions , taking into account the optical depth of the line . with ccd diagrams we found that the upper disk layers are subthermally excited for many molecular lines , in particular for the upper transitions . 2 . by comparing excitation temperatures , synthesized spectra , and spectral maps we demonstrate that the simple lte and fep methods work reasonably well for chemically stratified disks where molecules are located mainly in the warm intermediate layer or in the disk midplane , but are not always accurate for disk models with uniform abundances . in contrast , the lvg , vep , and especially vor methods take the radiative coupling in the disks into account and thus accurately reproduce the line intensities and profiles . since the lvg and vep methods are computationally inexpensive , their use is recommended . we found a possibility to accelerate the ray - tracing part of the rt solver in monte - carlo method for rotating disks . in this modification the ray tracing is only performed over radiatively coupled disk zones , leading to a computational speed gain by factors of 1050 . this work is only the first step in our systematic analysis of the line radiative transfer in protoplanetary disks . the excitation of molecules with complex level structure , including hyperfine line components , and the applicability of various approximate lrt methods for the @xmath79-minimization fitting of the observational data will be addressed in future studies . we thank michiel hogerheijde and jinhua he for useful and stimulating discussions . authors are thankful to the anonymous referee for valuable comments and suggestions . this research has made use of nasa s astrophysics data system . ; thick solid line ) compared to the individual ( circles ) and 3d spatially - averaged ( thin solid line ) coherence lengths ( art ) . the uniform model and the hco@xmath1(4 - 3 ) line are used.,scaledwidth=45.0% ] co , and hco@xmath1 . for each of these species , the ccd diagram shows the radial distribution of 3 column densities in the chemically stratified disk : the column density in the sub - critical layer ( solid line ) , in the super - critical layer ( dashed line ) , and the column density that is needed to reach an optical depth of unity for the ( 1 - 0 ) transition ( @xmath93 for h@xmath2co ) ( @xmath136 , thin solid line ) . the uniform hco@xmath1 model is shown in the lower right panel . note that for better presentation the column densities are restricted to values above @xmath137 cm@xmath29 . the ratio between the column densities in the sub- and super - critical layers , together with the value for @xmath92 , can be used for preliminary analysis of the line excitation conditions in disks and stimulate the choice of a proper lrt method.,scaledwidth=60.0% ] lll distance & @xmath138 & 140 pc + stellar temperature & @xmath139 & @xmath140 k + stellar radius & @xmath141 & @xmath8 + stellar mass & @xmath142 & @xmath9 + disk inner radius & @xmath143 & 0.037 au + disk outer radius & @xmath144 & 800 au + disk mass & @xmath145 & @xmath5 + surface density at 1 au & @xmath146 & @xmath147 gcm@xmath29 + density profile & @xmath148 & @xmath149 + lcccccc co & 0.7 & 0.0 & 2.0 & 0.0 & 8(-05)@xmath150 & 1(-04 ) + c@xmath0o & 0.7 & 0.0 & 2.0 & 0.0 & 2(-07 ) & 2(-07 ) + hco@xmath1 & 1.37 & -0.1 & 1.68 & -0.05 & 6(-10 ) & 1(-08 ) + dco@xmath1 & 0.0 & 0.0 & 0.4 & 0.0 & 5(-11)@xmath151 & 2(-10 ) + h@xmath2co & 1.37 & -0.1 & 1.2 & 0.0 & 3(-10 ) & 1(-08 ) + hcn & 1.75 & -0.16 & 1.92 & -0.07 & 1(-09 ) & 1(-09 ) + cs & 1.44 & -0.07 & 1.83 & -0.03 & 1(-09 ) & 1(-08 ) + lclc lte & local & super - critical & @xmath152 s + & & region & + fep & local & as above & optically & @xmath152 s + & & thin lines & + lvg & local & regions with strong & @xmath153 s + & & velocity gradients & + vep & local & regions with low & @xmath153 s + & & velocity gradients & + vor & non - local & regions with low & @xmath154 m + & & velocity gradients & + art & exact & everywhere in disks & @xmath155 h + llllll co & aab@xmath156 & aaa & aaa & aaa & abb + cs & acc & aaa & aba & aba & bcb + hco@xmath1 & acc & aaa & baa & cba & ccb + hcn & acc & aaa & aba & baa & ccb + h@xmath2co & acc & aaa & aaa & aaa & bca + c@xmath0o & aaa & aaa & aaa & aaa & aaa + dco@xmath1 & abc & aaa & aaa & aaa & aaa + llllll co & aaa@xmath156 & aaa & aaa & aaa & aaa + cs & aac & aaa & aaa & aaa & aaa + hco@xmath1 & abc & aaa & aaa & aaa & bba + hcn & acc & aaa & aaa & aaa & cba + h@xmath2co & bbc & aaa & aaa & aaa & aaa + c@xmath0o & aaa & aaa & aaa & aaa & aaa + dco@xmath1 & aaa & aaa & aaa & aaa & aaa +

the existence of nonbaryonic dark matter in modern cosmology is strongly supported by several independent signatures @xcite , including the cosmic microwave background ( cmb ) , gravitational lensing and large scale structure surveys . however , the nature of dark matter remains a mystery . to reproduce the correct relic density , it is naturally assumed that dark matter is composed of weakly interacting massive particles ( wimps ) such as the supersymmetric neutralino which is one of the most popular candidates @xcite . since the self - annihilation rate is proportional to the dark matter number density squared , potential signals from dark matter annihilation are most likely to be detected in highly dense regions such as the centers of galaxies . it is also expected that the energy released in dark matter annihilation in the early universe can influence cmb anisotropies . with the high precision cmb data from the wmap satellite @xcite , strong constraints on the properties of dark matter can be obtained @xcite . a multi - wavelength analysis of dark matter annihilations has been proposed recently in ref . @xcite . recently , several cosmic ray and gamma - ray experiments provided new windows to detect the signatures of dark matter . in particular , the pamela satellite @xcite reported an `` excess '' of the positron fraction above 10 gev . also the @xmath1 spectra above a few hundred gev from the fermi - lat @xcite and hess measurements @xcite are significantly harder than expected @xcite . these striking observations were often interpreted in terms of dark matter annihilation or decay which can contribute to the flux of high energy electrons and positrons @xcite . however , standard astrophysical sources , such as pulsars @xcite and supernova remnants @xcite also provide a possible explanation for these excesses . the high energy electrons and positrons induced by wimp annihilation produce gamma - rays through inverse compton scattering ( ics ) off the low energy background photons in the interstellar radiation field ( isrf ) and through bremsstrahlung emissions due to the interaction with the ionized interstellar medium . these components add to the prompt gamma - ray radiation accompanying pairs emitted in the annihilation event @xcite . gamma - ray dark matter signatures could be identified by the fermi large area telescope @xcite , due to its unprecedented angular and energy resolution , despite the fact that the gamma - ray spectrum is dominated by conventional astrophysical sources , such as pulsars , supernovas , blazars @xcite and perhaps structure formation shocks @xcite . many efforts @xcite have been devoted to extract indirect dark matter signatures or to set limits on a large class of wimp models by investigating the fermi - lat diffuse gamma - ray measurements . especially , a universal model - independent method can be used to place constraints on any dark matter model @xcite through convolving _ response functions _ of signal - to - background with a specific injection energy spectrum of electrons and positrons . in the wimp cold dark matter scenario , the dark halos can form at very high redshift , @xmath2 , with a minimum mass of @xmath3 @xcite determined by the free - streaming limit and collisional damping leading to a cutoff of the primordial power spectrum . this suggests that if they can survive until the present day , an enormous number of dark matter clumps ( subhalos ) are expected to be embedded in our galaxy . recent numerical simulations @xcite confirm this prediction , although the role of the tidal effects from the baryonic component has yet to be quantified and there is no direct test by observations . nevertheless , several references have discussed the resulting signals pointing out that subhalos can boost annihilation rates and produce a distinct radial distribution of emissions @xcite . alternatively , a statistical analysis of the full - sky emission map can be used to identify dark matter signatures , which can be promising if , e.g. , one can identify features in the anisotropy power spectrum that characterize uniquely the dark matter spatial distribution with respect to other astrophysical sources . in this respect , one is interested in quantifying the anisotropy signatures from annihilating dark matter which can be different from those of astrophysical origin particularly at small scales . the first calculations of this kind were performed analytically for gamma - ray background anisotropies induced by both annihilating extragalactic dark matter halos @xcite as well as galactic subhalos @xcite . analytic calculations were also performed for the anisotropic radio signatures from decaying dark matter @xcite and compared with the mean intensity and angular power spectrum of the astrophysical and cosmological radio background . more recently anisotropies of the diffuse gamma - ray background from annihilating dark matter in galactic subhalos have been computed using a numerical approach in ref . @xcite . according to their results , fermi - lat should be sensitive enough to constrain the amount of dark matter substructure through the anisotropy signal discussed above . the same numerical approach as in @xcite can be extended to the anisotropies from both the galactic and extragalactic dark matter contributions @xcite . in general , however , both the mean flux and anisotropy signal are contaminated by ordinary astrophysical sources including galactic and extragalactic resolved point sources , structure formation shocks @xcite , and normal galaxies @xcite . removal of such contributions @xcite , can greatly improve the detectability for searches of dark matter . all of the above analyses were based on the gamma - rays from direct annihilation into photons . however , the ics radiation from energetic electron and positron pairs produced by annihilation of dark matter also generates unique gamma - ray signatures , in an energy range that depends on the dark matter model . thus in the present paper we study the gamma - ray anisotropies produced by this `` secondary '' emission , which can be regarded as a new target for indirect searches of signatures induced by galactic substructure halos . the important point is that the properties of the angular anisotropies for emission from secondary electrons and positrons will be different in general than for prompt emission , because the former will be affected by propagation effects in the galaxy . in fact , by adopting a spatial diffusion model typical for high energy electrons and positrons in the galaxy , we show that the angular power spectrum is suppressed at small angular scales corresponding to the distance diffusively traveled by the charged particles during their energy loss time . as a result , for a typical dark matter model with mass of 1 tev and a canonical thermal freeze - out cross section @xmath4 , the angular power spectrum from inverse compton scattering peaks at large angular scales . the remainder of this paper is organized as follows . in sect . [ sec : formulation ] , we set up the formalism used in this paper for calculating the mean intensity and angular power spectrum of gamma - rays induced by dark matter annihilation and develop a robust numerical scheme to simulate the small subhalos with masses down to @xmath5 . we present the results of our calculations in sect . [ sec : result ] . finally , we summary our paper in sect . [ sec : conclusion ] . we will use natural units in which @xmath6 throughout . the propagation of high energy electrons and positrons in the turbulent galactic magnetic field can be described as a diffusion - energy loss equation @xcite for the electron - positron number density @xmath7 , neglecting the convection and re - acceleration terms which are only relevant for electrons and positrons below 10 gev @xcite : @xmath8 here , the spatial diffusion coefficient is taken as @xmath9 ; @xmath10 is the energy loss term and @xmath11 the source term , both of which are described in more details below . in the following we will often use the term electron as a short hand for both electrons and positrons . with the assumption that spatial diffusion and energy loss coefficients are spatially independent , for a given source distribution and boundary conditions , the propagation equation can be solved analytically . we assume the diffusion zone to be a cylinder with half - height @xmath12 of a few kpc and radius @xmath13kpc . in this paper , we adopt the widely used med model which , compared to other models , i.e. the min and max models , predicts intermediate values for the antiproton flux when fitted to reproduce the b / c ratio @xcite with @xmath14 , @xmath15 , and the half - height of the diffusion zone @xmath16kpc . the min and max propagation models would decrease and increase the predicted gamma - ray fluxes roughly by a factor of three and two , respectively . however , we verified that the shape of the angular power spectrum predicted by these three propagation models is basically the same . since in the present paper we mainly focus on a new `` suppression '' feature in angular power spectrum of the gamma - ray component from ics , we will use the med model in the following calculations . in this study , we investigate only the case in which electrons and positrons can escape freely at the boundaries of the diffusion zone . if the diffusion coefficient increases exponentially instead of the free - escape boundary condition , we expect that the angular power spectrum would be modified only at multipoles @xmath17 since the diffusion of electrons and positrons would flatten their distribution at length scales even larger than the whole diffusion zone . this implies that the angular power spectrum produced in the region of extremely large diffusion coefficient can not significantly influence the signature in the full - sky angular power spectrum we found at @xmath18 . in other words , the anisotropies weakly depend on different boundary conditions . at energies above 10 gev , the dominant energy losses are synchrotron radiation and inverse compton scattering ( ics ) on the interstellar radiation field . therefore , in the thomson limit we write @xmath19 [ for a more complicated treatment of energy loss see ref.@xcite ] , where @xmath20 for starlight ( sl ) , infrared ( ir ) , cmb photons and a magnetic field of @xmath21 . the inter - stellar radiation field ( isrf ) can be approximately characterized as a superposition of three blackbody - like spectra with different temperatures and normalization factors relative to a true black - body emitter : one for the cmb with @xmath22 , for the ir with @xmath23 and for the sl with @xmath24 @xcite . the typical normalization of the sl and ir fields of radiation depends on the position in the galaxy . the averaged normalizations of isrf photon densities per energy used in this paper are @xmath25 and @xmath26 for sl , ir and cmb , respectively @xcite . although these normalizations are valid only in the region with latitude @xmath27 , we have checked that changing the normalization of the isrf affects the gamma - ray emissions only weakly because an increased emission is partly compensated by a decrease of the density of electrons and positrons as energy losses increase . we also verified that the deviation from the detailed numerical simulation with galprop @xcite is less than a factor of two for a realistic spatial distribution of the isrf @xcite . in the annihilating dark matter scenario , the source term can be written as @xmath28 where @xmath29 is the annihilation spectrum into electrons and positrons at energy @xmath30 . for simplicity , we assume mono - energetic injection of the positron and electron in case of cp conservation , namely @xmath31 . we choose @xmath32 , which can well fit the pamela excess @xcite while not in conflict with gamma - ray observations by fermi - lat @xcite . finally , we use @xmath33 to reproduce the correct relic density for thermal freeze - out . by convolving our results for the gamma - ray spectra with the pair energy , our computational approach can be easily adapted to pair spectra different from mono - energetic injection , such as from dark matter annihilating into @xmath34 , and @xmath35 . as most of the electrons escape towards the @xmath36direction , we impose the dirichlet boundary condition @xmath37 , at which the particles can freely escape . we thus model the diffusion zone as an infinite slab of half thickness @xmath12 , and we take @xmath38 for the med diffusion model . the free - space green s function @xcite for eq . ( [ eq : diffeqn ] ) is @xmath39= \frac{1}{b(e)}\frac{1}{(\pi\lambda_d^2)^{3/2}}\e^{-({\bf r}-{\bf r}^\prime)^2/\lambda_d^2}\,,\ ] ] where we have defined the diffusion length as @xmath40 which is the average distance @xmath41 diffuse through during their energy loss time . then , the green s function satisfying appropriate boundary conditions can be obtained by considering a series of image charges at positions @xmath42 . one can verify that @xmath43 fulfills the dirichlet boundary condition . thus , the general solution to eq . ( [ eq : diffeqn ] ) , in the limit of time - independent sources and electron / positron number densities which already reached equilibrium , is given by @xmath44 for the diffuse gamma - ray emission we are more interested in the column density of electrons , @xmath45 than in the local space density of electrons . the observer is located at the solar system . in galactic coordinates , a point in cartesian coordinates @xmath46 at a distance @xmath47 from the observer is given by @xmath48 where @xmath49 and @xmath50 are the galactic longitude and latitude , respectively . we truncate the integral in eq . ( [ eq : column ] ) at the edge of the diffusion zone , beyond which particles are not confined , @xmath51 or @xmath52 . the line - of - sight integral can directly act on the free green s function @xcite , @xmath53/\lambda_d}}{2\pi \lambda_d^2b(e ) } \left[\erf\left(\frac{\ell_{\rm max}-{\bf n}\cdot{\bf r}^\prime}{\lambda_d}\right)- \erf\left(\frac{-{\bf n}\cdot{\bf r}^\prime}{\lambda_d}\right)\right]\,,\nonumber\end{aligned}\ ] ] where we have defined the unit - vector @xmath54 . the green s function satisfying the boundary condition is thus @xmath55 the column density of electrons therefore reads @xmath56 q({\bf r}^\prime , e^\prime)\,.\ ] ] in the limit of @xmath57 , where @xmath58 is the scale radius of the subhalo profile , the subhalo can be regarded as a point - like source . ( [ eq : sigma ] ) can then be simplified to @xmath59j_{k}({\bf r}_k),\ ] ] where @xmath60 for a given subhalo source @xmath61 located at @xmath62 . for the largest subhalos with masses larger than @xmath63 , their radius @xmath58 can be somewhat larger than the diffusion length @xmath64 . nevertheless , their contribution to the total flux is a factor @xmath65 smaller than the flux from the smaller subhalos . therefore , neither the mean intensity nor the dimensional angular power spectrum ( see sect . [ subsubsec : cl ] ) relies significantly on the distribution of electrons in the largest subhalos . as a result , for our purposes we can apply eq . ( [ eq : ssigma ] ) to all subhalos even if @xmath66 . alternatively , for primary electrons and positrons from the smooth host dark matter halo , the bessel - fourier scheme @xcite can require less computational time than the green s function . the electron and positron number density after propagation can be expressed as @xmath67 and @xmath68 is the halo function defined by @xmath69e^{-\left[\left(\frac{n\pi}{2l}\right)^2+\frac{\alpha_i^2}{r^2}\right]\frac{\lambda_d^2}{4 } } r_{i , n}\,.\ ] ] here , the @xmath70 are the zeros of the bessel function @xmath71 and @xmath72 are the coefficients of the bessel - fourier transform of the source term . for relativistic electrons and positrons with energy @xmath30 up scattering background photons from energy @xmath73 to @xmath74 , the emitted inverse compton power per energy interval is @xmath75 where @xmath76 is the differential isrf photon number density , while the differential cross section @xmath77 is given by the klein - nishina formula @xcite . folding @xmath78 with the spectral distribution of the equilibrium number density of electrons and positrons , we get the emissivity of ic photons of energy @xmath74 , @xmath79 which yields the ic intensity at energy @xmath74 by the line - of - sight integral @xmath80 according to the eq . ( [ eq : sigma ] ) , the ic intensity from the electrons and positrons can be simplified to @xmath81 there is a well known `` delta - function approximation '' where an electron with energy @xmath30 inverse compton scattering black - body photons with temperature @xmath82 emits photons with a characteristic energy @xmath74 in both the thomson and extreme klein - nishina limits @xcite , @xmath83 , @xmath84 $ ] , where @xmath85 is the total ic energy loss rate of the electron . the numerical calculations show that @xmath86 may be approximated by @xmath87 and in the thomson regime one has @xmath88 , where @xmath89 is the electron charge and @xmath90 is the background photon energy density . ( [ eq : ici ] ) can thus be simplified to @xmath91 where the electron / positron energy @xmath30 is related to the gamma - ray energy @xmath74 through the following condition : @xmath92 . this relation reproduces the known slope @xmath93 for an electron spectrum of @xmath94 . in the thomson limit and for our choice of monoenergetic injection of electron - positron pairs , the index @xmath95 in the stationary situation if the energy loss term dominates on the right hand side of eq . ( [ eq : diffeqn ] ) , as is usually the case for electron energies above 10 gev . for @xmath96tev we are always in the thomson limit . in the present section we discuss contributions to the source term @xmath97 both from the smooth host halo and the individual subhalos . the two most important features of the subhalos are their mass distribution and their spatial distribution . in the following , we briefly summarize the properties of subhalos used in our calculations for which we adopt the same description as in ref . @xcite , and references therein . there are two widely used scenarios for describing the subhalo radial distribution . one is unbiased relative to the galactic smooth component ( host halo ) with the nfw density profile @xcite given by @xmath98 with @xmath99 , where @xmath58 is a scale radius and @xmath100 is the characteristic density . for the case where the subhalo distribution is anti - biased compared to the smooth component , we use the fitting formula of the subhalo radial distribution from gao _ et al _ @xcite . the cumulative fraction of subhalos within a given radius is @xmath101 with @xmath102 , @xmath103 , @xmath104 , @xmath105 and @xmath106 is the host halo concentration . furthermore , @xmath107 is the total number of subhalos within virial radius @xmath108 of the host halo . for a host halo with the nfw density profile we adopted @xmath109 , @xmath110kpc and @xmath111 @xcite . recent simulations suggest that the cumulative number of subhalos above a given mass @xmath112 in units of the solar mass can be fitted simply by a power law @xcite , @xmath113 in the present paper , we simulate the subhalos with mass down to around one earth mass , @xmath114 and choose @xmath115 , as suggested by simulations . we find that choosing @xmath116 between @xmath117 and @xmath26 @xcite would change the mean intensity by roughly a factor of @xmath118 . more recent simulations suggest that the central structure of dark matter halos deviates from the nfw profile in the innermost regions , which can be well reproduced by an einasto density profile . therefore , here we use the einasto profile with the parameter @xmath119 given by gao _ @xcite , @xmath120\,.\ ] ] the relation of halo concentration and mass is given by @xmath121 the slope and normalization are consistent with those found by bullock _ @xcite . the angular power spectrum of the emission maps can be calculated by using the public healpix package @xcite . we define the dimensionless quantity @xmath122 as a function on the sphere , which can be expanded in spherical harmonics @xmath123 as @xmath124 where i(@xmath125 ) describes the gamma - ray intensity in the direction @xmath125 . the _ dimensionless _ angular power spectrum of @xmath126 is given by the coefficients @xmath127 note that the measured dimensional amplitude @xmath128 of the total angular power spectrum in units of intensity squared can be obtained by multiplying the dimensionless angular power spectrum of a given component , @xmath129 , with its mean intensity squared , @xmath130 and summing over the components , @xmath131 where in our case the sum basically runs over the contributions of the smooth host halo , the subhalo distribution and other astrophysical foregrounds and backgrounds . here we assumed that the different contributions are uncorrelated . in numerical calculations , taking into account subhalos down to masses @xmath5 in the angular power spectrum would require the generation of @xmath132 subhalos within the diffusion zone in a given monte carlo realization . this is not very practical and severely limits the number of realizations one can simulate . thus , we need to develop a sound scheme to circumvent these computational limits . here we take advantage of a few simple facts : ( i ) two terms contribute to the fluctuation anisotropies in units of average intensity squared ; the one - subhalo term ( @xmath133 ) which is poissonian noise and the two - subhalo term ( @xmath134 ) due to the cross - correlation of substructure arising from their radial distribution within the host halo ( see eqs . @xmath135 and @xmath136 of ref . @xcite ) ) . ( ii ) qualitatively speaking , the two - subhalo term scales linearly with the square of the average flux whereas the one - subhalo term is proportional to an integral of squared luminosities of subhalos . since subhalo luminosity @xmath137 is related to its mass @xmath112 via @xmath138 , by using eq . ( [ eq : c ] ) , the correlation between gamma - ray luminosity and mass thus is positive , i.e. , @xmath139 . one - subhalo term and two - subhalo term per decade of subhalo mass then can be estimated as @xmath140 and @xmath141 . therefore , large and relatively rare substructures dominate the one - subhalo term , but negligibly to the two - subhalo term , and the small but numerous substructure halos dominate the two - subhalo term . there is thus a dividing mass of substructures , @xmath142 , below which their contribution to one - subhalo term are negligible . the numerical value of this mass depends on the assumed radial distribution of the substructure halos ( see below ) . in any case , it turns out that halos with @xmath143 have a characteristic radius , @xmath58 , much smaller than the diffusion length of the gamma - ray emitting electrons , i.e. @xmath144 . thus , effectively all halos below @xmath142 have the same image on the sky in ics with a characteristic size given by @xmath64 . this further simplifies our calculations because besides minor poissonian noise , the angular power spectrum from subhalos with @xmath143 is simply obtained by convoluting the subhalo spatial distribution with the mass - independent ics subhalo image . thus the contribution from subhalos with @xmath143 is taken into account straightforwardly as follows : we generate a distribution of subhalos with mass in the decade @xmath145 to obtain the full - sky map . the number of generated subhalos is typically much smaller than the actual value , @xmath146 , but is sufficiently large that the two - subhalo term divided by the mean intensity squared has converged . the intensity of the map and the corresponding two - subhalo term thus can be rescaled to the value appropriate for @xmath147 halos . in this way , the simulated two - subhalo term of the angular power spectrum can fairly approach the actual one . since the one - subhalo term is not linearly proportional to the square of the average intensity and is contributed mostly by larger subhalos , we thus complete the calculation by adding the contribution of subhalos with mass @xmath148 , which can now be done with a direct monte carlo simulation . specifically , we find that our calculation of the angular power spectrum reaches convergence when we use at least @xmath149 subhalos for each mass decade from @xmath150 to @xmath142 within the diffusion zone , when we use @xmath151 for the unbiased radial distribution and @xmath152 for the anti - biased case . note that this method can not be applied to the angular power spectrum of the direct annihilation component since in this case the profile of the emitting region depends on the profile of each subhalo rather than on the identical diffusion length which just depends on the energy of electrons and positrons for all subhalo masses . in this section , we make use of the formalism derived in the previous section to compute the gamma - ray mean intensity map and associated anisotropy power spectrum due to both the smooth host halo component and the substructure halos . we also study how the map morphology and anisotropies depend on the radial distribution of subhalos and of the gamma - ray energy . we first consider the gamma - ray emission produced by dark matter annihilating into @xmath41 pairs scattering off the isrf both in the smooth host halo and the subhalos . using the formalism developed in sect . [ sec : formulation ] , the fluxes from the smooth host halo at 1 gev , 10 gev and 100 gev are shown in fig . [ fig : diff - host ] . here we use galactic coordinates , where an observer is located at 8.5 kpc from the galactic center . we adopt the nfw density profile with the parameters given in sect . [ subsec : subst ] , and find , as expected , that most of the signal comes from the central galactic region where the dark matter is highly concentrated . the values of the mean gamma - ray intensity from both the host halo and the substructure are also reported in tab . [ tab : i ] for the above three photon energies . we find that in both cases the mean gamma - ray intensity decreases faster than @xmath153 in eq . ( [ eq : simici ] ) . this is because there will be a high energy cut - off in @xmath74 once the required energy of the parent pairs @xmath154^{1/2}$ ] scattering off the low energy background photons exceeds the maximum energy @xmath155 produced by dark matter annihilation . the effect is even stronger when the delta - function approximation for ics is assumed , as in our case . as a consequence , all background photon energies at cmb energies and above contribute to the 1 gev ics photon flux , whereas only the ir and sl contribute to the 10 gev photon flux and only the sl contributes to the 100 gev photon flux , as also summarized in tab . [ tab : e_e ] . furthermore , using the relation between @xmath30 and @xmath74 and the expression for @xmath85 , for an electron spectrum @xmath156 eq . ( [ eq : simici ] ) gives the scaling @xmath157 . using the different normalizations for the cmb , ir and sl densities , then would suggest that the gamma - ray intensity originating from @xmath41 pairs scattering off cmb photons is about 5 times larger than the one from scattering off the ir photons and about 10 times lager than the contribution from scattering off sl photons , provided there is no restriction from the kinematics . however , the @xmath41 pairs produced by annihilating dark matter do have an absolute cutoff at the parent particle energy . as a result , the more detailed numerical results show that most of the gamma - ray intensity at 1 gev is produced by pairs with @xmath158 scattering off the cmb , whereas pairs with @xmath159 scattering off ir photons dominate the 10 gev gamma - ray flux and pairs with @xmath160 scattering off sl dominate ics photons at 100 gev . .the mean gamma - ray intensities @xmath161 at 1 gev , 10 gev and 100 gev , averaged over the sky , from subhalos with minimum mass @xmath162 for the unbiased and anti - biased radial distribution and from the smooth host halo . the unit is @xmath163 . [ cols="^,^,^,^",options="header " , ] in fig . [ fig : diff - unbias ] , we present the sky map of gamma - ray emission predicted by an unbiased radial distribution of subhalos : in order to avoid saturation from the dominant population of small halos , in the figure we only include halos with mass above @xmath164 . the most remarkable feature in fig . [ fig : diff - unbias ] is that the diffuse emission regions can extend to a few kpc in length scale , corresponding to a few tens of degrees on the sky . furthermore , the size of the brightest regions representing the largest intensity tends to increase with increasing @xmath74 , which could give rise to an increase of the angular power spectrum on corresponding angular scales . the size of the emitting region is basically determined by the diffusion length @xmath64 which is the distance that the @xmath41 diffuses during their energy loss time . this length scale can be estimated from eq . ( [ eq : d ] ) , giving @xmath165 for the med propagation model . this energy dependence of @xmath64 is shown in tab . [ tab : e_e ] . the typical diffusion length is @xmath166 kpc , @xmath167 kpc and @xmath168 kpc for electron energies leading to emission at @xmath169gev , 10 gev and 100 gev , respectively , corresponding to an angular scale @xmath170 with @xmath171 few kpc the typical distance to the dark matter annihilation source . these angular scales are roughly what one sees in fig . [ fig : diff - unbias ] and the above estimate also applies to the smooth host halo case ( see fig . [ fig : diff - host ] ) and to the anti - biased radial distribution of subhalos . how does the mean intensity depend on @xmath172 ? empirically , we find that the mean intensity roughly doubles with each decade of decreasing mass of subhalos , similar to what has been found by ref . @xcite . in light of the subhalo mass function @xmath173 in eq . ( [ eq : dndm ] ) and the concentration parameter @xmath174 in eq . ( [ eq : c ] ) , the annihilation rates per decade of subhalo mass can be approximated by @xmath175 , which is fairly consistent with our detailed numerical calculation . in order to determine the influence of the radial distribution of subhalos on the intensity , in fig . [ fig : diff - bias ] the gamma - ray full - sky maps at 1 gev , 10 gev and 100 gev for the un - biased distribution are compared with the anti - biased distribution . for the anti - biased radial distribution the mean intensities are roughly 50 times smaller than for the unbiased case since the mean distance of subhalos to us is much larger than in the unbiased case . one notices the important feature in fig . [ fig : diff - bias ] that the emission from the spatially anti - biased distribution is much less centrally concentrated , and apparently accumulates around the galactic plane compared with the unbiased case . this can be understood from the fact that only electrons and positrons confined within the diffusion zone of scale height @xmath176 kpc and radius @xmath177 kpc for the med model can efficiently produce gamma - rays by ics and the subhalos within the diffusion zone are distributed much more isotropically for the anti - biased case compared to the more central distribution of the unbiased case . furthermore , ics outside the diffusion zone contributes less than @xmath178 to the mean intensity @xcite . the dimensionless angular power spectra @xmath179 of gamma - ray emission due to the ics at photon energies 1 gev , 10 gev and 100 gev from the main components of galactic dark matter are presented in fig . [ fig : cl ] . shown are the power spectrum for subhalos with @xmath180 for the unbiased radial distribution and for the smooth host halo . to clearly illustrate the effects of diffusion , the angular power spectra for spatially unbiased subhalos with @xmath181 and for the host halo are also shown in absence of diffusion . we recall that in order to obtain for a given component the contribution to the angular power spectrum @xmath128 in units of intensity squared , according to eq . ( [ c_l ] ) one has to multiply the dimensionless power spectra @xmath179 shown in fig . [ fig : cl ] by the squared total intensity @xmath182 of the corresponding component from tab . [ tab : i ] . [ fig : cl ] demonstrates the remarkable feature that the power spectrum of the ics component of galactic dark matter annihilation is exponentially suppressed for @xmath183 compared with what one would obtain without diffusion . furthermore , the lower energy gamma - rays have more angular power at @xmath183 corresponding to small angular scale . this can be understood from the energy - dependence of the diffusion length : intensity fluctuations should be damped on length scales smaller than the diffusion length @xmath64 , corresponding to a multipole @xmath184 , where @xmath185 is the typical distance to the dark matter annihilation source . based on the discussion in sect.[subsec : dg ] , we can estimate the diffusion length of the electrons and positrons emitting a given gamma - ray energy . when doing so , one should keep in mind that the @xmath41 pairs interact with three different photon backgrounds and that , as it turns out , unlike the case of a single background , the lower the gamma - ray energy the higher the energy of the emitting electrons . since we found that the diffusion length decreases with the electron energy , @xmath165 , we expect suppression of anisotropies to occur at smaller scales , or larger values of the multipole @xmath49 , for lower photon energy . for other propagation models such as the min and max models which have a slightly different energy dependence of the spatial diffusion coefficient , the diffusion length would be slightly larger or smaller , respectively . this would shift the suppression scale in the angular power spectrum by less than a factor of two in the multipole @xmath49 . in fact , for the smooth host halo the typical distance is @xmath186 kpc , and the suppression due to diffusion should occur at @xmath187 and @xmath188 for gamma - ray energies of 1 gev , 10 gev and 100 gev , respectively . this analysis can be applied to the case of subhalos . for an unbiased spatial distribution of subhalos , the typical distance to a subhalo is @xmath189kpc which corresponds to diffusive suppression at relatively smaller @xmath49 compared to the host halo case . these simple estimates are consistent with our detailed calculations shown in fig . [ fig : cl ] . we also find that the amplitude of the dimensionless angular power spectrum @xmath179 from the smooth host halo is larger than that from the subhalos since the emissivity profile from annihilation in the smooth halo is proportional to the density squared and thus more peaked than annihilation in the subhalos which essentially follow the linear density profile of an nfw profile , as we see in fig . [ fig : cl ] . we note that although the smooth host halo has a large amplitude of the dimensionless angular power spectrum @xmath179 , its contribution to the total intensity is as small as @xmath190 for the unbiased subhalo distribution and @xmath191 for the anti - biased distribution , as seen in tab . [ tab : i ] . we can therefore safely neglect the contribution from the host halo both to the mean intensity and to the dimensional angular power spectrum @xmath128 . finally , we show in fig . [ fig : cl ] the dimensionless angular power spectrum @xmath179 for a smooth halo with emissivity tracing the density of the nfw profile instead of the squared density that would be relevant for the contribution of the host halo in decay scenarios . this shows that a smooth nfw profile describes the emission profile by a large number of subhalos following an unbiased radial distribution very well , at a level of @xmath190 . this is not surprising since the number of subhalos within the diffusion zone with masses below @xmath192 is sufficiently large , @xmath193 , for each mass decade to strongly suppress any deviation from a smooth distribution . this conclusion is also true for the anti - biased case . despite the fact that the contribution of large subhalos in the mass range of @xmath194 fluctuates strongly , their contribution to the total emission are three orders of magnitude smaller which leads to fluctuations at @xmath195 level in angular power spectrum as seen in fig . [ fig : cl ] . how does the radial subhalo distribution affect the angular power spectrum ? in fig . [ fig : bias_cl ] we compare results for the unbiased case and the anti - biased case . at small @xmath49 the angular power spectrum @xmath179 for the anti - biased distribution is suppressed relative to the unbiased case which is due to the more isotropic subhalo distribution at large angular scales seen in fig . [ fig : diff - bias ] . for @xmath183 , the angular power spectrum induced by an anti - biased distribution has more power than the unbiased case because the typical distance to subhalos is larger for the anti - biased distribution , shifting power to larger @xmath49 . we should note that , of course , the electrons and positrons of ordinary astrophysical sources , such as pulsars and supernova remnants , could also produce significant ics signals . at @xmath183 the corresponding angular power spectrum would be difficult to distinguish from the one induced by dark matter because the large-@xmath49 power spectrum is highly suppressed and the suppression scale is mostly determined by the diffusion length of electrons and positrons which is independent of wether their origin is astrophysical or from dark matter . in contrast , the power spectrum at @xmath196 should contain information on the sources of electrons and positrons due to the different spatial morphologies of dark matter and astrophysical sources on these scales which are not significantly influenced by diffusion . in addition , the ics energy spectrum is in general significantly different due to the harder pair spectrum from dark matter annihilation which would naturally explain the @xmath41 excesses observed by pamela and fermi - lat . although the nearby astrophysical sources such as pulsars can also reproduce these excesses , pulsars are mainly concentrated at the galactic plane and the primary @xmath41 injected spectrum have an exponential cutoff at high energies , both of which are quite different from annihilating dark matter scenarios . therefore , a combination of these two effects may allow to discriminate between astrophysical sources and at least some dark matter scenarios . for example , ref . @xcite proposed an energy dependence of an anisotropy signature to distinguish millisecond pulsars from heavy dark matter candidates . the disentanglement of dark matter signals from astrophysical backgrounds is , however , beyond the scope of the present paper . in this paper we have investigated the angular power spectrum of the gamma - ray emission from inverse compton scattering off low energy target photons of electrons and positrons produced by dark matter annihilation . we considered two extreme cases for the radial distribution of subhalos and simulated the full - sky gamma - ray maps at three energies through realizations of a large number of subhalos with masses down to @xmath150 . the contributions to the angular power spectrum and to the total flux from the smooth host halo were also calculated . 1 . we point out a new feature in the angular power spectrum of photons produced by inverse compton scattering of pairs from annihilating dark matter that does not occur in the power spectra of gamma - rays produced directly in annihilation , on which earlier work often has focused : in contrast to the direct annihilation component for which the poisson noise dominates at large @xmath49 , the angular power spectrum from inverse compton scattering is exponentially suppressed at @xmath183 , because the diffusion of the high energy parent electrons and positrons produced by dark matter annihilation strongly smears out anisotropies of gamma - rays at small angular scales . this suppression scale is determined by the diffusion length @xmath197 , evaluated at the energy @xmath30 of the emitting electrons and positrons . the energy @xmath30 of the dominant emitters is determined by the low energy target photon backgrounds and , in the presence of several background components , is not a monotonic function of the gamma - ray energy . for the backgrounds used in this paper , cmb , ir and sl , we find that the lower energy gamma - rays are produced by the higher energy @xmath41 pairs ( and vice - versa ) so that when the energy dependence of the spatial diffusion coefficient is taken into account , @xmath165 , power spectrum suppression is found to occur at larger multipole @xmath49 for lower photon energy . the contribution to the absolute amplitude of the angular power spectrum from the smooth host halo can be safely neglected which is due to its negligible mean intensity compared to that from the subhalos . for subhalos , the intensity from an unbiased subhalo distribution is @xmath198 times larger than from an antibiased distribution . at small @xmath49 the dimensionless angular power spectrum for the anti - biased distribution is suppressed relative to the unbiased case , whereas at large @xmath183 the opposite is true . an enormous number of subhalos with the masses down to @xmath199 can be well described by a smooth halo whose emissivity follows the spatial distribution of subhalos . that is not surprising , because diffusion of the emitting particles can smooth the whole sky map if the number of subhalos is sufficiently large . finally we remark that we have assumed homogeneous and isotropic magnetic fields and interstellar radiation field . however , for the more realistic case where the radiation and magnetic fields depend on the position in the galaxy @xcite , the angular power spectrum should be modified at least at large angular scales . in particular , the small scale fluctuations of the magnetic field could affect the distribution of electrons and positrons @xcite , influencing the angular power spectrum at all scales . moreover , a space - dependence of the diffusion length could also lead to a directional dependence of the angular power spectrum . obviously detecting the differential effect of the anisotropy signal across the galactic latitude would be even more difficult than detecting the signal integrated over the whole sky . however , it could introduce subtle effects when part of the sky is masked out , e.g. to avoid the contribution from the galactic plane . these complications should be investigated more thoroughly in the future by performing a real 3d simulation through numerically solving the transport equation for electrons and positrons . while the formalism developed in this paper can be extended to the anisotropies of the radio sky , the small scale structure of the magnetic field is more important for synchrotron emission and should be treated in more detail . in addition , our work could also be extended to inverse compton gamma - ray emission from extragalactic dark matter halos . we acknowledge use of the healpix @xcite software and analysis package for deriving the results in this paper . this work was supported by the deutsche forschungsgemeinschaft through the collaborative research centre sfb 676 particles , strings and the early universe : the structure of matter and space - time , and by the state of hamburg , through the collaborative research program connecting particles with the cosmos within the framework of the landesexzellenzinitiative ( lexi ) . we thank luca maccione and enrico borriello for reading the manuscript and for very useful comments . 99 g. bertone , d. hooper and j. silk , `` particle dark matter : evidence , candidates and constraints , '' phys . rept . * 405 * , 279 ( 2005 ) [ arxiv : hep - ph/0404175 ] . g. jungman , m. kamionkowski and k. griest , `` supersymmetric dark matter , '' phys . rep . * 267 * , 195 ( 1996 ) [ arxiv : hep - ph/9506380 ] . d. hooper and s. profumo , `` dark matter and collider phenomenology of universal extra dimensions , '' phys . rep . * 453 * , 29 ( 2007 ) [ arxiv : hep - ph/0701197 ] . e. komatsu _ et al . _ [ wmap collaboration ] , `` five - year wilkinson microwave anisotropy probe observations : cosmological interpretation , '' arxiv:0803.0547 [ astro - ph ] . x. l. chen and m. kamionkowski , `` particle decays during the cosmic dark ages , '' phys . rev . d * 70 * , 043502 ( 2004 ) [ arxiv : astro - ph/0310473 ] . l. zhang , x. chen , m. kamionkowski , z. g. si and z. zheng , `` contraints on radiative dark - matter decay from the cosmic microwave background , '' phys . d * 76 * , 061301 ( 2007 ) [ arxiv:0704.2444 [ astro - ph ] ] . l. zhang , x. l. chen , y. a. lei and z. g. si , `` the impacts of dark matter particle annihilation on recombination and the anisotropies of the cosmic microwave background , '' phys . rev . d * 74 * , 103519 ( 2006 ) [ arxiv : astro - ph/0603425 ] ; a. a. abdo _ et al . _ [ the fermi - lat collaboration ] , `` measurement of the cosmic ray e+ plus e- spectrum from 20 gev to 1 tev with the fermi large area telescope , '' phys . lett . * 102 * , 181101 ( 2009 ) [ arxiv:0905.0025 [ astro-ph.he ] ] . d. grasso _ et al . _ [ fermi - lat collaboration ] , on possible interpretations of the high energy electron - positron spectrum astropart . * 32 * ( 2009 ) 140 [ arxiv:0905.0636 [ astro-ph.he ] ] . g. di bernardo , d. gaggero , d. grasso and f. l. collaboration , `` the high energy cosmic ray electron spectrum measured by fermi gamma - ray space telescope : some possible interpretations , '' arxiv:0912.3887 [ astro-ph.he ] . m. cirelli , m. kadastik , m. raidal and a. strumia , `` model - independent implications of the e+ , e- , anti - proton cosmic ray spectra on properties of dark matter , '' nucl . phys . b * 813 * , 1 ( 2009 ) [ arxiv:0809.2409 [ hep - ph ] ] . i. cholis , l. goodenough , d. hooper , m. simet and n. weiner , `` high energy positrons from annihilating dark matter , '' arxiv:0809.1683 [ hep - ph ] . v. barger , w. y. keung , d. marfatia and g. shaughnessy , `` pamela and dark matter , '' arxiv:0809.0162 [ hep - ph ] . r. harnik and g. d. kribs , `` an effective theory of dirac dark matter , '' arxiv:0810.5557 [ hep - ph ] . a. e. nelson and c. spitzer , `` slightly non - minimal dark matter in pamela and atic , '' arxiv:0810.5167 [ hep - ph ] . y. bai , m. carena and j. lykken , `` the pamela excess from neutralino annihilation in the nmssm , '' phys . d * 80 * , 055004 ( 2009 ) [ arxiv:0905.2964 [ hep - ph ] ] . d. malyshev , i. cholis and j. gelfand , `` pulsars versus dark matter interpretation of atic / pamela , '' arxiv:0903.1310 [ astro-ph.he ] . s. profumo , `` dissecting pamela ( and atic ) with occam s razor : existing , well - known pulsars naturally account for the anomalous cosmic - ray electron and positron data , '' arxiv:0812.4457 [ astro - ph ] . t. kobayashi , y. komori , k. yoshida and j. nishimura , `` the most likely sources of high energy cosmic - ray electrons in supernova remnants , '' astrophys . j. * 601 * , 340 ( 2004 ) [ arxiv : astro - ph/0308470 ] . n. j. shaviv , e. nakar and t. piran , `` natural explanation for the anomalous positron to electron ratio with supernova remnants as the sole cosmic ray source , '' phys . lett . * 103 * , 111302 ( 2009 ) [ arxiv:0902.0376 [ astro-ph.he ] ] . m. cirelli and p. panci , `` inverse compton constraints on the dark matter e+e- excesses , '' nucl . b * 821 * , 399 ( 2009 ) [ arxiv:0904.3830 [ astro-ph.co ] ] . e. a. baltz and l. wai , `` diffuse inverse compton and synchrotron emission from dark matter annihilations in galactic satellites , '' phys . d * 70 * , 023512 ( 2004 ) [ arxiv : astro - ph/0403528 ] . q. yuan , p. f. yin , x. j. bi , x. m. zhang and s. h. zhu , `` gamma rays and neutrinos from dark matter annihilation in galaxy clusters , '' phys . d * 82 * , 023506 ( 2010 ) [ arxiv:1002.0197 [ astro-ph.he ] ] . c. boehm , t. delahaye and j. silk , `` can the morphology of gamma - ray emission distinguish annihilating from decaying dark matter ? , '' arxiv:1003.1225 [ astro-ph.ga ] . g. hutsi , a. hektor and m. raidal , `` implications of the fermi - lat diffuse gamma - ray measurements on annihilating or decaying dark matter , '' jcap * 1007 * , 008 ( 2010 ) [ arxiv:1004.2036 [ astro-ph.he ] ] . t. lin , d. p. finkbeiner and g. dobler , `` the electron injection spectrum determined by anomalous cosmic ray , gamma ray , and microwave signals , '' arxiv:1004.0989 [ astro-ph.co ] . a. loeb and m. zaldarriaga , `` the small - scale power spectrum of cold dark matter , '' phys . d * 71 * , 103520 ( 2005 ) [ arxiv : astro - ph/0504112 ] . e. bertschinger , `` the effects of cold dark matter decoupling and pair annihilation on cosmological perturbations , '' phys . d * 74 * , 063509 ( 2006 ) [ arxiv : astro - ph/0607319 ] . s. profumo , k. sigurdson and m. kamionkowski , `` what mass are the smallest protohalos ? , '' phys . lett . * 97 * , 031301 ( 2006 ) [ arxiv : astro - ph/0603373 ] . j. diemand , b. moore and j. stadel , `` earth - mass dark - matter haloes as the first structures in the early universe , '' nature * 433 * , 389 ( 2005 ) [ arxiv : astro - ph/0501589 ] . l. pieri , g. bertone and e. branchini , `` dark matter annihilation in substructures revised , '' mon . not . soc . * 384 * , 1627 ( 2008 ) [ arxiv:0706.2101 [ astro - ph ] ] . m. kamionkowski , s. m. koushiappas and m. kuhlen , `` galactic substructure and dark matter annihilation in the milky way halo , '' phys . d * 81 * , 043532 ( 2010 ) [ arxiv:1001.3144 [ astro-ph.ga ] ] . s. ando and e. komatsu , `` anisotropy of the cosmic gamma - ray background from dark matter annihilation , '' phys . d * 73 * ( 2006 ) 023521 [ arxiv : astro - ph/0512217 ] . s. ando , e. komatsu , t. narumoto and t. totani , `` dark matter annihilation or unresolved astrophysical sources ? anisotropy probe of the origin of cosmic gamma - ray background , '' phys . d * 75 * , 063519 ( 2007 ) [ arxiv : astro - ph/0612467 ] . s. ando , `` gamma - ray background anisotropy from galactic dark matter substructure , '' arxiv:0903.4685 [ astro-ph.co ] . m. fornasa , l. pieri , g. bertone and e. branchini , `` anisotropy probe of galactic and extra - galactic dark matter annihilations , '' arxiv:0901.2921 [ astro - ph ] . a. cuoco , j. brandbyge , s. hannestad , t. haugboelle and g. miele , `` angular signatures of annihilating dark matter in the cosmic gamma - ray background , '' phys . d * 77 * ( 2008 ) 123518 [ arxiv:0710.4136 [ astro - ph ] ] . b. s. hensley , j. m. siegal - gaskins and v. pavlidou , `` the detectability of dark matter annihilation with fermi using the anisotropy energy spectrum of the gamma - ray background , '' arxiv:0912.1854 [ astro-ph.co ] . j. zavala , v. springel and m. boylan - kolchin , `` extragalactic gamma - ray background radiation from dark matter annihilation , '' mon . not . * 405 * , 593 ( 2010 ) [ arxiv:0908.2428 [ astro-ph.co ] ] . f. miniati , `` numerical modelling of gamma radiation from galaxy clusters , '' mon . not . * 342 * , 1009 ( 2003 ) t. delahaye , f. donato , n. fornengo , j. lavalle , r. lineros , p. salati and r. taillet , `` galactic secondary positron flux at the earth , '' astron . astrophys . * 501 * , 821 ( 2009 ) [ arxiv:0809.5268 [ astro - ph ] ] . strong a.w . and moskalenko i.w . , `` propagation of cosmic - ray nucleons in the galaxy , '' astrophys . j. * 509 * ( 1998 ) 212 [ astro - ph/9807150 ] . t. a. porter , i. v. moskalenko and a. w. strong , `` inverse compton emission from galactic supernova remnants : effect of the interstellar radiation field , '' astrophys . j. * 648 * , l29 ( 2006 ) [ arxiv : astro - ph/0607344 ] . t. delahaye , r. lineros , f. donato , n. fornengo and p. salati , `` positrons from dark matter annihilation in the galactic halo : theoretical uncertainties , '' phys . d * 77 * , 063527 ( 2008 ) [ arxiv:0712.2312 [ astro - ph ] ] . d. maurin , f. donato , r. taillet and p. salati , `` cosmic rays below z=30 in a diffusion model : new constraints on propagation parameters , '' astrophys . j. * 555 * , 585 ( 2001 ) [ arxiv : astro - ph/0101231 ] . rybicki g.b . , lightman a.p . , radiative processes in astrophysics , 1979 , john wiley & sons . k. m. gorski , e. hivon , a. j. banday , b. d. wandelt , f. k. hansen , m. reinecke and m. bartelman , `` healpix a framework for high resolution discretization , and fast analysis of data distributed on the sphere , '' astrophys . j. * 622 * , 759 ( 2005 ) [ arxiv : astro - ph/0409513 ] . m. perelstein and b. shakya , `` comment on calculation of positron flux from galactic dark matter , '' arxiv:1002.4588 [ astro-ph.he ] . j. m. siegal - gaskins , r. reesman , v. pavlidou , s. profumo and t. p. walker , `` anisotropy constraints on millisecond pulsars in the diffuse gamma ray background , '' arxiv:1011.5501 [ astro-ph.he ] . g. dobler , d. p. finkbeiner , i. cholis , t. r. slatyer and n. weiner , `` the fermi haze : a gamma - ray counterpart to the microwave haze , '' astrophys . j. * 717 * , 825 ( 2010 ) [ arxiv:0910.4583 [ astro-ph.he ] ] .

correlation and interference distinguish quantum from classical mechanics . the former is manifest in the measurement of many - body coincidences predicted by a quantum joint probability density function ( pdf ) , which are not observed classically @xcite . the latter is most familiar as a probability density function for an outcome that can be achieved in at least two indistinguishable ways . however , interference can also be generated by superposing many - body states in indistinguishable ways @xcite . experimental confirmation of quantum correlation has involved photons @xcite , atoms @xcite , and josephson phase qubits @xcite . however , there is little experimental evidence of correlated interference between massive particles . the boundary between the classical and quantum regimes continues to be an area of interest both in attempts to explain why quantum effects are difficult to measure on a macroscopic scale @xcite and in attempts to measure quantum effects on a macroscopic object @xcite . reflection measurement methods used in an attempt to extend this boundary involve light interacting with stationary @xcite and oscillating mirrors @xcite and with microwaves interacting with an oscillating mirror @xcite . reflection of a single photon from an oscillating mirror @xcite and using a bose einstein condensate as a mirror from which photons reflect @xcite have been studied . to extend this boundary to larger masses , we model correlation of a mesoscopic / macroscopic ` mirror ' with a non - zero rest mass microscopic ` particle ' after elastic reflection . motion is in free space along one dimension and all states are unbound . measurements of particle reflection , but not associated with correlated interference , have involved mirrors that reflect atoms @xcite and bose - einstein condensates @xcite , atoms reflecting from a solid surface @xcite , neutrons @xcite and atoms @xcite reflecting from vibrating mirrors , and atoms reflecting from a switchable mirror @xcite . if the particle and mirror are initially in eigenstates of energy then the following two states interfere : the particle - mirror state before the particle reflects from the mirror and the particle - mirror state after reflection ( interference is a consequence of the incident and reflected states being indistinguishable for a measurement of position ) . this is similar to the standing wave interference of a harmonic electromagnetic wave reflecting from a stationary mirror @xcite . classically , however , the incident and reflected waves interfere while the mirror experiences only a continuous force due to radiation pressure . the two - body quantum analogy involves solving the schrdinger equation with an interaction potential modeling the reflection . interference is expected between the incident and reflected particle substates _ along with _ interference of the mirror substates which have and have not reflected the particle . their correlation is perhaps not expected , being a consequence of the solution to the schrdinger equation from which a joint probability density function is constructed . this joint pdf then describes the correlations in the two - body interference which are manifest as coincidence rates , e.g. a correlation in the simultaneous measurement of particle _ and _ mirror positions . asynchronous measurement in such a system has also been discussed @xcite . any experimental realization of such an interferometer will involve wavegroups . while the coherence length for the particle currently can be as large as @xmath0 m for atoms released from a bose - einstein condensate , that for a mesoscopic / macroscopic mirror will be many orders of magnitude smaller . however , the coherence length of the two - body state involves that of both the particle and mirror substates . the two - body interference in reflection from a mirror is then limited spatially ( to the overlap region ) and temporally ( to the time of overlap ) in a manner similar to that of a pulse of light reflecting from a moving mirror . classical interference in reflection is more robust when the pulse of light reflects from a moving thin film for two main reasons : first the reflected pulses from each thin film surface maintain overlap as they both travel in the backward direction and second they have the same doppler shift ( since they reflect from interfaces moving at the same speed ) . interference is then limited by a pulse coherence length which at its minimum is of order the film thickness . to achieve such robust two - body quantum interference we consider the analogous two - body situation , an example of which is a neutron reflecting from each of the two surfaces of a moving aluminum ` slab . ' it is shown below that this two - body interference has the same ` non - local ' characteristics as does a pulse of light reflecting from a moving thin film . treating both the particle and reflector as quantum objects yields the expected results of a one - body treatment ( analogous to interference for a pulse of light reflecting from a moving thin film ) but only as a special case . in addition , this treatment predicts novel reflection regimes depending on the coherence lengths of the particle and reflector . quantum two - body reflection differs from that classically . for example , the classical kinematics , associated with reflection of a wavegroup , involves accelerated mirror motion while the kinematics associated with the expectation value of the position of either the particle or mirror , during the interaction , depends on the two - body wavegroup parameters such as coherence lengths . this is discussed in section [ sec : summary ] . also , under conditions of destructive interference , the particle and slab will never simultaneously be observed . classically , only the pulse of light will never be observed . a simultaneous measurement , however , is difficult to perform . a much simpler method involves only measuring the reflected particle . such a one - body measurement can reveal the quantum state of the reflector due to the effects of the reflector s coherence length on the two - body pdf , from which the one - body marginal pdf is derived . this manifestation of the macroscopic quantum state of the reflector on the interference measured only on the particle may be a practical way to verify that the mesoscopic / macroscopic reflector is indeed in a quantum state and thereby extend the quantum - classical boundary to larger masses . finding examples of mesoscopic / macroscopic quantum phenomena has been a topic of interest since early in the development of quantum mechanics , particularly with regard to validation issues . this continues to be a subject of interest . consider next some of the arguments used to explain why such effects are not observed with mesoscopic / macroscopic objects , in relation to the two - body treatment of particle reflection : ( 1 ) in an introductory quantum course , the lack of such evidence is often justified via the double slit experiment , where the fringe spacing becomes imperceptible as the mass increases . in the comments after eqn . [ eq : fringespacing ] , an explanation is given as to why the mirror fringe spacing does not shrink as its mass increases . ( 2 ) another concern is if quantum correlation is indeed generated . the initial particle - mirror separable system becomes correlated in reflection as a consequence of conservation of energy and momentum . the discussion in section [ sec : summary ] demonstrates why interferometric quantum correlation is a requisite in satisfying these conservation laws . ( 3 ) the difference in wavevectors of the interfering mesoscopic / macroscopic mirror substates , due to reflecting a microscopic particle , is much smaller than the spread in wavevectors of which the mirror wavegroup is composed , due to thermal motion . this effect is shown to not destroy interference in section [ sec : summary ] . ( 4 ) more recently , decoherence of the macroscopic superposition via interaction with the environment has been used to explain the lack of interference . mitigation of this and other decoherence mechanisms are discussed in section [ sec : decoher ] . this work examines the assumptions of these arguments in the context of reflection of a microscopic particle from a mesoscopic / macroscopic mirror . a synopsis of this work is as follows : first the schrdinger equation is solved for the two - body energy eigenstates in reflection from a moving mirror . wavegroups formed from these solutions are then used to illustrate interference in reflection with an emphasis on variations in coherence lengths of the mirror . coherence transfer from the particle to the mirror is also demonstrated and related to classical reflection . two models are next developed for two - body reflection of a particle from a slab , whose classical analog is that of reflection of a pulse of light from a moving thin film . the effects of the coherence lengths of both the particle and slab are then discussed . measurements of either the particle but not the slab and the slab but not the particle are then considered . finally , decoherence issues , as well as methods to measure a mesoscopic / macroscopic reflector in a superposition state , are addressed . the goal of this paper is to illustrate the unique interferometric properties of correlation interferometry in reflection . these properties are most easily understood using the example of a microscopic particle reflecting from a mesoscopic / macroscopic mirror . however , the focus on this example is not intended as a practical experimental proposal . observation of such correlated interference will most likely first occur with a microscopic particle reflecting from a `` microscopic '' mirror . before reflection , the solution to the schrdinger equation for the non - interacting particle - mirror state is @xmath1 , \label{eq : psiseparable}\end{aligned}\ ] ] where @xmath2 and @xmath3 are the particle and mirror positions along the x - axis while @xmath4 and @xmath5 are the wavevectors for the particle and mirror respectively ; @xmath6 and @xmath7 with masses @xmath8 , @xmath9 , and initial velocities @xmath10 and @xmath11 , respectively . a wavegroup constructed from such uncorrelated particle - mirror states then leads to predictions about the probability of simultaneously finding the particle at @xmath2 and mirror at @xmath3 . the particle - mirror interaction is modeled as a moving delta function potential where reflection is assumed to occur at the center of mass ( cm ) of the mirror with the schrdinger equation given by @xmath12 \notag + i\partial_{t})\psi=0 . \label{eq : schreqn}\end{aligned}\ ] ] where square brackets are used to indicate the argument of a function . the mirror reflectivity , related to @xmath13 , goes to infinity for a lossless mirror . the solution yields an energy eigenstates for the particle - mirror interaction ( ` harmonic wavefunctions ' ) , for which neither the particle nor the mirror is localized . a separable solution to the two - body schrdinger equation results from a transformation to the center of mass ( cm ) and relative ( rel ) system ( not to be confused with the cm of the particle or mirror ) . this does not change the total energy , @xmath14 . the transformed schrdinger equation becomes @xmath15 \notag + i\partial_{t})\psi[x_{cm},x_{rel},t]=0 \label{eq : scheqncm}\end{aligned}\ ] ] where @xmath16 , @xmath17 , @xmath18 , and @xmath19 . using @xmath20=\psi_{cm}\psi_{rel } \notag \\ = e^{-i e_{cm } t/\hbar } u[x_{cm } ] e^{-i e_{rel } t/\hbar}u[x_{rel } ] \notag , \label{eq : schtot}\end{aligned}\ ] ] reduces the schrdinger equation to two ordinary differential equations : @xmath21}{dx_{cm}^{2 } } = e_{cm } u[x_{cm } ] \label{eq : scheqode1}\ ] ] @xmath22}{dx_{rel}^{2}}+\beta \delta[x_{rel } ] = e_{rel } u[x_{rel } ] . \label{eq : scheqode2}\ ] ] the particle - mirror solution must vanish at @xmath23 to satisfy the boundary condition at the mirror and not exist for @xmath24 ( or @xmath25 ) since the particle can not move through the mirror ( for the uncorrelated incident state , however , there is no interaction and the particle can move past the mirror ) . in this transformed system , a solution is constructed from the superposition of incident and `` reflected '' wavefunctions , @xmath26 , \label{eq : superposition}\ ] ] where @xmath27 $ ] is the unit step function . the only difference between the arguments of the two exponentials is the sign of the relative wavevector @xmath28 which , due to reflection in the relative coordinate , is negative . that is , @xmath29 where the initial velocities must allow reflection to occur . the solution to eqn . [ eq : scheqode1 ] is given by @xmath30 the complete solution is then @xmath31=\psi_{cm}\psi_{rel}$ ] . in this separable system the probability density is @xmath32 . \label{eq : separableinterference}\end{aligned}\ ] ] the cm - rel transformation does not give a solution for a physically realizable system . there exists no particle with a reduced mass , for example . nevertheless , its utility lies in inverting the solution from cm and relative substates into particle and mirror substates . this yields a correlated particle - mirror state . an example of this procedure is found in the solution to the hydrogen atom where the schrdinger equation is first transformed from the laboratory to the cm - relative coordinates yielding uncorrelated substates . however , expressing this result in the electron and proton substates reveals a correlation between them @xcite . this change of partition is accomplished by the following substitutions in the separable solution given in eqn . [ eq : separableinterference ] : @xmath33 , @xmath34 , @xmath19 , @xmath18 , @xmath35 , and @xmath36 . these constraints transform @xmath37 in eqn . [ eq : superposition ] into the non - interacting particle - mirror solution of eqn . [ eq : psiseparable ] . transformation of the energy component of each reflected substate , given by @xmath38 with @xmath39 for the reflected particle and @xmath40 with @xmath41 for the mirror , is consistent with that of reflection classically . with this change in coordinates eqn . [ eq : separableinterference ] becomes @xmath42 . \label{eq : entangledinterference}\end{aligned}\ ] ] this is similar to gottfried s result for the interference obtained in the correlation between two particles produced in a momentum - conserving decay after each has traversed separate double slits @xcite . note also that the interference given in eqn . [ eq : entangledinterference ] couples particle and mirror variables , illustrating how many - body states interfere with themselves rather than each substate interfering only with itself @xcite . measurement of such a prediction involves preparation of the initial state , adjusting the instruments to measure _ both _ positions @xmath2 and @xmath3 simultaneously , making the measurement , and then repeating this procedure over an ensemble to build a distribution . to illustrate how this leads to a measurement of the mirror fringe spacing , first the apparatus is set to always measure the particle at fixed position @xmath2 while the mirrors cm is measured at different positions @xmath3 for different members of the ensemble . using the approximation @xmath43 in eqn . [ eq : entangledinterference ] leads to interference fringes _ for the mesoscopic / macroscopic mirrors cm _ which vary from maximum to minimum through a distance @xmath44 similarly , by always measuring the mirror at fixed @xmath3 while varying the measurement position of the particle @xmath2 for different members of the ensemble , this approximation leads to interference fringes _ for the microscopic particle s cm _ which vary from maximum to minimum through a distance @xmath45 . for a static mirror both the mirror and particle fringes are spaced at half the debroglie wavelength of the _ particle _ , which can be up to @xmath0 m for ultra cold atoms @xcite . one issue predicted here , fundamental to extending the quantum - classical boundary , is that of interferometric effects which do not become imperceptible in the limit of large mirror mass . such effects are surprising in comparison with the imperceptible fringe spacing for a massive particle traversing a double slit . to understand why fringe spacing in reflection is robust with respect to variations in @xmath9 consider simplifying the two - body state into two one - body states . the first consists of the wavefunctions of the incident and reflected particle . the second consists of the mirror wavefunctions before and after reflection of the particle , both of which travel in the same direction but with different momenta . in the case of a mirror whose speed is close to zero , superposition of these particle states results in a `` standing wave '' with nodes or fringes spaced at about half the particle debroglie wavelength . superposition of the moving mirror states , on the other hand , forms a fringe spacing commensurate with the difference in momentum between the incident and reflected _ particle _ states . the larger the mass of the mirror the smaller the change in velocity of the mirror upon reflection . yet the difference in momentum between the mirror before and after reflection remains the same , depending only on the change in momentum of the particle and _ not _ on the mass of the mirror . it is this difference in mirror momenta which leads to the phase difference @xmath46 when superposing these two mirror wavefunctions . that is , the mirror fringe spacing is only determined by the change in particle momentum since conservation of momentum in reflection requires @xmath47 . while this simplified model , reducing a two - body state into two one - body states , demonstrates why interference for a mesoscopic / macroscopic mirror mass does not become imperceptible as @xmath48 , it does not account for the correlation given by the exact solution in eqn . [ eq : entangledinterference ] , which is inherently a two - body effect . double slit interference with a massive particle , on the other hand , superposes two one - body states with the same momentum whose difference in phase , @xmath49 , is due to the difference in path lengths from each slit to the measurement point times an extremely large wavevector . this results in an imperceptible fringe spacing . the probability of measuring the particle at @xmath50 and the mirror at @xmath51 is given by @xmath52 dx_{1 } dx_{2}$ ] with the joint pdf determined by the solution of equation [ eq : schreqn ] as @xmath53 . using this equation , conservation of probability can then be expressed locally as , @xmath54}{\partial t}+ \frac{\partial j_{1}[x_{1},x_{2},t]}{\partial x_{1}}+\notag \\ \frac{\partial j_{2}[x_{1},x_{2},t]}{\partial x_{2}}=0 , \label{eq : consprob}\end{aligned}\ ] ] where @xmath55=\hbar ( \psi^ { * } \partial_{x_{1 } } \psi-\psi \partial_{x_{1 } } \psi^{*})/(2 i m)$ ] and @xmath56=\hbar ( \psi^ { * } \partial_{x_{2 } } \psi-\psi \partial_{x_{2 } } \psi^{*})/(2 i m)$ ] . while the expressions for these current densities appear similar to that for one particle systems there are subtle but important differences for a two - body system @xcite . multiplying equation [ eq : consprob ] by @xmath57 , integrating over the segment from @xmath58 to @xmath59 along the x - axis ( @xmath60 and @xmath61 ) , and then rearranging terms yields a solution to equation [ eq : consprob ] if @xmath62 dx_{1 } dx_{2 } + \notag \\ \int_{a}^{b } ( j_{1}[b , x_{2},t]-j_{1}[a , x_{2},t ] ) dx_{2 } \notag \\ + \int_{a}^{b } ( j_{2}[x_{1},b , t]-j_{2}[x_{1},a , t ] ) dx_{1}=0 \notag . \label{eq : cons2}\end{aligned}\ ] ] this is most easily interpreted using an @xmath63 by @xmath63 rectangular region aligned along the @xmath2 and @xmath3 axes . the time rate of change in probability within this region is determined by the change in flux of probability out along the @xmath2 axis ( @xmath64-j_{1}[a , x_{2},t]$ ] ) plus that along the @xmath3 axis @xmath65-j_{2}[x_{1},a , t])$ ] . since these fluxes in general vary spatially they have to be integrated over this variation along the rectangular boundaries . numerical integration of the wavegroup results presented below are consistent with conservation of probability . to better understand the practical consequences of these results , wavegroups are next formed from a superposition of the incident and reflected ` harmonic wavefunctions ' ( given in eqn . [ eq : superposition ] ) expressed in terms of the correlated particle and mirror substates . an analytic expression for such wavegroups can be obtained for a gaussian distribution in wavevector components @xmath4 and @xmath5 . for the mirror , this is proportional to @xmath66/(2 \delta k^{2})$ ] , where the peak of the distribution is at @xmath67 and @xmath68 is its width , while for the particle , this is proportional to @xmath69/(2 \delta k^{2})$ ] , where the peak of the distribution is at @xmath70 and @xmath71 is its width . the incident wavegroup propagates in the @xmath72 plane along a line whose slope is determined by a ratio of the group velocities of each substate and spreads due to dispersion independently in each direction . . the lower pdf waveform moves toward the diagonal white line , corresponding to @xmath23 , then reflects in the middle snapshot where the incident and reflected two - body wavefunctions ` overlap ' , and finally it moves away from the diagonal in the upper snapshot . the upper left inset is a schematic of the ` classical ' analog before reflection while the upper right inset is that after reflection with initial and final particle and mirror velocities @xmath10 , @xmath11 , @xmath73 , and @xmath74 respectively . there is no classical analog for the middle snapshot . ] in fig . [ fig : interference ] snapshots of the two - body probability density function are shown at three times for @xmath75 , @xmath76 , and @xmath77 . a slice of fig . [ fig : interference ] for @xmath78 along the @xmath3 coordinate is shown in fig . [ fig : fringespacing ] ( the solid line ) along with a slice of this fig . for @xmath79 along @xmath2 ( the dashed line ) for different bandwidth wavegroups . illustrates that the fringe spacing for the particle and mirror substates with narrow bandwidth wavegroups , are essentially the same for the particle and mirror substates , as discussed for the approximation @xmath80 following eqn . [ eq : fringespacing ] . which show the fringe spacing along the @xmath3 axis for @xmath78 ( dashed lines ) and along the @xmath2 axis for @xmath79 ( solid lines ) . the @xmath2 axis has been inverted to display both the dashed and solid lines together . although each graph has @xmath76 the value of @xmath68 increases sequentially by a factor of @xmath81 from the front to the back . ] verification of correlated interference requires simultaneous measurement of the particle and cm of the mirror with instruments having a spatial resolution smaller than the fringe spacing . for a static mirror this spacing is half the debroglie wavelength of the particle , which at @xmath82 for ultracold atoms @xcite satisfies this requirement while it is dubious at @xmath83 for slow neutrons @xcite . another constraint on the interference illustrated in fig . [ fig : interference ] is that the fringe visibility function must be non - zero . that is , the incident and reflected particle - mirror wavegroups must ` overlap . ' the interference fringes are then determined predominately by a superposition of ` harmonic wavefunctions ' @xcite . for example , the interference shown in fig . [ fig : interference ] is determined predominately by eqn . [ eq : entangledinterference ] when the wavegroups ` overlap ' in the center snapshot . the longitudinal coherence lengths for both the particle and mirror are greater than the fringe spacing in this figure . in the upper snapshot there is neither ` overlap ' nor such interference . the fringe visibility function is non - zero if each wavegroup substate ` overlaps ' within approximately a longitudinal coherence length @xcite , which is given by @xmath84 @xcite . for particle substates this can be @xmath85 for ultracold atoms @xcite or @xmath86 for slow neutrons @xcite . if the uncertainty in the mirror velocity is determined by its thermal equilibrium with the environment then @xmath87 yielding @xmath88 . this expression is consistent with results for ultra - cold atoms in a bose - einstein condensate @xcite . in the interference region , simultaneous measurement is approximately confined both to a spatial region determined by the two coherence lengths and a temporal region given by the time during which the wavegroups overlap . the former is small for mesoscopic / macroscopic mirror masses while the duration of the interference is essentially determined by the speed of the particle and its coherence length which is approximately equal to @xmath89 for a static mirror when @xmath90 . one method , discussed below , to reduce these coherence length limitations is with multiple such two - body states . first however , the effect of variations in mirror coherence lengths on correlated interference is illustrated . [ fig : spreadvelocity ] illustrates how variation in the longitudinal coherence length of the mirror substate affects the particle - mirror interference with fixed particle substate coherence length . part a shows a longer mirror coherence length than is used in fig . [ fig : interference ] while parts [ fig : spreadvelocity]b through [ fig : spreadvelocity]d progressively reduce the coherence length of only the mirror substate . one might expect that the small coherence length associated with a mesoscopic / macroscopic mirror mass would not allow for the interference shown in fig . [ fig : spreadvelocity]a or [ fig : spreadvelocity]b . in fig . [ fig : spreadvelocity ] d a small mirror coherence length along with a large mirror recoil prevents wavefunction overlap over a range of @xmath2 values ( @xmath91 ) where it was present before . nevertheless , a slice along the @xmath3 axis for measurement of the particle at these values of @xmath2 indicates a splitting of the mirror substate into two states which do not overlap and are therefore distinguishable . this splitting is a consequence of two ways that the particle could have reached @xmath2 . it could have come from the incident _ or _ reflected particle wavegroup substates since the position of the particle is not known to within its large coherence length . as the _ mirrors _ coherence length increases , the wavefunctions associated with these two ways overlap and generate correlated interference as shown in fig . [ fig : spreadvelocity]a . as the mirrors coherence length decreases , the position of the mirror before reflection is distinguishable from that after reflection due to mirror recoil , which results in _ no interference of either the particle or mirror _ , as shown in fig . [ fig : spreadvelocity]d for @xmath91 . that is , measurement of the particle at @xmath92 and mirror at @xmath79 is consistent with conservation of momentum for the particle and mirror not to have reflected while measurement of the particle at @xmath92 and mirror at @xmath93 is consistent with conservation of momentum for the particle and mirror to have reflected . the short coherence length of the mirror relative to the recoil distance then results in distinguishable two - body states . the condition for interference in a two - body system therefore involves the coherence of both bodies . but with different spreads in mirror velocities for a fixed spread in particle velocities while all other parameters are the same . @xmath94 , and @xmath95 for @xmath96 in parts a , b , c , and d respectively . note the change of scale in graph d. ] after reflection the spatial width of the mirror wavegroup substate is exchanged with that of the particle wavegroup substate when @xmath97 . this is most easily seen by constructing a particle - mirror wavefunction with different bandwidths for the particle and mirror wavegroup substates , shown in fig . [ fig : coherencetransfer ] which is a contour plot of joint pdfs similar to figure [ fig : interference ] but without the middle snapshot in the interference region . the solid and dashed contours correspond to @xmath98 and @xmath99 respectively with the spread in velocities given by @xmath100 . this result can be understood by comparing classical and quantum reflection . in a one - dimensional classical collision , conservation of energy and momentum require an exchange of particle - mirror velocities independent of either velocity for @xmath101 . this is manifest quantum mechanically in the exchange of commensurate wavefunction parameters @xmath4 and @xmath5 between the incident and reflected two - body wavefunctions . if an incident particle substate , consisting of only one harmonic component ( corresponding to speed @xmath10 ) reflects from a mirror substate with many velocity components , then each harmonic component of the mirror substate ( corresponding to different values of @xmath11 ) reflects the particle substate and therefore acquires velocity @xmath10 while the reflected particle substate acquires different velocity values for each reflected component of the mirror substate . this results in the reduction of the mirror bandwidth and an increase in the particle bandwidth , which is manifest in fig . [ fig : coherencetransfer ] as the exchange of incident and reflected wavegroup shapes . it also is responsible for the distortion of the reflected wavegroup shape in fig . [ fig : interference ] . but without the interference or `` overlap '' region snapshot , illustrating coherence transfer for two different particle - mirror mass ratios . the solid and dashed lines are for a spread in velocities given by @xmath100 with @xmath98 and @xmath99 , respectively . ] experimentally confirming coherence transfer involves reflection with disparate coherence lengths . on a microscopic scale , this could be observed between an atom which has been cooled to ultracold temperatures retro - reflecting from the same type of atom which has a shorter coherence length . the transfer of coherence might then be monitored not by a measurement of the pdf but rather by its coherence properties via interferometry . a more useful implementation for increasing the coherence length of larger objects could involve a group of bose - einstein particles ( the condensate ) , whose total mass is similar to that of the mirror , retro - reflecting`coherently ' from a mirror as ` one particle ' rather than a collection of individual particles . to mitigate both spatial and temporal fringe localization in reflection due to small particle and mirror coherence lengths , consider interference of multiple such two - body states : a one - dimensional example of which is the interference from the reflection of a neutron by each of the two surfaces of a moving aluminum ` slab ' . the two - body pdfs associated with two models of such reflection are calculated below . the first , presented in the next section , simply approximates each weak surface reflection as being due to a delta function potential which allows for both transmission and weak reflection . these two delta function potentials are fixed to and symmetrically offset from the cm of the slab ( of length @xmath102 ) by a distance @xmath103 , with multiple reflections neglected . the results from this simple model then corroborate those from a more accurate model , given in section [ sec : slab ] , involving a two - body treatment of a particle interacting with a finite barrier or finite well . the delta function ` beamsplitter ' model is included to focus attention on the physics associated with generating multiple two - body states . the more accurate finite barrier or well treatments conceal the fundamental physics but yield similar results . in essence , the latter incorporates multiple reflections along with the appropriately modified wavevectors between the surfaces of the slab while the former does not . the results of the simpler delta function model approach those of the finite barrier - well model as the reflections from the slab interfaces become weaker . for completeness , the results of the particle trapped in a moving infinite well are included in subsection [ sec : well ] . while the finite barrier or well examples discussed below deal only with unbound states , the infinite well is an example of the two - body formalism for a bound state . the classical analogy for the neutron reflecting from each of the two surfaces of a moving aluminum ` slab ' is that of a short pulse of light reflecting from a moving thin film . as the film s thickness , speed , or the average wavelength of the incident pulse changes , the reflected pulse intensity then varies as a consequence of the interference between the reflected waves from the two interfaces . classically , such thin film interference mitigates both the spatial and temporal fringe localization associated with reflection of a pulse of light from a moving mirror ( the analog of the two - body quantum treatment of a particle reflecting from a moving mirror ) . interference between the incident and reflected two - body pulse from the moving mirror is localized spatially by the coherence length of the pulse . these fringes are also temporally constrained by the pulse duration . however , interference of the two reflected light pulses , one from each surface of the thin film , dramatically reduces this fringe localization for two reasons : first the two reflected light pulses travel in the same direction thereby maintaining spatial overlap and second they experience the same doppler shift in reflecting from the two interfaces which move at the same speed resulting in fringes which are not time dependent ( this is not the case for interference of the incident and reflected pulses from a moving mirror ) . for similar reasons , interference of the two - body wavegroup states associated with reflection from each surface of the slab mitigates the fringe localization associated with reflection from a mirror however , the two - body interference is now between two wavegroups , each analogous to the one shown in the upper snapshot of fig . [ fig : interference ] and each of which corresponds to a reflection from only one surface of the slab and both of which move in the same direction thereby maintaining overlap . the classical treatment yields interference only for the pulse of light and not the mirror from which it reflects . the two - body quantum treatment presented below yields predictions about correlations in the measurements of both the neutron and slab . destructive interference of these wavegroups then corresponds to no probability of measuring the neutron and slab along the overlap region . a change in the slab s thickness , speed , or the average wavelength of the incident particle then results in a variation of the probability of simultaneously measuring the neutron and slab between zero and some maximum value , corresponding to constructive and destructive interference . a pulse of light reflecting from a thin film can also generate a standing wave within the film under appropriate conditions . similarly , two - body wavefunctions also generate such a `` fabry - prot '' resonance which is illustrated below . modifying @xmath13 to split rather than totally reflect the incident wave while including an offset in the delta function argument to model the surface which is offset from the cm of the slab yields the two - body reflected wavefunction . this eigenstate of energy or ` harmonic solution ' , for each surface reflection , is then superposed to form wavegroups associated with refection from that surface . rather than the waveform of the center snapshot in fig . [ fig : interference ] , a waveform analogous to that in the upper snapshot is the focus of this discussion . there are now two such reflected neutron - slab wavegroups , one for each slab surface , which interfere . since the neutron either reflects from the first surface or transmits through the first to the second , from which it then reflects , both reflected neutron substates have the same distribution of velocities . since in each case momentum is transferred to the slab only once for each of these two possible paths ( reflection from either the first _ or _ second surface of the slab ) , two associated slab substates are generated , each with the same velocity distribution . this results in two reflected neutron - slab wavegroups traveling at the same speed and direction but offset . interference then requires wavegroup ` overlap ' within a coherence length of both the neutron and slab substates . an example of a lack of overlap in the mirror substate , resulting in no two - body interference , was discussed in section [ sec : bandwidth ] and illustrated in fig . [ fig : spreadvelocity]d . this requirement , applied to the slab substate , is that the cm of the two reflected slab substates separate no further than their coherence lengths . this offset is estimated by the extra slab speed , beyond @xmath11 , acquired from the first reflection multiplied by the difference in particle reflection times between the two slab surfaces . assuming @xmath43 and @xmath104 this offset , @xmath105 , then must be less than @xmath106 , which if determined by thermal equilibrium requires @xmath107 next consider the constraint on @xmath108 . when a neutron wavegroup of spatial width @xmath109 ( where dispersion in reflection is neglected ) retro - reflects a distance @xmath110 from the cm of the slab , with @xmath110 much greater than both the wavegroup width @xmath109 and the slab length @xmath102 ( @xmath111 ) , ` overlap ' of the neutron wavegroups which have reflected from both surfaces occurs . for a near stationary slab the reflected neutron wavegroups are offset by about a distance @xmath112 , interference from which requires a neutron coherence length exceeding @xmath112 . an example which satisfies this constraint is a neutron of coherence length @xmath113 m @xcite reflecting from the two surfaces of an aluminum slab of thickness @xmath114 m ( there is little constraint on the area of this slab and therefore its mass ) . these parameters easily satisfy the overlap constraints for both the mesoscopic / macroscopic slab at room temperatures and the neutron . having achieved overlap , the resulting interference is then determined essentially by the harmonic wavefunctions yielding the _ constant _ two - body reflected joint probability density @xmath115 . \label{eq : slabinterference}\ ] ] assuming @xmath43 and @xmath104 yields @xmath116 . \label{eq : slabintapprox}\ ] ] the slab _ and _ reflected neutron will never be simultaneously observed when @xmath117 @xcite . such non - local two - body interference is similar to that of a pulse of light retro - reflecting from a thin film where destructive interference depends neither on the locations of the detector nor the thin film . indeed , in the limits just given these two cases are mathematically identical if the photon wavelength is replaced with the particle s debroglie wavelength . the pdf for reflection from the slab is illustrated for wavegroups in fig . [ fig : slab ] using parameters similar to those above : @xmath113 m , @xmath118 m , @xmath119 m/s , @xmath120 kg , and a slab coherence length determined by @xmath121 at @xmath122 . the neutron reflects from the first slab surface at @xmath79 , @xmath78 , and @xmath123 . to illustrate the non - local nature of the interference , three snapshots are chosen for times which allow the center of the wavegroup to change position by four orders of magnitude in both the neutron and slab coordinates . since the spatial size of the neutron - slab pdf is limited by small coherence lengths , a graph with a spatial region encompassing all of these points , while still showing the wavegroup structure , can not easily be generated . instead ` blow - ups ' in the @xmath72 plane of small regions around these points are shown . the snapshots in the upper row have a span in @xmath2 of @xmath124 m and a span in @xmath3 of @xmath125 m with the approximate centers of each graph located at @xmath126 m and @xmath127 m . the middle row of graphs have a span in @xmath2 of @xmath128 m and a span in @xmath3 of @xmath129 m with the approximate centers of each graph located at @xmath130 m and @xmath131 m . the lower row of graphs have a span in @xmath2 of @xmath132 m and a span in @xmath3 of @xmath133 m with the approximate centers of each graph located at @xmath134 m and @xmath135 m . the increase in span is needed since the wavegroup expands over these time scales . the vertical scale therefore decreases from top to bottom rows but is the same within each row so that pdf heights can only be compared within a row . the only parameter that differs between the graphs in the left and right columns of fig . [ fig : slab ] is the neutron velocity before reflection . it is @xmath136 m / s for the left and @xmath137 m / s for the right column respectively . these two neutron speeds result in either constructive or destructive interference ( left or right columns respectively ) of the neutron - slab wavegroups which are offset by @xmath138 due to reflection from the two slab surfaces . as this offset decreases ( e.g. by decreasing @xmath102 ) or @xmath108 increases , the pdf goes to zero approaching the destructive interference predicted in eqn . [ eq : slabintapprox ] ( since @xmath43 and @xmath104 ) . similar calculations also confirm the predictions that there is little effect on the interference from increasing @xmath9 while changes in @xmath10 have a significant effect . note also that the small size of the slab coherence length , @xmath139 m , is much larger than the offset due to reflection from the two slab surfaces , @xmath140 m , and therefore has little effect on the destructive interference . increasing the mass of the slab changes only the span in the @xmath3 coordinate of these graphs but not the interference . the parameters in fig . [ fig : slab ] are _ not _ chosen to illustrate effective wavegroup destructive interference but rather to show coherence length effects on the interference using experimentally realizable values . and @xmath2 respectively ) after reflection . the left column graphs differ from the right only in the speed of the incident neutron , illustrating constructive and destructive interference respectively . ] the model for a neutron traversing a slab of matter which incorporates multiple reflections and variation in the wavevector inside the slab is determined from the schrdinger equation given by @xmath141 \notag \\ + i\partial_{t})\psi[x_{1},x_{2},t]=0 , \label{eq : scheqn}\end{aligned}\ ] ] where @xmath142 $ ] is now the potential energy associated with either the well or barrier . of interest is a solution which yields an energy eigenstate for the particle - well or barrier interaction , for which neither the particle nor the well or barrier is localized . wavegroups are then constructed from these states . a separable solution to this two - body schrdinger equation is obtained in a manner similar to that outlined in section [ sec : theoryparticlemirror ] . the center of mass and relative coordinate transformation yields solutions which are matched at the boundaries and then transformed to the particle - well coordinates . the infinite well potential is presented to illustrate the application of the two - body formalism to a bound state . then in subsection [ subsec : barrier ] calculations are outlined for a particle interacting both with a finite barrier and a finite well in unbound states . these provide a more accurate model for a neutron traversing a slab of matter than that just discussed . the boundary condition for the infinite potential well , of width @xmath102 , is @xmath143\rightarrow 0 $ ] . the relative wavefunction does not exist outside the well for @xmath144 and @xmath145 . a solution to eqn . [ eq : scheqn ] for a particle which has reduced mass is then @xcite @xmath146~\theta[x_{rel},d ] , \label{eq : psiwell2}\ ] ] where @xmath147=(\theta [ x_{rel}+d]-\theta [ x_{rel}-d])$ ] has value one within the well and zero everywhere else with n being the number of nodes . the solution to the center of mass ode ( the same as eqn . [ eq : scheqode1 ] ) is @xmath148 the complete solution is then @xmath31 \propto \psi_{cm}\psi_{rel}$ ] . again , we are interested in measurements of the particle and well rather than measurements of two transformed `` objects , '' one with a reduced mass and the other with the total mass of the system . expanding the sine function in eqn . [ eq : psiwell2 ] into exponential form results in wavefunctions traveling in opposite directions . transforming back to the particle - well coordinates , the momenta and energies of the particle and well differ in magnitude in these two directions . the modes of the well , characterized by the number of nodes @xmath149 , are determined in the cm - rel coordinates by @xmath150 . in the particle - well coordinates this constrains the incident values of particle and well velocities . that is , given an initial well velocity @xmath11 only the following initial particle velocities are allowed , @xmath151 in the particle - well coordinates the wavefunction s phase , @xmath152 , needs to have its temporal part parsed into the kinetic energy of the particle and the kinetic energy of the well . this is done as in the previous calculations . the resulting kinetic energies of the particle and well differ as the particle moves in opposite directions . these energies and their associated momenta are those expected of a classical particle reflecting from the moving well walls . the finite barrier or well ( referred to as only the barrier ) has potential energy @xmath153 ( positive for the barrier and negative for the finite well ) and extends over a distance @xmath102 . this divides space into three regions : before the barrier or `` before '' , in the barrier region or `` barrier , '' and after the barrier or `` after . '' solutions are first obtained for these three regions in the cm and rel coordinates by solving eqns . [ eq : scheqode1 ] and [ eq : scheqode2 ] . the solution to eqn . [ eq : scheqode2 ] before the barrier consists of incident and reflected wavefunctions given by @xmath154 where @xmath155 . the solution in the barrier - finite well region also consists of incident and reflected wavefunctions given by @xmath156 where @xmath157 . the solution after the barrier consists only of a transmitted wavefunction given by @xmath158 where @xmath159 . the boundary conditions are continuity of the wavefunctions and their derivatives with respect to @xmath160 at @xmath161 . these then constrain the coefficients @xmath162 , @xmath163 , @xmath164 , and @xmath165 with @xmath166 . again , the solution to eqn . [ eq : scheqode1 ] is given by eqn . [ eq : psiwell1 ] . the complete solution is then @xmath31 \propto \psi_{cm}\psi_{rel}$ ] . since we are interested in predictions about measurements of the particle and barrier rather than measurements of the reduced mass and total mass `` objects , '' a transformation from the relative and center of mass to the particle - barrier coordinates is required . in the particle - barrier coordinates the wavefunction s phase , @xmath152 , needs to have its temporal part parsed into the kinetic energy of the particle and the kinetic energy of the barrier . this procedure is the same as it is for the particle - mirror and particle - infinite well two - body states as described above . however , the product of the potential energy part of the total energy with time @xmath167 in the barrier region must be added to the phase . such a term has no measurable effect as will be shown next . the wavefunction in the barrier , expressed in the particle - well coordinates , is then @xmath168 where @xmath169 and @xmath170 contain the spatial terms in the phase and are functions of @xmath171 and @xmath172 . the temporal terms contain the kinetic energy for the particle and barrier moving to the right and left , @xmath173 and the potential energy @xmath153 . the kinetic energy terms are functions of @xmath174 , and @xmath172 . the potential energy term @xmath175 is a common factor of both the incident and reflected wavefunctions in the barrier . since the pdf is generated from the wavefunction multiplied by its complex conjugate , such common factors have no effect on the pdf . there is then no need to associate the potential energy part of the total energy with either the particle or the barrier or as a separate term in the phase , @xmath176 . the potential energy part of the total energy has observable consequences only parametrically within the momenta and kinetic energies of the particle and barrier . this simple division of the momentum and energy of the particle in the barrier can be contrasted with that of the stress - energy tensor for an electromagnetic wave traversing a dielectric slab @xcite . wavegroups are next formed from a gaussian superposition of the two - body energy eigenstates for the particle and infinite well or finite particle and barrier - well described above . correlated interference is a consequence of any such superposition . however , we focus the following discussion on the subsets of correlated interference effects which deal with the superposition of two such wavegroups . in particular , emphasis is given to the two - body quantum analogy of a pulse of light reflecting from a thin film : interference from the two - body wavegroups which have reflected from the two surfaces of the slab . at the risk of distracting from this main topic , correlated interference from multiple reflections when such a wavegroup interacts with an infinite barrier are also discussed . the wavegroup for the particle and barrier or finite well is calculated using a gaussian superposition of the energy eigenstates given in subsection [ subsec : barrier ] . unfortunately , the coefficients @xmath162 , @xmath163 , @xmath164 , and @xmath165 of eqns . [ eq : psibefore ] , [ eq : psibarrier ] , and [ eq : psiafter ] depend in a non - trivial manner on the variables of integration . the resulting integrals can not be determined in closed form . to facilitate the calculation , the following sums will replace these integrals : @xmath177 \notag \label{eq : barrierwavegroup}\end{aligned}\ ] ] where the peak of the barrier velocity distribution is at @xmath178 , @xmath179 is its width , and the sum is from an initial barrier velocity @xmath180 to a final velocity @xmath74 . summing over the particle velocity distribution yields the wavefunction for the wavegroup given by @xmath181 where the peak of the particle velocity distribution is at @xmath182 , @xmath183 is its width , and the sum is from an initial particle velocity @xmath184 to a final velocity @xmath73 . two cases are now treated : the unbound two - body wavegroup state of a particle and finite well and a similar unbound state of a particle and finite barrier . [ [ sec : barriergreater ] ] particle traversing a finite well : @xmath185 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + consider next a particle interacting with a finite well whose sum of initial kinetic energies in the relative coordinate system is greater than the well pe . the size of the particle substate wavegroup is chosen to be a few times larger than the finite well width @xmath102 . [ fig : barrier_negenergy ] shows results of the pdfs for three sequential snapshots taken at equal time intervals progressing along the dashed line from the lower left to upper right and upper left . the analogous classical positions of the particle and finite well for particular snapshots are illustrated in the insets . the barrier boundaries occur at the diagonal white lines , corresponding to @xmath186 . the parameters used in fig . [ fig : barrier_negenergy ] are @xmath187 , @xmath188 , and @xmath189 while the @xmath190 using the average value of @xmath191 for the wavegroup particle and well distributions . one category of correlated interference , which we call type i , occurs when the incident and reflected two - body wavefunctions , traveling in opposite directions in the @xmath192 plane , ` overlap . ' this is illustrated in fig . [ fig : barrier_negenergy ] for weak reflection by the fringes of the middle snapshot just to the left of the barrier or line @xmath193 and also by the fringes in fig . [ fig : interference ] . these fringes are spaced by about half the debroglie wavelength of the _ particle _ for @xmath194 and @xmath195 as are the similar correlated interference fringes in the two - body reflection of a particle from a mirror given in eqn . [ eq : entangledinterference ] . however , the interaction generates another form of correlated interference when the reflected wavegroups , one from each barrier surface , interfere as they travel along the _ same _ direction in the @xmath192 plane . this new category of correlated interference , which is referred to as type ii , is illustrated in fig . [ fig : barrier_negenergy ] by the peak labeled @xmath59 . this is the same type of interference from reflections at the two surfaces of the slab modeled as two delta function in section [ sec : twodelta ] . it is similar to the classical interference of a pulse of light reflecting from a thin film . changing only the barrier spacing generates an oscillation in the pdf for peak @xmath59 analogous to that found in the interference of a pulse of light reflecting from a thin film whose thickness varies . that is , this peak goes through constructive and destructive interference from the two barrier reflections when the wavegroup size is much larger than that of the barrier and the spacing @xmath102 is varied . as time progresses peak @xmath59 maintains this interference as it travels in the @xmath192 plane . these are the same characteristics of the interference found in the simple two delta function model of reflection from a slab given in section [ sec : twodelta ] and shown in fig . [ fig : slab ] . this differs from the type i correlated interference associated with the fringes shown in the middle snapshot of fig . [ fig : barrier_negenergy ] which is localized to a small temporal and spatial region . this figure illustrates yet another form of correlated interference , referred to as type iii : that from multiple reflections from the two barrier edges , which is shown as the fringes within the well of the middle snapshot of fig . [ fig : barrier_negenergy ] . these are analogous to the standing wave formed in a thin film or optical cavity . . the first snapshot generates peak @xmath58 while peaks @xmath59 and @xmath196 comprise the last snapshot . the classical analogs of these peaks are shown in the insets . the pdf progresses temporally along the dashed line . ] [ [ sec : barriermiddle ] ] particle traversing a finite barrier : @xmath197 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + consider next wavegroups for which some fourier components of the particle and barrier substates have a total initial kinetic energy in the relative coordinate system which exceeds the barrier potential energy while other components have a total relative initial kinetic energy which is less than the barrier potential . to illustrate the resulting pdfs the size of the particle and barrier substate wavegroups are chosen to be slightly larger than the barrier width @xmath102 . [ fig : barrier_posenergy ] shows the pdfs from such an interaction using three sequential snapshots , progressing along the dashed line from the lower left to upper right . the speed of the particle and well are illustrated for a classical system of two such particles in the insets next to each snapshot . again the diagonal white lines correspond to @xmath186 . the parameters used in fig . [ fig : barrier_posenergy ] are @xmath187 , @xmath188 , @xmath189 while the @xmath198 for the average value of the @xmath191 for the wavegroup distributions . this figure also illustrates type iii correlated interference but it differs from fig . [ fig : barrier_negenergy ] in that only the first mode is excited . additionally , the mode decays slowly enough to be visible in the third snapshot labeled as peak @xmath199 ( located between both @xmath186 and peaks @xmath59 and @xmath196 in the third snapshot ) . this peak is analogous to the buildup and decay of electromagnetic energy in a optical cavity . later snapshots ( not shown ) illustrate its decay . the peaks labeled @xmath59 in fig . [ fig : barrier_posenergy ] and fig . [ fig : barrier_negenergy ] are both of type ii . for the particle traversing the barrier . the only difference between the parameters used here and in fig . [ fig : barrier_negenergy ] is the pe which forms a barrier . ] the position of peak @xmath196 can be compared between figs . [ fig : barrier_negenergy ] and [ fig : barrier_posenergy ] since all parameters are the same except the pe . the location of this peak indicates the effect of the interaction on the relative transit times for the particle and finite well or barrier wavegroup substates . the infinite well - particle calculation differs from that of the barrier due both to the particle and well velocities being constrained by the resonance condition given in eqn . [ eq : quantizationvv ] and by the lack of coefficients , such as the @xmath162 , @xmath163 , @xmath164 , and @xmath165 used in the previous section , which depend on the parameters of integration . a closed form expression for the well substate wavegroup can be obtained from a gaussian distribution of the energy eigenstates parameterized in terms of velocity components , @xmath11 , of the well , given by @xmath200dv \notag \label{eq : psiwellwavegroup}\end{aligned}\ ] ] where the peak of the distribution is at @xmath178 and @xmath179 is its width . this is then summed over integral values of @xmath149 ( the number of nodes ) using the gaussian distribution @xcite @xmath201 \psi_{well}^{wavegroup } , \notag \label{eq : psiparticlewavegroup}\end{aligned}\ ] ] where the peak of the distribution is at @xmath202 and @xmath203 is its width . using these relations , the pdf is plotted with particle and well substate wavegroups whose spatial widths are less than the well spacing @xmath102 . [ fig : wellsynchfinal ] shows such pdf results for six snapshots taken at equal time intervals progressing from the lower left to upper right along the dashed line . ` reflection ' occurs at the diagonal white lines , corresponding to @xmath186 . the particle and well `` reflect '' from each other twice , in the second and fourth snapshots . the classical analogs for the particle and well positions for some snapshots are shown in the insets , labeled by a , b , and c. these correspond to the snapshots of the wavegroups labeled with the respective letters . while the insets are schematics of the ` classical ' analog between the wave and particle pictures , there is nothing similar for the correlated interference snapshots . interference occurs when the incident and reflected two - body wavefunctions ` overlap ' and is shown in higher spatial resolution for the first reflection in the right side inset of fig . [ fig : wellsynchfinal ] . the fringes are spaced by about half the debroglie wavelength of the _ particle _ for @xmath194 and @xmath195 , as are similar correlated interference fringes in two - body reflection of a particle from a mirror in fig . [ fig : interference ] . the parameters used in fig . [ fig : wellsynchfinal ] are @xmath204 , @xmath205 , @xmath206 , and @xmath207 . for a two - body wavefunction whose wavegroup size is less than the infinite well spacing . the diagonal white lines correspond to @xmath186 . the trajectory is indicated by the dashed line . the right inset is a blow - up of the correlated interference of the second snapshot . ] correlation measurements are more difficult to perform than measuring just one of the bodies in a many body system . to predict one body effects , the two - body pdf is integrated ( traced ) over the coordinates of one of the bodies thereby converting it into a ` one - body ' or marginal pdf for the other body @xcite . for example , consider a measurement of the particle but not the mirror for the two - body pdfs shown in fig . [ fig : spreadvelocity ] . averaging over the mirror coordinate in fig [ fig : spreadvelocity]a essentially eliminates any interference when measuring only the particle . averaging over the particle coordinate eliminates any interference in a measurement of only the mirror . yet there is interference in a correlated measurement . gottfried describes a related result using a different two - body interferometer @xcite . such averaging over the mirror coordinate of fig . [ fig : spreadvelocity]c , however , does not ` wash out ' interference in the particle - only measurement . the resulting particle - only ` standing - wave ' interference pattern in reflection from a static mirror is of course expected . yet the integral of this two - body pdf over the particle coordinate yields no mirror - only interference . this matches the classically expectation that there is interference for the wave reflecting from a mirror but not for the mirror . however , if the coherence lengths are switched ( it is difficult to increase the coherence length of a mesoscopic / macroscopic mirror ) then a particle - only measurement will yield no interference while a mirror - only measurement will yield interference . the two - body solution , when used in this manner to determine a one - body pdf , then reveals the limitations of starting with a solution to the one - body schrdinger equation where the potential ( mirror ) is assumed to be a classical rather than a quantum object . for example , a one - body pdf for the particle , determined by the one - body schrdinger equation , is a consequence of the solution to the two - body schrdinger equation only for coherence lengths used in fig . [ fig : spreadvelocity]c . the reflected one - particle pdfs obtained when averaging over the mirror coordinates of fig . [ fig : spreadvelocity]a , [ fig : spreadvelocity]b , and [ fig : spreadvelocity]d , however , _ can not _ be obtained without treating the mirror as a quantum object . the particle - only measurements , determined from this two - body treatment , then act as a probe of the mirrors quantum state . two regimes in which the quantum nature of the mirror is revealed are illustrated in fig . [ fig : spreadvelocity]a and d using fixed particle bandwidth . the result in part [ fig : spreadvelocity]a was discussed above . the other regime , shown in fig . [ fig : spreadvelocity]d , requires such small mirror coherence lengths that the incident and reflected two - body states are distinguishable , thereby also eliminating any interference . consider a simple estimate of the condition for which interference of the particle reflecting from the mirror disappears in the two - body pdf of fig . [ fig : spreadvelocity]d . this is done by approximating the incident two - body pdf of fig . [ fig : spreadvelocity]d as a rectangle with height @xmath208 and width @xmath209 @xcite . assuming a near stationary mirror with @xmath43 and @xmath104 , reflection then tilts this rectangle to an angle @xmath13 given by @xmath210 where @xmath211 and @xmath212 are the velocities after reflection of the mirror and particle . the overlap of the incident and reflected rectangles occurs at a distance @xmath213 from the point of maximum overlap given by @xmath214 . beyond @xmath213 no overlap or interference occurs . using the coherence length associated with thermal equilibrium of the mirror yields the following condition for the particle fringes to disappear in the two - body solution of fig . [ fig : spreadvelocity]d @xmath215 this condition generates a regime in which the quantum nature of the mirror is manifest via elimination of the expected one - body interference of the particle in reflection , which is caused by the incident and reflected paths becoming distinguishable as the mirror coherence length becomes smaller ( see section [ sec : bandwidth ] ) . although there is a lack of interference in the particle - only measurement , the two - body state is still split . this is evident in fig . [ fig : spreadvelocity]d as a bimodal distribution for which there is a finite probability of finding states of both reflection _ and _ no reflection . the effect is enhanced with reflection of ultra - cold atoms which have both larger coherence lengths and masses than neutrons . a yet larger mass could involve a bose - einstein condensate itself if it reflects as ` one particle ' rather than a collection of individual particles from the mirror @xcite . such ` standing - wave ' interference on reflection of a bose - einstein condensate ( @xmath216 m ) is eliminated with a mirror of @xmath217 kg ( the average mass of a mosquito ) and @xmath218 k. reflection of both atoms @xcite and bose - einstein condensates @xcite from mirrors has been studied but neither in the regimes discussed here nor in an attempt to measure such ` standing - wave ' interference . next consider measuring only the particle after reflection . for destructive interference of the joint pdf illustrated in fig . [ fig : slab ] ( or destructive interference of peaks b in figs . [ fig : barrier_posenergy ] and [ fig : barrier_negenergy ] ) , such averaging yields a low probability of measuring only the neutron even if the slab is _ not _ measured . in this case , destructive interference is maintained in the one - body neutron pdf . the analogy between the measurement of a short pulse of light reflecting from a thin film and a measurement of only the neutron then has a similar physical interpretation . in both cases the detector need only be placed in the path of the moving neutron or the pulse of light , which leads to a one - body measurement that is independent of the wavegroup size . there is no requirement that the spatial resolution of the measuring instrument be smaller than the coherence length of the object being measured . as in the mirror case , this example of destructive interference in a one - body measurement ( which is derived from the two - body wavefunction ) is valid only for a range of coherence lengths for both the slab and particle . such coherence length dependence can be illustrated assuming a small reflection coefficient and using the mirror pdf of fig . [ fig : spreadvelocity]d for reflection from the first slab surface where the small slab coherence length is due to its mesoscopic / macroscopic mass . consider a similar figure to represent the `` mirror '' reflection from the second slab surface but offset by a slab thickness , @xmath102 , which is sufficient to prevent overlap of the reflected two - body wavefunctions from these two figures . this is a case where the particle does not experience interference even after having reflected from both slab surfaces and having a coherence length greater than the slab thickness . a simple estimate of this condition for the slab is given by again assuming that the incident two - body pdf of fig . [ fig : spreadvelocity]d is a rectangle of height @xmath208 and width @xmath209 . for a near sationary mirror with @xmath43 and @xmath104 , the rectangles reflected from each surface are tilted at an angle @xmath13 given by @xmath210 . the positions of the slab surfaces are essentially vertical in the @xmath219 plane . the two reflected rectangles just separate when @xmath220 . using the coherence length associated with thermal equilibrium of the slab with its environment yields the temperature condition for which there is no interference @xmath221 note the similarity with eqn . [ eq : slabtemp ] where the condition on @xmath222 is for interference rather than for lack thereof . such elimination of interference is due to the two particle - slab reflected states becoming distinguishable as the slab coherence length decreases in a manner similar to that discussed for the particle - mirror system in section [ sec : bandwidth ] and is only a consequence of treating both the particle and slab as quantum objects . if the slab is not treated as a quantum object then particle interference will always occur as long as the particle coherence length is greater than the slab thickness . a similar effect is found in the particle - well model . consider now the opposite extreme in the quantum character of the slab , that of small slab bandwidth or large uncertainty in the slab position . unlike the case of large mirror uncertainty where averaging over the mirror position in fig . [ fig : spreadvelocity]a ` washes out ' interference , uncertainty in the slab position does not eliminate the one - body - only measurement of the interference for the particle reflecting from both slab surfaces . this is a consequence of that interference not depending on the slab location . the inequality in eqn . [ eq : slabtemp2 ] is most easily satisfied by increasing the mass of the reflected particle ( increasing the recoil of the slab ) and/or that of the slabs spacing ( requiring an increase in the coherence length of the particle ) while decreasing the mass of the reflector ( increasing the slab recoil ) . reflection of a neutron traveling at @xmath223 m/s from a slab at @xmath224k , as in the example used above , yields no interference for @xmath225 kg . elimination of ` thin - film ' interference for a mesoscopic / macroscopic slab requires larger @xmath8 and @xmath102 ( with the size of the mesoscopic / macroscopic mass decreasing quadratically with these parameters ) . interference in a typical one - body interferometer ` washes - out ' as the fringe spacing becomes imperceptible with increasing mass . another mechanism that eliminates interference is a measurement to determine along which path the object traveled through the interferometer . the former issue was mitigated using two - body reflection of a microscopic particle from a mesoscopic / macroscopic object , as described above , while the latter is the first topic of this section . a method to determine path information and therefore destroy interference utilizes a ` heisenberg microscope , ' where an additional particle scatters from the object traversing the interferometer . if the wavelength of this particle , @xmath226 , is smaller than the spatial separation along the two paths , @xmath203 , then path information can be obtained in principle from the state of this scattered probe particle and interference is destroyed . this was verified using a three grating mach - zehnder interferometer traversed by an atom and using a photon as the probe to determine path information @xcite . the contrast of the interference , as a function of @xmath203 , was shown to drop dramatically for @xmath227 . generating such decoherence for particle - mirror interference involves using a probe particle to determine the mirrors position during interference . the ` two paths ' exist only during the time that the incident and reflected two - body wavegroup states overlap for reflection from a mirror . during this time there are two distances that the mirror could have moved : that associated with and that without reflection . if the probe particle can resolve the difference in these distances then the ` path ' associated with reflection is distinguishable from that without reflection . for a near stationary mirror with @xmath43 and @xmath104 , the mean difference in these mirror distances is @xmath228 , which when equal to the wavelength of a probe particle of mass @xmath229 then requires that the probe have velocity @xmath230 . this calculation can also be used to approximate the effect of such decoherence on the mirror from the collisions with gas atoms forming the thermal bath which surrounds the mirror if these gas atoms are considered to be the probe which determines path information and they scatter from only one of the superposed mirror states . the interference in reflection described above is then maintained if there is a small probability of these atoms having the velocity , @xmath231 . for a mesoscopic / macroscopic mirror and both a microscopic particle and a probe , @xmath203 is too small to be resolved by typical low temperature gas atoms . next this decoherence mechanism is applied to slab reflection . the two surface reflections cause a difference in slab displacement which was given in section [ sec : twodelta ] for a near sationary mirror with @xmath43 and @xmath104 , as @xmath232 . determining from which slab surface the particle reflects then requires that the probe have velocity @xmath233 . this decoherence mechanism can also be applied to the emission of photons by the slab between the times of particle reflection from each slab surface . if the photons wavelength is small enough to allow for the location of the slab associated with one of the reflections ( @xmath140 m in the neutron - slab example given above ) , interference is destroyed since it can then be determined at which slab surface the reflection occurred @xcite . obviously the probability of such a thermal photon emission is small . these calculations indicate that environmental decoherence associated with obtaining path information can potentially be mitigated . this is due in part to the displacements associated with the interfering reflector substates being proportional to @xmath234 . to measure such small displacements requires scattering particles with proportionally smaller wavelengths which are typically not found in the environment . it should be pointed out that the calculations presented above assume that states for the particle , mirror , and slab have a minimum spatial profile ( minimum uncertainty gaussian wavepackets ) . this will be referred to as a phased state . introducing a random phase into the fourier components of either one or both of the substates of such an incident two - body phased state affects the shape and size of the wavepacket while maintaining the same bandwidth and therefore the same coherence length . this is referred to as a dephased state . such random phase structure is introduced as an initial condition but could have been imposed via an earlier measurement which collapsed the reflector state into a dephased state . consider an incident phased particle substate and incident dephased mirror substate which then reflect . a one - body particle measurement after reflection can then act as a sensitive probe of this mirror substate . to illustrate this consider how the pdf of fig . [ fig : spreadvelocity]c changes for such a dephased state . first , the pdf is spread spatially along the @xmath3 axis . second , the magnitude of the interference part of the pdf is greatly reduced but not eliminated . finally , the pdf is not spatially uniform . apart from its magnitude and spatial uniformity , this pdf is more similar to fig . [ fig : spreadvelocity]a than [ fig : spreadvelocity]c . this is not surprising since the localization of the dephased mirror substate is similar to that of the mirror in fig . [ fig : spreadvelocity]a . results of a particle - only measurement are predicted from an integral ( trace ) over the mirror coordinate of this pdf which then ` washes - out ' this interference . if the mirror is in a dephased state , a reflecting particle - only measurement will not exhibit interference in the overlap region . however , such interference is expected both when treating the mirror as a classical potential and for a phased mirror substate with the coherence lengths similar to those shown in fig . [ fig : spreadvelocity]c . the spatial extent of such a dephased mirror substate can vary between being unlocalized to being the minimum size determined by the bandwidth . this size can then be probed by varying the wavelength of the incident particle in a reflected particle - only measurement , in which case interference either disappears or reappears as this wavelength becomes smaller or larger than the spatial size of the dephased mirror substate . yet another decoherence mechanism is through interaction with the environment . zurek predicted exponential decay of the off - diagonal density matrix element terms for a macroscopic object in a superposition state , @xmath235 $ ] , with time constant @xmath236 where @xmath237 is the thermal relaxation time , @xmath203 is the difference in distance between these two states , and @xmath238 is the thermal de broglie wavelength of the macroscopic object @xcite . the decay is into a mixed rather than a pure state . it is interesting to note that the lack of interference predicted in section [ subsec : onlyparticle ] also depends on @xmath203 but for very different reasons . applying this decoherence model to the macroscopic slab states after reflection of the neutron , as described above , yields @xmath239 . for these neutron parameters this ratio is large ( @xmath240 for @xmath241 kg at @xmath242k ) due to the very small offset introduced from the recoil ( @xmath243 ) . the parameter of interest , @xmath244 , is also affected by @xmath237 which can vary dramatically depending on environmental and material properties along with the dimensional values of the slab . mesoscopic masses can have short thermal relaxation times . nevertheless , it appears that environmental decoherence for reflection from a slab can easily be made negligible in this model of decoherence . applying this decoherence model to the mesoscopic / macroscopic mirror states during correlated interference in reflection allows a greater variation of parameters since atoms or molecules can more easily reflect from a surface than from both surfaces of a ` thin film . ' the result , @xmath245 , therefore allows environmental decoherence for the mirror to more easily be made either negligible or non - negligible . consider next the consequences of such decoherence on correlated interference where the particle and mirror are in superposition states of both having and not having reflected . if this mirror substate decoheres into a mixed state then it seems reasonable to assume that the commensurate particle substate must also decohere into a mixed state since the two substates are correlated via conservation of energy and momentum . measurement of only the particle , after the mirror has decohered , therefore yields no interference since the particle can not then be in a superposition state . applying this result to a measurement of only the neutron flux reflecting from the slab , while varying the incident neutron s wavelength , results in no interference if the slab decoheres during the time between reflection from both surfaces . the details of actually executing such a measurement and of choosing parameters which straddle the region where decoherence is and is not applicable are of course much more complicated and are beyond the scope of this work . braginsky and khalili also treat environmental decoherence @xcite . however , they predict that for a macroscopic oscillator to exhibit quantum effects both the mechanical and thermal relaxation times are important while the parameter @xmath203 is not incorporated in their model . they argue that quantum effects can indeed be exhibited for macroscopic high q resonators . it has been shown that correlated interference in reflection does not disappear with increasing mirror mass , is difficult to eliminate via a path measurement , and can be made insensitive to environmental decoherence . it should also be pointed out that the robust character of interference for objects with many degrees of freedom is reinforced by measurements which demonstrate that even if the size of the object is larger than both the coherence length and debroglie wavelength , interference can still be observed @xcite . the use of reflection to generate perceptible interference in a mesoscopic / macroscopic object involves a microscopic particle interacting with a mesoscopic / macroscopic reflector to create two reflector substates which differ only by the microscopic momentum exchanged . unlike the fringes for a particle traversing a double slit , the interference fringes for the reflector substates do not vanish in the limit of large reflector mass . in addition , particle and reflector interference is required for conservation of energy and momentum to be satisfied . for example , consider reflection of a neutron from a slab . ` thin - film ' interference yields peak b in the two - body pdf of fig . [ fig : barrier_negenergy ] . destructive interference of the reflected neutron substate , which is certainly expected in a one - body treatment , must be associated with destructive interference of the slab substate ( peak b must not exist ) . that is , if the neutron can not be measured in reflection then the slab can never be measured at the position associated with it either having or not having reflected the neutron . similarly , by only changing the neutron wavelength to generate constructive thin - film interference in reflection , the slab must then have a commensurate probability to be found in a position of having reflected the neutron . measurement of interference in the reflected particle then requires that the reflector be in a superposition state which interferes in a manner correlated with that of the particle . this conclusion also holds for the interference shown in fig . [ fig : interference ] and presumably for all two - body interferometers . existence of such interference might be disputed since the difference in wavevectors of the interfering mirror substates , due to reflecting a microscopic particle , is much smaller than the spread in wavevectors of which the mirror wavegroup is composed , due to its thermal motion . this can be expressed as the ratio @xmath246 . interference , however , is not determined by @xmath247 but rather by the relation between the interferometer path difference and the coherence lengths of the interfering wavegroups . for a neutron reflecting from a mirror of @xmath248 kg , @xmath249 m / s , and @xmath218 k , the ratio @xmath250 yields a number which is of the same order of magnitude as is this ratio for an optical pulse of duration @xmath251 s ( from a mode locked ti - sapphire laser ) traversing a mach - zehnder interferometer , with the mirror in one arm moving at speed @xmath252 m / s ( all other interferometer components are static ) . optical pulse trains exhibit interference in similar interferometers @xcite with smaller @xmath247 values than in the particle reflection example just given . correlated interference in reflection is an example of one of the simplest interferometers , utilizing neither division of amplitude nor division of wavefront methods to generate interference . in addition , path lengths need not be carefully matched for interference to be manifest . the complexity of this interferometer is revealed as the coherence lengths of the particle and mirror substates are varied . examples are : ( 1 ) when both are small , classical reflection rather than interferometry essentially occurs . ( 2 ) when both are large , compared with the particle s wavelength , reflection becomes an interferometer yielding correlated interference , as shown in fig . [ fig : spreadvelocity]a , while no interference is observed in a one - body - only measurement of either the particle or mirror . due to uncertainty in the location of both objects , a measurement of the particle at @xmath2 then interferometrically correlates with a range of position measurements of the mirror . ( 3 ) when the mirror s displacement , due to recoil in reflection , is less than the mirrors coherence length and the coherence length of the particle is larger than that of the mirror , then a measurement of only the particle yields the expected interference from a one - body schrdinger equation treatment ( an example is shown in fig . [ fig : spreadvelocity]c ) . ( 4 ) when the coherence length of the particle is much larger than that of the mirror and mirror recoil is larger than its coherence length , reflection becomes a two - body beamsplitter . this is illustrated in fig . [ fig : spreadvelocity]d for @xmath91 . here two states are generated but do not interfere , one being the particle - mirror state before reflection and the other that after reflection , similar to a beamsplitter producing photon states which traverse spatially separate and therefore distinguishable paths . for example , at fixed @xmath2 there is a bimodal pdf along the @xmath3 axis , each peak of which corresponds to one of these two states . there is no interference since the short coherence length of the mirror relative to the recoil distance results in distinguishable two - body states ( the mirror has reflected the particle in the upper but not the lower pdf peak ) . the kinematics in the correlated interference regime can be compared with that of an electromagnetic wavegroup reflecting classically from a mirror . the latter involves accelerated motion due to a continuous force delivered to the mirror throughout the reflection process . to determine the former , consider the interference region shown in fig . [ fig : interference ] with @xmath253 , @xmath254 , and equal coherence lengths . if only the mirror is measured ( integrating the pdf over the particle coordinate ) then the expectation value for the mirrors position accelerates throughout the interaction time , as expected classically . if the mirror is not measured but the particle is , then the expectation value for the mirrors position decelerates at the same rate . however , these expectation value kinematics vary with parameters such as particle and mirror substate coherence lengths and masses . finally , the issue of how to determine that the mirror is in a superposition of states is considered . one method is to perform a correlation measurement of the positions of the particle and cm of the mirror in the interference region . in a region of destructive interference neither will be found . however , correlation measurements are difficult to perform . another possibility is to use a third particle as a probe , interacting only with the mirror to determine if the mirror is in a superposition state , in the following way : let this microscopic probe particle reflect from both mirror states ( reflecting from the other side of the mirror ) resulting in one - body interference of the probe . the coherence length of the probe need only be larger than the very small displacement of the mirror states for this probe interference to be exhibited , while the magnitude of the phase shift is determined by the mass of the probe . this is a method to measure that the mesoscopic / macroscopic mirror is in a superposition state , without any direct measurement of the mirror . however , calculations of such three - body interactions , which follow easily from the formalism discussed above , are beyond the scope of this already lengthy foundational work . methods to measure quantum effects on a mesoscopic / macroscopic reflector using only the reflected particle have been a focus of this work , rather than the methods just described . the reflected particle flux is then used to reveal properties of the quantum state of the reflector . one example is of a neutron reflecting from a slab . as with interference in reflection of light , the one - body schrdinger equation predicts a variation in reflected neutron flux as a function of incident neutron wavelength whenever the neutron coherence length is longer than the slab thickness . the calculation presented here , treating both the particle and reflector as quantum objects , also predicts such interference but only as a special case . in addition , this treatment predicts the following regimes under which no reflected interference occurs _ for a measurement of only the particle _ : ( 1 ) the reflection process acts as a beamsplitter and not an interferometer , generating two mirror or slab states which do not overlap . ( 2 ) the fourier components of the mirror substate have random phases sufficient to extend its size beyond that of the particle s wavelength . ( 3 ) the slab and mirror have coherence lengths long enough for correlated interference but not interference when only the particle is measured . in each of these cases , consider varying the incident particle s wavelength and observing no interference in the reflected particle flux . this indicates that the mesoscopic / macroscopic reflector is in a particular quantum superposition state . by varying parameters appropriately , as discussed in section [ sec : particleonlymirror ] , interference in this reflected particle flux then reappears . ironically , it is this lack of interference that then reveals the quantum nature of this mesoscopic / macroscopic ` thin film ' or mirror . in this case , measuring only the reflected particle flux yields significant experimental advantages over a measurement of the correlation between the particle and mirror . another possible method to study the state of the mesoscopic / macroscopic mirror is via the expected standing wave pattern generated between incident and retro - reflected states of a bose - einstein condensate as it interacts with a mirror . if this process can be modeled as described above , with the mass of the condensate replacing that of the particle , then the point at which condensate interference disappears can be extended to larger mirror masses . although interference between split condensates has been observed @xcite , as far as we are aware , there is no evidence of standing wave interference due to reflection . reflection of atoms from mirrors has been previously studied but not in treating both the particle and reflector as quantum objects nor in the reflector mass regimes discussed here nor in an attempt to measure either the resulting ` standing - wave ' interference or the effect of decoherence on such interference . there exist few examples of quantum correlated interference for non - zero rest mass particles . verification of the above predictions , even in the microscopic regime for both particle and reflector , is therefore of fundamental interest . although far from being comprehensive , these results indicate a direction , heretofore unexplored , for further research in understanding quantum correlation , probing decoherence , and extending quantum measurements to larger masses . bell , physics * 1 * 195 ( 1964 ) . k. gottfried , am . * 68 * 143 ( 2000 ) . d. salart , et . al . , phys . * 100 * 220404 ( 2008 ) . m. a. rowe , et . nature , * 409 * 791 ( 2001 ) . m. ansmann , et . al . , nature , * 461 * 504 ( 2009 ) . a. hagar , phil . r. soc . * 370 * 4425 ( 2012 ) ; m.schlosshauer , rev . of mod 76 * 1267 ( 2004 ) . a. d. cronin , j. schmiedmayer , and d. e. pritchard , rev . phys . * 81 * , 1051 ( 2009 ) . t. juffmann et . al . , nature nanotech * 7 * , 297 ( 2012 ) . b. abbott , et . new j. of phys . * 11 * 073032 ( 2009 ) . m. aspelmeyer , p. meystre , and k. schwab , phys . today * 65 ( 7 ) * , 29 ( 2012 ) . a.d . oconnell et.al . , nature * 464 * , 697 - 703 ( 2010 ) . a. nunnenkamp , k. borkje , and s. m. girvin , phys . lett . * 107 * , 063602 ( 2011 ) . f. de martini et.al . * 104 * , 050403 ( 2010 ) . d. kouznetsov and h. oberst , phys . a * 72 * , 013617 - 1 ( 2005 ) . t. a. pasquini , m. saba , g .- b . jo , y. shin , w. ketterle , and d. e. pritchard , phys . 97 * , 093201 - 1 ( 2006 ) . f. shimizu , phys . lett . * 86 * , 987 ( 2001 ) . t. hils et . al . , phys . a * 58 * , 4784 ( 1998 ) . y. colombe , b. mercier , h. perrin , and v. lorent , phys . a * 72 * , 061601 ( 2005 ) . szriftgiser , p. , d. guery - odelin , m. arndt , and j. dalibard , phys . lett . * 77 * , 4 ( 1996 ) . h. rauch , contemp . * 2 * 345 ( 1986 ) . kowalski , f.v . and browne , arxiv:1405.0031 [ quant - ph ] p. tommasini et.al . , am . * 66 * , 881 ( 1998 ) . siverman , _ more than one mystery _ , p-96 ( springer verlag , new york , 1995 ) . in the particle - mirror coordinates the incident and reflected wavevectors for the particle ( or mirror ) are not necessarily equal and opposite as they are in the cm / rel coordinates ( nor are they equal and opposite for a one particle system in which the mirror is fixed ) . the probability current density to the left of the mirror then exhibits spatial and temporal interference due to the effect this difference in wavevectors has on the incident and reflected wavefunctions . this results in a non - additive current density , a consequence of the space to the left of the mirror increasing as it moves . the net flux must account for the build - up of probability within this `` new '' space . pushin et . al . , phys . 100 * , 250404 ( 2008 ) . hamilton , a.g . klein , and g.i . opat , phys . rev . a , * 28 * , 3149 ( 1983 ) . the coherence length does not change even if the size of the wavepacket changes due to dispersion . klein , g.i . opat , and w.a . hamilton , phys . * 50 * , 563 ( 1982 ) . m. ferrero and a. van der merwe , _ fundamental problems in quantum physics _ , ( kluwer academic publishing , dordrecht , 1995 ) . the pauli exclusion principle is another example of destructive interference in a two - particle system . see r. c. liu , b. odom , y. yamamoto , and s. tarucha , nature 391 , 263 - 265 ( 1998 ) . c.u.serge and j.d . sullivan , am . * 44 * 729 ( 1976 ) . the coherence length is used here rather than the wavepacket size since the former rather than the later determines interference , op . s.m . barnett and r. loudon , phil . r. soc.a * 368 * 927 ( 2010 ) . chapman , et.al . , * 75 * , 3783 , ( 1995 ) . w. h. zurek , rev . * 75 * , 715 , ( 2003 ) . v.b . braginsky and f.y . khalili , _ quantum measurement _ , ( cambridge university press 1992 ) . l. lepetit , g. cheriaux , and m. joffre , j. opt . b * 12 * , 2467 ( 1995 ) . c. kohstall , s. riedl , e. s. guajardo , l. sidorenkov , j. h. denschlag , r. grimm , new j. phys . * 13 * , 065027 ( 2011 ) m. witkowski , r. gartman , b. nagrny , m. piotrowski , m. podzie , k. sacha , j. szczepkowski , j. zachorowski , m. zawada , and w. gawlik , phys . a * 88 * , 025602 ( 2013 ) if the condensate particles reflect individually from a receding mirror then they each acquire a different energy which decreases in their order of reflection . this is due to the energy exchanged with the mirror from the particles which have already reflected and is manifest as a distorted reflected wavegroup .

studies of electron - positron pair production in strong external fields are an important tool to reveal the structure of the qed vacuum . dynamical pair creation by two coulomb fields in relativistic heavy ion collisions has been analysed in great detail , both theoretically and experimentally ( see , e.g. , @xcite for a topical review ) . in recent years , due to the sustained progressing of laser technology , pair creation in strong laser fields is encountering a growing interest . in several theoretical papers pair production in the standing wave formed by two counterpropagating laser beams has been considered @xcite . in most cases a laser field strength of order or above the critical value @xmath0 v / cm is required in order to observe the process in experiment @xcite , which still is by four orders of magnitude larger than the highest laser field strengths achievable today @xcite . more promising in view of its experimental realization is pair production in the combination of a laser and a coulomb field . here one can exploit the possibility to let a nucleus or a highly charged ion collide at large value of the relativistic lorentz factor @xmath1 with the laser beam . then , in the nuclear rest frame , the laser s field strength @xmath2 and frequency @xmath3 are enhanced by a factor @xmath4 . observable pair creation rates should result for @xmath5 or @xmath6 in this frame ( here , @xmath7 , @xmath8 and @xmath9 are planck s constant , light velocity and electron mass , respectively ) . with the help of the upcoming novel laser @xcite and accelerator @xcite facilities these conditions can be met . the large hadron collider , that is presently under construction at cern ( geneva , switzerland ) , will accelerate protons to an energy of 7 tev ( @xmath10 ) @xcite . when such protons are brought into collision with the strongest laser beams available today ( @xmath11 at @xmath12 ev ) , then the laser field strength in the proton s rest frame approaches or even exceeds the critical value . on the other hand , the x - ray free - electron laser ( xfel ) facilities presently being developed at desy ( hamburg , germany ) and slac ( stanford , usa ) are proposed to provide spatially coherent and highly brilliant beams of synchrotron radiation with single - photon energies of up to @xmath13 kev at maximum field strengths of @xmath14 @xcite . when combined with a moderately relativistic ion beam , the laser photon energy in the projectile frame can reach the electron s rest energy . we should mention that a few years ago nonlinear electron - positron pair creation in the collision of an ultrarelativistic electron beam ( @xmath15 ) and an intense optical laser pulse was observed at slac @xcite . in this situation , there are two channels for pair production : a direct channel , via laser - photon absorption in the coulomb field of the electron , and an indirect two - step channel ( so - called breit - wheeler process ) , where first a high - energy photon is generated by compton backscattering which afterwards creates the pair via absorption of laser photons @xcite . in the experiment the contribution of the indirect channel was shown to be dominant . in the present paper , however , the indirect production mechanism will not be addressed . it is of importance only in the case of a light projectile beam . in fact , when the incoming electrons were replaced by a beam of heavy particles ( e.g. , protons ) the indirect production channel would be strongly suppressed and direct nonlinear pair production ( by a virtual photon from the nuclear coulomb field ) could be investigated . the latter process is the subject of this paper . the first calculation of pair creation by a coulomb and a strong laser field is due to yakovlev @xcite . he treated the case of circular laser polarization and derived asymptotic formulae for the production rate in certain parameter regimes . particularly , the total cross sections in the perturbative regime ( @xmath16 , with the intensity parameter @xmath17 given in eq.([xi ] ) below ) as well as in the overcritical - field limit ( @xmath18 and @xmath19 ) were obtained . twenty years later , mittleman @xcite considered the process in a linearly polarized laser wave . his calculation of the leading term for the total cross section in the perturbative multiphoton regime ( @xmath16 and @xmath20 ) led him to the conclusion that , for the laser intensities and frequencies available at that time , this cross section was the smallest one on record . despite of this pessimistic result , another twenty years later the great headway in laser technology that was following the invention of chirped - pulse amplification @xcite triggered newly revived interest in the process . dietz and prbsting @xcite investigated the influence of a highly charged ion of charge @xmath21 within a nonperturbative numerical approach . however , to make their coupled - channel calculations feasible they treated the laser field in dipole approximation . while in the earlier work the nucleus was always assumed at rest , mller , voitkiv , and grn explored in detail the improved opportunities arising from laser - nucleus collisions . in the articles @xcite they performed numerical calculations for tunneling pair production by an ultrarelativistic nucleus colliding with a superintense near - infrared laser beam and for multiphoton pair creation by a moderately relativistic ion beam in combination with an intense x - ray laser wave . moreover , they considered the variant of bound - free pair production where the electron is created in a bound state of the projectile nucleus @xcite . free pair creation in the collision of a relativistic nucleus with an ultrastrong x - ray laser of circular polarization was also treated by avetissian , avetissian , mkrtchian , and sedrakian in ref.@xcite , where analytical formulae in the high - intensity limit and some numerical results on the energy spectrum of the created particles in the nonperturbative multiphoton regime ( @xmath22 and @xmath23 ) are given . most of the treatments mentioned above fully account for the influence of the external laser field by using the volkov solutions to the dirac equation as basis states in a perturbative calculation of first order with respect to the nuclear coulomb field . within this framework , the fully differential pair creation rate can be expressed as a fourier series over the number of absorbed laser photons , the coefficients of which are essentially given by ordinary or generalized bessel functions for circular or linear laser polarization , respectively . because of the large photon numbers involved , taking this sum is one of the main obstacles in this kind of approach , in particular for a linearly polarized laser wave . as a consequence , it is difficult to get useable expressions for the total production rate . pictorial equation in terms of feynman diagrams describing @xmath24 pair production in combined laser and coulomb fields . in our approach , the coulomb field is treated within the first order of perturbation theory , while the effect of the laser wave is nonperturbatively taken into account to all orders . this is expressed by the feynman graph on the left - hand side , where the double lines represent the exact lepton wave - functions in the laser field ( volkov states ) , while the dashed line stands for their interaction with the coulomb field . expanding the volkov states with respect to the lepton - laser coupling results in a perturbation series , some typical low - order terms of which are shown on the right - hand side . the wavy lines symbolize the laser photons and the arrows indicate whether the respective laser photon is emitted or absorbed during the process . the first diagram on the right - hand side , e.g. , describes the leading order of pair production by the net - absorption of one laser photon ( with @xmath25 ) . for later reference , we note that the interference between the first and the fourth diagram leads to a @xmath26-correction to the probability for one - photon pair creation [ cf . eq.([delta ] ) below].,width=302 ] in the present paper a different approach based on the optical theorem is used to calculate the total rate for pair creation in combined laser and coulomb fields . this approach employs the explicit form of the polarization operator of a photon in a laser field found by baier , milstein , and strakhovenko @xcite by means of an operator technique . an alternative form of this polarization operator was derived independently by becker and mitter @xcite by means of another method . both results are in agreement with each other . similar to refs.@xcite the interaction of the leptons with the laser field is taken into account to all orders while the effect of the coulomb field is treated in first order ( cf . 1 ) . the main advantage of our method is the possibility to derive , independent of the laser s polarization state , in an analytical manner compact formulae for the total production rate that only involve low - dimensional integrals and no additional summations . based on these representations one can rather easily find all different kinds of asymptotics . in this way we will confirm and extend previous results for a circularly polarized laser wave and give for the first time corresponding expressions for elliptic laser polarization . as a main result , we show that there are essential differences between the general case of an elliptically polarized laser field and the special case of circular polarization . furthermore , because of its interest for near - future experiments , particular emphasis is placed on the nonlinear process of pair creation by the simultaneous absorption of two photons from the laser wave in a situation , where the energy of a single photon is not sufficient to produce a pair and where the contribution from higher photon orders is neglibly small . we note that , in general , contributions stemming from different net - numbers of absorbed laser photons can be distinguished via the momentum spectra of the produced particles . in addition , the strong - field regimes of pair production shall be examined . the paper is organized as follows . in the next section we describe how the polarization operator is related to the process of pair creation , and how it can be used to calculate the total pair production rate in the superposition of a coulomb and a laser field . general formulae are given in terms of two - fold integrals for both circular and elliptic laser polarization . in the following section we apply these expressions to various parameter regions and determine the corresponding asymptotic behaviour of the pair production rate . special attention is paid to the nonlinear process of pair production by two - photon absorption in the limit of low laser intensity , for which in particular the behaviour close to the energetic threshold is investigated . moreover , the pair - creation cross - section in the high - frequency limit ( @xmath27 ) is analysed , where we calculate the next - to - leading order correction to one - photon pair production at low intensity and analyse the limiting case of high laser intensity . finally , we consider the overcritical - field regime as well as the quasiclassical limit of high intensity and low frequency , where the rate shows an exponential tunneling behaviour . we finish with a conclusion where the significance of our results in view of possible experimental investigations is discussed . the polarization operator @xmath28 describes the propagation of a photon of four - momentum @xmath29 in a background field ( e.g. , the qed vacuum ) , including self - energy corrections . it is related to the corresponding exact photon propagator @xmath30 via dyson s equation @xcite @xmath31 where @xmath32 denotes the free photon propagator . in refs.@xcite the ( properly renormalized ) polarization operator was evaluated in the combined fields of the qed vacuum and a classical electromagnetic plane wave . the interaction with the vacuum was taken into account to first order in the fine - structure constant @xmath33 , with @xmath34 being the electron charge , while the interaction with the plane - wave field was included to all orders . in the following , the shape of the electromagnetic wave will be taken as @xmath35 with @xmath36 and the dimensionless intensity parameters @xmath37 here and henceforth we use relativistic units with @xmath38 and write @xmath39 for the product of two four - vectors . the respective result on the polarization operator of a photon in a constant electric field instead of a plane electromagnetic wave can be found in ref.@xcite . apart from its theoretical significance , the knowledge of the polarization operator allows for a number of applications . for example , the total probability for @xmath24 pair production by a photon of momentum @xmath40 in an external field ( e.g. , a plane laser wave ) is related to the imaginary part of the corresponding polarization operator via @xcite @xmath41 where @xmath42 denotes the photon s polarization four - vector . in fig.2 we give an intuitive , pictorial explanation of this formula . lowest order ( in @xmath43 ) feynman graph for the polarization operator of a photon in an external laser field . in our case the photon is a virtual one stemming from a coulomb field ( outer dashed lines ) . the diagram can be viewed as describing the elastic forward scattering of the photon through an intermediate laser - dressed @xmath24 state ( double line ) . by performing , as indicated , a cut along the central dashed line , it is seen that the graph represents the product of the amplitude for the production of a laser - dressed @xmath24 pair by the virtual photon and its complex conjugate , including a sum over all electron states . application of the optical theorem ( i.e. , the unitarity of the scattering matrix ) therefore relates the imaginary part of the polarization operator with the total probability for pair creation , as expressed by eqs.([ppph ] ) and ( [ w]).,width=94 ] the calculation in ref.@xcite was performed for an arbitrary photon momentum @xmath29 , including the case @xmath44 . therefore , the polarization operator found there not only applies to the combination `` photon + laser wave '' . instead , the photon may be replaced by any additional external field @xmath45 , e.g. a nuclear coulomb field . since a coulomb field can transfer momentum but not energy , here one has @xmath46 . to get , at a given value of @xmath47 , the differential probability ( per unit time ) for @xmath24 pair creation one has to replace the photon wave - function @xmath48 in eq.([ppph ] ) by the fourier transform of the coulomb field @xmath49 . here , @xmath50 denotes the nuclear charge number . the total production rate then is found by integration over all possible momenta : @xmath51 note that in eq.([w ] ) the additional factor of 1/2 compensates for the double - counting of the contributions from the momenta @xmath47 and @xmath52 . for the case of a constant homogeneous electromagnetic field the relation ( [ w ] ) was used in ref.@xcite , where the probability for pair production was found within the quasiclassical approximation for the polarization operator . according to the results of ref.@xcite , the required component of the polarization operator in eq.([w ] ) can be written as @xmath53 \right\ } \nonumber\\ & & \!\!\!\!\ ! \times{1\over\cos^2\theta } \left [ \sin^2\theta\ , ( d_3 \cos^2\phi + d_4 \sin^2\phi ) - d_5 \right],\end{aligned}\ ] ] where we introduced the angles @xmath54 and @xmath55 , and write @xmath56 . the coefficients @xmath57 ( @xmath58 ) in eq.([pi00 ] ) are given by eq.(2.32 ) in ref.@xcite ; they read @xmath59 \nonumber\\ & & + \xi_1 ^ 2\sin^2\rho{1+v^2\over 1-v^2}j_0(\zeta ) \nonumber\\ & & + { 1\over 4}\left[{q^2\over m^2}-i{|\lambda|(1-v^2)\over\rho}\right][j_0(\zeta)-{\rm e}^{iy } ] \nonumber\\ d_4 & = & d_1(\xi_1 \leftrightarrow \xi_2)\nonumber\\ d_5 & = & { q^2\over 4m^2}(1-v^2)[j_0(\zeta)-{\rm e}^{iy}]\end{aligned}\ ] ] with the abbreviations @xmath60 for simplicity , we first consider the case of a circularly polarized laser wave ( @xmath61 ) . then the coefficients in eq.([pi00 ] ) can be simplified to read @xcite @xmath62(1-{\rm e}^{iy})\nonumber\\ d_5 & = & { q^2\over 4m^2}(1-v^2)(1-{\rm e}^{iy})\end{aligned}\ ] ] with @xmath63 employing spherical coordinates such that @xmath64 , the integration over @xmath65 in eq.([w ] ) is trivial and the total rate becomes @xmath66\right ] \nonumber\\ & & \times \bigg\ { ( 1-t^2)\xi^2\sin^2\rho { 1+v^2\over 1-v^2 } { \rm e}^{-iy } + \bigg [ q^2(1-v^2 ) \nonumber\\ & & + { 1-t^2\over 2}\left [ 1-q^2(1+v^2 ) \right ] \bigg](1-{\rm e}^{-iy } ) \bigg\},\end{aligned}\ ] ] where we have introduced the dimensionless quantities @xmath67 , @xmath68 , and @xmath69 . going over to the new variables @xmath70 , @xmath71 considerably simplifies the @xmath72-dependence such that also this integral can be taken : @xmath73 \bigg\ { \xi^2\sin^2\rho \nonumber\\ & & \!\!\!\ ! \times\left[{2\over 3}(2+t^2){\sqrt{1-t^2}\over t^2}- \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\right ] { \rm e}^{-iy}\nonumber\\ & & \!\!\!\ ! 3}{(1-t^2)^{3/2}\over t^2}+{q^2\over 2 } \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\right](1-{\rm e}^{-iy } ) \bigg\ } \nonumber\\ \\end{aligned}\ ] ] where now @xmath74 . alternatively , the @xmath75-integration can be expressed by neumann functions @xmath76 : @xmath77 \nonumber\\ & & + { 2\over 3}\left [ y_1(\beta)-{1\over z}y_1(\beta z)\right]{(1-t^2)^{3/2}\over t^2 } \nonumber\\ & & + \big [ y_1(\beta)- z y_1(\beta z)\big ] \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right ) \bigg\}\end{aligned}\ ] ] with @xmath78 and @xmath79^{1/2}$ ] . equation ( [ circ2 ] ) is a very compact , general expression for the total pair production rate in a circularly polarized laser field . for applications , both eqs.([circ1 ] ) and ( [ circ2 ] ) will prove to be useful . in the general case of an elliptically polarized laser wave the integrations over @xmath65 and @xmath72 can be taken in the same manner as before yielding @xmath80 { \rm e}^{-iy } \nonumber\\ & & \!\!\!\ ! \times\bigg\ { \left(\xi_1 ^ 2+\xi_2 ^ 2\right)j_0(\zeta ) \bigg [ { 2\over 3}a_1(1-t^2)^{3/2 } \nonumber\\ & & \!\!\!\ ! + \sin^2\rho \bigg({2\over 3}(2+t^2)\sqrt{1-t^2 } - t^2 \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\bigg ) \bigg ] \nonumber\\ & & \!\!\!\ ! + i\left(\xi_1 ^ 2-\xi_2 ^ 2\right)j_1(\zeta ) \bigg [ { 2\over 3}a_1(1-t^2)^{3/2 } \nonumber\\ & & \!\!\!\ ! -\sin^2\rho \left(2\sqrt{1-t^2 } - t^2 \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\right ) \bigg]\nonumber\\ & & \!\!\!\ ! + \left[j_0(\zeta)-{\rm e}^{iy } \right ] \bigg[{2\over 3}(1-t^2)^{3/2}\left(q^2-i{qt\over a\rho}\right ) \nonumber\\ & & \!\!\!\ ! -q^2t^2\ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\ \bigg ] \bigg\}.\end{aligned}\ ] ] the integral over @xmath75 can again be taken analytically : @xmath81 \nonumber\\ & & \!\!\!\ ! \times \bigg [ { 2\over 3}a_1(1-t^2)^{3/2 } \nonumber\\ & & \!\!\!\ ! + \sin^2\rho \bigg({2\over 3}(2+t^2)\sqrt{1-t^2 } - t^2 \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\bigg ) \bigg ] \nonumber\\ & & \!\!\!\ ! - { \xi_1 ^ 2-\xi_2 ^ 2\over \sqrt{g_0 ^ 2-g_1 ^ 2 } } \left [ \varphi_- j_1'(\varphi_-)y_1(\varphi_+ ) - \varphi_+ j_1(\varphi_-)y_1'(\varphi_+ ) \right ] \nonumber\\ & & \!\!\!\ ! \times { \rm sign}(g_1 ) \bigg [ { 2\over 3}a_1(1-t^2)^{3/2 } \nonumber\\ & & \!\!\!\ ! -\sin^2\rho \left(2\sqrt{1-t^2}-t^2 \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\right ) \bigg]\nonumber\\ & & \!\!\!\ ! - \left [ \varphi_- j_1(\varphi_-)y_0(\varphi_+ ) + \varphi_+ j_0(\varphi_-)y_1(\varphi_+ ) - \beta y_1(\beta)\right ] \nonumber\\ & & \!\!\!\ ! \times\bigg[{2\over 3}(1-t^2)^{3/2 } - t^2 \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\bigg ] \nonumber\\ & & \!\!\!\ ! -{4\over 3 } \left [ j_0(\varphi_-)y_0(\varphi_+ ) - y_0(\beta ) \right ] ( 1-t^2)^{3/2 } \bigg\}\end{aligned}\ ] ] with @xmath82 , @xmath83 , and @xmath84 . the total production rate can be converted into a cross section according to the relation @xmath85 with @xmath86 being the photon flux . let us consider the case of small intensity ( @xmath16 ) and frequency @xmath87 . the leading term in the @xmath26-expansion corresponds to pair creation by a single photon in a coulomb field . under these circumstances the total production probability is independent of the laser s polarization state and it is convenient to derive the asymptotic behaviour by using the formula ( [ circ1 ] ) for circular laser polarization . assuming @xmath16 , we can expand the exponentials to lowest order with respect to @xmath88 . afterwards we perform the integration over @xmath89 with the result @xmath90 \nonumber\\ & & \times \left(1-{\gamma\over 2}\right ) \bigg\},\end{aligned}\ ] ] where @xmath91 and @xmath92 is the step function , which restricts the integration to the kinematically allowed region @xmath93 and @xmath94 with @xmath95 . integrating over @xmath75 gives : @xmath96 \nonumber\\ & & - 2\left ( 1+{a^2\over t^2}\right)\sqrt{{t^2\over a^2}-1 } \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right ) \nonumber\\ & & + { 2a\over t}\ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right ) \ln\left({{t\over a}+\sqrt{{t^2\over a^2}-1}}\right ) \bigg\}. \nonumber\\ \ \end{aligned}\ ] ] the corresponding expression for the cross section [ see eq.([flux ] ) ] agrees with the well - known result of bethe and heitler @xcite . accordingly , in the ultrarelativistic limit ( @xmath97 ) we find @xmath98\end{aligned}\ ] ] with the classical electron radius @xmath99 , and in the nonrelativistic limit ( @xmath100 ) @xmath101 it is interesting to consider the cross section for two - photon pair production at @xmath16 , which represents a nonlinear strong - field process of lowest possible order in the number of absorbed photons . we assume that @xmath102 , which excludes the possibilty of one - photon pair creation . going one step further in the @xmath88-expansion of eq.([circ1 ] ) and performing the @xmath89-integration , we obtain for the circular polarization case @xmath103 + { 1\over 2}\left(1-{\gamma\over 4}\right)^2{a^2\over q^2t^2 } \nonumber\\ & & \!\!\!\!\!\!\!\ ! \times \left[{2\over 3}{(1-t^2)^{3/2}\over t^2}+q^2 \ln\!\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\right]\bigg\}.\end{aligned}\ ] ] proceeding as before , we get @xmath104 \nonumber\\ & & + { ( 4t^2-a^2)^{5/2}\over 45at^4}{(1-t^2)^{3/2}\over t^2 } + { a\over 6t^4 } \nonumber\\ & & \times\ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right ) \bigg [ ( 2t^2+a^2)\sqrt{4t^2-a^2 } \nonumber\\ & & -3a^2t\,\ln\!\left({2t\over a}+\sqrt{{4t^2\over a^2}-1}\right)\bigg]\bigg\}.\end{aligned}\ ] ] after conversion into a cross section , eq.([2ph ] ) reads @xmath105 with @xmath106^{3/2}\over ( xt)^2 } \nonumber\\ & & \times \left[{2\over 3}(2+t^2){\sqrt{1-t^2}\over t^2}- \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\right ] \nonumber\\ & & + { 4\over 15}{[(xt)^2 - 1]^{5/2}\over ( xt)^4}{(1-t^2)^{3/2}\over t^2 } + { 1\over ( xt)^4 } \nonumber\\ & & \times\ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right ) \big [ \left(2+(xt)^2\right)\sqrt{(xt)^2 - 1 } \nonumber\\ & & -3xt\ln\left ( xt+\sqrt{(xt)^2 - 1}\right ) \big ] \bigg\}.\end{aligned}\ ] ] in the limit @xmath107 , eq.([f ] ) becomes @xmath108 hence , the threshold behaviour of the cross section for two - photon pair production in a circularly polarized laser field reads @xmath109 we note that , based on the results of their numerical calculations , the authors of ref.@xcite found that the rate for two - photon pair creation within the whole frequency range @xmath102 approximately scales as @xmath110 . however , eq.([2phlimit ] ) now shows that the exact scaling in the threshold region ( @xmath111 and @xmath112 ) is given by @xmath113 . further , in ref.@xcite a total production rate of @xmath114 sec@xmath115 ( in the nuclear rest frame ) was reported for @xmath116 , @xmath117 kev , and @xmath118 . applying eq.([2ph ] ) to these parameters , perfectly reproduces this result . in the general case of elliptic polarization , the corresponding asymptotic cross section reads @xmath119 where @xmath120 with the function @xmath121 from eq.([f ] ) and @xmath122 \nonumber\\ & & -\left [ ( xt)^2 - 4 \right]{\sqrt{(xt)^2 - 1}\over xt } \nonumber\\ & & \times \left ( 2\sqrt{1-t^2}-t^2\ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right ) \right ) \nonumber\\ & & + { 1\over 2(xt)^3 } \big [ \left ( ( xt)^2 - 16\right)\sqrt{(xt)^2 - 1 } \nonumber\\ & & + 12xt\ln\left(xt+\sqrt{(xt)^2 - 1}\right)\big ] \nonumber\\ & & \times \bigg[{2\over 3}(1-t^2)^{3/2 } - t^2\ln\!\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right)\bigg]\end{aligned}\ ] ] and the ellipticity parameter @xmath123 ^ 2 $ ] . this time the threshold behaviour ( @xmath124 ) is found to be @xmath125 hence , in the general case , the cross section for two - photon pair creation close to threshold raises like the second power of the excess energy @xmath126 . only in the special case of circular polarization ( @xmath127 ) the threshold cross section scales with the fourth power as given in eq.([2phlimit ] ) . we notice , however , that our first - order treatment of the coulomb field requires the parameter @xmath128 to be small , where @xmath129 denote the velocities of the created particles . for pair production very close to the threshold , coulomb corrections are expected to be important . a summary of our results on two - photon pair production for linear and circular polarization including various approximations is displayed in fig.3 . the function @xmath130 in the intervall @xmath131 for @xmath132 ( linear polarization , solid line ) and @xmath127 ( circular polarization , short - dashed line ) . the long - dashed and dotted lines show the respective threshold laws for @xmath133 according to eqs.([flim ] ) and ( [ fmu]).,width=226 ] ( 0,0)(0,0 ) ( -23,-5)@xmath134 ( -153,3)@xmath135 ( -195,50 ) when applying our general formula ( [ sigmagen ] ) to the same parameters as above but for a linearly polarized laser wave ( @xmath132 , @xmath136 , @xmath137 , @xmath117 kev , and @xmath116 ) , we get a total production rate of 290 sec@xmath115 , which is in rather good agreement with the corresponding value of 240 sec@xmath115 found in ref . the difference is most likely due to small inaccuracies in the 5-dimensional integration along with the numerical evaluation of the generalized bessel functions in ref . as is seen from fig.3 , the pair production rate for a linearly polarized laser wave is noticeably larger than the corresponding rate for circular laser polarization within the whole frequency range @xmath138 . in this subsection we analyse the high - energy behaviour of strong - field pair creation , i.e. , we shall assume @xmath97 but allow for arbitrary value of @xmath17 . let us first consider the circular polarization case and start from eq.([circ2 ] ) . we exchange the order of integration and first integrate over @xmath139 . since @xmath97 , we can introduce a splitting parameter @xmath140 with @xmath141 , and divide the @xmath139-integration into ranges from @xmath142 to @xmath140 and from @xmath140 to 1 . in the first region we have @xmath143 . substituting @xmath144 we get @xmath145 \bigg\}.\end{aligned}\ ] ] with the help of the identity @xmath146 this is readily integrated to give @xmath147 \nonumber\\ & & \!\!\!\!\!\!\!\!\!\!\!\!\!\!\ ! \times \left ( \ln{a\rho\over\epsilon } + c \right ) + \left ( 4\sin^2\rho - { 1\over\xi^2 } \right)\ln z \bigg\}\end{aligned}\ ] ] where the relation @xmath148 $ ] was used , which is valid for @xmath149 ; @xmath150 denotes euler s constant . in the second range we have @xmath151 and we can exploit the small - argument behaviour of the neumann function : @xmath152 . then the integration is easily performed : @xmath153 \nonumber\\ & & - { 19\over 3}\sin^2\rho - { 4\over 3}\left(1-{\sin^2\rho\over\rho^2}\right ) \bigg\}.\end{aligned}\ ] ] putting both parts together , @xmath140 drops out as it should and we arrive at @xmath154 \left ( \ln{2\over a\rho z } - c \right ) \nonumber\\ & & + { z^2\over\xi^2}\ln z - { 19\over 3}\sin^2\rho - { 4\over 3}\left(1-{\sin^2\rho\over\rho^2}\right ) \bigg\}.\end{aligned}\ ] ] applying an integration by parts to one of the terms , this can be rewritten as @xmath155 \left ( \ln{2\over a\rho z } - c \right ) \nonumber\\ & & + \left ( 1-{\sin 2\rho\over 2\rho}\right ) - { 19\over 3}\sin^2\rho - { 7\over 3}\left(1-{\sin^2\rho\over\rho^2}\right ) \bigg\}.\nonumber\\ \\end{aligned}\ ] ] at @xmath16 , the cross section resulting from eq.([rho ] ) in the leading order in an expansion with respect to @xmath26 coincides with eq.([bh ] ) . the next - to - leading order term for the cross section reads @xmath156.\end{aligned}\ ] ] the correction @xmath157 is negative . this is due to the fact that it not only contains the process of two - photon pair production but also includes the @xmath26-corrections to one - photon pair creation ( cf . the latter is negative and has a larger absolute value than the contribution from two - photon pair creation . now we consider the case @xmath18 , which corresponds to a supercritical laser field strength @xmath19 . note that , for the very high frequencies ( @xmath97 ) assumed , such extremely large field strengths are far beyond present and near - future technical capabilities . in this situation , the main contribution to the integral over @xmath89 in eq.([rho ] ) comes from the region @xmath158 . performing the according expansion of the integrand and taking the integral , we obtain @xmath159.\end{aligned}\ ] ] we notice that the leading logarithm in eq.([sigmahe ] ) has already been derived by yakovlev @xcite . further , eq.([sigmahe ] ) agrees with the corresponding asymptotic limits found in ref.@xcite for pair creation by a coulomb field and a constant crossed field and in ref.@xcite for pair creation in a coulomb field and a constant homogeneous field . it is worth stressing that the @xmath17-dependence in eq.([sigmahe ] ) can not simply be expressed in terms of the effective ( laser - dressed ) electron mass @xmath160 @xcite . for the general case of an elliptically polarized laser wave the expression corresponding to eq.([whe ] ) reads @xmath161 \nonumber\\ & & \times \bigg[{\xi_1 ^ 2+\xi_2 ^ 2\over g_2}(a_1 + 2\sin^2\rho ) - 2\ln{\chi\over 2 } \nonumber\\ & & + { 1\over a_0}\left({g_0\over g_2}-1\right)(a_1 - 3\sin^2\rho)\bigg ] \nonumber\\ & & + { \xi_1 ^ 2+\xi_2 ^ 2\over g_2 } \bigg[a_1\left(\ln{g_0+g_2\over 2g_2 } - { 4\over 3}\right ) \nonumber\\ & & + 2\sin^2\rho\left ( \ln{g_0+g_2\over 2g_2 } - { 19\over 12 } \right)\bigg ] \nonumber\\ & & -\ln{\chi\over 2}\left(\ln{\chi\over 2}-{5\over 3}\right ) -{1\over 2}l\left({g_2-g_0\over g_2+g_0}\right ) \nonumber\\ & & + { 1\over a_0 g_2}\bigg [ a_1\left(g_0\ln{g_0+g_2\over 2g_2}-{4\over 3}(g_0-g_2)\right ) \nonumber\\ & & -3\sin^2\rho\left(g_0\ln{g_0+g_2\over 2g_2}-{3\over 2}(g_0-g_2)\right)\bigg]\bigg\}\end{aligned}\ ] ] with @xmath162 , @xmath163 , and spencer s function @xmath164 in the limit @xmath165 , we again can expand the integrand for small values of @xmath89 and , within logarithmic accuracy , arrive at the following cross section : @xmath166 here , @xmath167 denotes a hypergeometric function , which smoothly depends on the ellipticity parameter @xmath168 and monotonously decreases from unity for @xmath127 ( circular polarization ) to @xmath169 for @xmath132 ( linear polarization ) . now we consider the limit @xmath18 . note that today s most powerful lasers achieve @xmath170 . in this limit , the main contribution to the integral over @xmath89 in eqs.([circ2 ] ) and ( [ gen2 ] ) comes again from the region @xmath158 . hence , for the case of circular polarization , we can write @xmath171 \nonumber\\ & & + { 2\over 3}\left [ y_1(\beta)-{1\over u}y_1(\beta u)\right]{(1-t^2)^{3/2}\over t^2 } \nonumber\\ & & + \big [ y_1(\beta)- u y_1(\beta u)\big ] \ln\left({1+\sqrt{1-t^2}\over 1-\sqrt{1-t^2}}\right ) \bigg\},\end{aligned}\ ] ] where now @xmath172 with @xmath173 and @xmath174 . concerning the laser field parameters , the total rate in eq.([circquasi ] ) only depends on the dimensionless ratio @xmath175 , which is the so - called quasiclassical parameter . in the limit @xmath176 we arrive at the production rate @xmath177,\end{aligned}\ ] ] which corresponds to the cross section ( [ sigmahe ] ) obtained in sec.iii.c . in the opposite limit @xmath178 , which corresponds to the existing high - power lasers in the optical or infrared frequency range , the integral in eq.([circquasi ] ) can be evaluated using laplace s method . the resulting pair creation rate @xmath179 exhibits a tunneling behaviour . equation([sigmaquasi ] ) is in agreement with corresponding results on pair production in a coulomb field and a constant crossed field @xcite or a coulomb field and a constant homegenous field @xcite . in a similar way we can derive from eq.([gen2 ] ) the @xmath178 limit for an elliptically polarized laser wave . to this end , we assume @xmath180 and @xmath181 , where @xmath182 for @xmath183 . the latter condition means , that the laser field is considerably different from a circularly polarized wave . then we find @xmath184 in comparison with the result for circular polarization in eq.([sigmaquasi ] ) , the production rate in eq.([ratebigeta ] ) is suppressed by an additional factor of @xmath185 . this suppression has the same nature as in the case of @xmath24 pair production by a real photon in a laser field @xcite . the reason for the suppression is that , in general , the modulus of the field strength is truely oscillating , while in a circularly polarized wave it has a constant value . we note that eq.([ratebigeta ] ) can also be derived by averaging eq.([sigmaquasi ] ) over one oscillation cycle . we have investigated in detail the total probability for electron - positron pair creation by a nuclear coulomb field and an intense laser field . employing the optical theorem in connection with the known polarization operator of a photon in an electromagnetic plane wave , we have derived compact , general expressions for the total pair production rate and found explicit formulae for various intensity and frequency regimes of interest . we have demonstrated that significant differences exist between the general case of an elliptically polarized laser wave and the special case of circular polarization [ see , e.g. , eqs.([fmu ] ) and ( [ ratebigeta ] ) versus eqs.([flim ] ) and ( [ sigmaquasi ] ) ] . particular emphasize was placed on the nonlinear process of two - photon pair creation . let us estimate the feasibility of @xmath24 pair production using an xfel beam @xcite and a beam of relativistic protons . we will compare the following possible regimes of interaction : 1 ) the one - photon perturbative regime , where @xmath16 and @xmath186 [ see eq.([bh ] ) ] , 2 ) the two - photon perturbative regime , where @xmath16 and @xmath102 [ see eq.([sigmagen ] ) ] , 3 ) the tunneling regime , where @xmath18 and @xmath187 [ see eq.([sigmaquasi ] ) ] , and 4 ) the overcritical - field regime , where @xmath18 and @xmath188 [ see eq.([rateloweta ] ) ] . we have to take into account that the mentioned formulae are valid in the rest frame of the proton , which moves with a large lorentz - factor @xmath1 . in the rest frame of the proton , the photon frequency is increased with respect to its value in the lab frame : @xmath189 , while the xfel pulse duration is decreased : @xmath190 . here , @xmath191 and @xmath192 denote the laser frequency and pulse duration in the lab frame , respectively . note that , nowadays , the proton beam at desy can be accelerated up to 920 gev , which corresponds to @xmath193 . for the following estimates , we will throughout assume an xfel pulse duration of @xmath194 fs . the xfel beam collides with a bunch of protons moving at @xmath195 , containing 10@xmath196 particles , and having a beam radius of 30 @xmath168 m . in the one - photon perturbative regime , assuming an xfel photon energy of @xmath197 kev , an intensity parameter of @xmath198 , and a beam radius of @xmath199 @xmath168 m , we find that @xmath200 pairs are produced per collision . for a given laser intensity , the number of created particles is independent of the laser s polarization state . in the second regime of two - photon pair creation , employing a linearly polarized xfel beam is more favorable . supposing the collision parameters @xmath201 kev , @xmath202 , and @xmath199 @xmath168 m , the process proceeds far above threshold and one pair can be produced per shot . in the overcritical - field regime , circular polarization of the xfel beam leads to higher production yields . in this regime , a large @xmath17-value is necessary , which can be achieved by focusing the xfel beam @xcite . for @xmath203 , @xmath197 kev and @xmath204 nm , one pair can be produced per collision . here one has to take into account that the xfel pulse will interact only with a very small fraction of the proton bunch . for the assumed parameters , the latter has a length of 10 cm and contains a particle density of 10@xmath205 @xmath206 . thus , we see that , in principle , experimental studies of the pair production process are possible by using an xfel and a relativistic proton beam . employing a non - focused xfel beam ( cf . the so - called design and available parameters in table 1 of ref.@xcite ) and an ultra - relativistic proton beam , the perturbative regimes of pair creation can be realized , including nonlinear pair production via two - photon absorption . the realization of the strong - field regimes with @xmath18 will require focusing of the xfel beam ( cf . the goal parameters in table 1 of ref.@xcite ) . in this case the interaction volume , i.e. the region of beam overlap , is extremely small which substantially decreases the total production probability . as a result , the pair creation seems unrealistic in the tunneling regime but , at least in principle , feasible in the overcritical - field regime . in summary , with the novel sources of spatially coherent x - rays along with the relativistic proton beams of desy , the results on pair production presented in this paper could be tested . a.i.m . thanks the max - planck - institute for nuclear physics for its kind hospitality during an extended guest researcher appointment when this work was done . u.d.j . acknowledges support from the deutsche forschungsgemeinschaft ( heisenberg program ) . see , e.g. , e. brezin and c. itzykson , phys . d * 2 * , 1191 ( 1970 ) ; v. s. popov , zh . . fiz . * 61 * , 1334 ( 1971 ) [ sov . jetp * 34 * , 709 ( 1972 ) ] ; yad . fiz . * 19 * , 1140 ( 1974 ) [ sov * 19 * , 584 ( 1974 ) ] ; pisma zh . . fiz . * 74 * , 151 ( 2001 ) [ jetp lett . * 74 * , 133 ( 2001 ) ] ; v. m. mostepanenko and v. m. frolov , yad . fiz . * 19 * , 885 ( 1974 ) [ sov . * 19 * , 451 ( 1974 ) ] ; a. di piazza , phys . rev . d * 70 * , 053013 ( 2004 ) . s. adler , ann . of phys . * 67 * , 599 ( 1971 ) ; v. o. papanyan and v. i. ritus , zh . . fiz . * 61 * , 2231 ( 1971 ) [ sov . jetp * 34 * , 1195 ( 1972 ) ] ; a. shabad , nuovo cim . lett . * 3 * , 457 ( 1972 ) ; v. n. baier , v. m. katkov , and v. m. strakhovenko , zh . . fiz . * 68 * , 405 ( 1975 ) [ sov . jetp * 41 * , 198 ( 1975 ) ] .

particles which locally strongly interact with their environment polarize this latter and finish up in self - trapped states . on a local level , the environment surrounding the particles can be described in terms of a coherent superposition ( in form of glauber states ) of the elementary bosonic excitations of the bare environment . as a result , new composite entities form , given by the charge carriers and their surrounding clouds of bosonic excitations of the environment . it is at present a matter of dispute whether in the normal state such composite entities can exist as itinerant bloch - like states or are purely diffusive . the retarded interaction between the charge carriers and the bosonic excitations of the environment favors the picture of diffusive motion caused by a dephasing of the dynamics of the two constituents . the question we want to address here is whether this process of dephasing can be blocked , leading to a delocalization when such self - trapped composite entities condense into a _ macroscopic _ coherent quantum state . we investigate such a possibility in terms of a model of very general form , given by bosonic charge carriers strongly coupled to local bosonic excitations of the environment . the bosonic charge carriers can , amongst others , represent electron pairs induced by strong electron - phonon or electron - spin fluctuation interaction , electron pairs in double valence fluctuation systems ( stabilized by their ligand environment ) , frenkel excitons ( stabilized by their interaction with the surrounding lattice deformations ) and , in a more remote sense , he@xmath0 atoms in a bath of he@xmath1 with whose excitations they interact . such bosonic entities can in principle condense into a superfluid state . our interest here is to show how in such a _ macroscopic _ coherent quantum state the internal structure of the _ local _ coherent quantum state of the bosonic excitations of the environment is modified . we shall investigate this problem on the basis of a model of itinerant electrons , coupled to some purely local vibrational modes . for moderately strong and strong coupling , it results in self - trapped composite entities , given by on - site electron pairs , surrounded by local lattice deformations which are described by clouds of phonons in local coherent quantum states . in the intermediate coupling limit these localized entities can be considered as two - particle resonant states of the underlying system of itinerant electrons . they form out of pairs of such uncorrelated itinerant electrons and ultimately decay into them . a model hamiltonian which captures such physics is given by the boson - fermion model ( bfm ) @xmath2 \nonumber \\ & & - \hbar \omega_0 \alpha \sum_i \left ( \rho_i^z + \frac{1}{2 } \right ) ( a_i+a_i^{+ } ) + \hbar \omega_0 \sum_i \left(a^{+}_i a_i + \frac{1}{2}\right ) \ , . \nonumber \\\end{aligned}\ ] ] here @xmath3 denote the creation and annihilation operators of hard - core bosons ( characterizing the self - trapped localized electron - pairs ) on some effective sites @xmath4 , which should be understood as of being made up of small clusters of atoms rather than single atomic sites . such hard - core bosons have spin-@xmath5 commutation relations @xmath6_-=2 \rho_i^z$ ] and @xmath7_+= \delta_{ij}$ ] . @xmath8 denote the fermionic operators referring to annihilation ( creation ) of itinerant electrons with spin @xmath9 , @xmath10 being the number operator for such fermions . @xmath11 denote annihilation ( creation ) operators of the excitations of the environment , related to the local lattice displacements @xmath12 where @xmath13 is some atomic mass characterizing the effective sites , and @xmath14 is the frequency of the local deformations of the lattice . the bare hopping integral for the itinerant electrons is given by @xmath15 , the bare electron half bandwidth by @xmath16 ( which will serve as energy unit ) , @xmath17 denoting the coordination number . @xmath18 denotes the coupling of the hard - core bosons to the deformations of the surrounding medium and @xmath19 the exchange coupling between the bosons and the pairs of itinerant fermions . the bare boson energy level is denoted by @xmath20 . the chemical potential @xmath21 is common to fermions and bosons such as to guarantee overall charge conservation . the model defined by the hamiltonian ( 1 ) was introduced shortly after the proposition of the concept of _ bipolaronic superconductivity _ @xcite ( appropriate to the extreme strong coupling anti - adiabatic regime ) in an attempt to extend this concept to the regime of intermediate electron - lattice deformation coupling . this model is based on a conjecture , rather than on a derivation from an underlying basic electron - phonon hamiltonian , such as for instance the holstein model@xcite . we shall in the following develop the ideas which had initially led us to this conjecture . let us start from a simple holstein model with purely local einstein modes for the lattice vibrations . the corresponding hamiltonian is given by @xmath22 the meaning of the various operators and coupling constants is the same as in the hamiltonian ( 1 ) , except for the value of @xmath23 which is equal to @xmath24 . the electron - phonon interaction with coupling strength @xmath23 is taken to be local , in view of the interaction with primarily local phonon modes which favours small polaron formation . @xmath25 denotes some effective coulomb repulsion - not to be confused with a hubbard - type on - site interaction - on the effective sites , which , we stress again , ought to be considered as being composed of small molecular clusters such as a diatomic molecular unit or similar . electrons described by this hamiltonian can gain energy either by becoming itinerant and remaining essentially uncoupled to the lattice or , alternatively , getting self - trapped . in the first case the energy gain is described by the intrinsic band dispersion @xmath26 @xmath27 denoting lattice vectors linking nearest - neighbor sites . for self - trapped polaronic states , the local part of the hamiltonian ( [ h ] ) can be transformed by a shift in the phonon variables @xmath28 upon which it acquires the form @xmath29 with the polaron binding energy @xmath30 denoting the gain in energy by the corresponding polaron level shift . the effective on - site interaction given by @xmath31 controls the possible formation of bipolaronic states @xmath32 when @xmath33 , with the single polaron states lying above such bipolaron states . for @xmath34 we have the inverse situation with the bipolaronic level lying above that of the single polarons . what we are interested here is the first case . let us begin by considering the situation of sufficiently strong interaction such that the bipolaronic level lies below the bottom of the unrenormalized electron band , i.e. , for @xmath35 , as illustrated in fig.1 . in that case bipolarons can acquire itinerancy by decaying into virtual states of pairs of itinerant electrons and reassembling subsequently on some neighbouring site . within the usual lang - firsov approximation , neglecting phonon creation as well as annihilation processes during the charge transfer from one site to the other , the corresponding effective bipolaron hopping integral is then given by @xmath36@xcite . within this approximation one obtains a system of itinerant bipolarons on a lattice which , at low temperatures , can condense into a superfluid state , giving rise to what has been termed _ bipolaronic superconductivity_. this theoretical proposition has ever since remained a matter of dispute , questioning if such a normal - state itinerant bose liquid can be achieved in real materials . the main arguments against it is that the stability of local bipolarons requires a relatively high value of @xmath37 for realistic values of @xmath38 , @xmath25 and @xmath14 and hence leads to exceedingly small values of the bipolaron bandwidth . a further hampering factor for the realization of such bipolaronic itinerant band states is related to the retardation effects between the motion of the charge carriers and the lattice deformations which surround them . this gives rise to dephasing between the two , as shown by exact diagonalization approaches@xcite and dmrg studies@xcite , indicating that incoherent processes in the optical conductivity are dominant . all these results suggest that coherent bipolaron motion is most unlikely to occur in real systems and it was for that reason that the boson - fermion model was initially proposed@xcite . the aim was to capture the physics outside the regime of the extreme strong coupling antiadiabatic limit and to describe the intermediate coupling case , where the bipolaron level lies inside the band of the itinerant electrons ( see fig.2 ) and where the superconducting state is controlled by the phase correlations of these resonant bosonic bipolaron states . quite generally , this model can be considered as a paradigm for cross - over situations . for the specific case we are interested here , it describes a cross - over between a bcs - type superconductor of cooper pairs in the weak coupling regime and the hypothetical , yet to be experimentally verified , _ bipolaronic superconductivity _ in the strong coupling regime . in a totally different physical situation , namely that of a strongly correlated electron system such as described by the hubbard model , it has been recently shown that the intermediate coupling case can be mapped onto such a boson - fermion scenario@xcite . such an effective boson - fermion hamiltonian was also derived recently for the exchange interaction between spinon singlets of the rvb electron - pairs and holons@xcite . the question to what extent such a bfm can emerge from a quite general class of interacting fermion systems , has been addressed from the point of view of a bosonization procedure of an intrinsically fermionic system@xcite . the presently much discussed molecular bose - einstein condensate , involving entangled atoms in squeezed states has also recently been analyzed@xcite in the framework of such a bfm . a derivation of the bfm for moderately strongly coupled electron - phonon systems , starting from the holstein model , is so far not available . in order to provide the physically intuitive picture which initially had led us to propose this model , let us consider the case where the local bipolaron states are stable relative to single polarons . this eventuality can only be decided on the basis of direct experimental evidence for such local bipolaronic entities . the classical cases which initially had led to consider such a situation are systems like amorphous semiconductors@xcite , ti@xmath39o@xmath40@xcite , na@xmath41v@xmath42o@xmath43@xcite , wo@xmath44@xcite and double valence fluctuation systems@xcite like pbte(tl ) , just to name a few of them , and all of which represent such bipolarons as _ exclusively _ localized self - trapped states . when these systems are doped , the electrons responsible for pairing are at the same time responsible for the - albeit diffusive - transport . it is hence tempting to hypothesize that such bipolaronic states can occur as resonant states inside a band of itinerant electrons . in more complex systems , such as ternary and quaternary compounds , different electrons come into play . some of them are prone to form bipolarons and others will remain itinerant , but being coupled together via some hybridization term . in systems like that , such a boson - fermion scenario imposes itself more directly . it is the eventuality of bipolaronic resonant states rather than localized states , which was at the origin of our proposal of the boson - fermion scenario@xcite . the level scheme for the local bipolaronic states inside the band of the bare itinerant electrons for this situation is depicted in fig.2 . filling up such a system gradually with electrons , it is evident from fig.2 that , as long as the fermi level remains below the level of the local bipolaronic states , the charge carriers will be essentially bare itinerant electrons . but upon further increasing the concentration of charge carriers , the itinerant electrons will eventually overlap in energy with the intrinsically localized bipolaronic level . the major supposition leading to this bfm is now to assume that this overlap can be simulated by some effective on - site hybridization describing a charge exchange between the bipolarons and pairs of itinerant electrons in the form of @xmath45 \ ; , \ ] ] with the itinerant electrons and the localized bipolarons , eq.(5 ) , treated as commuting fields . such a supposition is supported by the fact that the spectrum of electrons , moderately strongly coupled to local modes , is known to be given by the bare spectrum for all @xmath46-vectors , except where their energy crosses the level of localized selftrapped electrons . in this regime of energies the electron spectrum is strongly modified , as shown by exact diagonalization studies@xcite as well as variational methods@xcite , with the spectrum becoming rather flat and approaching the local level of the self - trapped polaronic states . these are features which are characteristic of systems where itinerant charge carriers are hybridized with localized charge carriers , and this was the center of our initial assumption of an effective hybridization between localized bipolarons and pairs of itinerant electrons , as given by eq.(6 ) . on a local level the dynamics of a polaronic system , involving two - particle resonant states , is controlled by the atomic limit of such a system , which we thus propose to describe by the following hamiltonian @xmath47 + \sum_i \hbar \omega_0 \left(a^{+}_ia_i + \frac{1}{2}\right ) \\ & \equiv & ( u-2\mu ) \sum_i \left(\rho_i^z+\frac{1}{2}\right)+ v\sum_i \left[\rho^+_ic_{i\downarrow}c_{i\uparrow } + \rho_i^- c^+_{i\uparrow}c^+_{i\downarrow}\right]- \hbar \omega_0 \alpha \sum_i \left(\rho_i^z + \frac{1}{2}\right ) \left(a_i+a_i^{+}\right)+\sum_i \hbar \omega_0 \left(a^{+}_ia_i + \frac{1}{2}\right ) , \nonumber\end{aligned}\ ] ] with a bare bipolaron level given by @xmath48 . the equivalent formulation of this hamiltonian [ the second part of eq.(7 ) ] is obtained after employing an inverse lang - firsov - type transformation @xmath49 with @xmath50 . adding to this local hamiltonian the term describing the kinetic energy of the itinerant electrons , leads immediately to the bfm hamiltonian given by eq.(1 ) . the basic physics contained in this bfm model originates from a competition of the local electron pairing , acting as a prerequisite for superconductivity , with the delocalizing effect of the itinerant electrons which favors non - local cooper pairing as the temperature is lowered and the superconducting phase is approached . the change - over from local pairing to cooper pairing is not only manifest in the electronic properties of such systems but is expected to be accompanied by corresponding lattice deformations , the study of which is the issue of the present paper . we shall examine this feature for the two following extreme limits : \i ) the normal - state phase , characterized by the existence of a pseudogap in the density of states ( dos ) of the itinerant electrons , which will be treated here in an approximate fashion within the atomic limit of the bfm , \ii ) the superconducting phase , which will be treated on a mean - field level , taking into account the phase locking of the bound electron pairs . throughout this study we choose as a representative example a set of parameters presenting a situation close to the fully symmetric case . this implies an effective bipolaron level ( corresponding to the energy level for the bound pairs ) given by @xmath51 , resulting in a final state bipolarons level lying close to the center of the band of the itinerant electrons , after the boson - phonon coupling has been switched on . we moreover want to consider the case @xmath52 , giving a concentration of electrons @xmath53 and a concentration of bound pairs @xmath54 and choose @xmath55 , @xmath56 and @xmath57 . let us start from the purely local physics , described by the atomic limit ( @xmath58 ) of the bfm corresponding to the hamiltonian @xmath59 , eq.(7 ) . this limit has been studied by us previously@xcite in connection with the pseudogap phenomenon and the incoherent background in the electron spectral function , which , in such a scenario , arise from phonon shake - off processes . for the choice of parameters adopted here , the physics controlling the change - over from the normal into the superconducting phase can essentially be represented by the one- , two- and three - particle eigenstates of @xmath59 @xmath60 with @xmath61 denoting the @xmath4-particle @xmath62-th eigenstate . the opening of the pseudogap at a certain temperature @xmath63 shows up in the rapid spectral weight increase , with decreasing temperature below @xmath63 , of the scattering amplitude involving transitions from the lowest energy two - particle bonding state @xmath64 to the lowest energy single - particle states @xmath65 and from the lowest energy three - particle states @xmath66 to @xmath64 . these contributions to the dos of the itinerant electrons are manifest by the wings ( see fig.3 ) for energies below and above that of the single particle non - bonding contributions , which have spectral weight around the chemical potential ( equal to @xmath67 ) and which arise from transitions not involving any of the two - particle states . decreasing the temperature below @xmath63 , the spectral weight at @xmath67 decreases rapidly and transfers to the wings of the dos , thus resulting in the opening of a pseudogap . taking into account the dispersion of the itinerant electrons in an approximate fashion , such as within a coherent potential approximation ( cpa ) approach@xcite , the main features seen in fig.3 are maintained , simply leading to a smearing of the dos . ) , as the temperature decreases below @xmath63.,width=302 ] the two - particle eigenstates of @xmath59 , eq.(9 ) , are obtained by expanding the local oscillator states in terms of the eigenstates @xmath68 of the undeformed harmonic oscillator at a given site @xmath4 . @xmath69 and @xmath70 denote respectively the vacuum states of the itinerant electrons and of the bound electron - pairs on such a site . the coefficients @xmath71 and @xmath72 are determined by exact diagonalization of this atomic limit problem within a restricted hilbert space of those excited oscillator states . it is generally sufficient to take into account up to @xmath73 such states in order to get converging results . the qualitative changes in the lattice deformations expected to occur when going from the pseudogap phase into the superconducting phase are essentially described by the lowest energy one- , two- and three - particle states , which contain the information on the relative weight of the bound electron - pairs ( @xmath74 ) and of the induced pairing in the subsystem of the itinerant electrons ( @xmath75 ) . the first ones are intrinsically coupled to the local lattice deformations while the second ones acquire such local lattice deformations due to the charge exchange mechanism acting between the two . as we shall see below , the efficiency of this transfer of polaronic features is noticeably different in the normal and in the superconducting phase . it shows up in the composition @xmath76 of the coherent phonon states describing the local phonon clouds surrounding the charge carriers on a given site @xmath4 , which in the normal state ( atomic limit ) is given by @xmath77 @xmath78 denoting the partition function and @xmath79 the eigenvalues of @xmath59 associated with the eigenstates @xmath80 . upon decreasing the temperature , @xmath81 shows a redistribution of weight from high to small values of the phonon number @xmath82 [ see fig.4(a ) ] . in the atomic limit and in the superconducting state for different temperatures @xmath83.,width=302 ] in the temperature regime of interest here , i.e. , from slightly above @xmath63 all the way down to @xmath84 , the contributions to @xmath76 come essentially from the lowest energy one- , two- and three - particle states , eqs.(8 - 10 ) . as illustrated in fig.5 , these various contributions have noticeably different features as far as the phonon distributions are concerned . the one - particle states are totally decoupled from the lattice and hence have a contribution only for @xmath85 , while the three - particle contribution are maximally coupled to the lattice showing a broad peak in this distribution which corresponds to that of a shifted oscillator ground state . the two - particle bonding state @xmath64 has features of a displaced oscillator which is particularly noticeable at low temperatures where it is the main contribution , together with a much reduced contribution from the two - particle antibonding state @xmath86 . that latter contribution to @xmath76 , considered without thermal weighting factor , clearly shows strong contributions for low values of @xmath82 , which is indicative of weak correlations between the electronic degrees of freedom and those of the lattice . in the normal state ( full squares ) into contributions coming from the one- , two- and three - particle states ( @xmath87 , @xmath88 and @xmath89 , respectively ) . @xmath90 and @xmath91 represent the contributions coming respectively from the lowest energy two - particle state ( bonding ) and from the first excited two - particle state ( antibonding).,width=264 ] the onset of the local coherence between bound pairs and of the induced pairing in the itinerant subsystem can be tracked in the pair distribution function ( pdf ) @xmath92 , given in the normal state ( atomic limit ) by @xmath93 which describes the probability of finding a local deformation of size @xmath94 and which can be studied by exafs and pulsed neutron scattering experiments@xcite . @xmath95 is obtained by expressing the eigenstates of @xmath59 in terms of a real - space representation involving the harmonic oscillator wavefunctions . in fig.6(a ) we illustrate the temperature evolution of the pdf . its decomposition into the contributions arising from , measured by the dimensionless parameter @xmath96 , in the atomic limit and in the superconducting phase for different @xmath83.,width=264 ] , @xmath97 and @xmath98 , respectively ) . @xmath99 and @xmath100 represent the contributions coming respectively from the two lowest bonding and antibonding two - particle eigenstates of @xmath59.,width=302 ] the one- , two- and three- particle states given in eqs.(8 - 10 ) are illustrated in fig.7 . the one- and three - particle states contribute to the pdf in form of peaks centered around @xmath101 and around some finite value of @xmath94 , respectively . this expresses the fact that the one - particle states ( 8) are completely decoupled from the lattice , while the three - particle states ( 10 ) contain a bound electron - pair and thus give a maximal shift in the pdf . the two - particle bonding and anti - bonding states , eq.(9 ) with @xmath102 , respectively , give contributions to the pdf with two peaks lying very close together , which is a result of the strong correlation between the bound electron pairs and the induced pairs in the itinerant electron system , characterizing the normal - state phase . at high temperatures , the pdf is thus given by a very broadened smeared out double - peak structure . but as the temperature is lowered below @xmath63 , local coherence between the bound electron pairs and the pairs of the itinerant electrons sets in . as a result , the opening of the pseudogap in the local density of states is accompanied by a change - over in the pdf from this very smeared out double - peak structure to a rather well defined single - peak structure . the latter is essentially due to the lowest energy bonding two - particle eigenstate contributing to a pdf with a maximum at some finite value of @xmath94 . this is a signature of the strong correlations , which gradually build up as we go below @xmath63 , between the bound pairs and the induced pairing in the itinerant electron subsystem leading to roughly equal weight coming from the coefficients @xmath75 and @xmath74 . let us now put such single units , as described by @xmath59 , together in an infinite lattice , and let them interact with each other via the hopping term of the itinerant electrons . when the system undergoes a transition into a superfluid state of the charge carriers , the on - site correlations between bound pairs and pairs of itinerant electrons will be weakened , since the @xmath103-space pairing of the itinerant electrons enters into competition with the strong on - site pairing which controls the physics in the normal state . as a result , upon entering the superconducting phase , a substantial modification of the coherent phonon states of the individual units will occur . we shall , for a moment , make the assumption that the deformation of the medium surrounding the bosonic charge carriers is not influenced by boson - fermion charge fluctuations . in that case the coupling of the bound electron - pairs to the lattice deformations can be eliminated , along the lines usually adopted in studies of the polaron problem , leaving us with an effective hamiltonian given by the first four terms of the hamiltonian ( 1 ) and with a renormalized hybridization coupling @xmath104 . such a hamiltonian has been studied in detail within different schemes , such as self consistent diagramatic techniques@xcite , dynamical mean - field theory@xcite and renormalization group procedures@xcite , and gives rise to the following physics . as the temperature is lowered below a certain characteristic temperature @xmath63 , pairing of the itinerant electrons sets in , leading to an emptying out of the single electron spectrum near the chemical potential and thus to the opening of a pseudogap . upon further decreasing the temperature these electron - pairs become itinerant@xcite , eventually condensing into a macroscopic superfluid ground state of the form @xmath105|0)_i\ ] ] ( @xmath106 being some model dependent parameters ) and exhibiting soundwave - like collective excitations as a consequence of the rigidity of the local phases @xmath107 . in order to study how such a superfluid state affects the surrounding lattice deformations , we now investigate this problem on the basis of a mean - field treatment of the full hamiltonian with the coupling of the phonons to the bound pairs explicitly taken into account . to this purpose , we introduce the following order parameters for the charge carriers@xcite @xmath108 which we assume to be homogeneous in space . the corresponding mean field hamiltonian is given by : @xmath109 \nonumber \\ h_b & = & ( \delta_b-2\mu ) \sum_i \left ( \rho_i^z + \frac{1}{2 } \right ) + v x \sum_i [ \rho^+_i \ ; + \rho_i^- ] - \hbar \omega_0 \alpha \sum_i \left ( \rho_i^z + \frac{1}{2 } \right ) ( a_i+a_i^{+ } ) + \hbar \omega_0 \sum_i a^{+}_i a_i \ ; .\end{aligned}\ ] ] on the level of this approximation , the effect of the charge fluctuations between bound pairs and pairs of itinerant electrons is diminished , being taken into account in a global , spatially independent , fashion . the itinerancy of the intrinsically localized bound electron pairs , induced by the exchange coupling @xcite is not contained in the present approximation and therefore we can not account for collective soundwave - like excitations and the spatially correlated lattice deformations which result from them . nevertheless , the present study does permit to investigate local excitations of the lattice , controlled by the local fluctuations of the order parameters to which they are coupled , and which can be related to a number of experiments . considering itinerant bosons on a deformable lattice with weak boson - phonon coupling@xcite , leads to a hybridization of the bogoliubov modes ( coming from the repulsive interaction between the bosons ) and the einstein - like phonon modes . at long wavelengths this results in a modified sound velocity of the bogoliubov mode , with little spectral weight carried by the phonons , and a moderately renormalized einstein mode , principally of phonon character . in such a context , what the present mean - field study permits us to do , is to investigate the effect of the superfluid state on such an einstein - like mode . due to the strong coupling between the bound electron pairs and the phonons , such an einstein mode will undergo a substantial renormalization , as it will be shown below . the fermionic part @xmath110 of @xmath111 , being dynamically decoupled from the bosonic part @xmath112 and hence from the phonons , can be cast into the standard form @xmath113 . the @xmath18 s and @xmath114 s denote the operators of bogoliubov quasi - particles having the dispersion @xmath115 which differs from the bare electron dispersion @xmath116 by showing a gap of size @xmath117 . the corresponding eigenstates of @xmath110 are given by the standard bcs expression @xmath118| 0 \rangle \ ; .\ ] ] the bosonic part @xmath112 of @xmath111 , showing intrinsic coupling between the bosonic charge carriers and the phonons , is diagonalized in terms of a set of states of the form @xmath119 | 0 ) _ i|n>_i\ ] ] having eigenvalues @xmath120 . the eigenstates of the mean - field hamiltonian , eq.(15 ) , are thus given by @xmath121 as in the atomic limit , the diagonalization of @xmath112 is carried out in a truncated hilbert space spanned by the eigenstates @xmath122 of the undeformed harmonic oscillator . we ultimately arrive at the following set of self - consistent equations determining the temperature dependence of the order parameters @xmath123 \ ; , \end{aligned}\ ] ] @xmath124 $ ] denoting the partition function corresponding to @xmath112 . this set of equations must be solved with the constraint that the total density of bosons and fermions remains fixed , i.e. @xmath125 , where @xmath126 \exp\,\left[-\beta e^{sc}_l(x)\right ] \ ; .\end{aligned}\ ] ] let us now examine the modification of the composition of the local coherent phonon states on the individual lattice sites ( discussed above on the basis of the atomic limit for the normal state ) when a macroscopic phase - coherent state of the charge carriers emerges and the bound elelctron - pairs condense into a superfluid state . the probability distribution of the phonons in the superconducting phase is then described by @xmath127 \left [ ( u^{sc}_{ln})^2 + ( v^{sc } _ { ln})^2\right ] \ ; .\ ] ] the evaluation of @xmath128 in the superfluid phase follows closely the calculation for @xmath81 in the atomic limit . in fact , the secular equation determining the coefficients @xmath129 and @xmath130 is formally the same as in the atomic limit , with the states @xmath131 and @xmath132 and @xmath133 , respectively . the only modifications which then appear in this set of equations are ( @xmath4 ) a renormalized hybridization constant @xmath134 , instead of @xmath19 in the atomic limit , and ( @xmath135 ) an energy level for the diagonal elements involving the states @xmath136 being equal to @xmath137 , instead of @xmath138 for the diagonal elements involving the states @xmath139 . this renormalization of @xmath19 is important since it reduces the bare @xmath19 by roughly an order of magnitude ( for the set of parameters chosen throughout this paper ) , expressing the fact that the superfluidity leads to a considerable reduction of the on - site correlation between the configurations with bound electron - pairs present and with such pairs being absent . this has significant consequences on the composition of the local phonon clouds surrounding the individual sites and thus on their corresponding pdfs . just below the superconducting @xmath140 ( @xmath141 0.0067 in our present case ) , @xmath128 is determined by an incoherent superposition of essentially the two lowest eigenstates of the mean - field hamiltonian , eq.(15 ) , one giving rise to a broad peak at some finite @xmath82 and the other to a sharp maximum at @xmath85 . at low temperatures , however , the ground state alone is sufficient to describe @xmath128 . the contribution to @xmath128 leading to a peak at some finite value of @xmath82 ( associated with the coefficients @xmath142 ) arises from the deformation induced by the presence of the bound electron pairs , while the contribution to @xmath128 leading to a maximum for @xmath85 ( associated with the coefficients @xmath143 ) arises from the absence of such an induced deformation for configurations where no bound pairs are present , with the oscillators being essentially undeformed . the coherent superposition of deformed and undeformed oscillator states in the ground state is a direct consequence of the macroscopic phase locking of the composite bosonic states in the superconducting phase , resulting from the variational ground state of @xmath112 of the form @xmath144 the operators @xmath145 are defined as @xmath146 and differ significantly from the usual polaron shift operators @xmath147 , characterizing the deformations of a simply shifted oscillator state . _ macroscopic _ phase locking and homogeneity imply that the coefficients @xmath148 and @xmath149 as well as phases @xmath107 are the same for all sites . eq.(22 ) is a direct generalization of the superfluid ground state of a system of hard - core bosons on a lattice , which is of the form given in eq.(13 ) , and thus shows explicitly the locking together of the charge @xmath150 and lattice ( @xmath145 ) degrees of freedom on a local level . ( squares ) in the superconducting phase in terms of the contributions @xmath151 coming from the ground state ( triangles ) and @xmath152 coming from the first excited state ( circles ) of @xmath112.,width=264 ] we shall next address ourselves to the question of how such a _ macroscopic _ coherent quantum state of the phonons can be tracked experimentally and , in particular , what difference there is to be expected with respect to the normal state features of the pdf , examined in the previous section . the pdf in the superconducting phase is given by @xmath153 in the superconducting phase ( full lines ) into the contributions @xmath154 coming from the ground state ( dashed ) and @xmath155 coming from the first excited state ( pointed ) of @xmath112.,width=264 ] in fig.[pdf0 ] we compare the pdf in the normal state and in the superconducting phase for different temperatures . the features examined above on the basis of the composition of @xmath128 are perfectly reflected in the pdf . in the normal state one observes a single peak at some finite @xmath94 , indicative of well defined shifted oscillator states . in the superconducting phase , on the contrary , one observes a weakly temperature dependent double - peak structure , which , as we can see from fig.[pdf2 ] , arises essentially from two contributions associated with the ground state and the first excited state of @xmath112 . at temperatures just below @xmath140 this double - peak structure results from an incoherent superposition of these two states , corresponding to a situation where localized bipolarons are randomly distributed over the lattice . the statistical distribution of empty and doubly occupied sites will thus give rise to that double - peak structure . in the limit of low temperature , on the contrary , the pdf arises @xmath156 from the ground state , with a peak centered near @xmath101 ( being determined by the coefficients @xmath143 ) and a peak centered around a finite value of @xmath94 ( being determined by the coefficients @xmath142 ) . the appearence of a double - peak structure in the pdf has long since been recognized as a signature of delocalization of the charge carriers in polaronic systems@xcite . the present study shows that , as the superconducting state sets in and is strengthened upon lowering the temperature , the _ macroscopic _ phase coherence of the system not only involves the charge carriers but also the deformations which surround them . this is only possible if at @xmath84 these deformations are controlled by a single quantum state , i.e. , the ground state @xmath157 . the superconductivity - induced delocalization of the self - trapped entities which characterize the normal state , is caused by the locking together of the local lattice deformations in a coherent superposition of deformed and essentially undeformed oscillator states , resulting in the developement of a double - peak structure in the pdf . at least in the scenario presented here , this result does not bear out any undressings@xcite of the charge carriers , stripping off partially the phonon clouds , as one enters the superconducting phase . such an effect would show up in a pdf which , in the superconducting phase , would be given by essentially a single - peak structure centered around @xmath101 . the discussion in the previous section demonstrated the qualitative difference which exists between some basic lattice properties in the normal and the superconducting phase , and largely being controlled by the induced electron pairing in the subsystem of the intrinsically uncorrelated itinerant electrons . what results is a competition between the local pairing , characterizing the normal - state phase , and the * k*-space pairing in the superconducting phase , expected to be particularly relevant around the superconducting transition where the spatial phase fluctuations play a major role and residual local electron pairing@xcite persists in the superconducting phase . the qualitative difference in the structure of the pdf in the two phases leads us to expect squeezing effects of the local coherent phonon quantum states which are stronger in the pseudogap phase than in the superconducting phase . associated with it , one should expect larger fluctuations of the local deformations of the environment and of the momenta of the ions representing it . these fluctuations are given by @xmath158 with @xmath159 the expectation values in eq.([xp ] ) have to be calculated with respect to the eigenvectors of @xmath59 and @xmath160 for , respectively , for the normal and the superconducting phase . the results for these fluctuations in the two phases are illustrated in fig.[dxdp ] . in particular , the product @xmath161 clearly shows that the coherent phonon states in the normal state are considerably squeezed , i.e. , this product is for @xmath162 very close to the lower limit of the heisenberg uncertainty principle , equal to @xmath163 . in the superconducting phase , on the contrary , we observe practically no squeezing effect with decreasing temperature , the product @xmath164 being always bigger than what would be expected in the normal state in this temperature regime . a possible experimental verification of these features comes from neutron absorption measurements@xcite which measure the kinetic energy @xmath165 of the local lattice vibrations . indeed , the experimental results shown that , as the temperature is reduced , @xmath166 saturates in the superconducting phase and is distinctly above the values expected in the normal state at the same temperature . this fact is born out in our study , as can be seen by inspection of fig.[dxdp](b ) . ( @xmath167 ) and with @xmath168 . ( c ) degree of the squeezing of the coherent phonon states as a function of @xmath169.,width=264 ] finally , an issue which ought to be of crucial interest in studying the lattice properties connected with the onset of the superconducting order or of the electron pairing at @xmath63 in the normal state , is the question of the changes in the phonon spectral properties upon going from one phase to the other . strong changes in the frequency of specific local modes , induced by strong electron - phonon coupling , have been well established by photoinduced infrared absorption measurements@xcite and been related to strong local displacements of the ions . in this context , the appearence of new additional modes occuring with changes from the normal into the superconducting phase have also been reported , although this is still a matter of debate@xcite . more established are the strong intensity changes of certain raman active modes and the strong frequency renormalizations and lifetime broadening occuring at such transitions@xcite . such effects can best be studied by energy loss spectra in inelastic neutron scattering or raman spectroscopy where phonons are emitted . the cross - section for it is given by @xmath170 with @xmath171 here @xmath172 and @xmath173 are the eigenvalues and eigenvectors of the hamiltonians @xmath59 and @xmath112 , respectively . we illustrate the results of such phonon emission spectra for the two phases in figs . [ ph1 ] and [ ph2 ] . quite generally , one observes a quasi - elastic contribution at frequency zero , which is the standard signature of coherent phonon states . as the temperature is lowered below either @xmath63 or @xmath140 and the itinerant electrons get correlated by local pairing in the normal state or by cooper pairing in the superconducting phase , the local phonon mode first splits into several modes and finally evolves into a well defined mode which is softer in the pseudogap phase and harder in the superconducting phase . these features of the phonon spectral properties are expected to hold qualitatively true . a quantitative link - up between the normal phase and the superconducting phase as the temperature is decreased , would require a full description of the local boson - fermion exchange correlations , incorporating the electron itinerancy in the normal state and allowing for local electron pairing in the superconducting phase . these are presently largely unresolved fundamental problems in such system where amplitude and phase order occur independently , as in the high - temperature superconductors with its pseudogap phase above @xmath140 where local electron pairing persists in an incoherent fashion . the present study is an attempt to examine the condensed state of a system of composite particles such as charge carriers surrounded by clouds of bosonic excitations of the neighboring environment to which they are coupled and to see to what extent a macroscopic coherent quantum state can be induced in the system of these bosonic excitations . as a specific example we chose a system of electrons , moderately strongly coupled to local lattice deformations , which can be described in terms of resonant local - pair states ( localized bipolarons ) inside a fermi sea of essentially uncorrelated itinerant electrons with which they interact via a charge exchange term . we find noticeable qualitative differences in the structure of the local coherent quantum states of the phonons in the superconducting phase and in the phase above it which corresponds to a normal state characterized by a pseudogap in the dos of the single - particle spectrum . several measurable quantities related to static and dynamic lattice properties were examined , such as the pair distribution function , the fluctuations in the lattice positions and the corresponding ion kinetic energies of the atoms making up such a lattice and , finally , the phonon spectral properties given by energy - loss spectroscopy ( measurable by inelastic neutron scattering or raman spectroscopy ) . the difference between local electron pairing in the normal state and * k*-space pairing in the superconducting phase has noticeable consequences on such measurable quantities , resulting in ( @xmath4 ) a single - peak structure of the pdf in the normal state and a double - peak structure in the superconducting phase , ( @xmath135 ) distinctively larger positional fluctuations and kinetic energy of the atomic clusters ( making up the lattice ) in the superconducting phase than in the normal state and , ( @xmath174 ) associated with it , a more effective squeezing of the phonon coherent quantum states in that latter . for the superconducting phase , the present mean - field study is restricted to local phase fluctuations of the superconducting state given by a generalization of a phase - locked state of bosons on a lattice , eqs.(17,18,22 ) , in which these bosons correspond to composite particles in form of bipolarons [ see eq.(5 ) ] . the examination of soundwave - like collective excitations of such a system is planned for some future work . there , the question arises if the local dynamical deformations do get spatially correlated upon entering the superconducting state , as possibly suggested by ion channeling experiments@xcite . a possible approach to solve this problem could be to integrate out the fermionic sector of the bfm hamiltonian eq.(1 ) , resulting in a gross - pitaevskii lagrangian@xcite for the effective itinerant bosonic charge carriers coupled to the phonons , followed by a treatment similar to that adopted for bosons on a defromable lattice@xcite . alexandrov and j. ranninger , phys . b , * 23 * , 1796 ( 1981 ) . t. holstein , ann . ( ny ) * 8 * , 343 ( 1959 ) . e. de mello and j. ranninger , phys . b * 55 * , 14872 ( 1997 ) . c. zhang , e. jeckelmann and s. white , phys . rev . b * 60 * , 14092 ( 1999 ) . j. ranninger and s. robaszkiewicz , physica b * 135 * , 468 ( 1985 ) . e. altman and a. auerbach , cond - mat/0108087 . e. kochetov and m. mierzejewski , cond - mat/0204420 . r. friedberg , t.d . lee and h.k . ren , phys . b * 50 * , 10190 ( 1994 ) . v.a . yurovski and a. ben - reuven , cond - mat/0205267 . h. fehske , j. loos and g. wellein , z. phys . b * 104 * , 619 ( 1997 ) . w. stephan , phys . b , * 54 * , 8981 ( 1997 ) . romero , d.w . brown and k. lindenberg , phys . rev . b * 59 * , 13728 ( 1999 ) . anderson , mater . * 8 * , 153 ( 1976 ) . s. lakkis , c. schlenker , b.k . chakraverty , r. buder and m. marezio , phys . rev . b * 14 * , 1429 ( 1976 ) . chakraverty , j.m . sienko and j. bonnerot , phys . b * 17 * , 3781 ( 1978 ) . o.f . schirmer and e. salje , j. phys . c * 13 * , l1067 ( 1980 ) . chernick , s.n . lykov and n.i . grechko , sov . solid state * 24 * , 1661 ( 1982 ) . j. ranninger and a. romano , phys . lett . * 80 * , 5643 ( 1998 ) . t. domanski , molecular physics reports * 34 * , 39 ( 2001 ) . t. egami and s.j.l . billinge , in _ physical properties of high temperature superconductors _ , edited by d. ginsberg ( world scientific , singapore , 1996 ) vol . v , p. 265 ; j. rhler _ et al . _ , in _ workshop in high-@xmath140 superconductivity 1996 : ten years after the discovery _ , edited by e. kaldis _ et al . _ ( kluwer , dordrecht , 1997 ) p. 469 . j. ranninger , j.m . robin and m. eschrig , phys . lett . * 74 * , 4027 ( 1995 ) . robin , a. romano and j. ranninger , phys . lett . * 81 * , 2755 ( 1998 ) . t. domanski and j. ranninger , phys . b * 63 * , 134505 ( 2001 ) . p. devillard and j. ranninger , phys . lett . * 84 * , 5200 ( 2000 ) . j. ranninger and l. tripodi , solid state commun . * 112 * , 349 ( 1999 ) . g. jackeli and j. ranninger , phys . b * 63 * , 184521 ( 2001 ) ; _ ibid . _ * 64 * , 104513 ( 2001 ) . shore and l.m . sander , phys . b * 7 * , 4537 ( 1973 ) . j. hirsch , phys . b * 62 * , 14487 ( 2000 ) . i. kosztin , q. chen , b. janko and k. levin , phys . b * 58 * , r5936 ( 1998 ) . c. taliani _ et al . _ , in _ electronic properties of high-@xmath140 superconductors and related compounds _ , edited by h. kuzmany , m mehring and j. fink ( springer - verlag , berlin , series of solid state science , 1990 ) , p. 208 r. j. mcqueeney _ et al . _ , phys . rev . * 82 * , 628 ( 1999 ) . l. pintschovius and m. braden , phys . b * 60 * , r15039 ( 1999 ) . g. ruani and p. ricci , phys . b * 55 * , 93 ( 1997 ) . et al . _ , b * 59 * , 11495 ( 1999 ) . et al . _ , phys . rev . lett . * 65 * , 2712 ( 1990 ) . sharma , t. venkatesan , z.h . zhang , l.r . liu , r. chu and w.k . chu , phys . * 77 * , 4624 ( 1996 ) . geschkenbein , l.b . ioffe and a.i . larkin , phys . b * 55 * , 3173 ( 1997 ) .

for many years , the dominant method for discovering massive clusters of galaxies at @xmath11 , has been via x - ray emission from the hot gas in their dark matter potential wells . over the years , the _ rosat _ and _ xmm - newton _ telescopes , in particular , have yielded a handful of well - studied examples e.g. , rdcs j0910 + 5422 @xcite , rdcs j1252.9 - 2927 @xcite , rx j0848 + 4452 ( lynx e ; @xcite ) , xmmu j2235 - 25 @xcite , and xmmxcs j2215.9 - 1738 @xcite . in recent years , it has been realised that , by incorporating deep infrared ( ir ) observations , existing optical imaging techniques can be adapted to successfully detect clusters at redshifts competitive to the existing x - ray surveys . however , because ir observations covering only modest areas ( 120 arcmin@xmath1 - 7.25 deg@xmath1 ) have been available , the ( cluster or candidate ) systems detected to date have been less massive than those discovered from the x - ray surveys ( @xcite ; although see @xcite ) . massive clusters of galaxies are rare , and one requires as widefield a survey as possible to detect them . the largest area _ space telescope survey is the 50 square degree seven passband ( 3.6 , 4.5 , 5.8 , 8.0 , 24 , 70 , 160 @xmath12 ) spitzer wide - area infrared extragalactic survey ( swire ) legacy survey @xcite . we have obtained deep @xmath13band imaging @xcite , and combined this with the pre - existing infrared array camera ( irac ; @xcite ) observations from the swire survey , aiming to select clusters at @xmath14 using a two filter @xmath15 $ ] infrared adaptation of the well - proven optical cluster red sequence ( rs ) method @xcite . the sparcsgillianw / sparcs ] survey @xcite , is now complete and has an effective area ( defined as the usable overlap with swire after excluding chip gaps , regions near bright stars , etc ) , of 41.9 deg@xmath16 . the sparcs catalog contains several hundred cluster candidates at @xmath14 , by far the largest , homogeneously selected sample of its kind . in @xcite and @xcite we presented an overview of the sparcs survey , and reported spectroscopic confirmation of three clusters at @xmath17 and @xmath18 . in this paper , we report spectroscopic confirmation of three additional clusters at @xmath19 and @xmath20 . the paper is organized as follows : in [ observations ] , we describe the imaging and spectroscopic observations , and the data reduction ; in [ analysis ] , we describe the spectroscopic catalog ; in [ results ] , we present our results for the three individual clusters , and estimate the cluster velocity dispersions and dynamical masses ; and in [ discussion ] , we present a discussion and the main conclusions based on our results . we assume a @xmath21-cdm cosmology with @xmath22 and @xmath23 , and h@xmath24 km s@xmath5 mpc@xmath5 . the sparcs / swire survey is comprised of six fields ( see table 1 of * ? ? ? * ) . the three clusters described in this paper were identified as high significance cluster candidates in the 7.9 deg@xmath1 elais - n1 field . a full description of the sparcs data reduction , cluster candidate detection algorithm and catalogs will appear in muzzin et al . 2010 , in preparation . we provide only a brief overview of the main details here . sparcs @xmath0 observations of the northern fields were observed by cfht / megacam for a total integration time of 6000s per pointing . photometry was performed on both the @xmath25 and irac mosaics using the sextractor photometry package @xcite . the total 3.6 @xmath12 magnitude of all sources in the field was computed using a large aperture , equal to the geometric mean radius of the sextractor isophotal aperture . further details of the photometric pipeline may be found in @xcite and @xcite . the depth of the @xmath0 data varies from pointing to pointing depending on the seeing and the sky background , however , the mean 5 @xmath26 depth for extended sources in the elais - n1 field was @xmath27 vega ( 24.2 ab ) . the @xmath28 color of all sources was computed using an aperture of diameter three irac pixels ( @xmath29 ) . no aperture corrections were applied to the @xmath25 photometry because the @xmath30 color aperture was much larger than the median seeing . however , the irac point spread function ( psf ) has broad wings compared to a typical ground - based seeing profile , and it was necessary to apply aperture corrections to the measured 3.6 @xmath12 magnitude before computing the color of each galaxy . galaxy clusters were identified in the sparcs survey , using an algorithm very similar to that described in detail in @xcite , as applied to the spitzer first look survey ( fls ; @xcite ) . for cluster detection in both the fls and sparcs surveys , we used an infrared adaptation of the cluster red - sequence ( crs ) technique @xcite . this algorithm maps the density of galaxies in a survey within narrow color slices , giving greater weight to brighter galaxies , and flagging the highest overdensities as candidate clusters . the one important difference here , compared to @xcite , is that the latter paper used an @xmath31 $ ] color to detect clusters at @xmath32 in the fls . the deeper irac data in swire ( @xmath33s frames ) , compared to the fls ( @xmath34s frames ) , combined with the @xmath15 $ ] color choice , allow sparcs to detect clusters to higher redshift than was possible with the fls dataset . from analysis of the richnesses and colors , sparcs j161641 + 554513 and sparcs j161037 + 552417 were identified as rich cluster candidates at @xmath11 , and sparcs j161315 + 564930 as an unusually rich cluster candidate at @xmath35 ( table [ tab_clusters ] ) . spectroscopic observations of sparcs j161315 + 564930 , sparcs j161641 + 554513 and sparcs j161037 + 552417 were obtained with the low resolution imaging spectrograph ( lris ; @xcite ) on the keck i telescope on the nights of april 3th and april 4th , 2008 . the seeing on both nights was about @xmath36 . lris has beam - splitters that separate the light between two sides : red @xcite and blue @xcite . on the red side , we used the 400 line mm@xmath5 grating which gives a coverage of @xmath37 centered at 8500 , with a dispersion of 1.86 / pixels . all the masks were designed with slits of 11 width , which gave a resolution of @xmath387 - 9 ( fwhm ) over the wavelength interval @xmath386600 - 9600 . for the redshift range of these particular clusters , all of the prominent spectral features e.g. , the 4000 - rest - frame break or [ o@xmath39 emission line , fall in the lris red side wavelength window . we observed one mask each in the case of clusters sparcs j161315 + 564930 and sparcs j161641 + 554513 , and two in the case of sparcs j161037 + 552417 . the total exposure time was 6000s ( @xmath40 ) for sparcs j161315 + 564930 , 8400s ( @xmath41 ) for sparcs j161641 + 554513 , and 8400s ( @xmath42 ) and 7000s ( @xmath43 ) for mask @xmath44 and mask @xmath45 of sparcs j161037 + 552417 ( table 1 ) . the masks contained about 30 slits each ( five of these were alignment stars and the remaining 25 were galaxies , selected according to a sliding scale of priority ) . to reduce the number of slits placed on obvious foreground galaxies , we prioritized slits on galaxies with colors near each cluster s red - sequence . to avoid significant selection bias , we used a very broad cut around the red - sequence , intended to include both star forming and non - star forming systems . slits were placed on galaxies with priorities from 1 to 4 . priority 1 was galaxies with colors within 0.2 magnitude of the rs , and with @xmath46 < 17.0 $ ] ( vega ) . priority 2 was galaxies with colors between 0.2 and 0.6 magnitudes bluer than the rs , and with @xmath47 < 18.7 $ ] . priority 3 was galaxies with same colors as priority 1 , but with @xmath47 < 18.7 $ ] , and priority 4 was galaxies with same colors as priority 2 , but with @xmath47 < 18.7 $ ] . priorities 1 to 4 roughly correspond to bright red - sequence , blue cloud , faint red - sequence , and faint blue cloud galaxies respectively . approximately ten priority 1 galaxies could be accomodated per mask for sparcs j161315 + 564930 , and five per mask for sparcs j161641 + 554513 and sparcs j161037 + 552417 . to reduce the lris data , we adapted custom software , based on that developed by @xcite to reduce vlt / fors2 data . overscan and bias corrections were applied to both the calibration ( bias , flats and lamp arcs ) and to the science ( clusters and standard stars ) frames using the iraf _ lrisbias _ and _ lccdproc _ tasks developed by g. d. wirth and c. fassnacht . after these corrections had been applied , the data were next corrected for geometric distortions across the dispersion axis , and then separated into individual slitlets and reduced using standard long - slit techniques . each individual slitlet was processed using an algorithm similar to that implemented in _ bogus _ ( developed by d. stern , a. j. bunker and s. a. stanford)stern / homepage / bogus.html ] . the flatfield slitlets were normalized and applied to the corresponding science slitlets . the individual 2d spectra were then background - subtracted and fringe - corrected before being stacked to produce a final co - added 2d spectrum . slits longer than 90 pixels ( @xmath48 ) were further corrected for long - slit " distortions . the 1d spectrum of the source ( or sources ) in each slitlet was then extracted using standard iraf tasks . a set of near lamp exposures were used to wavelength - calibrate the observations . lris spectra can suffer significant flexure which may introduce wavelength offsets as large as @xmath49 in the wavelength solution , in the most extreme cases . the offsets in the wavelength calibration due to flexure were corrected by comparing to the wavelengths of known skylines . we estimate the residual uncertainties in wavelength calibration to be at most 1 , which correspond to a redshift uncertainty of @xmath50 at @xmath51 . finally , a relative flux calibration was performed , using a sensitivity function spanning the range 5600 - 9400 which was obtained from long - slit observations of the standard stars feige 34 and feige 67 @xcite . a redshift solution was sought for each object by cross - correlating @xcite each of the lris spectra with observed galaxy templates @xcite using the task _ xcsao _ in the iraf _ rvsao _ package @xcite . if a redshift could be obtained for a spectrum , it was assigned a quality flag , `` q '' , with one of two possible values : 0 or 1 . high - confidence redshifts , unambiguously determined by two or more clear features in the continuum or in absorption , or by obvious emission lines such as [ o@xmath39 ( @xmath52 ) were assigned @xmath53 . lower - confidence redshifts based on ambiguous identification of a few spectral features , often in low s / n spectra , were assigned @xmath54 . in addition to a quality flag , an emission line flag , `` e '' , was also assigned to each galaxy , based on the presence or absence of emission line features . spectra displaying one or more emission feature such as h@xmath55 , [ o@xmath56(@xmath574959,@xmath575007 ) or h@xmath58 in the case of low - redshift ( @xmath59 ) galaxies , or [ o@xmath39(@xmath573727 ) in the case of higher redshift galaxies , were assigned @xmath60 . spectra showing no sign of excess emission above the stellar continuum , were assigned @xmath61 . the total number of galaxies ( cluster members and foreground or background galaxies , excluding stars ) extracted from the fields of sparcs j161315 + 564930 , sparcs j161037 + 552417 and sparcs j161641 + 554513 were 22 , 19 and 34 ( see tables [ tab_en1_240 ] , [ tab_en1_349 ] and [ tab_en1_381 ] ) . occasionally , one or more serendipitous sources fell within a slit . the breakdown by quality flag was 20 , 13 and 14 @xmath53 galaxies , and 2 , 6 and 20 @xmath54 galaxies , respectively . example spectra for a subsample of cluster members ( see section [ cl240 ] for the definition of a `` cluster member '' ) are shown in figure [ cl_specs ] . the left column shows examples of sparcs j161315 + 564930 members , the center column shows examples of sparcs j161641 + 554513 members , and the right column shows exampels of sparcs j161037 + 552417 members . the i d of each object is indicated ( tables [ tab_en1_240 ] , [ tab_en1_349 ] and [ tab_en1_381 ] ) . fluxes are in relative units , and smoothed by 7 pixels ( 1 pix @xmath62 ) . prominent spectral features are indicated by vertical lines . [ results ] in determining cluster membership , and calculating a velocity dispersion and mass , only high confidence , @xmath53 , galaxies were considered . definitive cluster membership was determined using the code of @xcite , which is based on the shifting - gap technique of @xcite . this procedure uses both galaxy angular position and radial velocity information to exclude near - field interlopers . the squares in the left panel of figure [ z_histo_240 ] show the 14 @xmath53 galaxies identified as members of cluster sparcs j161315 + 564930 ( see table [ tab_en1_240 ] ) by the shifting - gap technique . for sparcs j161315 + 564930 , these @xmath53 cluster members fall in the range @xmath63 , indicated by the vertical dashed lines shown in the right panel of figure [ z_histo_240 ] , which shows the redshift histogram for the sparcs j161315 + 564930 field . the properties of the cluster members and non - members are summarized in table [ tab_en1_240 ] . the total 3.6 @xmath12 magnitude ( column 4 in table [ tab_en1_240 ] ) , and @xmath28 color ( column 5 in table [ tab_en1_240 ] ) were calculated as described in section [ photometry ] . in determining membership , we did not require that any @xmath54 galaxies satisfy the shifting - gap criteria ; if their redshifts fell in the range @xmath63 we included them as cluster members " in table [ tab_en1_240 ] , but re - emphasize that we do not utilize them in estimating the cluster mean redshift or velocity dispersion . the total number of cluster members is 16 ( 14 with @xmath53 and 2 with @xmath54 ) , and six foreground / background galaxies . all confirmed members of cluster sparcs j161315 + 564930 were passive ( @xmath61 ) galaxies . in some cases , a redshift was obtained which did not correspond to any galaxy in our photometric catalog . in the case of faint galaxies , this was because a spectroscopic redshift was obtained for a strong emission line galaxy whose continuum fell below the detection threshold of the @xmath25 catalog . in the case of bright galaxies , this was because of blending issues with the irac @xmath64 @xmath65 psf . these galaxies were assigned both a @xmath46$]-band magnitude and a @xmath66 $ ] color of 99 in table [ tab_en1_240 ] . figure [ proj_dist_240 ] shows @xmath67 $ ] color composites of cluster sparcs j161315 + 564930 . the @xmath68 data were obtained from wfc on the isaac newton telescope , and are available with the swire public data release @xcite . the white squares ( green circles ) overlaid on the right panel show the 16 cluster members ( and foreground / background galaxies in the @xmath69 fov ) with spectroscopically - confirmed redshifts from keck / lris ( see table [ tab_en1_240 ] ) . once cluster membership was established , both the redshift and rest - frame velocity dispersion , @xmath70 , of sparcs j003550 - 431224 were then calculated iteratively using the `` robust estimator '' , @xmath71 @xcite . the robust estimator has been shown to be less sensitive than the standard deviation to outliers which may persist even after rejecting interlopers using the shifting - gap technique . the actual estimator used depends on the number of cluster members and is either the biweight estimator for datasets with at least 15 members , or , as here in the case of sparcs j161315 + 564930 with 14 @xmath53 members , the gapper estimator . the gapper estimator is discussed more fully in @xcite , @xcite and @xcite . the line - of - sight rest - frame velocity dispersion , @xmath72 , was calculated directly from the vector of spectroscopic redshift measurements , , as @xmath73 where @xmath74 is the speed of light , and @xmath75 is the estimated dispersion of the measured redshifts with respect to the center of the distribution , @xmath76 . a mean redshift of @xmath77 and a velocity dispersion of @xmath78 km s@xmath5 were calculated for sparcs j161315 + 564930 ( table [ tab_clusters ] ) . the uncertainty on the latter was determined using jackknife resampling of the data . for comparison , a gaussian with an rms of 1230 km s@xmath5 has been overlaid on the redshift histogram in the right panel of figure [ z_histo_240 ] . the line - of - sight rest - frame velocity dispersion can be used to calculate a dynamical estimate of @xmath79 , the radius at which the mean interior density is 200 times the critical density , @xmath80 , and @xmath81 , the mass contained within @xmath79 . in the spherical collapse model , at redshift @xmath82 , @xmath83 can be calculated from @xmath84 where @xmath85 , is the hubble parameter at redshift @xmath82 , and @xmath86 based on its velocity dispersion of @xmath78 km s@xmath5 , we estimate an @xmath87 mpc and a dynamical mass of @xmath88 m@xmath89 for sparcs j161315 + 564930 ( table [ tab_clusters ] ) . although this mass is preliminary ( see section [ richness ] ) , it seems likely that sparcs j161315 + 564930 is an unusually massive cluster , perhaps the most massive cluster in the entire sparcs survey , and comparable in mass to that of cluster ms 1054 - 03 at @xmath90 @xcite . the @xmath66 $ ] vs. @xmath0 color - magnitude diagram for all galaxies ( gray circles ) within a radius of @xmath83 ( @xmath91 mpc ) of the center of sparcs j161315 + 564930 is shown in figure [ colmag_240 ] . this radius is approximately equal to the virial radius . spectroscopically - confirmed galaxies are shown by colored symbols . in this figure ( and in figures [ colmag_349 ] and [ colmag_381 ] ) , cluster members ( @xmath53 or @xmath54 ) are shown by red circles and foreground / background galaxies by blue squares . note that there are several cluster members or foreground / background galaxies shown in table [ tab_en1_240 ] , which fall within a projected radius of @xmath83 of the cluster center , but for which a color could not be determined . these galaxies do not appear in figure [ colmag_240 ] ( or in figures [ colmag_349 ] or [ colmag_381 ] , in the cases of tables [ tab_en1_349 ] and tables [ tab_en1_381 ] ) . seven @xmath53 galaxies were determined to be cluster members of sparcs j161641 + 554513 by the shifting - gap technique . these galaxies are shown by squares in the left panel of figure [ z_histo_349 ] . these galaxies lie in the redshift range @xmath92 , indicated by the vertical dashed lines in the right panel of figure [ z_histo_349 ] . table [ tab_en1_349 ] summarizes the cluster members " , the ten @xmath53 and @xmath54 galaxies with redshifts in this range , and the nine foreground / background galaxies . the right panel of figure [ z_histo_349 ] shows the redshift histogram for the field of sparcs j161641 + 554513 . a gaussian with an rms of 950 km s@xmath5 has been overlaid . figure [ proj_dist_349 ] shows @xmath67 $ ] color composites of cluster sparcs j161641 + 554513 . the white squares ( green circles ) overlaid on the right panel show the cluster members ( foreground / background galaxies ) with spectroscopically - confirmed redshifts ( see table [ tab_en1_349 ] ) which fall within the @xmath93 fov of the image . of the seven @xmath53 cluster members , five were classified as passive ( @xmath61 ) , one was classified as emission line ( @xmath60 ) , and one was classifed as an agn ( @xmath94 ) . the upper left white square in figure [ proj_dist_349 ] corresponds to the agn . there were also three passive @xmath54 cluster members ( table [ tab_en1_349 ] ) . the spectrum of the confirmed agn ( object i d @xmath95 854892 in table [ tab_en1_349 ] ) is shown in the lowest panel of the center column in figure [ cl_specs ] . this source shows nev(@xmath573346,@xmath573426 ) and neiii(@xmath573869,@xmath573968 ) in emission , which are common in agn , as well as prominent [ oii](@xmath573727 ) emission and high - order balmer lines ( h@xmath96 , h@xmath97 and h6 ) also in emission . the cluster mean redshift and velocity dispersion , were estimated iteratively from the seven @xmath53 members , using the gapper estimator . the mean redshift of sparcs j161641 + 554513 was calculated to be @xmath98 . the velocity dispersion was calculated to be @xmath99 km s@xmath5 , which corresponds to @xmath100 mpc , and a dynamical mass of @xmath101 m@xmath89(table [ tab_clusters ] ) . the @xmath66 $ ] vs. @xmath0 color - magnitude diagram for all galaxies ( gray circles ) within a radius of @xmath83 ( @xmath102 mpc ) of the center of sparcs j161641 + 554513 is shown in figure [ colmag_349 ] . the red circles and blue squares indicate those cluster members and foreground / background galaxies which lie within a projected distance of @xmath83 of the cluster center , and for which a color could be determined . seven @xmath53 galaxies were also determined to be cluster members of sparcs j161037 + 552417 by the shifting - gap technique . these galaxies are indicated by squares in the left panel of figure [ z_histo_381 ] . two galaxies , indicated by crosses in figure [ z_histo_381 ] ( i d @xmath95 s 727869 and 734082 in table [ tab_en1_381 ] ) , were identified as near - field interlopers . the seven @xmath53 members lie in the redshift range @xmath103 , shown by the vertical dashed lines in the right panel of figure [ z_histo_381 ] . table [ tab_en1_381 ] summarizes the cluster members " , the ten @xmath53 or @xmath54 galaxies with redshifts in this range , and the 24 foreground / background galaxies . the right panel of figure [ z_histo_381 ] shows the redshift histogram for the field of sparcs j161037 + 552417 . a gaussian with an rms of 410 km s@xmath5 has been overlaid . figure [ proj_dist_381 ] shows @xmath67 $ ] color composites of cluster sparcs j161037 + 552417 . the white squares ( green circles ) overlaid on the right panel show the cluster members ( foreground / background galaxies ) with spectroscopically - confirmed redshifts ( see table [ tab_en1_381 ] ) which fall within the @xmath104 fov of the image . of the seven @xmath53 cluster members , all were classified as emission line ( @xmath60 ) . of the three @xmath54 cluster members , one was classified as passive ( @xmath61 ) , and two as emission line ( @xmath60 ) . the cluster mean redshift and velocity dispersion , were calculated iteratively from the seven @xmath53 members , using the gapper estimator . the mean redshift of sparcs j161037 + 552417 was estimated to be @xmath105 and the velocity dispersion , @xmath106 km s@xmath5 , which corresponds to @xmath107 mpc and a dynamical mass of @xmath108 m@xmath89 ( table [ tab_clusters ] ) . sparcs j161037 + 552417 is significantly less massive than sparcs j161315 + 564930 or sparcs j161641 + 554513 . the much smaller @xmath83 estimated for sparcs j161037 + 552417 , combined with the geometric limitations of lris with respect to the redshift number density yield measurable from a single mask , resulted in a yield of only four spectroscopically confirmed members within @xmath83 from two masks . moreover , because of blending issues with irac s @xmath64 @xmath65 psf , a reliable @xmath66 $ ] color could not be determined for one of these galaxies ( i d @xmath95 4000013 in table [ tab_en1_381 ] ) . figure [ colmag_381 ] shows the @xmath66 $ ] vs. @xmath0 color - magnitude diagram for all galaxies ( gray circles ) within a radius of @xmath109 ( @xmath110 kpc ) of the center of cluster sparcs j161037 + 552417 . the three red circles denote the cluster members with i d @xmath95 s 723814 , 722784 , and 722712 in table [ tab_en1_381 ] . color - magnitude diagrams for sparcs j161315 + 564930 , sparcs j161037 + 552417 and sparcs j161037 + 552417 were presented in figures [ colmag_240 ] , [ colmag_349 ] , and [ colmag_381 ] . column 2 of table [ tab_properties ] shows the @xmath15 $ ] color of the red sequence for sparcs j161315 + 564930 , sparcs j161037 + 552417 and sparcs j161037 + 552417 . the color is calculated from the mean color of the spectroscopically confirmed red - sequence cluster members . also shown in column 2 of table [ tab_properties ] is the color of the rs for three additional clusters , previously reported in @xcite and @xcite . these clusters are sparcs j163435 + 402151 at @xmath111 , sparcs j163852 + 403843 at @xmath112 , and sparcs j003550 - 431224 at @xmath113 ( column 6 ) . for all six clusters a systematic uncertainty in the rs color of 0.15 magnitude has been assumed ( this uncertainty reflects the fact that , at present , we are using the zeropoints provided by elixir for the @xmath0 observations . we expect , in the future , to be able to reduce these photometric uncertainties , using our own internal calibration ) . columns 3 , 4 and 5 of table [ tab_properties ] show the redshift that would be estimated for each cluster based on the measured rs color ( column 2 ) , assuming a solar metallicity single burst bc03 model and a formation redshift of either @xmath114 , 4 or 10 . as can be seen from table [ tab_properties ] ( and the left panel of figure [ zsandmasses ] ) , the photometric redshift inferred from the measured @xmath66 $ ] color has a slight dependence on one s choice of formation redshift , although differences in color between the models at @xmath115 are fairly small ( @xmath116 between @xmath117 and @xmath118 ) . utilizing the colors from the @xmath119 model , the three new clusters presented here were assigned preliminarily redshift estimates of @xmath120 , @xmath121 , and @xmath122 ( table [ tab_properties ] ) . these photometrically estimated redshifts are very similar to , albeit slightly lower than , the spectroscopically determined values of @xmath123 , @xmath124 , and @xmath125 ( column 6 in table [ tab_properties ] ) . the @xmath126 $ ] color vs. spectroscopic redshift for all six sparcs clusters in table [ tab_properties ] is plotted in the left panel of figure [ zsandmasses ] . the solid , dotted and and dashed lines show the bc03 model colors as a function of redshift for formation redshifts of @xmath114 , 4 and 10 . it is clear from figure [ zsandmasses ] that the agreement between the model colors and the observations is very good . with a larger sample of clusters , there may turn out to be small but real discrepancies between the models and the measured rs colors . in the left panel of figure [ zsandmasses ] , the @xmath119 model can be seen to be slightly redder than the observations in the case of five clusters ( or equivalantly , the inferred photometric redshift can be seen to be slightly lower than the spectroscopic redshift ) , but slightly bluer in the case of one cluster ( sparcs j003550 - 431224 ) . these small offsets in color between the models and the observations , can also be seen directly from figures [ colmag_240 ] , [ colmag_349 ] , and [ colmag_381 ] . the solid lines in these three figures show the rs color predicted by the bc03 @xmath119 model _ at the spectroscopic redshift of the cluster _ , and can be seen to be slightly redder than the observed color . a more detailed comparison between the model predictions and the observations will be made in a future paper employing a larger sample of sparcs clusters . despite the aforementioned caveats , and the issue of degeneracies between the photometric redshift of the clusters and the formation redshifts of their galaxies , our overall conclusion is that the inferred one - color photometric redshifts and the spectroscopic redshifts are in excellent ( @xmath127 ) general agreement . in addition to estimating the masses for sparcs j161315 + 564930 , sparcs j161641 + 554513 , and sparcs j161037 + 552417 from the galaxy line - of - sight velocity dispersion , we also estimated the masses from the richness of the clusters , using the @xmath128 richness parameter . @xcite introduced @xmath128 , an adaptation of the @xmath129 richness parameter , intended to utilize two - band photometry to increase the contrast of the cluster with the background , and therefore provide a measurement of the richness that is less sensitive to foreground / background large scale structures . @xmath129 is the amplitude of the three - dimensional , cluster center - galaxy spatial correlation function , @xmath130 @xcite . instead of counting galaxies in a single passband , @xmath128 is obtained by counting galaxies in a color slice centered on the location of each cluster s red - sequence in the @xmath66 $ ] color - magnitude diagram . in computing @xmath128 , we used a slice bounded in color by @xmath66 = \pm0.3 $ ] of the best - fit rs color returned by the cluster finding algorithm ( @xcite ) , and bounded in magnitude by @xmath131 , where @xmath132 is the bc03 @xmath119 model prediction of the characteristic magnitude of a galaxy _ at the photometric redshift _ corresponding to that rs color ( table [ tab_properties ] ) . the background galaxy counts were determined from the color distribution in the entire 7.9 deg@xmath1 elais - n1 field , minus the regions known to contain galaxy clusters . the @xmath128 richnesses of the three clusters were computed to be @xmath133 mpc@xmath134 ( sparcs j161315 + 564930 ) , @xmath135 mpc@xmath134 ( sparcs j161641 + 554513 ) , and @xmath136 mpc@xmath134 ( sparcs j161037 + 552417 ) . based on the empirical calibration of @xmath137 vs. @xmath138 determined by @xcite in the k - band for 15 cnoc1 clusters at @xmath139 , these richnesses correspond to @xmath140 for sparcs j161315 + 564930 , @xmath141 for sparcs j161641 + 554513 , and @xmath142 for sparcs j161037 + 552417 . for comparison , columns 7 and 8 of table [ tab_properties ] show the dynamical mass , @xmath143 , and richness mass , @xmath144 , estimates for all six sparcs clusters . although the uncertainties associated with both of the mass estimators are large , they are consistent with each other at the 1@xmath145 level for five out of the six clusters and at the 2@xmath145 level for the sixth . the agreement between the two mass estimators can be seen in the right panel of figure [ zsandmasses ] . based on all six sparcs clusters spectroscopically confirmed to date , our conclusion is that , in addition to there being excellent agreement between the photometric and spectroscopic redshifts , there is also reasonable agreement between the cluster dynamical and richness mass estimates . the dynamical masses estimates should be considered preliminary at this stage . uncertainties will reduce as more data becomes available from a large spectroscopic follow - up program of sparcs clusters currently being carried out at the gemini telescopes . collectively , the six sparcs clusters confirmed to date ( the three clusters presented in @xcite and @xcite , plus the three clusters presented here ) , demonstrate that , given the availability of infrared observations , the rs technique is an efficient and effective method of detecting _ bona fide _ massive galaxy clusters at @xmath146 . moreover , our studies of these six clusters are showing that it is possible to infer fundamental parameters such as cluster redshift and mass _ from the survey data itself _ ( see also @xcite ) . at @xmath147 , both the optical red - sequence cluster surveys , rcs-1 @xcite and rcs-2 @xcite , and the sloan digital sky survey ( sdss ; @xcite ) have shown that it is feasible to measure cosmological parameters from the evolution of the cluster mass function @xcite . in order to do this efficiently , the survey data themselves are used to detect clusters , and also to estimate the redshift and the mass of those clusters . the redshift is estimated from the red sequence color @xcite , and the mass is estimated from the optical richness @xcite . the fact that sparcs is also now demonstrating the practicality of estimating redshifts and masses at @xmath146 from the survey data alone is heartening for the current generation of surveys aiming to utilize optical - infrared high redshift cluster observations to constrain cosmological parameters e.g. , sparcs , the ukirt infrared deep - sky survey deep extragalactic survey ( ukidss dxs ; @xcite),and the irac shallow cluster survey ( iscs ; @xcite ) . these optical - ir surveys will provide complementary samples to those selected using the sunyaev - zeldovich effect , e.g. , the south pole telescope survey ( spt ; @xcite ) , the atacama cosmology telescope ( act ; @xcite ) , and the atacama pathfinder experiment ( apex ; @xcite ) . the complete sparcs catalog contains several hundred cluster candidates at @xmath14 . with new large , homogeneous , reliable @xmath11 catalogs becoming available from sparcs and other surveys in the very near future , the prospects look bright for high redshift cluster and cluster galaxy evolution studies in the coming years . + we thank the referee for useful comments . the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community . we are most fortunate to have the opportunity to conduct observations from this mountain . this work is based in part on archival data obtained with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa . support for this work was provided by an award issued by jpl / caltech . gw also gratefully acknowledges support from nsf grant ast-0909198 , and from the college of natural and agricultural sciences at ucr . andreon , s. , maughan , b. , trinchieri , g. , & kurk , j. 2008 , , 507 , 147 becker , m. r. , et al . 2007 , , 669 , 905 beers , t. c. , flynn , k. , & gebhardt , k. 1990 , , 100 , 32 bertin , e. , & arnouts , s. 1996 , , 117 , 393 blindert , k. 2006 , ph.d . thesis , university of toronto , canada bremer , m. n. , et al . 2006 , , 371 , 1427 brodwin , m. , et al . 2006 , , 651 , 791 bruzual , g. , & charlot , s. 2003 , , 344 , 1000 carlstrom , j. e. , et al . 2009 , arxiv:0907.4445 demarco , r. , et al . 2005 , , 432 , 381 demarco , r. , et al . 2007 , , 663 , 164 dobbs , m. , et al . 2006 , new astronomy review , 50 , 960 eisenhardt , p. r. m. , et al . 2008 , , 684 , 905 fadda , d. , girardi , m. , giuricin , g. , mardirossian , f. , & mezzetti , m. 1996 , , 473 , 670 fazio , g. g. , et al . 2004 , , 154 , 10 gilbank , d. g. , yee , h. k. c. , ellingson , e. , gladders , m. d. , barrientos , l. f. , & blindert , k. 2007 , , 134 , 282 gioia , i. m. , braito , v. , branchesi , m. , della ceca , r. , maccacaro , t. , & tran , k .- v . 2004 , , 419 , 517 girardi , m. , biviano , a. , giuricin , g. , mardirossian , f. , & mezzetti , m. 1993 , , 404 , 38 girardi , m. , demarco , r. , rosati , p. , & borgani , s. 2005 , , 442 , 29 gladders , m. d. , yee , h. k. c. , majumdar , s. , barrientos , l. f. , hoekstra , h. , hall , p. b. , & infante , l. 2007 , , 655 , 128 gladders , m. d. , & yee , h. k. c. 2000 , , 120 , 2148 gladders , m. d. , & yee , h. k. c. 2005 , , 157 , 1 goto , t. , et al . 2008 , , 60 , 531 hoekstra , h. , franx , m. , & kuijken , k. 2000 , , 532 , 88 jee , m. j. , white , r. l. , ford , h. c. , blakeslee , j. p. , illingworth , g. d. , coe , d. a. , & tran , k .- v . h. 2005 , , 634 , 813 kinney , a. l. , calzetti , d. , bohlin , r. c. , mcquade , k. , storchi - bergmann , t. , & schmitt , h. r. 1996 , , 467 , 38 koester , b. p. , et al . 2007 , , 660 , 239 kosowsky , a. 2003 , new astronomy review , 47 , 939 koyama , y. , kodama , t. , tanaka , m. , shimasaku , k. , & okamura , s. 2007 , , 382 , 1719 krick , j. e. , surace , j. a. , thompson , d. , ashby , m. l. n. , hora , j. l. , gorjian , v. , & yan , l. 2008 , , 686 , 918 kurk , j. , et al . 2008 , astronomical society of the pacific conference series , 399 , 332 kurtz , m. j. , mink , d. j. , wyatt , w. f. , fabricant , d. g. , torres , g. , kriss , g. a. , & tonry , j. l. 1992 , astronomical data analysis software and systems i , 25 , 432 lacy , m. , et al . 2005 , , 161 , 41 lawrence , a. , et al . 2007 , , 379 , 1599 lonsdale , c. j. , et al . 2003 , , 115 , 897 mccarthy , j. k. , et al . 1998 , , 3355 , 81 mccarthy , p. j. , et al . 2007 , , 664 , l17 mullis , c. r. , rosati , p. , lamer , g. , bhringer , h. , schwope , a. , schuecker , p. , & fassbender , r. 2005 , , 623 , l85 muzzin , a. , yee , h. k. c. , hall , p. b. , & lin , h. 2007 , , 663 , 150 muzzin , a. , wilson , g. , lacy , m. , yee , h. k. c. , & stanford , s. a. 2008 , , 686 , 966 muzzin , a. , et al . 2009 , , 698 , 1943 oke , j. b. 1990 , , 99 , 1621 oke , j. b. , et al . 1995 , , 107 , 375 rosati , p. , stanford , s. a. , eisenhardt , p. r. , elston , r. , spinrad , h. , stern , d. , & dey , a. 1999 , , 118 , 76 rosati , p. , et al . 2004 , , 127 , 230 rozo , e. , et al . 2009 , , 703 , 601 rozo , e. , et al . 2009 , arxiv:0902.3702 ruhl , j. , et al . 2004 , , 5498 , 11 shapley , a. e. , steidel , c. c. , pettini , m. , & adelberger , k. l. 2003 , , 588 , 65 shupe , d. l. , et al . 2008 , , 135 , 1050 stanford , s. a. , elston , r. , eisenhardt , p. r. , spinrad , h. , stern , d. , & dey , a. 1997 , , 114 , 2232 stanford , s. a. , holden , b. , rosati , p. , eisenhardt , p. r. , stern , d. , squires , g. , & spinrad , h. 2002 , , 123 , 619 stanford , s. a. , et al . 2005 , , 634 , l129 stanford , s. a. , et al . 2006 , , 646 , l13 , j. , shupe , d. l. , fang , f. , lonsdale , c. j. , & gonzalez - solares , e. 2005 , ssc website release tran , k .- v . h. , franx , m. , illingworth , g. d. , van dokkum , p. , kelson , d. d. , blakeslee , j. p. , & postman , m. 2007 , , 661 , 750 tonry , j. , & davis , m. 1979 , , 84 , 1511 van breukelen , c. , et al . 2007 , , 382 , 971 wilson , g. , et al . 2009 , , 698 , 1934 yee , h. k. c. , & ellingson , e. 2003 , , 585 , 215 yee , h. k. c. , & lpez - cruz , o. 1999 , , 117 , 1985 yee , h. k. c. , gladders , m. d. , gilbank , d. g. , majumdar , s. , hoekstra , h. , ellingson , e. , & the rcs-2 collaboration 2007 , arxiv : astro - ph/0701839 york , d. g. , et al . 2000 , , 120 , 1579 zatloukal , m. , rser , h .- j . , wolf , c. , hippelein , h. , & falter , s. 2007 , , 474 , l5 sparcs j161315 + 564930 & 16:13:14.6 & + 56:49:29.9 & 1 & 16 & 14 & @xmath148 & @xmath149 & @xmath150 & @xmath151 + sparcs j161641 + 554513 & 16:16:41.3 & + 55:45:12.5 & 1 & 10 & 7 & @xmath152 & @xmath153 & @xmath154 & @xmath155 + sparcs j161037 + 552417 & 16:10:36.5 & + 55:24:16.6 & 2 & 10 & 7 & @xmath156 & @xmath157 & @xmath158 & @xmath159 + [ tab_clusters ] 1070485 & 243.32851 & 56.821579 & @xmath160 & @xmath161 & @xmath162 & 0 & 0 + 1071085 & 243.31200 & 56.828140 & @xmath163 & @xmath164 & @xmath165 & 0 & 0 + 1071707 & 243.34280 & 56.835220 & @xmath166 & @xmath167 & @xmath168 & 0 & 0 + 1072162 & 243.36000 & 56.840450 & @xmath169 & @xmath170 & @xmath171 & 0 & 0 + 1068630 & 243.33870 & 56.802010 & @xmath172 & @xmath173 & @xmath174 & 0 & 0 + 1072641 & 243.30150 & 56.845970 & @xmath175 & @xmath176 & @xmath177 & 0 & 0 + 2000012 & 243.31107 & 56.826047 & @xmath178 & @xmath178 & @xmath179 & 0 & 0 + 1070775 & 243.31090 & 56.824970 & @xmath180 & @xmath181 & @xmath182 & 0 & 0 + 1066451 & 243.39220 & 56.778770 & @xmath183 & @xmath184 & @xmath185 & 0 & 1 + 2000007 & 243.34483 & 56.772531 & @xmath178 & @xmath178 & @xmath186 & 0 & 0 + 1066059 & 243.35040 & 56.775040 & @xmath187 & @xmath188 & @xmath189 & 0 & 0 + 1069805 & 243.34230 & 56.814430 & @xmath190 & @xmath191 & @xmath192 & 0 & 0 + 1074041 & 243.28720 & 56.862221 & @xmath193 & @xmath194 & @xmath195 & 0 & 1 + 1072364 & 243.29550 & 56.842720 & @xmath196 & @xmath197 & @xmath198 & 0 & 0 + 1069395 & 243.32159 & 56.810188 & @xmath199 & @xmath200 & @xmath201 & 0 & 0 + 1073458 & 243.30260 & 56.855709 & @xmath202 & @xmath173 & @xmath203 & 0 & 0 + 1068289 & 243.33870 & 56.797932 & @xmath204 & @xmath205 & @xmath206 & 1 & 0 + 1071481 & 243.32381 & 56.832401 & @xmath207 & @xmath167 & @xmath208 & 0 & 0 + 1070036 & 243.27380 & 56.817139 & @xmath209 & @xmath210 & @xmath211 & 0 & 0 + 1067689 & 243.27271 & 56.791660 & @xmath212 & @xmath213 & @xmath214 & 1 & 0 + 1065709 & 243.34480 & 56.771900 & @xmath215 & @xmath216 & @xmath217 & 1 & 0 + 2000006 & 243.34517 & 56.771469 & @xmath178 & @xmath178 & @xmath218 & 1 & 0 + [ tab_en1_240 ] 848097 & 244.17630 & 55.748390 & @xmath219 & @xmath220 & @xmath221 & 0 & 1 + 853440 & 244.18610 & 55.764381 & @xmath166 & @xmath222 & @xmath223 & 0 & 0 + 849838 & 244.17340 & 55.753590 & @xmath224 & @xmath225 & @xmath226 & 0 & 0 + 849110 & 244.19260 & 55.751389 & @xmath227 & @xmath228 & @xmath229 & 0 & 1 + 851083 & 244.17180 & 55.757141 & @xmath230 & @xmath231 & @xmath232 & 0 & 0 + 837239 & 244.26920 & 55.715519 & @xmath233 & @xmath234 & @xmath235 & 1 & 0 + 846319 & 244.25529 & 55.743149 & @xmath236 & @xmath237 & @xmath238 & 0 & 1 + 3000009 & 244.26935 & 55.715914 & @xmath178 & @xmath178 & @xmath239 & 0 & 0 + 839226 & 244.27161 & 55.721340 & @xmath240 & @xmath241 & @xmath242 & 0 & 0 + 854892 & 244.21670 & 55.768921 & @xmath243 & @xmath244 & @xmath245 & 2 & 0 + 833647 & 244.16330 & 55.704411 & @xmath246 & @xmath247 & @xmath248 & 1 & 0 + 828383 & 244.22971 & 55.688580 & @xmath249 & @xmath250 & @xmath251 & 1 & 0 + 840274 & 244.26221 & 55.724548 & @xmath252 & @xmath253 & @xmath254 & 1 & 0 + 856147 & 244.18530 & 55.772751 & @xmath255 & @xmath256 & @xmath257 & 1 & 0 + 844301 & 244.21809 & 55.736950 & @xmath258 & @xmath259 & @xmath260 & 1 & 0 + 842789 & 244.15030 & 55.732430 & @xmath261 & @xmath262 & @xmath263 & 0 & 1 + 857305 & 244.18820 & 55.776299 & @xmath264 & @xmath265 & @xmath266 & 0 & 0 + 851794 & 244.17329 & 55.759270 & @xmath267 & @xmath234 & @xmath268 & 0 & 1 + 860365 & 244.18150 & 55.786140 & @xmath209 & @xmath269 & @xmath270 & 0 & 1 + [ tab_en1_349 ] 4000013 & 242.65226 & 55.405353 & @xmath178 & @xmath178 & @xmath271 & 1 & 0 + 723814 & 242.65221 & 55.404598 & @xmath272 & @xmath273 & @xmath274 & 1 & 0 + 722784 & 242.63251 & 55.401890 & @xmath275 & @xmath276 & @xmath277 & 0 & 1 + 718577 & 242.61940 & 55.391060 & @xmath278 & @xmath279 & @xmath280 & 1 & 0 + 722712 & 242.66769 & 55.401711 & @xmath281 & @xmath282 & @xmath283 & 1 & 1 + 4000026 & 242.60676 & 55.475206 & @xmath178 & @xmath178 & @xmath284 & 1 & 1 + 710318 & 242.60490 & 55.370651 & @xmath285 & @xmath286 & @xmath287 & 1 & 0 + 4000010 & 242.67026 & 55.393053 & @xmath178 & @xmath178 & @xmath288 & 1 & 0 + 728525 & 242.69780 & 55.416550 & @xmath289 & @xmath290 & @xmath291 & 1 & 0 + 4000032 & 242.60455 & 55.371836 & @xmath178 & @xmath178 & @xmath292 & 1 & 0 + 744358 & 242.57840 & 55.457119 & @xmath293 & @xmath294 & @xmath295 & 1 & 0 + 4000012 & 242.65180 & 55.404111 & @xmath178 & @xmath178 & @xmath296 & 1 & 0 + 723736 & 242.64661 & 55.404339 & @xmath297 & @xmath298 & @xmath299 & 0 & 1 + 709346 & 242.61549 & 55.368011 & @xmath300 & @xmath301 & @xmath302 & 0 & 1 + 714998 & 242.59080 & 55.381821 & @xmath303 & @xmath304 & @xmath305 & 1 & 1 + 710721 & 242.60280 & 55.371490 & @xmath306 & @xmath307 & @xmath308 & 0 & 0 + 751852 & 242.60730 & 55.476471 & @xmath309 & @xmath310 & @xmath311 & 0 & 1 + 734249 & 242.66650 & 55.430779 & @xmath312 & @xmath313 & @xmath314 & 0 & 1 + 747661 & 242.62151 & 55.465469 & @xmath315 & @xmath316 & @xmath317 & 1 & 1 + 740510 & 242.69310 & 55.447041 & @xmath255 & @xmath318 & @xmath319 & 0 & 1 + 708779 & 242.61410 & 55.366550 & @xmath320 & @xmath321 & @xmath322 & 1 & 1 + 725337 & 242.69600 & 55.408329 & @xmath323 & @xmath324 & @xmath325 & 0 & 1 + 4000015 & 242.60952 & 55.453100 & @xmath178 & @xmath178 & @xmath326 & 0 & 1 + 730800 & 242.66510 & 55.422260 & @xmath327 & @xmath328 & @xmath329 & 0 & 0 + 720317 & 242.64140 & 55.395630 & @xmath330 & @xmath331 & @xmath332 & 0 & 1 + 745152 & 242.66029 & 55.459110 & @xmath333 & @xmath334 & @xmath335 & 0 & 1 + 753137 & 242.63040 & 55.479912 & @xmath336 & @xmath337 & @xmath338 & 0 & 1 + 738292 & 242.63560 & 55.441189 & @xmath339 & @xmath340 & @xmath341 & 0 & 0 + 727869 & 242.68291 & 55.414661 & @xmath333 & @xmath342 & @xmath343 & 1 & 0 + 734082 & 242.71609 & 55.430450 & @xmath344 & @xmath345 & @xmath343 & 1 & 0 + 712763 & 242.63080 & 55.376591 & @xmath346 & @xmath347 & @xmath348 & 1 & 1 + 742510 & 242.67841 & 55.452202 & @xmath349 & @xmath350 & @xmath351 & 1 & 1 + 737166 & 242.70931 & 55.438370 & @xmath352 & @xmath353 & @xmath354 & 1 & 1 + 735789 & 242.69299 & 55.434792 & @xmath355 & @xmath356 & @xmath357 & 1 & 1 + [ tab_en1_381 ] sparcs j161315 + 564930 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath148 & @xmath151 & @xmath362 + sparcs j161641 + 554513 & @xmath363 & @xmath364 & @xmath365 & @xmath366 & @xmath152 & @xmath155 & @xmath367 + sparcs j161037 + 552417 & @xmath368 & @xmath369 & @xmath370 & @xmath371 & @xmath156 & @xmath159 & @xmath372 + sparcs j163435 + 402151 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath377 & @xmath378 & @xmath379 + sparcs j163852 + 403843 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & @xmath384 & @xmath385 & @xmath386 + sparcs j003550 - 431224 & @xmath387 & @xmath388 & @xmath389 & @xmath390 & @xmath391 & @xmath392 & @xmath379 + [ tab_properties ]

cells exist in a crowded environment consisting of macromolecules ( lipids , mrna , ribosome , sugars , etc . ) , making it critical to investigate protein folding in the presence of crowding agents @xcite . if the interactions between the crowding agents and the protein of interest are short - ranged and non - specific ( as is often the case ) , then the volume excluded by the crowding agents prevents the polypeptide from sampling extended conformations . as a consequence , the entropy of the denatured state ensemble ( dse ) decreases relative to the case when the crowding agents are absent . these arguments suggest that excluded volume of crowding agents should enhance the stability of the folded state provided that the crowding - induced changes in the native state are negligible @xcite . the entropic stabilization mechanism , described above , has been used in several theoretical models to quantitatively describe the extent of folded protein as a function of the volume fraction , @xmath13 , of the crowding agents @xcite . more recently , a theory whose origins can be traced to the concept of intra - protein attraction due to depletion of crowding agents near the protein @xcite , predicts that the enhancement in stability , @xmath14 , where @xmath15 is the folding temperature at @xmath13 and @xmath16 is related to the flory exponent that characterizes the size of the protein in the dse @xcite . from this prediction it follows that crowding affects the dse to a greater extent than the folded state . although the precise theoretical predictions of the power law change in @xmath17 as @xmath13 changes have not been verified , several experiments using a number of proteins have confirmed that indeed @xmath18 increases with @xmath13 @xcite . it can not be emphasized enough that the theory described here applies only to cases when the crowding interactions between crowding agents and proteins and between crowding particles themselves are purely repulsive . while much less is known about the effects of crowding on the folding kinetics , cheung et al . @xcite predicted that the entropic stabilization also suggests that the folding rates should increase at moderate values of @xmath13 . they suggest that crowding can enhance folding rates by a factor @xmath19 where @xmath20 @xmath21 is the decrease in the entropy of the dse relative to its value in the bulk . from the arguments of cheung et al . @xcite it follows that the equilibrium changes in the entropy ( @xmath20 ) of the dse , with respect to the bulk , should also determine rate enhancement provided that neither the barriers to folding @xcite nor the native state is perturbed significantly by crowding particles . single molecule force spectroscopy , such as atomic force microscopy ( afm ) and laser optical tweezers , have been used to monitor the behavior of biopolymers under tension are ideally suited to probe the enhancement in crowding - induced stability by a direct measurement of @xmath22 . indeed , ping et al . @xcite have recently investigated the effect of dextran molecules on the mechanical stability of @xmath23 . the 8 ub ( fig . [ fig:1]*a * ) modules were n - c linked ( i.e. , modules i and i+1 were chemically linked together in a head - to - tail manner ) . they found that the average force required to unfold a module , @xmath24 , increased by 21@xmath25 as the dextran concentration , @xmath26 , was increased from 0 to 300 g / l at @xmath27 pn / s . similar results have been obtained recently by yuan et al . at @xmath28 pn / s @xcite . motivated in part by experiments @xcite , we used simulations to investigate the effects of crowding agents on the mechanical stability of a protein subject to external tension . we focused on ubiquitin ( ub ) , a 76-residue protein composed of 5 @xmath29-strands and 2 @xmath30-helices ( fig . [ fig:1]*a * ) , and confined our investigation to non - equilibrium ` force - ramp ' experiments @xcite . the primary data recorded during such an experiment is a trace of the force exerted on the tip as a function of the extension of the molecule ; a force - extension curve ( fec ) . when the force exceeds some critical value , the fec displays a sudden increase in length and is often accompanied by a concomitant sharp decrease in force . presumably , the sharp change corresponds to the unfolding of the protein . typical afm experiments use tandem arrays of proteins which are chemically linked together ( often through genetic engineering ) . we use the term module to denote a protein of the array . the fec resulting from such an experiment reveals several equally spaced peaks punctuated by sharp increases in the extension of the molecule corresponding to the unfolding of individual modules . the height of these force peaks and their shape depend on the loading rate , @xmath31 , where @xmath32 is the cantilever s spring constant and @xmath33 is the ( constant ) speed at which the stage is retracted away from the cantilever @xcite . in order to compare to experiments our simulations are performed using coarse - grained models for which simulations can be done at @xmath34 that are comparable to those in afm experiments . our work has led to a number of testable results : ( 1 ) at @xmath35 the average unfolding force for ub increases by at most only 7% compared to @xmath36 . we find that @xmath24 in small crowding agents is greater than in larger particles . ( 2 ) in the presence of crowding agents , secondary structural elements reform multiple times even after initial rupture . ( 3 ) although large crowding particles are predicted to have a smaller effect on @xmath24 ( for a given @xmath13 ) , they can profoundly affect the unfolding of poly ub . we predict that @xmath24 for a given subunit depends on the number of already unfolded portions of the poly protein . this result is important because many naturally occurring proteins that are subject to tensile stresses exist as tandem arrays of modules . it further suggests that the existence of such redundancy can more properly be understood in the context of a crowded cellular milieu . we used a coarse - grained model for proteins to investigate crowding effects on the mechanical stability of ub and @xmath37 at loading rates that are comparable to those used in afm experiments @xcite . we assumed that ub could be described using the self - organized polymer ( sop ) model ; a model that has been successfully used to make a number of predictions regarding the unfolding of proteins and rna @xcite , allosteric transitions in enzymes @xcite , and movement of molecular motors on polar tracks @xcite . previous studies @xcite have used more standard go - models @xcite to probe various aspects of forced unfolding of ub . the sop energy function ( @xmath38 ) for a protein with n amino - acids , specified in terms of the c@xmath39 coordinates @xmath40 @xmath41 , is @xmath42+\sum _ { i=1}^{n-3 } \sum _ { j = i+3}^n \varepsilon _ h \left[\left(\frac{r_{ij}^0}{r_{ij}}\right)^{12}-2\left(\frac{r_{ij}^0}{r_{ij}}\right)^6\right]\delta _ { ij}\\ & + \sum _ { i=1}^{n-2 } \sum _ { j = i+2}^n \varepsilon _ l\left(\frac{\sigma } { r_{ij}}\right)^6\left(1-\delta _ { ij}\right ) , \end{split}\ ] ] where @xmath43 , @xmath44 is the value of @xmath45 in the native structure , @xmath46 kcal/(mol@xmath47@xmath48 ) , @xmath49 kcal / mol , @xmath50 kcal / mol , and @xmath51 nm . note that @xmath52 kcal / mol @xmath53 pn@xmath54 nm . in eq . ( [ eqn:1 ] ) @xmath55 if @xmath56 nm , and @xmath57 otherwise . native coordinates corresponded to those of the c@xmath39 atoms of the 1.8 resolution protein data bank crystal structure 1ubq @xcite . for ub @xmath58 and @xmath59 for @xmath37 . the first term in eq . ( [ eqn:1 ] ) is the fene potential @xcite that accounted for chain connectivity . the second ( lennard - jones ) term accounted for the non - bonded interactions that stabilize the native state , and the final ( soft - sphere ) term accounted for excluded - volume interactions ( including those of an angular nature ) . the sop model is different from the go - model because there are no angular terms in sop , and the connectivity is enforced differently as well . the sop representation of the polypeptide chain is in the same spirit as other coarse - grained models used in polymers @xcite . we assumed that the crowding particles are spherical with diameter @xmath60 . ( @xmath61 nm in some simulations , while @xmath62 nm in others . ) crowders interacted amongst themselves and with the protein , respectively , via the following lj potentials : @xmath63 @xmath64 where @xmath65 , @xmath66 nm , @xmath67 , and @xmath68 . the heaviside functions truncate the potentials at their minima and thereby ensured that only the repulsive portions of eqs . ( [ eqn:2 ] ) and ( [ eqn:3 ] ) contributed to interactions involving crowding agent . _ mimics of crowding using asakura - oosawa theory : _ even using a coarse - grained sop representation of proteins , it is difficult to carry out converged simulations in the presence of crowding agents . the reason is that the number of crowding agents can be large . moreover , the separation in the spatial and temporal scales of the protein and the crowding particles has to be carefully considered to obtain reliable results . in light of these difficulties , it is of interest to consider the effective attraction between the sites on the protein using the implicit pairwise potential computed by asakura and oosawa . the intramolecular attraction arises due to the depletion of crowding particles near the protein . to probe the efficacy of these models we employed in some simulations the asakura - oosawa model @xcite of crowding effects . for these simulations , we added the following term to the bare sop hamiltonian ( eqn . ( [ eqn:1 ] ) ) : @xmath69 where @xmath51 nm , @xmath61 nm , @xmath35 , @xmath70 pn@xmath54 nm , and @xmath45 is the distance separating protein beads @xmath71 and @xmath72 . for simulations involving @xmath37 , residues in different modules interacted via : @xmath74 where r is the distance separating the two beads . note that this potential is short - ranged and purely repulsive and that it is the same potential used for non - native intra - protein interactions @xcite . _ @xmath36 _ : hundreds of simulations of @xmath75 steps ( @xmath76s ) at @xmath77 k were used to generate initial structures for use in the pulling simulations . the protein was completely free in solution ( i.e. , no forces were applied to either terminus ) , and no crowders were present during the equilibrations . the n - terminus of the protein was subsequently translated to the origin , and the protein was rotated such that its end - to - end vector , @xmath78 , ( i.e. , the vector pointing from the n - terminal bead to the c - terminal bead ) coincided with the pulling ( + z ) direction . an unfolding trajectory was initiated by selecting a random initial structure from amongst the set of thermally equilibrated structures , and tethering a harmonic spring to the c - terminal bead . the n - terminal bead was held fixed throughout the simulations . tension was applied to the protein by displacing the spring along the + z axis and resulted in application of the following force to the c - terminal bead : @xmath79-\left[z_s(t)-z_s(0)\right]\right),\ ] ] where @xmath32 is the spring constant , @xmath80 , and @xmath81 corresponds to the displacement of the end of the spring . note that @xmath32 was also used to constrain the simulation to the z - axis ; @xmath82 $ ] and @xmath83 $ ] . the displacement of the spring was updated at every timestep . we simulated forced - unfolding of monomeric ub at four different @xmath34 ( @xmath84 pn / s , @xmath85 pn / s , @xmath86 pn / s , and @xmath87 pn / s ) , while simulations on n - c - linked @xmath37 were performed at @xmath88 pn / s . all overdamped force - ramp simulations were performed at the same speed @xmath89 nm / s , and spring constants were varied over a range from 0.3879 pn / nm - 31.032 pn / nm to achieve the aforementioned @xmath34 ( via the relation @xmath90 ) . our simulations were realistic because they maintained loading rates consistent with experiment and because @xmath34 is the prime determinant of unfolding pathway @xcite . _ @xmath91 0.0 _ : simulations involving explicit crowders were carried out at a fixed volume fraction @xmath35 and with a fixed number @xmath92 of crowding spheres . ( @xmath93 was fixed to render the problem computationally tractable ) . using the relation @xmath94 , we adjusted the length of a side of the cubic simulation box ( @xmath95 ) to maintain @xmath35 . thus , @xmath96 nm when @xmath97 nm and @xmath98 nm when @xmath62 nm . explicit crowders were added to the simulation after loading an equilibrated structure but before the application of tension . initial crowder positions were chosen randomly and in a serial manner from a uniform distribution . if the distance between an initial crowder position and that of another crowder or protein bead did not exceed the sum of their radii , then the prospective position was rejected and another random position chosen to avoid highly unfavorable steric overlaps . periodic boundary conditions ( pbc ) and the minimum image convention @xcite were employed in the simulations . two sets of coordinates were stored for protein beads at every timestep ; pbc were applied to one set and the other was propagated without pbc . distances between protein beads were calculated from the uncorrected set of coordinates without minimum imaging , while protein - crowder distances were calculated from the pbc coordinates with minimum imaging . to improve simulation efficiency , a cell list @xcite was used to calculate crowder - crowder and protein - crowder interactions . the entire simulation volume was partitioned into 64 subvolumes , and it was only necessary to calculate interactions within a subvolume and between beads of the subvolume and those of 13 of its 26 neighbors . the cell - list was updated at every timestep to ensure the accuracy of the simulations . the equations of motion in our overdamped simulations ( used in all force - ramp simulations ) were integrated with a timestep @xmath99 ( @xmath100 ps ) using the method of ermak and mccammon @xcite . the friction coefficient of the crowders , @xmath101 , was determined via the relation @xmath102 , where @xmath60 is the crowder diameter , @xmath51 nm is the diameter of a protein bead , and @xmath103 g / s is the friction coefficient associated with a protein bead of mass @xmath104 g. simulated - times were translated into real - times using @xmath105 @xcite . at @xmath77 k , @xmath106 ps , and since @xmath107 the real - time per step is 5.4306 ps . at @xmath108 , forced unfolding of ub was simulated at four different @xmath34 ( @xmath87 pn / s , @xmath86 pn / s , @xmath85 pn / s , @xmath84 pn / s ) , where the lowest value corresponds approximately to the value used in the pulling experiments of ping et al . @xcite , and all @xmath34 are experimentally accessible . _ force profiles _ : fig . [ fig:2](*a * and * b * ) provides examples of fec s collected at the highest and lowest @xmath34 . we used a nominal contour length of ( @xmath109 ) @xmath110 nm and unfolding forces , @xmath111 , to determine contour - length increments , @xmath112 , for each trajectory at @xmath113 pn / s ( fig . [ fig:2]*a * ) . we identified @xmath111 with the peak of the fec before the stick - slip transition @xcite . the average extension @xmath114 nm is in excellent agreement with the experimental result of 24 @xmath115 5 nm found by carrion - vazquez et al . the projection ( @xmath116 ) of the end - to - end vector at @xmath111 in the z - direction varied between 4.1 nm and 4.7 nm , depending on @xmath34 . since the native end - to - end distance , @xmath117 nm , @xmath118 ranges from 0.4 - 1.0 nm . the lower end of this range is slightly larger than the 0.25 nm transition - state distance for the mechanical unfolding of the structurally similar titin immunoglobulin domains @xcite . indeed , we expect @xmath119 nm , because of the non - equilibrium nature of the simulations . larger @xmath34 typically lead to larger @xmath120 ( @xmath121 ) . average unfolding forces , @xmath122 , depended approximately logarithmically on @xmath34@xcite ( fig . [ fig:2]*c * ) , and @xmath123 pn is in fair agreement with the experimental value of @xmath124 pn observed by ping et al . @xcite at @xmath27 pn / s . the @xmath125 also showed a logarithmic dependence on @xmath34 , but the difference between the value calculated at @xmath113 pn / s and that calculated at @xmath126 pn / s is small ( @xmath127 2 ) . although the underlying free - energy landscape is time - dependent in a non - equilibrium force - ramp pulling experiment , these results suggest that the distance from the native state to the transition state is small . it is likely that at loading rates that are achieved in laser optical tweezer experiments ( @xmath128 10 pn / s ) , the location to the transition state would increase because the response of biopolymers to loading rate changes from being plastic ( low @xmath34 ) to brittle ( high @xmath34 ) @xcite . _ unfolding pathways _ : in the dominant pathway , unfolding proceeded in a fairly markovian fashion with the primary order of events following the sequence @xmath1291/@xmath1295 @xmath130 @xmath1293/@xmath1295 @xmath130 @xmath1293/@xmath1294 @xmath130 @xmath1291/@xmath1292 ( figs . [ fig:1]*b * and [ fig:12 ] ) . this is precisely the same sequence seen in the simulations by li et al.@xcite . alternative pathways , reminiscent of kinetic partitioning @xcite observed in forced - unfolding of gfp @xcite and lysozyme @xcite , were also infrequently sampled . for example , at @xmath126 pn / s , @xmath131 of the trajectories unfolded as follows : @xmath1291/@xmath1295 @xmath130 @xmath1291/@xmath1292 @xmath130 @xmath1293/@xmath1295 @xmath130 @xmath1293/@xmath1294 ( fig . [ fig:1]*b * ) , while the remaining @xmath132 followed the dominant pathway . at the highest loading rate , one frequently observed the following sequence of events @xmath1291/@xmath295 @xmath130 @xmath1293/@xmath1295 @xmath130 @xmath1291/@xmath1292 @xmath130 @xmath1291/@xmath1292 ( reform ) @xmath130 @xmath1293/@xmath294 @xmath130 @xmath1291/@xmath1292 , where @xmath1291/@xmath1292 ruptured but then reformed prior to the rupture of @xmath1293/@xmath1294 . as illustrated in fig . [ fig:2](*a * and * b * ) , unfolding events at smaller @xmath34 tended to result in larger molecular extensions . the non - equilibrium character of a pulling experiment decreased with decreasing @xmath34 , and smaller @xmath34 resulted in smaller force - drops after the unfolding force , @xmath111 , is reached . since the force applied by the spring to the end of the protein did not fall off as sharply at lower @xmath34 , more of the protein was extended during an unfolding event . the smaller extensions following unfolding events at higher @xmath34 ( relative to those observed at lower @xmath34 ) were responsible for the @xmath1291/@xmath1292 unfolding / refolding events mentioned above because the applied tension was very low after the initial rupture event ( fig . [ fig:2](*a * and * b * ) ) . at lower @xmath34 , this situation no longer held because the initial unfolding event resulted in a chain extension that was a significant fraction of the chain s contour length ( fig . [ fig:2 ] * b * ) . depletion forces stabilize proteins and shift the folding equilibrium towards more compact states @xcite . these forces result from an increase in the entropy of the crowding agents that more than compensates for an increase in the free - energy of a protein molecule upon compaction . simulations of forced - unfolding of ub in the presence of explicit crowders of diameters @xmath61 nm and @xmath62 nm were used to assess the contribution of the depletion forces to mechanical stability . sixteen trajectories were collected for each @xmath34 investigated . three @xmath34 ( @xmath86 pn / s , @xmath85 pn / s , @xmath84 pn / s ) , were explored for the @xmath61 nm sized depletants . only the two highest @xmath34 ( @xmath85 pn / s , @xmath84 pn / s ) , were explored for the @xmath62 nm sized crowders . _ small crowding particles increase unfolding forces _ : example fec s collected in the presence of crowders of diameter @xmath61 nm and @xmath62 nm are illustrated in fig . [ fig:9](*a * and * b * ) . the @xmath61 nm curves in fig . [ fig:9 ] look qualitatively very similar to those seen at @xmath108 , while the fec s collected at @xmath62 nm look qualitatively different . for example , larger @xmath133 than those observed at either @xmath108 or in the presence of the @xmath61 nm crowders are apparent in the fec s collected at @xmath62 nm ( fig . [ fig:9](*a * and * b * ) ) . indeed , fig . [ fig:9]*c * reveals that this observation is quantitatively accurate . although the @xmath133 in the presence of the @xmath61 nm crowders were statistically indistinguishable from the @xmath133 at @xmath36 ( compare fig . [ fig:9]*c * with fig . [ fig:2]*c * ) , the average unfolding forces in the presence of the @xmath134 nm crowders were statistically greater than those measured at @xmath108 . at @xmath113 pn / s , the @xmath135 in the presence of the @xmath62 nm crowders exceeded that at @xmath108 by 3@xmath25 , while at @xmath136 pn / s the increase was 4@xmath25 ( compare fig . [ fig:9]*c * with fig . [ fig:2]*c * ) . in one respect these results are not surprising ; it follows from the ao theory and eq . ( [ eqn:6 ] ) that smaller crowders stabilize ub more than larger ones . on the other hand , the extent of stabilization ( as measured by increases in @xmath133 ) was small . we should emphasize that although the increase in the unfolding force is small , the stability change upon crowding is significant . the enhancement in stability is @xmath137 @xmath138 using a 3% increase in the unfolding force , and @xmath139 the location of the unfolded basing @xmath140 nm . _ crowding leads to transient local refolding _ : the unfolding pathways in the presence of crowding agents of both sizes were very similar to those seen at @xmath108 . nevertheless , at @xmath35 there tended to be more unfolding / refolding events as the molecule extended past @xmath116 ( fig . [ fig:12 ] ) than at @xmath108 . as illustrated , strand - pairing between @xmath1291 and @xmath1292 and between @xmath1293 and @xmath1294 persisted to a greater extent after the initial unfolding event at @xmath35 than at @xmath108 . as ub passed through the point ( @xmath116,@xmath111 ) , its termini were extended to distances greater than the diameter of even the larger crowders . depletion forces resulting from the presence of 6.4 nm crowders act on these larger length scales . assuming that the tension applied to the c - terminus is small enough ( e.g. , after rupture events at higher loading rates ) , then such depletion forces can promote reformation of contacts between secondary structural elements several times during the course of a trajectory . if this is indeed the case , then it suggests that unfolding polyub may be different from the unfolding of monomeric ub , because depletion effects should increase with number of modules in the tandem ( see below ) . _ ao model for forced - unfolding in the presence of crowders _ : we used the asakura - oosawa ao model @xcite ( eq . ( [ eqn:8 ] ) ) of the depletion interaction to model the effects of a crowded environment on ub . the ao theory has been successfully used to model the effects of a crowded environment on polymers and colloids @xcite , and we found that it gives qualitatively accurate results for ub . we used eq . ( [ eqn:8 ] ) to approximate the effective interaction between spherical protein beads immersed in a crowded solution of volume fraction @xmath35 . from the form of the ao potential between protein beads , it follows that the range of the potential is proportional to @xmath60 , but the strength is greater for smaller crowders . indeed , as revealed in the previous section , simulations in the presence of explicit crowders showed that 1.0 nm crowders resulted in larger average unfolding forces than 6.0 nm crowders ( fig . [ fig:9]*c * ) . although the ao - potential yields qualitatively accurate results , use of the ao - potential of eq . ( [ eqn:8 ] ) to implicitly model non - bonded interactions in ub did not yield quantitatively accurate results . at @xmath113 pn / s @xmath141 pn , while @xmath142 pn . thus , simulations with the ao potential led to a mean unfolding force that is roughly 50@xmath25 greater than in its absence . this disagrees sharply with the @xmath143 pn resulting from our own simulations in the presence of explicit crowding agent at @xmath35 at @xmath113 pn / s ( fig . [ fig:9]*c * ) . indeed , the result also stands in marked contrast to the experimental results of ping et al . @xcite on octameric ub , which saw a maximum increase in @xmath133 of 21@xmath25 ( at @xmath144 , @xmath145 pn / s , and with @xmath146 nm ) over the @xmath147 pn at @xmath36 @xcite . there are a couple of origins to the discrepancy . first , the ao - potential was derived to understand the equilibrium of colloidal spheres and plates in the presence of smaller - sized spherical crowding agents . our experiments were of a non - equilibrium nature , so it is somewhat unreasonable to expect such simulations to yield quantitatively accurate unfolding forces or dynamics . second , as pointed out by shaw and thirumalai @xcite , three - body terms are required to properly model depletion effects even in good solvents let alone in the poor - solvent conditions of our simulations . to elaborate , let us consider the volume excluded to crowders by a ub molecule , v@xmath148(ub ) , to be the volume enclosed by a union of spheres of radii @xmath149 = @xmath150 . with an ao potential , the volume excluded to the spherical crowding agents is : @xmath151 where @xmath152 is the number of residues in monomeric ub ( 76 ) , @xmath153 is the volume of @xmath149 and @xmath154 is the volume associated with the overlap of @xmath149 and @xmath155 . ( [ eqn:9 ] ) neglects the overlap of three or more spheres . the importance of such overlaps , for soft - spheres , increases as the crowders become much larger than the protein beads , and the neglect of such overlaps is the reason for the quantitative inaccuracy . as the size of the crowders increases ( i.e. , as the thickness of the depletion layer surrounding the protein beads increases ) , the surface enclosing ub becomes more spherical , loses detail , and undoubtedly changes less in response to changes in the conformation of the molecule . since depletion forces are proportional to the change in @xmath156(ub ) with respect to changes in ub conformation , eq . ( [ eqn:9 ] ) overestimates the size of depletion forces in the presence of large crowders . despite these limitations the ao model , which is simple , can be used to provide qualitative predictions . finally , we note that experiments typically use polyproteins to study force induced unfolding ( e.g. , ping et al . @xcite used ( ub)@xmath157 in their experiments ) . for polyproteins the quantitative accuracy of the ao theory for unfolding in the presence of crowders of diameter @xmath158 nm is likely to increase , because the volume excluded to the crowders will change significantly with changes in the conformation of the polyprotein . from arguments based on volume exclusion that lead to crowding - induced entropic stabilization of the folded structures , it follows that crowding effects should be more dramatic on poly ub than the monomer . in order to illustrate the effect of crowding on stretching of ( ub)@xmath0 we chose @xmath61 nm , which had negligible effect on @xmath2 for the monomer . however , we found significant influence of the large crowding particles when ( ub)@xmath0 was forced to unfold in their presence . the fec ( fig . [ fig:14 ] ) shows three peaks that corresponded to unfolding of the three domains , when simulations were performed at crowder volume fractions @xmath108 and @xmath35 at @xmath88 pn / s . [ fig:13 ] presents average unfolding forces as a function of the unfolding event number . although the first two unfolding events were statistically indistinguishable at @xmath108 and @xmath35 , the final event occurred at much larger @xmath133 in the presence of crowders than in their absence . _ order of unfolding was stochastic : _ @xmath37 has 3 chemically identical modules . in the pulling simulations the n - terminus of module @xmath159 was held fixed , while force was applied to the c - terminus of module @xmath160 . figure [ fig:20 ] illustrates the time - dependence of contacts between secondary structure elements of modules @xmath159 , @xmath161 , @xmath160 . it is clear from fig . [ fig:20]*a * that module @xmath160 unfolded first , followed by module @xmath161 , and finally by module @xmath159 . on the other hand , the order of events in fig . [ fig:20]*b * was @xmath162 . the frequency with which the @xmath163 possible permutations of these orders at @xmath36 and at @xmath35 were observed is presented in table [ tbl : moduleorderfrequencies ] . it is clear that at both @xmath36 and at @xmath35 , the most probably order of events was @xmath164 . at @xmath36 this order was only marginally more probable than the order @xmath165 , while at @xmath35 @xmath164 became overwhelmingly more probable than any other unfolding order . in none of the simulations at @xmath36 or @xmath35 , did the rupture of module @xmath160 occur as the final event . _ unfolding within a module depended on proximity to the point of force application : _ at @xmath166 , @xmath37 is fairly brittle , and the rupture of contacts within a module occured nearly simultaneously . nevertheless , by carefully examining time - dependent contact maps such as those illustrated in fig . [ fig:20 ] , we were able to determine ( 1 ) at both @xmath36 and @xmath35 , @xmath167 contacts were the first to rupture , and ( 2 ) only for module @xmath160 was this rupture event _ invariably _ followed by the loss of @xmath168 contacts . when other modules ruptured , loss of @xmath167 contacts was occasionally followed by loss of the @xmath169 strand - pair contacts . @xmath37 _ must achieve a larger _ @xmath7 _ to rupture at _ @xmath35 : figure [ fig:21]*a * illustrates the time dependence of @xmath170 at @xmath36 and at @xmath35 . interestingly , the plot reveals that after the second rupture event , the @xmath171 increased more rapidly in the presence of crowders than in their absence . this is likely a reflection of the fact that at @xmath3 modules @xmath159 and @xmath160 were the first two modules to unfold in 44% of the trajectories , while at @xmath35 these two modules were the first to unfold in only 19% of trajectories . thus , @xmath171 increased more rapidly in the presence of crowders , because the @xmath7 of @xmath37 with two adjacent modules unfolded is larger than that with two unfolded but non - adjacent modules . interestingly , these differences are masked in the time - dependent increase of the end - to - end distance ( fig . [ fig:21]*b * ) . figure [ fig:21]*a * also reveals that the horizontal inflection points marking the third unfolding event occur at different times at @xmath172 and in the absence of crowders . the difference between these two times was responsible for the difference in average unfolding forces of @xmath173 pn illustrated in fig . [ fig:13 ] . thus , it is clear that despite the highly non - equilibrium nature of this pulling experiment , depletion effects were substantial and it required a much greater force to reach @xmath111 in the presence of these crowders than in their absence . we predict that systematic experiments will reveal that polyub molecules composed of larger numbers of modules will show greater increases in @xmath24 relative to @xmath174 than polyub molecules composed of fewer repeats . the size of these differences should increase with decreasing loading rate . finally , it may even be possible to observe differences in the @xmath133 as a function of unfolding event number ( as in fig . [ fig:13 ] ) . the increase in @xmath24 , at a fixed @xmath13 , for poly ub is likely to be even more significant for small crowding agents as shown in eq . ( [ eqn:6 ] ) ( see below for further discussion ) . general theory , based on the concept of depletion effects ( see eqs . ( [ eqn:8 ] ) and ( [ eqn:6 ] ) ) , shows that crowding should enhance the stability of proteins , and hence should result in higher forces to unfold proteins . however , predicting the precise values of @xmath24 is difficult because of the interplay of a number of factors such as the size of the crowding agents and the number of amino acid residues in the protein . despite the complexity a few qualitative conclusions can be obtained based on the observation that , when only excluded volume interactions are relevant , then the protein or polyprotein would prefer to be localized in a region devoid of crowding particles @xcite . the size of such a region @xmath175 . if @xmath176 then the crowding would have negligible effect on the unfolding forces . the condition @xmath176 can be realized by using large crowding particles at a fixed @xmath1 . in the unfolded state , @xmath177 nm @xcite which for ub leads to @xmath178 nm . thus , @xmath179 . these considerations suggest that the crowder with @xmath158 nm would have negligible effect on the unfolding force , which is in accord with the simulations . on the other hand , @xmath180 when @xmath181 nm , and hence we expect that the smaller crowders would have measurable effect on the unfolding forces . our simulations are in harmony with this prediction . we expect that for the smaller crowding agent @xmath2 would scale with @xmath1 in a manner given by eq . ( [ eqn:6 ] ) . in general , appreciable effect of crowding on the unfolding forces can be observed only for large proteins or for polyproteins using relatively small crowding agents . although we have only carried out simulations for ub and ( ub)@xmath0 at one non - zero @xmath13 , theoretical arguments can be used to predict the changes in @xmath24 as @xmath13 increases . the expected changes in the force required to unfold a protein can be obtained using a generalization of the arguments of cheung et al.@xcite in the presence of crowding agents the protein is localized in a region that is largely devoid of the crowding particles @xcite . the most probable size of the region is @xmath182 where @xmath4 is the size of the crowding agent . if the structures in the dse are treated as a polymer with no residual structure then the increase in entropy of the dse upon confinement is @xmath183 where @xmath7 is the dimension of the unfolded state of the protein . if native state stabilization is solely due to the entropic stabilization mechanism we expect : @xmath184 where @xmath22 is the critical force for unfolding the protein , @xmath185 is the gain in contour length at the unfolding transition , and @xmath186 relates @xmath7 to the number of amino acids through the relation @xmath187 ( @xmath188 varies between 2 - 4 ) . a few comments regarding eq . ( [ eqn:6 ] ) are in order . ( 1 ) the @xmath13 dependence in eq . ( [ eqn:6 ] ) does not depend on the nature of the most probable region that is free of crowding particles . as long as the confining region , which approximately mimics the excluded volume effects of the macromolecule , is characterized by a single length d , we expect eq . ( [ eqn:6 ] ) to be valid . ( 2 ) the additional assumption used in eq . ( [ eqn:6 ] ) is that @xmath189 , and hence there may be deviations due to finite size effects . ( 3 ) the equivalence between crowding and confinement breaks down at large @xmath13 values . consequently , we do not expect eq . ( [ eqn:6 ] ) to fit the experimental data at all values of @xmath13 . ( 4 ) it follows from eq . ( [ eqn:6 ] ) that , for a given @xmath7 , small crowding agents are more effective in stabilizing proteins than large ones . thus , the prediction based on eq . [ eqn:6 ] is supported by our simulations . ( 5 ) from the variation of @xmath24 with @xmath13 ping et al . @xcite suggest that @xmath190 . however , the large errors in the measurements can not rule out the theoretical prediction in eq . ( [ eqn:6 ] ) . we have successfully fit their experimental results using eq . ( [ eqn:6 ] ) ( fig . [ fig:21]*c * ) . additional quantitative experiments are required to validate the theoretical prediction . it is difficult to map the concentrations in g / l used in the study of ping et al . @xcite to an effective volume fraction because of uncertainties in the molecular weight of dextran used in the study . hence , a quantitative comparison between theory and experiments is challenging . a naive estimate may be obtained using the values reported by weiss et al . given that the dextran used in the study is thought to have an average molecular weight of 40 kda and an estimated average hydrodynamic radius of 3.5 nm @xcite , we find that @xmath191 g / l corresponds to a volume fraction @xmath192 , which is very large . nevertheless , @xmath13 must be large when @xmath191 alternatively , we estimated @xmath193 for dextran using @xmath194 where @xmath152 is the number of monomers in a 40 kda dextran is @xmath195 . if the monomer size @xmath196 nm , then we find @xmath197 . if we assume that @xmath191 g / l corresponds to @xmath198 and that @xmath199 ( eq . ( [ eqn:6 ] ) ) , then at @xmath35 we would expect a nearly 18@xmath25 increase in @xmath133 . similarly , if @xmath191 g / l corresponds to @xmath200 then we would expect an increase in @xmath133 of approximately 21@xmath25 . in any case , we can say that at physiologically relevant volume fractions ( @xmath13 @xmath201 [ 0.1,0.3 ] ) , the percent increase in @xmath133 is likely to be @xmath202 . our simulations for @xmath203 nm predict an increase of 3 - 4% , which shows that a more detailed analysis is required to obtain an accurate value of @xmath4 for dextran before a quantitative comparison with experiments can be made . the larger increase seen in experiments may also be a reflection of the use of ( ub)@xmath157 rather than a monomer . regardless of the crowder size we find that the unfolding pathways are altered in the presence of crowding agents . it is normally assumed that the rupture of secondary structure elements is irreversible if the applied force exceeds a threshold value . however , when unfolding experiments are carried out in the presence of crowding particles , that effectively localize the protein in a smaller region than when @xmath36 , reassociation between already ruptured secondary structures is facilitated as shown here . thus , forced - unfolding can not be described using one dimensional free energy profiles with @xmath116 as the reaction coordinate @xcite . we find that the average unfolding force for the final rupture event of the unfolding of @xmath37 occurred at much larger values in the presence of crowders than in their absence . with @xmath61 nm , which has practically no effect on the unfolding force of the monomer , and @xmath35 even with unfolding of two modules the interactions between the stretched modules and protein are small ( @xmath204 ) . only upon unfolding of the third ub do crowding effects become relevant , which leads to an increase in @xmath24 . our results suggest that @xmath205 should increase with the number of modules in the array and that it may be possible to detect differences in @xmath133 which are conditional on the unfolding event number . we speculate that naturally occurring polyproteins that are subject to mechanical stress have evolved to take advantage of precisely such enhanced depletion effects . this work was supported by a grant from the national science foundation ( che 05 - 14056 ) . dlp is grateful for a ruth l. kirschstein postdoctoral fellowship from the national institute of general medical sciences ( f32gm077940 ) . computational time and resources for this work were kindly provided by the national energy research scientific computing ( nersc ) center . 10 zhou , h. x. ; rivas , g. ; minton , a. p. _ annu . biophys . _ * 2008 , * _ 37 , _ 375 - 397 . minton , a. p. _ biophys . j. _ * 2005 , * _ 88 , _ 971 - 985 . cheung , m. s. ; klimov , d. ; thirumalai , d. _ proc . * 2005 , * _ 102 , _ 4753 - 4758 . zhou , h. x. _ acc . chem . * 2004 , * _ 37 , _ 123 - 130 . asakura , s. ; oosawa , f. _ j. chem . phys . _ * 1954 , * _ 22 , _ 1255 - 1256 . asakura , s. ; oosawa , f. _ j. polym . sci . _ * 1958 , * _ 33 , _ 183 - 192 . vrij , a. _ pure appl . chem . _ * 1976 , * _ 48 , _ 471 - 483 . sasahara , k. ; mcphie , p. ; minton , a. p. _ j. mol . biol . _ * 2003 , * _ 326 , _ 1227 - 1237 . stagg , l. ; zhang , s. q. ; cheung , m. s. ; wittung - stafshede , p. _ proc . * 2007 , * _ 104 , _ 18976 - 18981 . homouz , d. ; perham , m. ; samiotakis , a. ; cheung , m. s. ; wittung - stafshede , p. _ proc . * 2008 , * _ 105 , _ 11754 - 11759 . cheung , m. s. ; thirumalai , d. _ j. phys . b _ * 2007 , * _ 111 , _ 8250 - 8257 . ping , g. ; yang , g. ; yuan , j .- m . _ polymer _ * 2006 , * _ 47 , _ 2564 - 2570 . yuan , j. m. ; chyan , c. l. ; zhou , h. x. ; chung , t. y. ; haibo , j. ; ping , g. ; yang , g. _ prot . _ published online before print * sep 9 , 2008 * doi:10.1110 . fisher , t. ; marszalek , p. ; fernandez , j. _ nature structural biology _ * 2000 , * _ 7 , _ 719 - 724 . evans , e. ; ritchie , k. _ biophys . j. _ * 1997 , * _ 72 , _ 1541 - 1555 . hyeon , c. ; thirumalai , d. _ biophys . j. _ * 2006 , * _ 90 , _ 3410 - 3427 . hyeon , c. ; dima , r. ; thirumalai , d. _ structure _ * 2006 , * _ 14 , _ 1633 - 1645 . hyeon , c. ; lorimer , g. ; thirumalai , d. _ proc . natl . * 2006 , * _ 103 , _ 18939 - 18944 . chen , j. ; dima , r. ; thirumalai , d. _ j. mol . biol . _ * 2007 , * _ 374 , _ 250 - 266 . hyeon , c. ; onuchic , j. _ proc . natl . acad . * 2007 , * _ 104 , _ 17382 - 17387 . best , r. b. ; paci , e. ; hummer , g. ; dudko , o. k. _ j. phys . chem . b _ * 2008 , * _ 112 , _ 5968 - 5976 . karanicolas , j. ; brooks iii , c. l. _ prot . sci . _ * 2002 , * _ 11 , _ 2351 - 2361 . li , m. ; kouza , m. ; hu , c. _ biophys . j. _ * 2007 , * _ 92 , _ 547 - 561 . vijay - kumar , s. ; bugg , c. ; cook , w. _ j. mol . _ * 1987 , * _ 194 , _ 531 - 544 . kremer , k. ; grest , g. _ j. chem . * 1990 , * _ 92 , _ 5057 - 5086 . pincus , d. l. ; cho , s. s. ; hyeon , c. ; thirumalai , d. minimal models for proteins and rna : from folding to function . in _ progress in nucleic acid research and molecular biology _ ; conn , p. m. , ed . ; ( in press ) . allen , m. p. ; tildesley , d. j. _ computer simulation of liquids ; _ oxford university press : 1987 . ermak , d. ; mccammon , j. _ j. chem . phys . _ * 1978 , * _ 69 , _ 1352 - 1360 . veitshans , t. ; klimov , d. ; thirumalai , d. _ folding and design _ * 1996 , * _ 2 , _ 1 - 22 . carrion - vazquez , m. ; li , h. ; lu , h. ; marszalek , p. ; oberhauser , a. ; fernandez , j. _ nature structural biology _ * 2003 , * _ 10 , _ 738 - 743 . bockelmann , u. ; essevaz - roulet , b. ; heslot , f. _ phys . e _ * 1998 , * _ 58 , _ 2386 - 2394 . carrion - vazquez , m. ; oberhauser , a. ; fowler , s. ; marszalek , p. ; broedel , s. ; clarke , j. ; fernandez , j. _ proc . usa _ * 1999 , * _ 96 , _ 3694 - 3699 . hyeon , c. ; thirumalai , d. _ j. phys . matt . _ * 2007 , * _ 19 , _ 113101 . guo , z. ; thirumalai , d. _ biopolymers _ * 1995 , * _ 36 , _ 83 - 102 . mickler , m. ; dima , r. ; dietz , h. ; hyeon , c. ; thirumalai , d. ; rief , m. _ proc . * 2007 , * _ 104 , _ 20268 - 20273 . peng , q. ; li , h. _ proc . * 2008 , * _ 105 , _ 1885 - 1890 . shaw , m. r. ; thirumalai , d. _ phys . a _ * 1991 , * _ 44 , _ r4797-r4800 . verma , r. ; crocker , j. ; lubensky , t. ; yodh , a. _ phys . lett . _ * 1998 , * _ 81 , _ 4004 - 4007 . toan , n. ; marenduzzo , d. ; cook , p. ; micheletti , c. _ phys . _ * 2006 , * _ 97 , _ 178301(4 ) . thirumalai , d. _ phys . rev . a _ * 1988 , * _ 37 , _ 269 - 276 . kohn , j. e. ; millet , i. s. ; jacob , j. ; zagrovic , b. ; dillon , t. m. ; cingel , n. ; dothager , r. s. ; seifert , s. ; thiyagarajan , p. ; sosnick , t. r. ; hasan , m. z. ; pande , v. s. ; ruczinski , i. ; doniach , s. ; plaxco , k. w. _ proc . * 2005 , * _ 101 , _ 14475 - 14475 . weiss , m. ; elsner , m. ; kartberg , f. ; nilsson , t. _ biophys . j. _ * 2004 , * _ 87 , _ 3518 - 3524 . humphrey , w. ; dalke , a. ; schulten , k. _ j. mol . graph . _ * 1996 , * _ 14 , _ 33 - 38 . .module unfolding order frequencies at @xmath36 and @xmath35 . [ cols="^,^,^",options="header " , ] [ tbl : moduleorderfrequencies ] fig . [ fig:1 ] : ( * a * ) cartoon representation of the native structure of ubiquitin ( pdb accession i d 1ubq ) in the presence of spherical crowding agents . the five beta - strands , labeled @xmath1291 through @xmath1295 , are colored in yellow . the two alpha - helices ( @xmath301 and @xmath302 ) are shown in purple . the n- and c - terminal beads are represented as spheres . in our simulations the n - terminal bead was held fixed while the c - terminal bead was pulled via a tethered spring . ( * b * ) snapshots from an unfolding trajectory illustrating the main ubiquitin ( ub ) unfolding pathway ( brown - dashed arrows ) and an alternate unfolding pathway ( green - dotted arrow ) . in both pathways the initial unfolding event corresponds to separation of the c - terminal strand @xmath1295 from the n - terminal strand @xmath1291 . along the main pathway , this is quickly followed by separation of @xmath1295 from @xmath1293 . the penultimate rupture event along the main pathway corresponds to disruption of the @xmath1293/@xmath1294 strand - pair , while the n - terminal @xmath1291/@xmath1292 strand pair is the last to break . the trajectory illustrated here was generated at @xmath108 and @xmath113 pn / s . an alternate pathway was observed at @xmath108 and @xmath126 pn / s . along this pathway separation of @xmath1295 from @xmath1291 is followed by separation of @xmath1291/@xmath1292 . the final two rupture events correspond to those of the @xmath1295/@xmath1293 contacts and @xmath1293/@xmath1294 strand - pair respectively . ( figures generated with vmd @xcite ) fig . [ fig:2 ] : force - extension curves ( fec s ) at two different loading rates , ( * a * ) @xmath206 pn / s , and ( * b * ) @xmath126 pn / s . data from the simulation is presented as a red trace . for each trajectory a black arrow points to the unfolding force , @xmath111 . @xmath116 corresponds to the extension of the molecule along the pulling ( i.e. , z- ) direction evaluated at @xmath111 . @xmath112 ( dotted blue line ) is the contour length increment , and is a measure of the amount of chain released in an unfolding event . we measured @xmath112 as @xmath207 , where @xmath208 nm is a nominal contour length of 28.5 nm . stars in each subfigure mark the minimum force observed after an unfolding event , and chain conformations corresponding to the starred points in figures * a * and * b * are illustrated at the center of the figure . unfolding events at smaller @xmath34 resulted in larger molecular extensions before significant resistance was encountered . ( yellow arrows correspond to beta - strands and purple cylinders correspond to @xmath30-helices ) . ( figures generated with vmd @xcite ) . ( * c * ) @xmath133 vs. @xmath34 evaluated at @xmath36 . the red curve corresponds to a linear - least squares fit to the set of basis functions @xmath209 and demonstrates that @xmath210 . ( note that the abscissa is a log - scale ) . each point is labeled mean @xmath115 standard error . statistics at @xmath113 pn / s , @xmath85 pn / s , @xmath86 pn / s , and @xmath87 pn / s were calculated from 50 , 49 , 50 , and 16 trajectories respectively . [ fig:12 ] : rupture events at @xmath108 ( * a * ) and at @xmath35 and @xmath61 nm ( * b * ) . the figure illustrates that after the initial rupture event at step 0 , subsequent unfolding was affected by the crowding agent . @xmath1291/@xmath1292 contacts and @xmath1293/@xmath1294 persisted longer and there were many more local refolding events in the presence of the crowders ( * b * ) than in their absence ( * a * ) . ( * c * ) snapshots from an unfolding trajectory at @xmath35 and @xmath126 pn / s illustrate the primary difference between unfolding at @xmath35 and at @xmath108 . the brown - dotted arrow shows unfolding without any local - refolding events , while the sequence of black arrows illustrates local - rupture and -refolding events . the hallmark of forced - unfolding in a crowded environment is repeated breaking and reforming of contacts after an initial rupture event . ( figures generated with vmd @xcite ) fig . [ fig:9 ] : ( * a * ) examples of force - extension traces resulting from simulation in the presence of spherical crowding agents of diameter @xmath61 nm and obtained at @xmath136 pn / s . ( * b * ) examples of force - extension traces resulting from simulation in the presence of spherical crowding agents of diameter @xmath62 nm and obtained at @xmath136 pn / s . both subfigures are labeled as in fig . [ fig:2](*a * and * b * ) . ( * c * ) @xmath133 vs. @xmath34 evaluated at @xmath35 . black triangles and red circles correspond to spherical crowders of diameter @xmath62 nm and @xmath61 nm respectively . each point is labeled mean @xmath115 standard error . statistics for each point were calculated from 16 independent trajectories . only the @xmath62 nm had an appreciable effect on @xmath133 when compared to those obtained at identical @xmath34 and at @xmath108 ( see fig . [ fig:2]*c * ) fig . [ fig:14 ] : fec s for @xmath37 forced unfolding at @xmath108 ( * a * ) and at @xmath35 ( * b * ) . trajectories were generated at @xmath88 pn / s and with crowders of diameter @xmath61 nm . black arrows mark each trajectory s three unfolding events . [ fig:13 ] : @xmath133 vs. unfolding event number for the unfolding of @xmath37 in the presence of spherical crowders of diameter @xmath61 nm ( @xmath35 , black triangles ) and in their absence ( @xmath108 , red circles ) . the individual modules of the polyub tandem were n - c - linked and the loading rate was @xmath211 pn / s . each point is labeled mean @xmath212 standard error . statistics for each point were calculated from 16 independent unfolding trajectories . an unfolding event corresponded to the unfolding of an individual module . note that although the crowders had little effect on @xmath133 for the first and second unfolding events , they had a substantial effect on the last unfolding event . @xmath133 increased by @xmath173 pn for the last unfolding event . [ fig:20 ] : illustration of the stochastic nature of module unfolding . in ( * a * ) module @xmath160 ( the most proximal to the applied force ) unfolds first , followed by module @xmath161 , and finally by rupture of module @xmath159 . in ( * b * ) the order of module unfolding events is @xmath162 . the two trajectories illustrated here were both collected at @xmath36 and @xmath88 pn / s . table [ tbl : moduleorderfrequencies ] provides the frequencies at which the different possible orders were observed at @xmath36 and at @xmath35 . the diameter of the crowders is @xmath158 nm . fig . [ fig:21 ] : ( * a * ) @xmath213 versus time at @xmath36 ( red ) and at @xmath35 ( black ) . although the @xmath170 increases more rapidly with time after the second rupture event at @xmath35 than at @xmath36 , the inset reveals that a larger @xmath7 must be achieved to initiate the final rupture event in the presence of crowding particles with @xmath158 nm . ( see text for additional discussion ) . ( * b * ) @xmath214 versus time at @xmath36 ( red ) and at @xmath35 ( black ) . @xmath215 can not discriminate between unfolding at the different volume fractions and hence is less suitable ( than @xmath7 ) as a potential reaction coordinate . ( * c * ) the experimental results of ping et al . @xcite ( red circles ) for unfolding - force , @xmath133 as a function of dextran concentration , @xmath216 . the black line is a fit assuming @xmath217 and the blue assuming @xmath218 . although both fits are consistent with the data , based on theoretical considerations ( see eq . ( [ eqn:6 ] ) ) we prefer the blue fit ( see text ) . ( standard deviations taken from ping et al .

since its first synthesis in 2004 , @xcite graphene has been attracting much interest due to its unique electronic properties arising from its singular energy spectrum , where in the vicinity of the points labelled as @xmath0 and @xmath1 in reciprocal space , the charge carriers behave as massless quasi - particles and exhibit an almost linear dispersion . @xcite these quasi - particles obey the dirac - weyl equations and therefore should be subject to quasi - relativistic effects , such as zitterbewegung , i.e. , a trembling motion caused by interference between positive and negative energy states . @xcite the phenomenon of zitterbewegung was predicted in 1930 by schrdinger @xcite and has been subject of renewed interest over the past years . previous theoretical works have suggested few ways of experimentally observing zitterbewegung , e. g. in narrow - gap semiconductors @xcite , in iii - v zinc - blende semiconductor quantum wells with spin - orbit coupling @xcite and , more recently , in monolayer @xcite and bilayer @xcite graphene . an experimental simulation of the zitterbewegung of free relativistic electrons in vacuum was performed by gerritsma et al . @xcite by using trapped ions . strain engineering in graphene has recently become a widely studied topic . @xcite the elastic properties of graphene nanoribbons were theoretically investigated by cadelano _ _ et al.__@xcite , which studied the in - plane stretching and out of plane bending deformations by combining continuum elasticity theory and tight - binding atomistic simulations . later , cocco _ _ et al.__@xcite and lu and guo @xcite showed that a combination of shear and uniaxial strain at moderate absolute deformations can be used to open a gap in the graphene energy spectrum . it has been shown recently that specific forms of strain produce a pseudo - magnetic field in graphene , which does not break the time reversal symmetry and which points in opposite directions for electrons moving around the @xmath0 and @xmath1 points . @xcite the strain - induced magnetic field is expected to produce landau levels and , consequently , the quantum hall effect , even in the absence of an external magnetic field . @xcite guinea _ et al . _ , @xcite showed theoretically that an in - plane bending of the graphene sheet leads to an almost uniform field . landau levels as a consequence of strain - induced pseudo - magnetic fields greater than 300 tesla were recently observed with scanning tunnelling microscopy ( stm ) in nanometer size nanobubbles . @xcite although previous works have studied wave packet propagation for the dirac - weyl hamiltonian of graphene in the absence of external fields and potentials , @xcite or in the presence of uniformly applied external magnetic fields @xcite , there is still a lack of theoretical works on the wave packet propagation through potential and ( pseudo ) magnetic field step barriers . moreover , the time evolution of a wave packet in graphene within the tight - binding ( tb ) model , where the intra - valley scattering to higher energy states and inter - valley scattering due to defects appear naturally , is still hardly found in the literature . it is also interesting to see whether the results from dirac and tb approaches for graphene differ or are similar . in this paper , we investigate the time - evolution of wavepackets in graphene within the tb model , based on the split - operator technique for the expansion of the time - evolution operator . we trace a parallel between the results from the tb model and those obtained from the dirac approximation for particles with momentum close to one of the dirac cones of the brillouin zone of graphene . the proposed method is then applied to the study of the dynamics of gaussian wave packets in graphene under external magnetic fields . the effects of the pseudo - magnetic field induced by bending the graphene sheet into an arc of circle are analyzed as well . our results show that for an appropriate choice of strain and external magnetic field strength , the system exhibits a strong effective magnetic field for particles in one of the dirac cones , whereas in the other cone the external and pseudo - magnetic fields cancel each other and the effective magnetic field is practically zero . we show that this effect is manifested as a transient ( permanent ) zitterbewegung for electrons in the cone where the effective magnetic field is zero ( non - zero ) , which can be verified experimentally by detecting the electric field radiation emitted by the trembling wave packet . @xcite moreover , our results show that with such a choice of external and strain - induced magnetic fields , one can construct an efficient valley - filter , which can be useful for future valley - tronic devices . by solving the time - dependent schrdinger equation , one obtains that the propagated wavefunction after a time step @xmath2 can be calculated by applying the time - evolution operator on the wave packet at any instant @xmath3 @xmath4 where we assumed that the hamiltonian @xmath5 is time - independent . different techniques have been suggested for the expansion of the exponential in eq . ( [ eqtimeevolutionop ] ) , e.g. the chebyschev polynomials method @xcite and the second order differencing scheme @xcite . the numerical method that we use for the application of the time evolution operator in this work , namely , the split - operator method , @xcite is the subject of this section , where we will show how this technique can be adapted for the study of the wave packet dynamics in graphene within the tight - binding and dirac approximations . graphene consists of a layer of carbon atoms forming a honeycomb lattice , which can be described by the hamiltonian @xmath6,\ ] ] where @xmath7(@xmath8 ) annihilates ( creates ) an electron at the site @xmath9 , with on - site energy @xmath10 , and the sum is taken only between nearest neighbor sites @xmath9 and @xmath11 , with hopping energy @xmath12 . the effect of an external magnetic field can be calculated by including a phase in the hopping parameters according to the peierls substitution @xmath13 $ ] , where @xmath14 is the vector potential describing the magnetic field . @xcite in a strained graphene sheet , the distance between two adjacent sites @xmath9 and @xmath11 is changed by @xmath15 , where @xmath16 is the lattice parameter of unstrained graphene and @xmath17 is the distance between the sites after the stress . the change in the inter - sites distance affects the hopping energy between the sites , which becomes @xcite @xmath18 . a similar expression can be obtained by expanding eq . ( 13 ) of ref . in taylor series and neglecting higher order terms in @xmath19 , i.e. , considering small lattice deformations . the strain - induced change in the hopping energies leads to an effective pseudo - magnetic field , which points to opposite directions in the valley @xmath0 and @xmath1 . @xcite notice that the pseudo - magnetic field in our model is not introduced artificially by considering an additional vector potential in the peierls phase , but it rather appears naturally after the changes in the inter - site distances due to the strain . and @xmath20 . the atoms are labelled as @xmath21 according to their line and column numbers @xmath22 and @xmath23 , respectively . ( b ) reciprocal lattice of graphene , with @xmath0 ( black ) and @xmath1 ( gray ) dirac points , where the area defined by the reciprocal vectors @xmath24 and @xmath25 represents the first brillouin zone . the numbers close to each dirac point are explained in the text . ] let us label the sites @xmath9 of the graphene lattice according to their line and column numbers @xmath22 and @xmath23 , respectively , as shown in fig . [ fig : honeycomb](a ) . the basis vector state @xmath26 represents an electron confined on the site of line @xmath22 and column @xmath23 . in a lattice consisting only of non - interacting sites , each @xmath26 is an eigenstate of @xmath27 with energy @xmath28 , i.e. @xmath29 . limiting ourselves to nearest neighbors interactions , we find @xmath30 where @xmath31 and @xmath32 are equivalent to the hopping energies @xmath12 between the site @xmath33 and the adjacent sites @xmath34 and @xmath35 , respectively . ( [ htbfull ] ) can be rewritten as @xmath36 where the operators @xmath37 and @xmath38 are defined as [ htbbroken ] @xmath39 and @xmath40 the wavefunction at any instant @xmath3 is then written as a linear combination of the basis vector states @xmath41 . the advantage of following the procedure described by eqs . ( [ htbfull]-[htbbroken ] ) lies in the fact that the operators @xmath37 and @xmath38 in eq . ( [ htbbroken ] ) can be represented by tri - diagonal matrices , which are easier to handle than the matrix representing the full hamiltonian , eq . ( [ htbfull ] ) . the split - operator technique can now be applied to the hamiltonian eq . ( [ htbbest ] ) , so that the time evolution operator is approximated by @xmath42 where the error comes from the non - commutativity between the operators @xmath37 and @xmath38 . we drop the @xmath43 terms by considering a small time step @xmath44 fs . the propagated wavefunction is then obtained from eq . ( [ eqtimeevolutionop ] ) , which in this case reads @xmath45 this equation is solved in three steps : [ splitoperatorhtb ] @xmath46 @xmath47 @xmath48 using the cayley form for the exponentials , @xcite we can rewrite eq . ( [ splitoperatorhtb]a ) as @xmath49 which leads to @xmath50 as the wavefunction @xmath51 is already known , the matrix equation in eq . ( [ cayley ] ) can be straightforwardly solved to obtain @xmath52 . we repeat this procedure for the other two exponentials in eqs . ( [ splitoperatorhtb]b ) and ( [ splitoperatorhtb]c ) , and eventually obtain @xmath53 . in fact , the form in eq . ( [ cayley ] ) can also be applied to the full hamiltonian @xmath27 , i.e. without splitting the @xmath38 and @xmath37 terms . however , this would lead to matrix equations involving penta - diagonal matrices , which are harder to handle than the tri - diagonal matrices in eq . ( [ cayley ] ) . as the error produced by the splitting in eq . ( [ eqtbsplitted ] ) is smaller than the error produced by the ( necessary ) expansion of the exponential given by eq . ( [ cayley ] ) , it is worthy to split these terms in order to simplify the numerical calculations . from the tb hamiltonian eq . ( [ tbhamiltonian ] ) , considering an infinite graphene sheet in the absence of external potential and magnetic fields , one obtains the energy bandstructure of graphene as @xmath54 @xmath55 which is gapless in six points of the reciprocal space where @xmath56 , from which only two are inequivalent , labelled as @xmath0 and @xmath1 , as shown in fig . [ fig : honeycomb](b ) . @xcite in the vicinity of each of these points , the dependence of the energy spectrum on the wave vector @xmath57 is almost linear and the electron can be described as a massless fermion by the dirac hamiltonian @xmath58e^{-i\phi } , \ ] ] where @xmath59 is the fermi velocity , @xmath14 is the electromagnetic vector potential , @xmath60 is an external potential , @xmath61 is the identity matrix , @xmath62 is the pauli vector and the wavefunctions are pseudo - spinors @xmath63^t$ ] , with @xmath64 as the probability of finding the electron in the sub - lattice @xmath65 that composes the honeycomb lattice of graphene . @xcite the angle @xmath66 is different for electrons around the @xmath0 and @xmath1 dirac cones . in the vicinity of the @xmath67-th dirac point ( see labels for each dirac point in fig . [ fig : honeycomb](b ) ) , one obtains @xmath68 , with @xmath69 . from here onwards , we will refer to the coordinates in the dirac ( tb ) model as @xmath70 ( @xmath71 ) and @xmath72 ( @xmath73 ) . the exponential term in eq . ( [ dirachamiltonian ] ) is usually dropped , because it can be considered as a phase in the state vectors in the dirac model . however , this term has an important meaning when comparing with the tb model : this exponential can be identified as a rotation operator with angle @xmath66 . notice that an infinite graphene hexagonal lattice has c@xmath74 symmetry , i.e. it is symmetric only for rotation angles @xmath75 ( @xmath67 - integer ) . as a consequence , the dirac hamiltonian eq . ( [ dirachamiltonian ] ) without the exponential term is not symmetric in the momenta @xmath76 and @xmath77 , as already pointed out previously . @xcite so , what would be the meaning of the direction - dependent observables in the dirac description of graphene , when they are not symmetric under rotation , exhibiting a privileged direction ? defining @xmath73 ( @xmath71 ) as the zigzag ( armchair ) direction , as in fig . [ fig : honeycomb](a ) , the results obtained by the dirac approximation for the @xmath70 and @xmath72 components of any observable are compared to the armchair and zigzag directions , respectively , after performing a rotation @xmath66 in the coordinates of the dirac model . from the possible values of @xmath66 , one deduces straightforwardly that at any dirac cone , the coordinate @xmath70 ( @xmath72 ) of the dirac model is always related to one of the zigzag ( armchair ) directions of the real graphene lattice . on the other hand , for finite rectangular samples the different angles @xmath66 represent two distinguishable situations , since the rectangle does not share the c@xmath74 symmetry of the infinite graphene lattice : the @xmath70 direction in the dirac model for the odd ( even ) cones in fig . [ fig : honeycomb](b ) represents a diagonal ( vertical ) direction in the rectangle . the comparison between the tb model for a rectangular graphene flake and the dirac approximation will be discussed in details further , in the following section . in a recent work , maksimova _ et al . _ @xcite presented an analytical study of the time evolution of a gaussian wave packet in graphene in the _ absence _ of external potentials and/or magnetic fields within the continuum model , i. e. using the dirac - weyl hamiltonian for electrons in the vicinity of the dirac point @xmath0 . in this paper , we will use an alternative and more general way of calculating the dynamics of a wavepacket in such a system , @xcite based on the split - operator technique , which can be applied for systems under arbitrary external potentials and magnetic fields . the dirac - weyl hamiltonian @xmath78 in eq . ( [ dirachamiltonian ] ) can be separated as @xmath79 , where @xmath80 keeps only the terms which depend on the wave vector @xmath81 , whereas @xmath82 depends only on the real space coordinates @xmath70 and @xmath72 . following the split - operator method , the time evolution operator for the hamiltonian @xmath78 can be approximated as @xmath83 \approx \nonumber \quad\quad\quad\quad \\ t}{2\hbar}h_r\right]\exp\left[-\frac{i \delta t}{\hbar}h_k\right]\exp\left[-\frac{i \delta t}{2\hbar}h_r\right]\end{aligned}\ ] ] with an error of the order of @xmath43 , due to the non - commutativity of the operators involved . we invoke a well - known property of the pauli vectors @xmath84 for any vector @xmath85 , where @xmath86 , and rewrite the exponentials in real and reciprocal space , respectively , in matrix form @xmath87e^{-\frac{i\delta t}{2\hbar}v}\ ] ] @xmath88 where @xmath89 , @xmath90 and @xmath91 , so that the time evolution of a wave packet @xmath92^t\psi(x , y)$ ] can be calculated as a series of matrix multiplications @xmath93 the matrix multiplication by @xmath94 is made in reciprocal space by taking the fourier transform of the functions . in the absence of magnetic fields and external potentials , one has * * m**@xmath95 = * i * and @xmath96 where the matrix multiplication in reciprocal space gives the exact result for the time evolution of the wave packet , since there is no error induced by non - commutativity of operators or matrices in this case . this shows that the split - operator method provides a way to study the dynamics of wavepackets in graphene within the continuum model where , in the presence of magnetic fields and/or external potentials , one can control the accuracy of the results by making @xmath2 smaller , while in their absence , the problem is solved exactly by a simple matrix multiplication for any value of @xmath2 . we shall now discuss the results obtained for a graphene lattice with 2000@xmath973601 atoms , with armchair ( zigzag ) edges in the @xmath71(@xmath73)-direction . the nearest neighbors hopping parameter and the lattice constant of graphene are taken as @xmath98 ev and @xmath99 , respectively . as initial wave packet , we consider a gaussian centered at @xmath100 in real space and @xmath101 in reciprocal space : @xmath102,\ ] ] where @xmath103 is the normalization factor . notice that we have included a pseudo - spinor @xmath104^t$ ] in the initial wave packet , where @xmath105 is the probability of finding the electron in the triangular sub - lattice @xmath65 that composes the graphene hexagonal lattice . one can also rewrite the pseudo - spinor as @xmath106^t$ ] , where the pseudo - spin polarization angle @xmath107 is shown explicitly . the pseudo - spin is a concept normally attributed to the dirac description of graphene . indeed , the pseudo - spin of a wavefunction in the dirac model is related to the expectation values of the pauli matrices @xmath108 , which can involve integrals of the product between wavefunctions for sub - lattices @xmath109 and @xmath20 . such a definition fails for the tb wavefunctions , since in this case they are defined in different points of the lattice , so that any integral that mixes functions of both sub - lattices gives zero . even so , the study of the pseudo - spin related to the initial discrete wave packet helps to understand the wave packet dynamics obtained from the tb model , as we will see further in this section . in this subsection , we will use the tb model to review some of the main properties of the wave packet dynamics in graphene . within the tb model , we consider the initial wave packet as a discrete form of the gaussian distribution in eq . ( [ initial ] ) for the graphene hexagonal lattice , where we multiply the gaussian function by @xmath105 in the sites belonging to the triangular sub - lattice @xmath65 . from eq . ( [ bandstructure ] ) , it is clear that in momentum space , the region of interest is the vicinity of the dirac points @xmath0 and @xmath1 , since the energy corresponding to wave vectors out of this region is very high . fs , for three different values of the dimensionless parameter @xmath110 . ( b ) expectation value of the position and ( c ) velocity in @xmath73-direction as a function of time . the results obtained from the tb ( dirac ) model are presented as curves ( symbols ) , for @xmath111 ( solid , circles ) , 2 ( dashed , triangles ) and 4 ( dotted , squares ) ] in the tb model for two - dimensional crystals , one usually considers the same gaussian distribution for all the sites of the lattice . @xcite this is equivalent as choosing @xmath112 in eq . ( [ initial ] ) . figure [ fig : timeposition](a ) shows the contour plots of the squared modulus of the propagated wavefunctions at @xmath113 fs for these values of @xmath7 , considering an initial wave vector @xmath114 , i.e. in the vicinity of the @xmath0 point labelled as 2 in fig . [ fig : honeycomb](b ) . as shown in ref . , the wave packet dynamics near the dirac cones in graphene does not depend separately on the momentum @xmath115 or on the width @xmath116 , but rather on the dimensionless quantity @xmath110 . this result was obtained from the dirac model for graphene , i.e. considering that even high energies states exhibit linear dispersion . within the tb model we expect that wave packets with the same @xmath110 behave alike only if @xmath115 is not too far from the dirac cone and if @xmath116 is not too small , so that the packet is well localized in energy space . within these conditions , fig . [ fig : timeposition ] shows the time evolution for different values of this dimensionless quantity : @xmath117 and 2 , with @xmath118 , and @xmath119 , with @xmath120 . we observe that the dispersion of the wave packet is stronger for smaller values of @xmath110 , where it becomes distorted into an arc - like shape . for larger @xmath110 , on the other hand , the wave packet keeps its circularly symmetric shape for longer times . ) , with pseudo - spin @xmath121 and momentum @xmath114 as a function of ( a ) the momentum @xmath115 , for widths @xmath122 , 100 and 200 , and ( b ) the width @xmath116 , for momenta @xmath123 @xmath124 , 0.02 @xmath124 and 0.04 @xmath124 . the symbols ( curves ) are obtained from the tb ( dirac ) model . in ( b ) , the results from the dirac model for @xmath125 0.02 @xmath124 ( dashed ) and 0.04 @xmath124 ( solid ) are multiplied by 0.985 and 0.97 , respectively . ] as explained in the previous subsection , in order to obtain the dirac hamiltonian eq . ( [ dirachamiltonian ] ) , one must shift the origin of the wave vectors @xmath57 to one of the six dirac points shown in fig . [ fig : honeycomb](b ) . besides , one must also rotate the @xmath71 and @xmath73 axis by an angle @xmath66 which depends on the @xmath0 or @xmath1 point that is considered as the origin in momentum space . for the @xmath126 point , labelled as 2 in fig . [ fig : honeycomb](b ) , the dirac hamiltonian eq . ( [ dirachamiltonian ] ) is obtained by rotating the axis by 90@xmath127 , with other words , by a transformation of the coordinates as @xmath128 and @xmath129 . the pseudo - spinor @xmath121 represents a wave packet polarized in @xmath70-direction , i.e. @xmath130 and @xmath131 . from the heisenberg picture , we obtain the velocity in the @xmath70-direction for the proposed wave packet as @xmath132 = v_f \sigma_x.\ ] ] performing the appropriate coordinate transformations , the velocity obtained from the tb model for the @xmath73-direction must be consistent with the prediction from the dirac approximation , namely , @xmath133 . this suggests that such a wave packet propagates towards the positive @xmath73-direction , but with non - constant velocity , since @xmath134 does not commute with @xmath78 . the expectation value of the @xmath73 position of the packet @xmath135 is shown as a function of time by the curves in fig . [ fig : timeposition](b ) , for @xmath117 ( solid ) , 2 ( dashed ) and 4 ( dotted ) , where the results obtained by the dirac equation are shown as symbols for comparison . a different linear behavior is already observed for each wave packet at large time , implying that they have different velocities , which is kind of counter - intuitive , since low - energy electrons in graphene are expected to propagate always with the same fermi velocity @xmath136 . figure [ fig : timeposition](c ) shows the velocity @xmath137 , calculated by taking the derivative of the tb results for @xmath135 with respect to time , which exhibits clear oscillations that are damped as time evolves , converging to a final value @xmath138 that depends on the initial wave packet s width @xmath116 and wave vector @xmath115 . the velocities obtained by the dirac model are shown by symbols , where the same qualitative behavior is observed as obtained from the tb model , though for higher wave packet momentum and width , a small quantitative difference is observed , which is a consequence of the different energy - momentum dispersion . fs , for three initial configurations of pseudo - spin @xmath139^t$ ] and momentum @xmath140 : 1 ) @xmath141^t$ ] , @xmath142 and @xmath119 ; 2 ) @xmath143^t$ ] , @xmath142 and @xmath119 ; and 3 ) @xmath143^t$ ] , @xmath144 , @xmath145 . ( b ) expectation value of @xmath71 obtained by the tb model for the initial wavepackets 1 ( solid ) , 2 ( dashed ) and 3 ( dotted ) as a function of time . the results obtained by the dirac model , after the appropriate coordinates rotation ( see text ) , are shown as circles , triangles and squares , respectively . the inset shows the trajectory of the wavepacket obtained from the tb model for the initial wavepacket 3 . ] the oscillatory behavior of the velocity is a manifestation of the zitterbewegung , i.e. a trembling motion of the wave packet due to the interference between positive and negative energy states that makes up the initial gaussian wave packet . @xcite this effect is well - known for relativistic particles , which are described by the dirac equation , and is also observed for electrons in graphene in the vicinity of the @xmath0 and @xmath1 points , since they can be described as massless quasi - particles by the dirac equation as well . the velocity wiggles with shorter period and smaller amplitudes for larger values of @xmath110 . the convergence of the velocities demonstrates that the zitterbewegung is not a permanent , but a transient effect . @xcite figure [ fig : velocityvsdandk ] shows the converged velocity @xmath146 as a function of ( a ) the momentum @xmath147 and ( b ) the width @xmath116 of the gaussian wave packet . the tb results ( symbols ) are compared to those calculated from eq . ( 31 ) in ref . ( curves ) , which was obtained analytically from the dirac approximation in the @xmath148 limit and is repeated here just for completeness : @xmath149 within the dirac model , one can observe that increasing @xmath116 or @xmath115 in eq . ( [ eqcompleteness ] ) , the final velocity increases monotonically and approaches @xmath136 , which is reasonable , since a wider packet in real space leads to a narrower distribution in @xmath67-space , whereas a higher value of the wave vector makes the center of the packet lay far from @xmath150 . in both cases , the interference with negative energy states is reduced and , consequently , zitterbewegung effects are less significant . however , since the analytical formula eq . ( [ eqcompleteness ] ) does not take into account any effect such as the curvature of the energy bands for higher energy states or trigonal warping effects , this formula is not expected to give accurate results for larger @xmath147 . indeed , fig . [ fig : velocityvsdandk](a ) shows that a very good agreement between tb and dirac results can be observed only for small values of @xmath147 , whereas for large @xmath147 , the final velocities obtained from the tb model are lower than those obtained from the dirac model and do not increase monotonically , but decreases slowly for very large @xmath147 , as a consequence of the curvature of the energy bands . on the other hand , in fig . [ fig : velocityvsdandk](b ) we observe that varying the wave packet width for a fixed momentum , good _ qualitative _ agreement with the dirac model is obtained for almost any value of @xmath116 . the curves for larger values of @xmath147 ( solid and dashed ) are just _ quantitatively _ different from those obtained by the tb model , and they are comparable to the tb results after multiplication by a factor 0.985 ( 0.970 ) for @xmath151 @xmath124 ( 0.04 @xmath124 ) . worse qualitative agreement between tb and dirac results in this case is observed only for very small @xmath116 , where the gaussian width in energy space , given by @xmath152 , incorporates higher energy values , leading to deviations in @xmath146 obtained from the tb model as compared to those from the dirac model . and pseudo - spinor @xmath153^t$ ] ( solid ) and @xmath143^t$ ] ( circles ) at the dirac point @xmath126 ( point 2 in fig . [ fig : honeycomb](b ) ) , and @xmath153^t$ ] ( triangles ) at @xmath154 ( point 1 in fig . [ fig : honeycomb](b ) ) . the @xmath71 and @xmath73 components of the velocity in the latter case are shown as dashed and dotted curves , respectively . ] in fig . [ fig : timeposition10](a ) we show contour plots of the squared modulus of the wave function at @xmath113 fs for three different choices of wave vectors @xmath155 and initial pseudo - spinors : 1 ) @xmath141^t$ ] , with @xmath142 and @xmath119 , 2 ) @xmath143^t$ ] , with @xmath142 and @xmath119 , and 3 ) @xmath143^t$ ] , with @xmath156 and @xmath145 . the curves ( symbols ) in fig . [ fig : timeposition10](b ) show the expectation value @xmath157 for each case obtained by the tb ( dirac ) model . in case 1 ( solid , circles ) , the pseudo - spinor points in the @xmath158-direction , so that @xmath159 and , consequently , the velocity for both in - plane directions must be zero . indeed , the wave packet splits into two equal parts that propagate in opposite @xmath73 directions , leading to @xmath160 . in the @xmath71-direction , although there is still a small zitterbewegung , @xmath157 rapidly converges to a constant , leading to @xmath161 . in case 2 ( dashed , triangles ) , the pseudo - spinor points in the @xmath72-direction , but the momentum of the wave packet in this direction is zero , so that the packet splits in the @xmath73-direction , since @xmath162 , but drifts slowly in the @xmath163 direction ( or , equivalently , @xmath72 ) . in case 3 ( dotted , squares ) , both the pseudo - spin and the momentum are in the @xmath72-direction , so that the wave packet propagates in this direction without any splitting . this situation is comparable to the one in fig . [ fig : timeposition](a ) , since in both cases the pseudo - spin and momentum are in the same direction and , as a consequence , the wave packet propagates in this direction practically preserving its circular symmetry . however , in the case 3 , the packet still presents a very weak oscillation in the @xmath73-direction , and also drifts very slowly in this direction , as one can see from the trajectory of the packet in the @xmath164 plane for this case , shown in the inset of fig . [ fig : timeposition10](b ) . this oscillation and drift are related to the contributions of higher energy states in the wave packet : a wave packet centered around @xmath165 and @xmath166 , as in fig . [ fig : timeposition](a ) , has a symmetric distribution of momenta in @xmath71-direction even for higher energies and , consequently , there is no additional oscillation in this direction . on the other hand , a packet centered around @xmath167 and @xmath168 , as in fig . [ fig : timeposition10](b ) , does not have a symmetric distribution of momenta in the @xmath73-direction due to the trigonal warping for higher energies and , consequently , some oscillations are observed in this direction . as the standard dirac hamiltonian @xmath78 for graphene does not take trigonal warping into account , this effect is not observed in the dirac model . in the numerical work of thaller , @xcite as well as in the analytical work of maksimova , @xcite it is demonstrated within the dirac model that even when @xmath169 , wave packet motion is still observed due to zitterbewegung effects . the velocities of the wave packet obtained from our tb model of graphene for wave packet momenta exactly at the @xmath1 and @xmath0 , i.e. points 1 and 2 in fig . [ fig : honeycomb](a ) , respectively , are shown in fig . [ fig : velkkprime ] . the velocities exhibit a damped oscillation with the same time - dependent modulus for any pseudo - spin and dirac point , though they point in different directions : for @xmath153^t$ ] ( solid ) and @xmath143^t$ ] ( circles ) in @xmath0 , the velocity is in the @xmath73 and @xmath163 directions of the lattice , respectively , which are exactly the directions of polarization of these pseudo - spinors after the @xmath170 rotation required by the @xmath0 cone 2 . in the @xmath1 cone 1 , the rotation angle is @xmath171 and , accordingly , the velocity points in this direction , as one can see by the decomposition of the velocity in the components @xmath172 ( dashed ) and @xmath173 ( dotted ) , which obey exactly the relations @xmath174 and @xmath175 . notice that the velocities converge exactly to @xmath176 , a value that can also be obtained analytically by making @xmath177 in eq . ( [ eqcompleteness ] ) . henceforth , we will consider initial wave vectors @xmath140 around the dirac points 2 and 5 of fig . [ fig : honeycomb](b ) , namely @xmath178 respectively . this choice is very convenient , since the rotation angles for these points are @xmath170 and @xmath179 , respectively , so that the pseudo - spinor @xmath153^t$ ] points to the @xmath73 ( -@xmath73 ) direction in the former ( latter ) case . hence , with this pseudo - spinor , wave packets in @xmath0 ( @xmath1 ) will propagate with positive ( negative ) velocity in the vertical zig - zag direction . recently,@xcite it was shown theoretically that bending a graphene sheet into an arc of a circle produces a strong and almost uniform pseudo - magnetic field profile . figure [ fig : sketcharc](a ) illustrates such a strained system , where the rectangular graphene sample of width @xmath180 and height @xmath181 is bent into an arc of a circle with inner radius @xmath182 . as the ( pseudo ) magnetic field points in the same direction ( opposite directions ) at each @xmath0 and @xmath1 points , @xcite the combination of both external and strain - induced magnetic field effects provides a valley - dependent magnetic field . if one applies the appropriate external magnetic field for some configuration of the strained graphene , one can obtain an almost perfect suppression of the effective magnetic field at one of the dirac cones , while the effective field in the other cone is enhanced . this leads to a complicated system to be studied within the dirac approximation , since one has two completely different systems for the @xmath0 and @xmath1 valleys . namely , landau levels would be present only around one of the cones ( though one can not expect a perfect landau level spectrum , since the strain - induced magnetic field is not perfectly uniform in space ) , whereas in the other cone , the usual continuum spectrum would be observed . this motivated us to analyze the trajectories of a wave packet in such a system within the tb model , where we do not need to include the pseudo - magnetic fields artificially in the dirac cones , since they appear naturally when we consider the effect of the strain - induced changes of the inter - site distances on the hopping energies , as explained in the previous section . in this subsection , we investigate the dynamics of a wave packet with width @xmath120 and initial wave vector @xmath142 and @xmath183 @xmath124 around the dirac points @xmath0 and @xmath1 of eq . ( [ diracpointskkl ] ) in the presence of external and strain - induced magnetic field barrier steps . as in the @xmath1 valley the pseudo - spinor @xmath153^t$ ] is polarized in the negative @xmath73-direction of the graphene lattice , we choose @xmath184^t$ ] for this case , so that a wave packet in this valley will also propagate in the positive @xmath73-direction . in order to obtain a pseudo - magnetic field barrier step , we consider that the graphene layer is strained only in the @xmath185 region , as sketched in fig . [ fig : sketcharc](b ) . we also consider an external magnetic field @xmath186 , where @xmath187 is the heaviside step function , which leads to a magnetic barrier step for @xmath185 , described by the vector potential @xmath188 . in order to avoid effects due to zitterbewegung in the ( pseudo ) magnetic field region , the wave packet starts at the position @xmath189 , @xmath190 , so that it can travel for some time in the magnetic field - free region @xmath191 until its velocity converge to a time independent value . plane , obtained by the tb method for such a system , for initial momentum @xmath183 @xmath124 around @xmath0 ( symbols ) and @xmath1 ( curves ) points , for ( a ) non - strained graphene with magnetic barrier height @xmath192 t ( solid , circles ) , 7 t ( dashed , triangles ) and 10 t ( dotted , squares ) , and for ( b ) a graphene sheet bent into an arc of circle with radius @xmath193 @xmath194 m ( solid , circles ) , 0.8 @xmath194 m ( dashed , triangles ) and 0.6 @xmath194 m ( dotted , squares ) , considering @xmath195 t. in ( b ) , symbols and curves coincide for each value of @xmath182 . ] the influence of the external and strain - induced magnetic field barriers on the trajectories of the wave packet are analyzed separately in fig . [ fig : filterscheme ] , which shows the trajectory of the centroid of the wave packets in @xmath0 ( symbols ) and @xmath1 ( curves ) points , calculated as @xmath196 , ( a ) in a non - strained graphene sheet with magnetic field barriers @xmath192 t ( solid , circles ) , 7 t ( dashed , triangles ) and 10 t ( dotted , squares ) and ( b ) in a strained graphene sheet with radius @xmath193 @xmath194 m ( solid , circles ) , 0.8 @xmath194 m ( dashed , triangles ) and 0.6 @xmath194 m ( dotted , squares ) . all the trajectories form semi - circles in the @xmath185 region , which is due to the lorentz force produced by the ( pseudo ) magnetic field . as the external magnetic field ( radius of the strained region ) increases ( decreases ) , the radii of these semi - circular trajectories are reduced , since a higher ( pseudo ) magnetic field produces a stronger modulus of the lorentz force . notice that the radii of trajectories in the external and pseudo - magnetic fields cases are comparable , which means that for radii @xmath193 @xmath194 m - 0.6 @xmath194 m of the strained graphene , the generated pseudo - magnetic field is also within @xmath197 5 t and 10 t. indeed , the strain induced pseudo - magnetic field distribution for the bend graphene ribbon is given by @xcite @xmath198 \nonumber \\ \times\left[1 - \frac{r + y}{l } \arcsin\left(\frac{l}{2r}\right)\right],\hspace{0.5 cm } \label{strainguinea}\end{aligned}\ ] ] where @xmath199 and @xmath200 is a dimensionless constant which depends on the details of the atomic displacements . @xcite considering @xmath201 in eq . ( [ strainguinea ] ) the pseudo - magnetic field can be approximated as @xmath202 . using the value @xmath203 t estimated numerically in ref . , one obtains pseudo - magnetic fields within @xmath204 t - 7.5 t for @xmath193 @xmath194 m - 0.6 @xmath194 m , which are of the same order of magnitude as the external magnetic fields that we considered . for the external magnetic field barrier , the trajectories of wave packets in @xmath0 and @xmath1 points form circles in opposite directions , as shown in fig . [ fig : filterscheme](a ) , which is reasonable , since these packets have opposite momentum , which causes a sign change in the lorentz force . conversely , considering the strain - induced magnetic barrier illustrated in fig . [ fig : sketcharc](b ) , the trajectories of wave packets in @xmath0 and @xmath1 curve in the same direction , since , although their momenta have opposite signs , the pseudo - magnetic fields also point in opposite directions at each dirac cone @xmath0 and @xmath1 . let us consider the strained sample in fig . [ fig : sketcharc](b ) with @xmath193 @xmath194 m . by comparing the radius of the semi - circular trajectory of the wave packet in such a system with those obtained for different intensities of the external magnetic field barrier , one obtains the strain - induced magnetic field for this value of @xmath182 as @xmath205 t. figure [ fig : filterresult](a ) shows the trajectories in the @xmath164 plane of the centroid of the wave packets in a system where we combine a @xmath193 @xmath194 m strain for @xmath185 with an external magnetic field barrier @xmath195 t ( solid , open ) and 4.9 t ( dashed , full ) , for wave packets in the @xmath0 ( symbols ) and @xmath1 ( curves ) dirac points . in the absence of the external magnetic field , both the @xmath0 and @xmath1 packets exhibit the same semi - circular trajectory , as discussed earlier . however , when we combine the effect of the strain - induced and external magnetic field barriers , the wave packet in @xmath1 undergoes a stronger lorentz force and is readily reflected , whereas the one in the @xmath0 point performs a practically straight trajectory , as if this packet is not influenced by any lorentz force . this is a consequence of the fact that combining the effects of a pseudo - magnetic field produced by a @xmath206 m strain and a @xmath207 t external magnetic field produces a stronger magnetic field in the @xmath1 point , while in the @xmath0 point these fields equilibrate , producing a practically magnetic field - free region for particles in this cone . in this situation , the system works as a valley filter , where only wave packets in the @xmath0 dirac cone are allowed to pass through the strained region , whereas the wave packets in @xmath1 are reflected . the results for the wave packet in @xmath0 for two other values of the external magnetic field are shown as thin solid lines , showing that within a range of @xmath208 t around @xmath209 t , which is a reasonable range for magnetic field intensities in experiments , only a weak lorentz force is observed and the valley filter works fine . plane for wave packets with initial momentum @xmath183 @xmath124 around @xmath0 ( symbols ) and @xmath1 ( curves ) points , considering a graphene sheet bent into an arc of circle with radius @xmath193 @xmath194 m and an external magnetic field @xmath195 t ( open , solid ) and 4.9 t ( full , dashed ) . the thin solid curves show the results for two other magnetic field intensities for the @xmath0 packet . ( b ) probability of finding the particle in @xmath185 as a function of time , for wave packets with the same configurations as in ( a ) . ] the results of fig . [ fig : filterresult ] are obtained for both external and pseudo - magnetic field barriers starting at the same position @xmath210 . it is straightforward to verify that if there is a mismatch between the starting points of the strained and external field regions , some deviations will occur in the trajectories of the wave packets but , provided the length of the mismatch is much smaller than the magnetic length , the filtering effect is still stable . as an example , a 30 mismatch between the external and pseudo - magnetic field barriers in the system analyzed in fig . [ fig : filterresult ] would produce a @xmath197 5@xmath211 deviation in the otherwise vertical trajectory of the wave packet in @xmath0 , whereas the wave packet in @xmath1 is still readily reflected by the combination of magnetic fields in the filter region . the probability @xmath212 of finding the particle in the strained @xmath185 region , calculated as @xmath213 where @xmath214 represents the lines of atomic sites with @xmath185 , is shown as a function of time in fig . [ fig : filterresult](b ) . in the @xmath195 t case , both wave packets in @xmath0 ( open circles ) and @xmath1 ( solid ) stay in the strained @xmath185 region until @xmath215 fs , when they turn back into the @xmath191 region , reflected by the lorentz force induced by the strain . however , for @xmath209 t , @xmath212 already approaches zero at @xmath216 fs for the packet in @xmath1 ( dashed ) , whereas for @xmath0 ( full circles ) , it remains close to 1 even for large @xmath3 . the efficiency of the proposed valley filter is confirmed by fig . [ fig : filterefficiency ] , where we present @xmath212 as a function of time for initial wave packets given by a combination of gaussians around the @xmath0 and @xmath1 points in eq . ( [ diracpointskkl ] ) : @xmath217 where @xmath218 is the gaussian wave packet in eq . ( [ initial ] ) with momentum @xmath219 around the @xmath220 dirac point . the results are presented for three different values of @xmath221 , where one can easily see that the probability of finding the packet in the strained region exhibits a peak at @xmath222 fs but , as the @xmath1 part of the packet is reflected by the magnetic barrier , this probability decays , reaching exactly @xmath223 for large @xmath3 . such a system proves to be a perfect valley filter , which is able to reflect all the components of the incoming packet that are in the @xmath1 point and transmit a wave packet that is composed only of particles with momentum in the vicinity of @xmath0 . we point out that when a wave packet reaches the edges of a graphene nanoribbon , it can be scattered to a different dirac cone . @xcite consequently , the efficiency of the valley filter could be compromised if one considers a thin nanoribbon , so that the filtered wave packet could still reach its boundaries and scatter back to the other valley . in order to avoid such an effect , we have considered a wide nanoribbon , so that for the time intervals we study in this work , boundary effects are not significant . in a previous paper , rusin and zawadzki @xcite used the dirac hamiltonian for graphene to show that the zitterbewegung , which is transient for @xmath20 = 0 , as discussed earlier , is permanent for @xmath224 . furthermore , the authors showed that for a gaussian wave packet , the time evolution of the average value of the current is different in @xmath71 and @xmath73 directions , which they explain as due to the fact that the dirac hamiltonian is not symmetric in the momenta @xmath76 and @xmath77 . although the same authors say in their subsequent paper @xcite that this effect is unphysical , because it violates the rotational symmetry of the @xmath164 graphene plane , we believe this result is still physical : one must remember that the dirac model of graphene comes from the tight - binding approach for this system , which describes a honeycomb lattice of atoms that is not symmetric in the @xmath164 plane by definition , exhibiting c@xmath74 symmetry , as mentioned in previous section . for each @xmath0 and @xmath1 point , the coordinates @xmath70 and @xmath72 in the dirac hamiltonian represent different directions in the real honeycomb lattice of graphene , where for an infinite sample the @xmath70 ( @xmath72 ) coordinate in the dirac equation is related to one of the zigzag ( armchair ) directions of the real sample . in this subsection , we use our tb model of graphene to extend the previous study of rusin and zawadzki @xcite to different situations . we now study the dynamics of a wave packet with width @xmath120 , pseudo - spinor @xmath225 and @xmath226 and initial wave vector @xmath227 , i.e. exactly at one of the dirac points @xmath0 and @xmath1 in eq . ( [ diracpointskkl ] ) , in the presence of an uniform applied external magnetic field @xmath228 , instead of the magnetic field barrier step considered in the previous subsection . we also consider a pseudo - magnetic field obtained by bending the whole rectangular graphene sample into an arc of a circle with radius @xmath182 , as illustrated in fig . [ fig : sketcharc](a ) . the radius of the circle is considered as @xmath193 @xmath194 m , which produces a @xmath205 t pseudo - magnetic field , as demonstrated in the previous subsection . accordingly , we consider the external uniform magnetic field as @xmath209 t. a few experimental techniques have been suggested in the literature for the observation of zitterbewegung . @xcite an interesting one @xcite is based on the fact that the wave packet @xmath229 exhibits an electric dipole moment @xmath230 and , consequently , the zitterbewegung yields an oscillation of this dipole moment , which is a source of electromagnetic radiation , described classically @xcite by the equation @xmath231 where @xmath232 and @xmath233 are the distance and angle of observation , respectively , and @xmath234 is the vacuum permittivity . figure [ fig : efield ] shows the effects of external and strain - induced magnetic fields on the electric field radiation produced by the zitterbewegung , written in units of @xmath235 . only weak oscillations are expected in the armchair ( @xmath71 ) direction , since the pseudo - spin @xmath153^t$ ] points in the @xmath236-direction in the dirac model which , as mentioned earlier , is related to the zigzag ( @xmath73 ) direction of the honeycomb lattice . indeed , in fig . [ fig : efield](a - d ) , the @xmath71-component of the electric field ( dashed ) is always close to zero . in fig . [ fig : efield](a ) , we present the results in the absence of strain and magnetic fields , for comparison . in this case , the oscillations on the electric field are suppressed for larger time , which is due to the transient character of the zitterbewegung . besides , the results for @xmath237 ( solid ) in the @xmath0 ( thin curves ) and @xmath1 ( thick curves ) points have opposite signs , since for these points the axis of the dirac cones are rotated by angles which differ by @xmath238 difference . in fig . [ fig : efield](b ) , the uniformly applied magnetic field @xmath209 t in an unstrained sample leads to persistent oscillations in @xmath237 , which is related to the discrete landau level spectrum created by this field . each landau level that is populated by the gaussian distribution contributes with a different frequency . @xcite figure [ fig : efield](c ) shows that such a persistent behavior is also obtained by bending the graphene sheet into an arc of circle with radius @xmath206 m , which produces a pseudo - magnetic field of the same order of magnitude . notice that the amplitude of oscillations in this case is four times smaller than those found in fig . [ fig : efield](b ) for the unstrained sample under an external magnetic field . in fact , these two cases are not expected to produce the same zitterbewegung , because , although both samples exhibit approximately the same magnitude of magnetic field , the strained sample has not only the pseudo - magnetic field , but also a different distribution of atomic sites . thus , in the strained case , there is an additional change in the direction of the pseudo - spin polarization as the wave packet drifts , due to the lattice distortion itself . as we have demonstrated in sec . iii a , the zitterbewegung strongly depends on the pseudo - spin polarization and hence , the different interplay between the strained atomic sites and the initial pseudo - spin polarization produces a different zitterbewegung for the strained case , as compared to the one of the unstrained sample under an external field . in fig . [ fig : efield](d ) , we combine the effects of the @xmath239 @xmath194 m strain and @xmath209 t external field to produce a system where the magnetic field is practically zero in the @xmath0 point , but is non - zero in the @xmath1 point , so that only the packet in the @xmath1 point exhibits persistent oscillations . for the @xmath0 point , the external field compensates only the effect of the pseudo - magnetic field , namely , the persistent zitterbewegung , but it does not remove the effect of the lattice distortion . as a result , comparing the results for @xmath0 ( thin curves ) in figs . [ fig : efield](a ) and ( d ) , one observes that the oscillations are transient in both cases , since there is no effective magnetic field , but they still exhibit a different oscillation profile at small @xmath3 , due to the lattice distortion in the latter case , which is absent in the former . we presented a study of the dynamics of gaussian wave packets in graphene under external and strain - induced magnetic fields , where the latter is obtained by bending the graphene sheet into an arc of a circle . the dependence of the zitterbewegung on the initial pseudo - spin of the wave packet is investigated , and the results obtained by means of the tight - binding model and the dirac equation are compared . we demonstrate that the combination of the pseudo - magnetic field , induced by bending the graphene sheet , along with an external magnetic field with appropriate strength can be used as an efficient valley filter . an incoming wave packet composed of momenta around the @xmath0 and @xmath1 dirac points is scattered such that all its components in one of the dirac cones undergoes a strong lorentz force and are readily reflected , while the components in the other cone are allowed to pass through the device with only small distortions in their trajectory , due to the very weak residual lorentz force . our results also show that in the absence of external or strain - induced magnetic fields , the zitterbewegung is a transient effect , whereas in the presence of any of these fields , the oscillations persist in time . in a strained sample under an external magnetic field with the appropriate strength , the effective magnetic field in one of the dirac cones is enhanced , whereas in the other cone it is practically cancelled . in this situation , a permanent zitterbewegung is observed only for wave packets in one of the dirac cones . the wave packet oscillations produce electric field radiation , which can be detected experimentally . finally , we believe the present work contributes to a better understanding of the relation between the results obtained from tb and dirac approaches for graphene and those to be observed in future experiments on strained graphene - based structures . this work was financially supported by cnpq , under contract nanobioestruturas 555183/2005 - 0 , pronex / cnpq / funcap , capes , the bilateral programme between flanders and brazil , the belgian science policy ( iap ) and the flemish science foundation ( fwo - vl ) . k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , y. zhang , s. v. dubonos , i. v. grigorieva , and a. a. firsov , science * 306 * , 666 ( 2004 ) . a. h. castro neto , f. guinea , n. m. r. peres , k. s. novoselov , and a. k. geim , rev . phys . * 81 * , 109 ( 2009 ) . b. thaller , arxiv : quant - ph/0409079v1 ( 2004 ) . w. zawadzki , t. m. rusin , phys . a * 374 * , 3533 ( 2010 ) . g. dvid and j. cserti , phys . b * 81 * , 121417(r ) ( 2010 ) . e. schrdinger , sitzungsber . * 24 * , 418 ( 1930 ) . w. zawadzki , phys . b * 72 * , 085217 ( 2005 ) . j. schliemann , d. loss , and r. m. westervelt , phys . 94 * , 206801 ( 2005 ) . t. m. rusin and w. zawadzki , phys . b * 80 * , 045416 ( 2009 ) . y .- wang , z. yang , and s .- j . xiong , europhys . lett . * 89 * , 17007 ( 2010 ) . r. gerritsma , g. kirchmair , f. zhringer , e. solano , r. blatt , and c. f. roos , nature ( london ) * 463 * , 68 ( 2010 ) . v. m. pereira and a. h. castro neto , phys . * 103 * , 046801 ( 2009 ) . seon - myeong choi , seung - hoon jhi , and young - woo son , phys . b * 81 * , 081407(r ) ( 2010 ) . f. m. d. pellegrino , g. g. n. angilella , and r. pucci , phys . rev . b * 81 * , 035411 ( 2010 ) . o. bahat - treidel , o. peleg , m. grobman , n. shapira , m. segev , and t. pereg - barnea , phys . lett . * 104 * , 063901 ( 2010 ) . t. fujita , m. b. a. jalil , and s. g. tan , appl . 97 * , 043508 ( 2010 ) . z. wu , f. zhai , f. m. peeters , h. q. xu and k. chang , arxiv:1008.4858v3 ( 2010 ) . e. cadelano , s. giordano , and l. colombo , phys . b * 81 * , 144105 ( 2010 ) . g. cocco , e. cadelano , and l. colombo , phys . b * 81 * , 241412 ( 2010 ) . y. lu and j. guo , nano res . * 3 * , 189 ( 2010 ) . m. a. h. vozmediano , m. i. katsnelson , and f. guinea , phys . rep . * 496 * , 109 ( 2010 ) . f. guinea , m. i. katsnelson , and a. k. geim , nat . phys . * 6 * , 30 ( 2010 ) . t. low and f. guinea , nano lett . * 10 * , 3551 ( 2010 ) . f. guinea , a. k. geim , m. i. katsnelson , and k. s. novoselov , phys . b * 81 * , 035408 ( 2010 ) . n. levy , s. a. burke , k. l. meaker , m. panlasigui , a. zettl , f. guinea , a. h. castro neto , m. f. crommie , science * 329 * , 544 ( 2010 ) . g. m. maksimova , v. ya . demikhovskii , and e. v. frolova , phys . b * 78 * , 235321 ( 2008 ) . t. m. rusin and w. zawadzki , phys . rev . b * 78 * , 125419 ( 2008 ) . e. romera and f. de los santos , phys . b * 80 * , 165416 ( 2009 ) . v. krueckl and t. kramer , new j. phys . * 11 * , 093010 ( 2009 ) . a. rycerz , j. tworzydo , and c. w. j. beenakker , nat . * 3 * , 172 ( 2007 ) . a. weisse , g. wellein , a. alvermann , and h. fehske , rev . 78 * , 275 ( 2006 ) . s. mahapatra and n. sathyamurthy , j. chem . soc . * 93 * , 773 ( 1997 ) . r. kosloff , j. phys . chem . * 92 * , 2087 ( 1988 ) . a. chaves , g. a. farias , f. m. peeters , and b. szafran , phys . b * 80 * , 125331 ( 2009 ) . d. a. bahamon , a. l. c. pereira , and p. a. schulz , phys . b * 79 * , 125414 ( 2009 ) . m. governale and c. ungarelli , phys . rev . b * 58 * , 7816 ( 1998 ) . v. m. pereira , a. h. castro neto , and n. m. r. peres , phys . rev . b * 80 * , 045401 ( 2009 ) . n. watanabe and m. tsukada , phys . e * 62 * , 2914 ( 2000 ) . j. m. pereira , a. chaves , g. a. farias , and f. m. peeters , sem . sci . tech . * 25 * , 033002 ( 2010 ) . h. n. nazareno , p. e. de brito , and e. s. rodrigues , phys . rev . b * 76 * , 125405 ( 2007 ) . chen , w. g. cullen , c. jang , m. s. fuhrer , and e. d. williams , phys . lett . * 102 * , 236805 ( 2009 ) . j. d. jackson , _ classical electrodynamics _ ( wiley , new york , 1975 ) .

the chemical composition of extremely metal - poor stars is expected to reflect the yields from a quite small number of nucleosynthesis processes . recent abundance analyses for extremely metal - poor stars have provided quite valuable information on the origin of the elements , ( in particular when combined with galactic chemical evolution studies , e.g. , * ? ? ? * ; * ? ? ? * ) , and the individual nucleosynthesis processes involved . the rapid neutron capture process ( r - process ) is known to be responsible for about half of the abundances of elements heavier than the iron - group in solar system material . although observational data have been rapidly increasing in the past decade , the astrophysical site of the r - process is still unclear . previous nucleosynthesis studies suggest several possibilities , e.g. , neutrino - driven winds @xcite or prompt explosions @xcite of core - collapse ( type ii / ibc ) supernovae , neutron star mergers @xcite , jets from gamma - ray burst accretion discs @xcite . all the scenarios proposed above involve , however , severe problems that remain to be solved , and no consensus has yet been achieved . models of the r - process nucleosynthesis are usually examined by comparison with the abundance pattern of the r - process component in solar - system material . recent measurements for abundances of neutron - capture elements in very metal - poor stars have been providing useful constraints on these models . @xcite have studied the chemical abundances of the extremely metal - poor star cs 22892052 , the first example of a small but growing class of metal - poor stars that exhibit very large excesses of r - process elements relative to iron ( [ r - process / fe ] @xmath0 ) . an important result of their work is that the relative abundance pattern of the neutron - capture elements from the 2nd to the 3rd peak ( 56 @xmath1 _ z _ @xmath1 76 ) in this star is identical , within observational errors , to that of the ( inferred ) solar system r - process component . this phenomenon is sometimes referred to as `` universality '' of the r - process , having a large impact on the studies of the nature of the r - process , and its astrophysical site . however , the abundance patterns of light neutron - capture elements ( 38 @xmath1 z @xmath1 48 ) in r - process enhanced stars exhibit clear deviations from that of the solar system r - process component @xcite . this suggests the existence of another r - process site which has contributed to the light r - process elements in solar - system material . this process is sometimes called as `` weak r - process '' , while the process that is responsible for heavy neutron - capture elements ( ba @xmath2 u ) is referred to `` main r - process '' . following the previous studies ( e.g. , * ? ? ? * ; * ? ? ? * ) , @xcite and @xcite have investigated the correlation of sr and ba abundances in very metal - poor stars ( [ fe / h ] @xmath3 ) in some detail . they showed that the dispersion of the sr abundances clearly decreases with increasing ba abundance . this correlation indicates the existence of ( at least ) two processes : one enriches both heavy and light neutron - capture elements , while the other yields light neutron - capture elements with only small production of heavy ones . @xcite discuss the contribution of a primary process which has produced light neutron - capture elements in our galaxy . the abundance patterns produced by the main r - process have been studied previously based on high resolution spectroscopy of metal - poor stars ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) as mentioned above . by way of contrast , the abundance pattern of neutron - capture elements produced by the weak r - process is still unclear . since the abundances of heavy neutron - capture elements yielded by this process are very low , it is quite difficult to accurately measure their abundances in metal - poor stars . in addition , most spectral lines of light neutron - capture elements , in particular those with @xmath4 exist in the near - uv region , in which atmospheric extinction makes the observation with ground - based telescopes difficult . we have been investigating light neutron - capture elements in very metal - deficient stars from near - uv spectroscopy with the 8.2 m subaru telescope . ishimaru et al . ( 2006 , in preparation ) studied the abundances of pd and ag , which exist between the first and second abundance peaks produced by neutron - capture processes , for several metal - poor stars . one of the targets is hd 122563 , a well - studied bright metal - poor ( [ fe / h ] @xmath5 ) giant . this object has very low abundances of heavy neutron - capture elements ( e.g. [ ba / fe ] @xmath6 * ? ? ? * ) , while light neutron - capture elements show large excesses relative to heavy ones ( e.g. [ sr / ba ] @xmath7 ) . therefore , the abundance pattern of neutron - capture elements in this star might well represent the yields of the process producing light neutron - capture elements at low metallicity . in this paper we report the abundance pattern of light and heavy neutron - capture elements in hd 122563 . in 2 we present our near - uv spectroscopy with subaru / hds , while we describe in detail the abundance analysis of neutron - capture elements in 3 . our results are compared with the previous studies in 4 . we discuss the derived abundances of hd 122563 , and the source of neutron - capture elements in 5 . finally , our conclusions are presented in 6 . high dispersion spectroscopy of hd 122563 was carried out with the subaru telescope high dispersion spectrograph ( hds : * ? ? ? * ) in 30 april , 2004 . our spectrum covers the wavelength range from 3070 to 4780 with a resolving power of @xmath8 = 90,000 . the total exposure time is 5400 seconds ( one 600 s exposure plus four 1200 s ones ) . the reduction was carried out in a standard manner using the iraf echelle package . the signal to noise ratio ( s / n ) of the spectrum ( per 0.9 km s@xmath9 pixel ) estimated from the photon counts is 140 at 3100 ; 480 at 3500 ; 860 at 3900 ; 1080 at 4200 and 1340 at 4500 . in order to measure ba lines at 5853 and 6141 , a spectrum obtained during the hds commissioning covering 5090 6410 was used . the s / n ratio of this spectrum is 350 at 6000 . we adopted the line list used in previous studies for line identifications and analyses of neutron - capture elements ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? in addition , the recent measurements of transition probabilities and hyperfine splitting reported for some elements ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) are incorporated . the line list used in the analysis is given in table [ tab : table1 - 4 ] . + equivalent widths of clean , isolated lines are measured by fitting gaussian profile . the results are given in table [ tab : table1 - 4 ] . for our quantitative abundance measurements , we used the analysis program sptool developed by y. takeda , based on kurucz s atlas9/width9 @xcite . sptool calculates synthetic spectra and equivalent widths of lines on the basis of the given model atmosphere , line data , and chemical composition , under the assumption of lte . we adopted the model atmosphere parameters ( effective temperature : @xmath10 , gravity : @xmath11 , micro - turbulent velocity : @xmath12 , and metallicity : [ fe / h ] ) derived by @xcite : @xmath10 = 4570 k , @xmath13 = 1.1 , @xmath12 = 2.2 km s@xmath9 , and [ fe / h ] @xmath14 . the estimated uncertainties are @xmath15 = 100 k , @xmath16 = 0.3 dex , @xmath17 = 0.5 km s@xmath9 , and @xmath18[fe / h ] = 0.19 dex ( honda et al . 2004 ) . the abundance analyses were attempted for 24 neutron - capture elements from sr to th . the effects of hyperfine and isotopic splitting are taken into account in the analysis of ba , la , eu , and yb ( see references in table 1 ) . the derived abundances are given in table [ tab : table2 ] . the abundances determined from individual lines are given in table [ tab : table1 - 4 ] . we used the solar system abundances obtained by @xcite to derive [ x / fe ] values . the size of the random errors are estimated from the standard deviation ( 1@xmath19 ) of the abundances derived from individual lines for elements that have three or more lines available for the abundance analysis . for the abundances of elements based only on one or two lines , we employ the mean of the random errors estimated from those elements with three or more lines available . the uncertainties of model atmosphere parameters result in systematic errors in the abundances of neutron - capture elements . the effects of these uncertainties on the abundance measurements are given in table [ tab : error ] . the behavior of the errors is slightly different for abundances derived from neutral species and those from ionized ones . this should be borne in mind in the discussion on the abundances of light neutron - capture elements , some of which were determined using lines of neutral species . by contrast , such differences do not significantly affect our discussion of the abundance ratios of heavy neutron - capture elements , which were all measured from ionized species . the details of the analysis for individual neutron - capture elements are presented below . for comparison purposes , the abundances of cu and zn were also determined . in addition to the cu 5105 and zn 4810 lines , the uv lines ( 3250 - 3350 ) are also used in the abundance analyses ( table [ tab : table1 - 4 ] ) . although the abundances of sr , y , and zr have been determined by a number of previous studies for hd 122563 , measurements for other light neutron - capture elements are quite limited so far . we have newly detected 5 elements ( nb , mo , ru , pd , and ag ) in our spectrum , and derived an upper limit for a sixth element ( rh ) . two strong lines of ( 4077 and 4215 ) and the weak line of ( 4607 ) were measured . the two resonance lines are so strong that the abundance derived from these lines is sensitive to the microturbulent velocity and treatment of damping . however , the agreement of the abundances derived from and ( see table [ tab : table1 - 4 ] ) suggests the reliability of our measurements . two ( @xmath21 ) lines are detected . the detection of the 3163 line is reported only for canopus by @xcite and for the sun . this line would be useful for abundance analyses because of its strength ( figure [ fig : spectra ] ) . we note , however , that the abundance determined from this line is slightly higher ( by 0.25 dex ) than that from the 3215 line , possibly suggesting a small contamination by other spectral features . however , here we simply adopt the mean of the abundances derived from the two lines . the ( @xmath22 ) 3864 line is detected in cs 22892 - 052 by @xcite . although a blend of a cn line is reported by these authors , that is not severe in hd 122563 , because the carbon abundance of this star is more than 1 dex lower than that of cs 22892052 . we have detected two lines of ( @xmath23 ) . the equivalent width of the 3498 line is measurable due to there being no ( apparent ) blend with other lines . since a strong line exists in the bluer region of the other line at 3728 , we applied the spectrum synthesis technique to the analysis of that line . the ( @xmath24 ) 3692 line exists in the wing of h i line , but is not detected in our spectrum . we determined only an upper limit on the abundance of this element . we have detected two lines of pd . the 3242 line blends with oh lines . the contamination was estimated using the oh line list of @xcite , adjusting the oxygen abundance to match the strengths of neighboring oh lines ( [ o / fe ] @xmath250.4 ) . the 3404 line is clearly detected with no severe blend ( figure [ fig : spectra ] ) . two lines have been measured in some metal - poor stars . the 3280 line is detected , while the 3383 line is not . since nh lines blend with the 3280 line @xcite , we included the nh lines adopted from @xcite in the spectrum synthesis , calibrating the n abundance to reproduce the nh features around the line ( [ n / fe]@xmath26 ) . we note that the n abundance adopted is much lower than those derived by previous works ( e.g.,[n / fe]@xmath7 * ? ? ? this discrepancy is partially due to the inaccurate gf - values used in our analysis . recent studies of n abundances based on nh lines applied corrections of gf - values to the kurucz s list ( e.g. , * ? ? ? in addition to the nh lines , some unidentified feature appears in the red part of this line ( figure [ fig : spectra ] ) . if such unknown lines also affect the line , the values derived from this line must be regarded as an upper limit on the ag abundance . the abundances of ba , la , eu , and yb were derived by a spectrum synthesis technique , taking into account hyperfine splitting . the abundances of other elements were determined by applying a single line approximation , which is justified by the fact that the absorption lines are quite weak in general . in addition to the two resonance lines at 4554 and 4934 , two ba lines in the red region were measured . the line data of @xcite were used for the analysis of ba lines , assuming the isotope ratios of r - process component in solar - system material . there is a small discrepancy between the ba abundances from the two resonance lines and others . a possible reason for this discrepancy is the non - lte effect ( e.g. , * ? ? ? however , we here simply adopt the average of the ba abundances from the four lines . we detected three lines at 3819 , 4129 and 4205 . however , the 4205 line blends with a line , while the 3819 one is affected by a wing component of a strong fe line at 3820 . since these eu lines are quite weak in the spectrum of hd 122563 , in contrast to those in r - process enhanced stars , the effect of the blending is significant . hence , we adopt the result from the 4129 . we note that the eu abundances derived from the excluded lines , within their relatively large errors , agree with the value from the 4129 line . the effect of isotope shift for gd lines was included in the analysis by @xcite for the s - process enhanced star cs 31062 - 050 . this effect is , however , neglected in the present analysis , because the gd lines of hd 122563 are very weak . we have detected three gd lines ( table [ tab : table1 - 4 ] ) . the gd abundance derived from the 3481 line is significantly higher than those from the other two , and an upper limit estimated from another line ( 3331 ) . we regard this as a result of contamination of the 3481 line by some unidentified lines , and exclude this line in the determination of the final gd abundance . yb is the heaviest element detected in our spectrum . we measured two yb lines at 3220 and 3694 . a spectrum synthesis technique was applied to the analysis including the effect of hyperfine splitting for these lines ( sneden et al . 2003 , private communication ) . agreement of the results from the two lines is fairly good . the upper limits of ir and th abundances are obtained from the analyses of 3800 and 4019 line . the line of 4019 is affected by contamination by other lines ( figure [ fig : th ] ) . we analyzed this line using the line list of @xcite for this wavelength region . we adopted @xmath28c/@xmath29c = 5 to estimated the blending of @xmath29ch lines . since the blends of fe and co lines are severe , no clear th feature is identified , although the quality of the spectrum is very high ( s / n = 950 at 4020 ) . we derived the upper limit of the th abundance ( log @xmath30 ( th ) @xmath31 ) from the fitting of synthetic spectra by eye . neutron - capture elements have been studied for this object by many authors . recently , @xcite , @xcite , @xcite , and @xcite performed detailed abundance analyses based on high resolution spectroscopy ( @xmath32 ) for this object . @xcite also determined the abundances of eight neutron - capture elements , though their resolving power ( @xmath33 ) is not as high as those in the above studies . recently , @xcite derived the abundances of ge , zr , la and eu from the uv region ( 2410 @xmath2 3070 ) of this object using hubble space telescope , and also uv - blue region ( 3150 @xmath2 4600 ) with keck i hires . figure [ fig : comp ] shows comparisons of the abundances of neutron - capture elements derived by the present analysis with those of previous work . our results are in good agreement with @xcite , @xcite and @xcite . however , the abundances of neutron - capture elements determined by the present analysis are systematically lower than those by @xcite and @xcite . here we inspect the discrepancy between the results of @xcite and ours in some detail , because their work determined abundances of a larger number of elements than @xcite , and they reported the equivalent widths used in the abundance analysis for some lines ( unfortunately , the equivalent widths were not given for the lines for which spectrum synthesis technique was applied ) . the atmospheric parameters adopted in our analysis are in good agreement with theirs : the differences ( our results minus those of @xcite ) of @xmath10 , @xmath13 , [ fe / h ] and @xmath12 are + 70 k , 0.2 dex , 0.07 dex and 0.3 km s@xmath9 , respectively . these differences results in differences of abundances smaller than 0.1 dex for heavy neutron - capture elements ( see table [ tab : error ] ) . therefore , the difference of the adopted atmospheric parameters is not a reason for the abundance discrepancy . we also found good agreements in the equivalent widths of the lines reported by both works . however , the @xmath34 value of the nd 4061 line used by @xcite is lower by 0.25 dex than ours . though the @xmath34 values of la adopted by the two works are quite similar , no information on the treatment of the hyperfine splitting was given by @xcite , while the effect is fully included in our analysis . we adopted recent line data for @xcite and @xcite , which should be more accurate than those used by @xcite . thus , the discrepancy between the results of the two work might be partially explained by the difference of the adopted line data . however , there seems to exist small ( @xmath20.2 dex ) systematic differences , for which no clear reason is identified . we investigated a high quality uv - blue spectrum of hd 122563 , and have detected 19 neutron - capture elements including nb , mo , ru , pd , ag , pr , and sm , which are detected for the first time by our study for this object , and upper - limits for 5 elements including th . the derived abundances are given in table [ tab : table2 ] . in this section , we discuss the abundance pattern of the neutron - capture elements , along with cu and zn , to investigate the origin of neutron - capture elements in this object . the abundance ratio between ba ( or la ) and eu is used as an indicator of the origin of neutron - capture elements . the value of [ ba / eu ] of the solar system r - process component is 0.81 , while that of the s - process component is + 1.45 @xcite . the value of the [ ba / eu ] is 0.50 in hd 122563 . in addition , the value of [ la / eu ] of the solar system r - process component is 0.59 , while that in hd 122563 is 0.50 . these results indicate that the heavy ( @xmath35 @xmath36 56 ) neutron - capture elements of this object are principally associated with the r - process , and the contribution of the ( main ) s - process is small if any . we recall that this object is very metal - poor ( [ fe / h]@xmath37 ) with no excess of carbon ( [ c / fe]=@xmath38 ) . figure [ fig : r ] shows the abundance pattern of hd 122563 , comparing with that of the solar - system r - process component . the solar system r - process abundance pattern is scaled to the eu abundance of hd 122563 . this figure clearly shows that the abundances of light neutron - capture elements ( 38 @xmath1 z @xmath1 47 ) are much higher than those of heavy ones in hd 122563 , compared to the solar - system r - process pattern . this behavior is very different from that found in r - process enhanced stars ( e.g. cs 22892 - 052 , cs 31082 - 001 : * ? ? ? * ; * ? ? ? although we indicated above that the heavy neutron - capture elements are principally associated with the r - process , figure [ fig : r ] indicates that even in the range 56 @xmath39 @xmath35 @xmath39 59 the abundances in hd 122563 exceed those of the scaled solar r - process , with only la falling on the eu - normalised curve . this suggests that the tendency for the light neutron - capture elements ( 38 @xmath1 @xmath35 @xmath1 47 ) to fall above the scaled solar r - process curve is part of a general , atomic - number dependent , trend . in order to demonstrate this behavior more clearly , figure [ fig : f7 ] shows the logarithmic difference of the abundances of this object from the solar system r - process pattern as a function of atomic number . we also show that of the r - process enhanced star cs 22892 - 052 @xcite in the same way . we find good agreement between the abundance pattern of cs 22892 - 052 and the solar - system r - process one , at least for elements with 56@xmath40 . by way of contrast , our new measurements for hd 122563 clearly demonstrate a quite different abundance pattern from that of the solar - system r - process component and of cs 22892052 . the excess of light neutron - capture elements ( e.g. sr ) with respect to the heavy ones ( e.g. ba ) was known for hd 122563 by previous studies ( see section 1 ) . however , the abundances of elements having intermediate mass ( e.g. mo , pd ) were measured in detail for the first time by the present work . the abundances of these elements are intermediate , and , hence , the abundances of neutron - capture elements gradually and continuously decrease with increasing atomic number from @xmath41 to @xmath42 . this trend is a key to investigating the nucleosynthesis process that is responsible for neutron - capture elements in this object . for comparison purposes , we also attempted to compare the abundances of hd 122563 with the solar system s - process distribution ( figure [ fig : s ] ) . in this case , ba , la , ce , and eu show large deviations from the solar system s - process curve , although the overall abundance pattern from sr to yb seems to be similar . since the s - process abundance pattern in the solar - system is a result of a combination of some individual processes ( at least the weak and main components of the s - process ) , we presume this apparent `` agreement '' is physically not important . we also have detected four neutron - capture elements heavier than eu ( gd , dy , er , yb : @xmath43 64 ) . the abundances of these elements also show a decreasing trend as a function of atomic number , with respect to the solar - system r - process pattern . this also has an impact on the understanding of the origin of neutron - capture elements in this object . th is a radioactive element and is synthesized only by the r - process . the upper limit of the th abundance determined by the present work indicated that this object does not show an excess of th with respect to eu ( and other heavy neutron - capture elements ) , compared to most r - process enhanced stars . the cu and zn abundances of hd 122563 are typical values found in very metal - poor stars ( e.g. , * ? ? ? * ; * ? ? ? that is , no clear difference is found in the abundances of these elements between the stars having high and low abundances of light neutron - capture elements . this result may give a hint for the origins of cu and zn , as well as of light neutron - capture elements , which are still unclear . here we consider the origin of neutron - capture elements in this object based on the abundance pattern determined by the present work . the abundance ratios between light and heavy neutron - capture elements ( e.g. , sr / ba ) in hd 122563 are clearly different from those in r - process enhanced stars , as already mentioned in 1 . therefore , the r - process responsible for the heavy neutron - capture elements in the solar system ( the so - called `` main '' r - process ) is not an important source of , at least , light neutron - capture elements in hd 122563 . the main s - process is also excluded from the possible source of light neutron - capture elements , because that yields even lower abundance ratios between light and heavy neutron - capture elements at low metallicity ( e.g. * ? ? ? * ; * ? ? ? the weak component of the s - process was introduced to interpret the light s - process nuclei in the solar system . however , the gradually decreasing trend of elemental abundances from sr to ba found in hd 122563 does not resemble that of the weak s - process , which predicts a rapid drop of abundances at @xmath44 ( i.e. y or zr ) . it should be noted in addition that the weak s - process is expected to be inefficient at low metallicity , because of the lack of the neutron source @xmath45ne ( e.g. * ? ? ? * ; * ? ? ? * ) in addition to the lack of iron seeds . thus , another process that has efficiently yielded light neutron - capture elements at low metallicity is required to explain the abundance pattern of neutron - capture elements in hd 122563 . the presence of such a process has been indicated by recent observational studies of metal - deficient stars with no excess of heavy neutron - capture elements ( * ? ? ? * and references therein ) . this process was called the `` light element primary process '' ( lepp ) by @xcite , who estimated its contribution to the solar abundances . @xcite have shown that such light r - process nuclei ( up to @xmath46 ) are produced in neutrino winds as a result of weak " ( or failed ) r - processing ( see also * ? ? ? it is likely , therefore , that these light neutron - capture elements originate from the core - collapse supernovae , although a contribution from other sources can not be excluded . the abundance pattern of light to heavy neutron - capture elements measured for hd 122563 provides a unique constraint on such model calculations . the abundance pattern of heavy neutron - capture elements of hd 122563 shows a significant departure from the r - process component in solar - system material , though the departure is not as large as that found for light neutron - capture elements ( figure [ fig : r ] ) . figure [ fig : r+s ] shows a comparison between the abundances of elements with @xmath47 measured for hd 122563 ( filled circles ) and the solar - system r - process abundance pattern @xcite , which is scaled to achieve the best fit to the observed data ( the solid line ) . ( recall that the scaling in figure [ fig : r ] was to eu . ) significant disagreements are found for ce and pr , as well as for dy , er and yb . given the fact that the abundance patterns of heavy elements measured for r - process enhanced stars show excellent agreement with that of the solar - system , the different abundance pattern of hd 122563 must be significant . some process that produces a different abundance pattern of heavy neutron - capture elements from that of the main r - process is at least required to explain the abundance pattern of this star . one possibility is to assume a small contribution of the ( main ) s - process . in order to examine this possibility , we adopted the r - process and s - process abundance patterns in the solar - system as determined by @xcite , and searched for the combination that gives the best fit to the abundance pattern of hd 122563 . the result is shown by the dashed line in figure [ fig : r+s ] . in this case , the contributions of the s - process to ba , la , and ce are significant ( @xmath2 70% ) , while those to elements heavier than nd are small ( @xmath48 10% ) . it should be noted that such a small contribution of the main s - process does not affect the abundance pattern of light neutron - capture elements discussed in the previous subsection . the root - mean - square ( rms ) of the logarithmic abundance difference between the observed and calculated abundance patterns is 0.29 dex for the case that a contribution from the s - process is introduced , compared with 0.34 dex if only the r - process is assumed . thus , the fit is slightly better if an s - process contribution is introduced . however , the improved fit comes at the cost of an additional free parameter ( the relative contribution of the s- and r - processes ) , and it is not clear that assuming an s - process contribution is justifiable from the point of view of the star s nucleosynthetic history . in particular , the discrepancy found in the abundance pattern of elements heavier than sm can not be explained by this approach , because the production of such heavy elements by the s - process is very small . this discrepancy leads us to consider an alternative possibility that the heavy neutron - capture elements of this star are also a result of the unidentified ( single ) process that is responsible for the large enhancement of the light neutron - capture elements discussed in 5.2.1 . @xcite have suggested that the small ( but non - negligible ) amount of heavy r - process nuclei ( @xmath49 ) with non solar r - process abundance pattern can be produced in neutrino winds if the entropy of the neutrino - heated matter is enough high ( but smaller than that required for the main r - process ) , in addition to the light r - process nuclei . if this is true , the abundance pattern of heavy neutron - capture elements in this object may give another hint for , or constraint on , the modeling of this process . the differences between the abundance patterns of heavy neutron - capture elements in hd 122563 and the solar - system r - process abundance pattern are not large . further observational data are clearly required to derive definitive conclusions . however , it would be very difficult to improve the measurement of abundances for hd 122563 , because the quality of the spectrum used in the present work is extremely high . one possibility is to reconsider the atomic line data to reduce the uncertainties in abundance measurements . another important observational study is to apply similar analysis to other metal - poor stars having high abundance ratios of light to heavy neutron - capture elements . in particular , measurements for stars with lower metallicity than hd 122563 are important , because no s - process contribution is expected for such stars . we obtained a high resolution , high s / n uv - blue spectrum of the very metal - poor star hd 122563 with subaru / hds . the abundances were measured for 19 neutron - capture elements , among which seven elements are detected for the first time in this star . our new measurements for hd 122563 clearly demonstrate that the elemental abundances gradually decrease relative to the scaled solar r - process abundances with increasing atomic number , at least from sr to yb . we have considered whether the higher abundances of lighter elements can be explained by contributions from the weak s- or main s - processes , and find these not to provide satisfactory explanations . the abundance pattern of elements with @xmath50 does not agree with any prediction of neutron - capture process models known . the abundance pattern of a wide range of neutron - capture elements determined for hd 122563 therefore provides new strong constraints on models of nucleosynthesis for very metal - poor stars , in particular those with excesses of light neutron - capture elements but without enhancement of heavy ones . this work was supported in part by a grant - in - aid for the japan - france integrated action program ( sakura ) , awarded by the japan society for the promotion of science , and scientific research ( 17740108 ) from the ministry of education , culture , sports , science , and technology of japan . most of the data reduction was carried out at the astronomical data analysis center ( adac ) of the national astronomical observatory of japan . aoki , w. , et al . 2002 , apj , 580 , 1149 aoki , w. , et al . 2005 , apj , 632 , 611 asplund , m. , grevesse , n. , & sauval , a. j. 2005 , in asp conf . 336 , cosmic abundances as records of stellar evolution and nucleosynthesis , ed . t. g. barnes iii & f. n. bash ( san francisco : asp ) , 25 asplund , m. 2005 , ara&a , 43 , 481 burris , d. l. , pilachowski , c. a. , armandroff , t. e. , sneden , c. , cowan , j. j. , & roe , h 2000 , apj , 544 , 302 busso , m. , gallino , r. , & wasserburg , g. j. 1999 , ara&a , 37 , 239 cameron , a. g. w. 2001 , apj , 562 , 456 cayrel , r , et al . 2004 , a&a , 416 , 1117 cowan , j. j. , et al . 2002 , apj , 572 , 861 cowan , j. j. , et al . 2005 , apj , 627 , 238 den hartog , e. a. , lawler , j. e. , sneden , c. & cowan , j. j. 2003 , apjs , 148 , 543 ecuvillon , a. , israelian , g. , santos , n. c. , mayor , m. , garcia lopez , r. j. , & randich , s. 2004 , a&a , 418 , 703 freiburghaus , c. , rosswog , s. , & thielemann , f. -k . 1999 , , 525 , l121 hill , v. , et al . 2002 , a&a , 387 , 560 honda , s. , et al . 2004 , apj , 607 , 474 ishimaru , y. & wanajo , s. 1999 , , 511 , l33 ishimaru , y. , wanajo , s. , aoki , w. , & ryan , s. g. 2004 , , 600 , l47 johnson , j. a. & bolte , m. 2001 , apj , 554 , 888 johnson , j. a. 2002 , apjs , 139 , 219 johnson , j. a. & bolte , m. 2002 , apj , 579 , 616 johnson , j. a. & bolte , m. 2004 , apj , 605 , 462 kurucz , r. l. , & bell . b 1995 , kurucz cd - rom , no.23 ( harvard - smithsonian center for astrophysics ) kurucz , r. l. 1993 , kurucz cd - rom , no.13 ( harvard - smithsonian center for astrophysics ) lawler , j. e. , sneden , c. , & cowan , j. j. , 2004 , apj , 604 , 850 lawler , j. e. , bonvallet , g. , & sneden , c. 2001a , apj , 556 , 452 lawler , j. e. , wickliffe , m. e. , den hartog , e. a. , & sneden , c. 2001b , apj , 563 , 1075 mcwilliam , a , preston , g. w. , sneden , c , & searle , l 1995 , aj , 109 , 27 mcwilliam , a 1998 , aj , 115 , 1640 noguchi , k.,et al . 2002 , pasj , 54 , 855 prantzos , n. , hashimoto , m. , & nomoto , k. 1990 , a&a , 234 , 211 sneden , c. , mcwilliam , a. , preston , g. w. , cowan , j. j. , burris , d. l. , & armosky , b. j. 1996 , apj , 467 , 819 sneden , c , cowan , j. j. , ivans , i. i. , fuller , g. m. , burles , s , beers , t. c. , & lawler , j. e. 2000 , apj , 533l , 139 sneden , c. , et al . 2003 , apj , 591 , 936 sumiyoshi , k. , et al . 2001 , apj , 562 , 880 surman , r. , mclaughlin , g. c. , & hix , w. r. 2005 , submitted to ( astro - ph/0509365 ) raiteri , c. m. , gallino , r. , busso , m. , neuberger , d. , & kappeler , f. 1993 , apj , 419 , 207 reynolds , s. e. , hearnshaw , j. b. , & cottrell , p. l. 1988 , mnras , 235 , 1423 rosswog , s. , liebendorfer , m. , thielemann , f .- k . , davies , m. b. , benz , w. , & piran , t. 1999 , a&a , 341 , 499 travaglio et al . 2004 , apj , 114 , 1293 truran , j. w. , cowan , j. j. , pilachowski , c. a. , & sneden , c. 2002 , pasp , 114 , 1293 wanajo , s. , kajino , t. , mathews , g. j. , & otsuki , k. 2001 , , 554 , 578 wanajo , s. , tamamura , m. , itoh , n. , nomoto , k. , ishimaru , y. , beers , t. c. , & nozawa , s. 2003 , apj , 593,968 wanajo , s. & ishimaru , i. 2005 , , in press westin , j. , sneden , c. , gustafsson , b. , & cowan , j. j. 2000 , apj , 530 , 783 woosley , s. e. , wilson , j. r. , mathews , g. j. , hoffman , r. d. , & meyer , b. s. 1994 , , 433 , 229 lcccccc wavelength & l.e.p.(ev ) & log@xmath34 & log@xmath30 & _ w_(m ) & ref + , @xmath51 & & & & & + 3247.53 & 0.000 & 0.060 & 1.04 & 115.1 * & 8 + 3273.95 & 0.000 & 0.360 & 1.24 & 112.3 & 8 + 5105.55 & 1.390 & 1.520 & 0.66 & 3.3 & 2 + , @xmath52 & & & & & + 3302.98 & 4.030 & 0.057 & 2.07 & 20.3 * & 1 + 3345.02 & 4.078 & 0.246 & 1.78 & 19.5 * & 1 + 4722.15 & 4.030 & 0.390 & 2.09 & 14.5 & 5 + 4810.54 & 4.080 & 0.170 & 2.10 & 20.5 & 5 + , @xmath41 & & & & & + 4607.33 & 0.000 & 0.280 & 0.14 & 2.4 & 7 + , @xmath41 & & & & & + 4077.71 & 0.000 & 0.170 & 0.18 & 163.3 & 6 + 4215.52 & 0.000 & 0.170 & 0.03 & 155.6 & 6 + , @xmath53 & & & & & + 3327.88 & 0.410 & 0.130 & 0.98 & 49.3 * & 8 + 3549.01 & 0.130 & 0.280 & 0.97 & 49.0 & 7 + 3584.52 & 0.100 & 0.410 & 0.98 & 44.1 * & 10 + 3600.74 & 0.180 & 0.280 & 1.06 & 68.2 & 7 + 3611.04 & 0.130 & 0.010 & 1.06 & 59.0 & 7 + 3628.70 & 0.130 & 0.710 & 0.90 & 31.8 & 10 + 3710.29 & 0.180 & 0.460 & 0.98 & 82.3 & 10 + 3747.55 & 0.100 & 0.910 & 0.98 & 22.8 & 7 + 3774.33 & 0.130 & 0.210 & 0.91 & 79.0 & 6 + 3788.70 & 0.100 & 0.070 & 0.95 & 66.6 & 6 + 3818.34 & 0.130 & 0.980 & 0.74 & 29.4 & 6 + 3950.36 & 0.100 & 0.490 & 0.94 & 47.4 & 6 + 4398.01 & 0.130 & 1.000 & 0.82 & 28.0 & 6 + 4883.69 & 1.080 & 0.070 & 0.82 & 25.0 & 6 + 5087.43 & 1.080 & 0.170 & 0.91 & 13.8 & 6 + , @xmath54 & & & & & + 3438.23 & 0.090 & 0.420 & 0.57 & 76.8 * & 7 + 3457.56 & 0.560 & 0.530 & 0.17 & 27.3 & 7 + 3479.02 & 0.530 & 0.690 & 0.34 & 16.7 & 7 + 3479.39 & 0.710 & 0.170 & 0.40 & 40.2 * & 7 + 3499.58 & 0.410 & 0.810 & 0.48 & 13.6 * & 7 + 3505.67 & 0.160 & 0.360 & 0.42 & 47.0 * & 7 + 3536.94 & 0.360 & 1.310 & 0.33 & 7.6 * & 7 + 3551.96 & 0.090 & 0.310 & 0.35 & 57.3 * & 8 + 3573.08 & 0.320 & 1.040 & 0.26 & 16.4 & 7 + 3578.23 & 1.220 & 0.610 & 0.29 & 3.8 * & 7 + 3630.02 & 0.360 & 1.110 & 0.28 & 12.8 & 7 + 3714.78 & 0.530 & 0.930 & 0.25 & 13.6 & 7 + 3836.77 & 0.560 & 0.060 & 0.31 & 47.3 & 6 + 3998.97 & 0.560 & 0.670 & 0.07 & 36.1 & 7 + 4050.33 & 0.710 & 1.000 & 0.09 & 11.0 & 7 + 4208.98 & 0.710 & 0.460 & 0.16 & 27.1 & 6 + 4317.32 & 0.710 & 1.380 & 0.08 & 5.3 & 6 + , @xmath21 & & & & & + 3163.40 & 0.376 & 0.260 & 1.30 & 24.2 * & 1 + 3215.59 & 0.440 & 0.190 & 1.65 & 4.5 * & 6 + , @xmath22 & & & & & + 3864.10 & 0.000 & 0.010 & 0.87 & 3.3 & 8 + , @xmath23 & & & & & + 3498.94 & 0.000 & 0.310 & 0.90 & 4.1 & 6 + 3728.03 & 0.000 & 0.270 & 0.82 & 4.7 * & 6 + , @xmath24 & & & & & + 3692.36 & 0.000 & 0.174 & @xmath48 - 1.20 & syn & 6 + , @xmath55 & & & & & + 3404.58 & 0.810 & 0.320 & 1.31 & 6.7 * & 6 + , @xmath56 & & & & & + 3280.68 & 0.000 & 0.050 & 1.88 & 6.5 * & 6 + , @xmath42 & & & & & + 4554.04 & 0.000 & 0.170 & 1.76 & 95.4 * & 6 + 4934.10 & 0.000 & 0.150 & 1.76 & 82.2 * & 6 + 5853.70 & 0.604 & 1.010 & 1.54 & 8.5 * & 6 + 6141.70 & 0.704 & 0.070 & 1.55 & 40.6 * & 6 + @xmath57 & & & & & + 3794.77 & 0.240 & 0.210 & 2.35 & 5.9 * & 7 + 3988.52 & 0.400 & 0.210 & 2.75 & 1.7 * & 4 + 3995.75 & 0.170 & 0.060 & 2.65 & 2.1 * & 4 + 4086.71 & 0.000 & 0.070 & 2.53 & 4.6 * & 4 + 4123.23 & 0.320 & 0.130 & 2.70 & 2.0 * & 4 + @xmath58 & & & & & + 4222.60 & 0.120 & 0.180 & 1.90 & 2.0 * & 6 + 4523.08 & 0.520 & 0.080 & 1.62 & 1.6 * & 10 + 4539.78 & 0.330 & 0.080 & 2.03 & 1.1 & 10 + 4562.37 & 0.480 & 0.190 & 2.01 & 1.4 & 10 + 4572.28 & 0.680 & 0.290 & 1.98 & 1.1 * & 8 + @xmath59 & & & & & + 4179.40 & 0.200 & 0.480 & 2.15 & 7.8 * & 7 + 4189.48 & 0.370 & 0.380 & @xmath48 - 2.15 & & 8 + @xmath60 & & & & & + 3784.25 & 0.380 & 0.150 & 2.14 & 2.0 & 9 + 3826.42 & 0.064 & 0.410 & 2.00 & 2.0 * & 9 + 4061.08 & 0.471 & 0.550 & 2.00 & 5.5 & 9 + 4232.38 & 0.064 & 0.470 & 1.89 & 2.3 * & 9 + @xmath61 & & & & & + 4318.94 & 0.280 & 0.270 & 2.16 & 1.8 * & 7 + 4642.23 & 0.380 & 0.520 & 2.16 & 0.8 & 7 + @xmath62 & & & & & + 3819.67 & 0.000 & 0.510 & 2.95 & syn & 5 + 4129.70 & 0.000 & 0.220 & 2.77 & 9.0 * & 5 + 4205.05 & 0.000 & 0.210 & 2.92 & syn & 5 + @xmath63 & & & & & + 3331.40 & 0.000 & 0.140 & @xmath48 - 2.15 & syn & 7 + 3481.80 & 0.490 & 0.230 & 1.79 & 4.0 * & 7 + 3549.37 & 0.240 & 0.260 & 2.40 & 2.1 & 7 + 3768.40 & 0.080 & 0.360 & 2.48 & 3.8 * & 7 + @xmath64 & & & & & + 3460.97 & 0.000 & 0.070 & 2.42 & 3.2 * & 7 + 3531.71 & 0.000 & 0.770 & 2.81 & 8.3 * & 8 + @xmath65 & & & & & + 3398.94 & 0.000 & 0.410 & @xmath48 - 2.00 & syn & 11 + @xmath66 & & & & & + 3499.10 & 0.060 & 0.136 & 2.75 & 2.7 & 10 + 3692.65 & 0.050 & 0.138 & 2.57 & 4.5 * & 6 + @xmath67 & & & & & + 3701.36 & 0.000 & 0.540 & @xmath483.00 & syn & 8 + @xmath68 & & & & & + 3289.37 & 0.000 & 0.020 & 2.70 & 27.5 * & 8 + 3694.19 & 0.000 & 0.300 & 2.86 & 13.3 * & 8 + @xmath69 & & & & & + 3220.76 & 0.350 & 0.510 & @xmath48 - 1.12 & syn & 7 + 3800.12 & 0.000 & 1.450 & @xmath48 - 1.60 & syn & 7 + @xmath70 & & & & & + 4019.12 & 0.000 & 0.270 & @xmath483.05 & syn & 7 + lccccc species & _ z _ & log@xmath30 & @xmath19 & [ x / fe ] & n + cu & 29 & 0.98 & 0.29 & 1.15 & 3 + zn & 30 & 2.01 & 0.16 & + 0.18 & 4 + sr & 38 & 0.11 & 0.08 & 0.26 & 3 + y & 39 & 0.93 & 0.09 & 0.37 & 15 + zr & 40 & 0.28 & 0.16 & 0.10 & 17 + nb & 41 & 1.48 & 0.17 & 0.13 & 2 + mo & 42 & 0.87 & 0.17 & 0.02 & 1 + ru & 44 & 0.86 & 0.17 & + 0.07 & 2 + rh & 45 & @xmath481.20 & & @xmath48 + 0.45 & 1 + pd & 46 & 1.36 & 0.17 & 0.28 & 2 + ag & 47 & 1.88 & 0.17 & 0.05 & 1 + ba & 56 & 1.62 & 0.12 & 1.02 & 4 + la & 57 & 2.66 & 0.10 & 1.02 & 5 + ce & 58 & 1.83 & 0.18 & 0.64 & 6 + pr & 59 & 2.15 & 0.17 & 0.09 & 1 + nd & 60 & 2.01 & 0.10 & 0.69 & 4 + sm & 62 & 2.16 & 0.17 & 0.40 & 2 + eu & 63 & 2.77 & 0.17 & 0.52 & 1 + gd & 64 & 2.44 & 0.17 & 0.76 & 2 + dy & 66 & 2.62 & 0.17 & 0.99 & 2 + ho & 67 & @xmath482.00 & & @xmath48 + 0.26 & 1 + er & 68 & 2.66 & 0.17 & 0.82 & 2 + tm & 69 & @xmath483.00 & & @xmath480.23 & 1 + yb & 70 & 2.78 & 0.17 & 1.09 & 2 + ir & 77 & @xmath481.60 & & @xmath480.21 & 2 + th & 90 & @xmath483.05 & & @xmath480.34 & 1 + lccccccccccc species & & & & & & & + & @xmath71 & @xmath72 & & @xmath73 & @xmath74 & & @xmath75 & @xmath76 & & @xmath76 & @xmath75 + & @xmath770.22 & @xmath780.17 & & @xmath780.06 & @xmath770.07 & & @xmath770.15 & @xmath780.06 & & @xmath780.23 & @xmath770.26 + & @xmath770.05 & @xmath780.04 & & @xmath770.05 & @xmath780.04 & & 0.00 & @xmath770.01 & & @xmath780.03 & @xmath770.02 + & @xmath770.13 & @xmath780.12 & & @xmath780.02 & @xmath770.03 & & @xmath770.04 & 0.00 & & @xmath780.01 & 0.00 + & @xmath770.12 & @xmath780.11 & & @xmath770.05 & @xmath780.03 & & @xmath770.03 & @xmath780.05 & & @xmath780.31 & @xmath770.39 + & @xmath770.08 & @xmath780.07 & & @xmath770.08 & @xmath780.08 & & @xmath780.02 & @xmath770.01 & & @xmath780.16 & @xmath770.09 + & @xmath770.08 & @xmath780.07 & & @xmath770.09 & @xmath780.08 & & @xmath780.03 & @xmath770.02 & & @xmath780.08 & @xmath770.05 + & @xmath770.09 & @xmath780.09 & & @xmath770.10 & @xmath780.10 & & @xmath780.04 & @xmath770.02 & & @xmath780.03 & @xmath770.02 + & @xmath770.19 & @xmath780.17 & & @xmath780.02 & @xmath770.01 & & @xmath770.03 & @xmath780.00 & & @xmath780.00 & 0.00 + & @xmath770.20 & @xmath780.17 & & @xmath780.01 & @xmath770.02 & & @xmath770.04 & @xmath780.00 & & @xmath780.01 & 0.00 + & @xmath770.20 & @xmath780.17 & & @xmath780.01 & @xmath770.03 & & @xmath770.04 & @xmath770.01 & & @xmath780.01 & @xmath770.01 + & @xmath770.21 & @xmath780.17 & & @xmath780.00 & @xmath770.01 & & @xmath770.04 & @xmath780.00 & & @xmath780.01 & @xmath770.01 + & @xmath770.09 & @xmath780.07 & & @xmath770.09 & @xmath780.07 & & @xmath780.01 & @xmath770.01 & & @xmath780.08 & @xmath770.06 + & @xmath770.08 & @xmath780.07 & & @xmath770.10 & @xmath780.09 & & @xmath780.03 & @xmath770.02 & & @xmath780.00 & @xmath770.00 + & @xmath770.08 & @xmath780.07 & & @xmath770.09 & @xmath780.09 & & @xmath780.02 & @xmath770.02 & & @xmath780.00 & 0.00 + & @xmath770.10 & @xmath780.08 & & @xmath770.10 & @xmath780.09 & & @xmath780.03 & @xmath770.02 & & @xmath780.01 & @xmath770.01 + & @xmath770.08 & @xmath780.08 & & @xmath770.09 & @xmath780.10 & & @xmath780.04 & @xmath770.02 & & @xmath780.00 & 0.00 + & @xmath770.09 & @xmath780.07 & & @xmath770.09 & @xmath780.08 & & @xmath780.03 & @xmath770.02 & & @xmath780.00 & @xmath780.00 + & @xmath770.09 & @xmath780.09 & & @xmath770.10 & @xmath780.10 & & @xmath780.04 & @xmath770.02 & & @xmath780.00 & @xmath770.00 + & @xmath770.10 & @xmath780.08 & & @xmath770.11 & @xmath780.09 & & @xmath780.03 & @xmath770.02 & & @xmath780.00 & 0.00 + & @xmath770.09 & @xmath780.10 & & @xmath770.10 & @xmath780.10 & & @xmath780.05 & @xmath770.02 & & @xmath780.01 & @xmath770.01 + & @xmath770.09 & @xmath780.09 & & @xmath770.10 & @xmath780.09 & & @xmath780.04 & @xmath770.03 & & @xmath780.00 & @xmath780.00 + & @xmath770.09 & @xmath780.08 & & @xmath770.10 & @xmath780.10 & & @xmath780.04 & @xmath770.02 & & @xmath780.03 & @xmath770.02 +

einstein s theory of general relativity is described by a complicated set of coupled , nonlinear partial differential equations . like the maxwell equations of classical electromagnetism , the einstein equations are over - determined , and can be separated into hyperbolic evolution and elliptic constraint equations , as in the arnowitt deser misner ( adm ) decomposition @xcite . the complexity of these equations is such that analytic solutions have been found for only very special configurations , and numerical studies arguably provide the only means to explore a wide variety of astrophysical and theoretically significant problems @xcite . unfortunately , these numerical solutions are often prone to various instabilities , which we divide into three classes : ( 1 ) continuum instabilities , ( 2 ) numerical instabilities and ( 3 ) artificial long - term `` instabilities '' ( alti ) . _ continuum instabilities _ exist in the formulation of the continuum problem ( and thus are naturally reflected in any consistent numerical scheme ) , and are characterized by the solution itself or some perturbation of it either `` blowing up '' at a finite time or growing fast in time . these instabilities may reflect physical phenomena , such as turbulence in fluids or the threshold of black hole formation @xcite , or may be an artifact of the formulation of the problem . for instance , ill - posed problems suffer from these instabilities , which also manifest themselves as numerical ones ( see @xcite , and @xcite for an illustration in numerical relativity ) . _ numerical instabilities _ are characterized by errors that , at a fixed time , _ increase _ as the discretization scale is decreased . these instabilities are present in many discretization schemes for well - posed problems that appear `` natural '' , such as the forward - time , centered - space scheme for the advection equation . finally , _ alti _ are sometimes exhibited even by numerically _ stable _ implementations , whenever errors grow too fast for the desired simulation time - scale and available computational resources . although these may not be instabilities in the strict sense , given that as defined here they are not present at the continuum and do go away with resolution , they are problematic in practice as they limit the time - length of a reliable simulation . in this paper we focus on this long - term error growth together with the application of the techniques described in @xcite , which throughout this work we refer to as paper i. the stability of finite difference approximations ( fdas ) for simple initial value problems ( ivp ) with constant coefficients and periodic boundaries is often analyzed in discrete fourier space , by examining the norm of the amplification matrix . this method , however , does not generalize in a straightforward way to more complicated problems , such as those with variable coefficients or those with boundaries . a more powerful technique is based on the energy method @xcite , also used to identify well - posed continuum problems . this method can be used in all problems for which the continuum energy method applies , e.g. , quasi - linear parabolic and symmetric hyperbolic partial differential equations with appropriate boundary conditions . the discrete formulation reproduces the continuum analysis , including the calculation of a discrete energy estimate , which can be used to meet sufficient conditions for a stable numerical method . combining these estimates with the consistency of the discrete system with the continuum problem , stability of the linear problem is equivalent to convergence via the lax equivalence theorem @xcite . while our focus in general relativity is on the nonlinear einstein equations , stability of the linearized equations is obviously a necessary step for solving non - linear problems , in particular those with smooth solutions . in the absence of matter , smooth solutions are indeed expected in general relativity when appropriate slicing conditions are used @xcite . as described in paper i , a discrete energy estimate can be obtained when using finite difference operators that satisfy _ summation by parts _ ( sbp ) @xcite , a discrete version of integration by parts , as well as appropriate boundary conditions @xcite and time integrators @xcite . paper i presents and extends these techniques in the context of numerical relativity . in this paper we demonstrate these techniques using some common test problems employed in the field . we evolve klein gordon scalar fields and maxwell fields on fixed schwarzschild spacetimes . we also consider gauge wave spacetimes , obtained via coordinate transformations of flat spacetime . we show that we naturally obtain stable and convergent numerical schemes by construction , even when an inner boundary is used to remove the black hole singularity ( black hole excision ) @xcite . a second goal of this paper is to understand some causes of long - term error growth ( alti ) , and ways of eliminating or minimizing it by achieving _ strict stability _ @xcite . numerical solutions of the einstein equations are frequently plagued by such long - term errors . since at least some of these appear in numerically stable implementations , their cause is often attributed to the formulation of the equations . while this may be the case in a number of problems , as discussed below , some numerical experiments of simple differential equations with no continuum instabilities also display this same artificial long - term instability even when discretized in a numerically stable way. hence , it is important to devise _ numerical _ techniques to control this spurious growth in the numerical error . we find that this spurious error growth is allowed in the absence of sharp energy estimates for the discrete problem . this can arise , from among other reasons , because the discrete derivatives fail to satisfy the leibnitz rule . with this observation in mind , we discuss here a method to remedy some of these problems by constructing schemes that suppress excitations of alti . this work is organized as follows : we first summarize , in section ii , some results from paper i , in order to introduce some notation . section [ strict ] discusses the notion of strict stability , and we present simple examples of `` natural '' schemes that are numerically stable but not strictly stable , for which the errors grow fast in time . we then explain the basic strategy to achieve strict stability , and show how this eliminates alti . in section [ equations ] we present some test problems from numerical relativity : three - dimensional scalar and electromagnetic fields propagating on a fixed schwarzschild black hole geometry , and one - dimensional perturbations of a dynamical slicing of flat spacetime ( `` gauge wave '' ) . the first cases allow us to test black hole excision , while the gauge wave has a time dependent background without a sharp energy estimate . however , a sharp energy estimate does exist for the subsidiary constraint system , and it can be used to understand the nature of the alti and suppress them . although more work would need to be done , these observations and ideas could be useful when evolving einstein s equations in more complicated situations . in section [ simulations ] we present numerical simulations of the systems discussed in section [ equations ] . finally , in appendix [ smallness ] we discuss cubical excision boundaries for black hole spacetimes . we first review the construction of numerically stable finite - difference schemes using the discrete energy method for linear , symmetric hyperbolic initial - boundary value problems ( ibvps ) with maximally dissipative boundary conditions . while the discussion here is brief , further information can be found in the standard text by gustaffsson , kreiss , and oliger @xcite , works by olsson @xcite , and paper i , for their particular application in numerical relativity . sufficient conditions for a stable discretization are fulfilled when : 1 . the continuum problem can be shown to be well posed using the energy method . 2 . spatial difference operators that satisfy sbp on the computational domain are used . boundary conditions are imposed via orthogonal projections @xcite . the semi - discrete equations are integrated with a _ locally stable _ time integrator , e.g. , third or fourth order runge kutta @xcite . optionally , explicit dissipation may be added using operators that do not spoil the energy estimates , and different ways of writing the discrete equations may be explored to achieve strict stability . as an example , consider a set of linear , symmetric hyperbolic equations on a one - dimensional domain , @xmath0 $ ] . this domain is discretized with points @xmath1 , @xmath2 , where @xmath3 , and a discrete scalar product , @xmath4 is introduced for some positive definite matrix with elements @xmath5 . in the limit @xmath6 this scalar product approaches the continuum one given by @xmath7 the standard continuum derivative operators and scalar products satisfy integration by parts : @xmath8 similarly , a discrete difference operator @xmath9 is said to satisfy sbp , if there is a scalar product with respect to which @xmath10 holds . the simplest difference operator and scalar product that satisfy sbp on this domain are @xmath11 where @xmath12 , @xmath13 , @xmath14 , and the scalar product is diagonal : @xmath15 for @xmath16 . higher order operators satisfying sbp have been constructed by strand @xcite . domains with inner boundaries in several dimensions introduce additional complexities , and we refer the reader to paper i for more information . we sometimes use explicit numerical dissipation . for that purpose we modify near boundaries the standard kreiss oliger dissipative operator for second order derivatives , @xmath17 , so that it does not spoil the semi - discrete estimate , i.e. , so that it satisfies @xmath18 . the operator @xmath17 , which is fully described in paper i , is added to the right - hand side of the evolution equations . in the absence of inner boundaries , it is ( on each direction ) @xmath19 we discretize the equations in time using the method of lines , and use third - order runge kutta to integrate the equations . more precisely , for the ordinary differential system @xmath20 we integrate the solution @xmath21 , defined at time @xmath22 , to the advanced time , @xmath23 , obtaining @xmath24 , through @xmath25 where @xmath26 throughout this paper we call artificial long - term instabilities ( alti ) those instabilities that are _ neither _ numerical the errors at a fixed time do not become larger when the resolution is increased , but rather converge to zero_nor _ true instabilities of the continuous solution . these instabilities , as shown in some examples below , can appear quite generically in long - term simulations . there are least two possible ways to suppress alti s : 1 . constructing _ strictly stable _ discretizations with respect to a conveniently defined energy . the equations are written such that a sharp continuum estimate also holds for the semi - discrete equations . the semi - discrete energy will then closely mirror the continuum energy estimate . explicit numerical dissipation can be used to eliminate high frequency errors , shifting the spectrum of the amplification matrix . although the latter solution is easy to implement , it may not be effective if the alti excites low frequency modes . the first solution , on the other hand , does not introduce artificial damping , but such a discretization may be difficult or even impossible to find for generic systems of equations . for example , no general sharp estimate for the full einstein equations is known at this time . thus , these two possibilities can be seen as complementary to each other . the energy of a solution , @xmath27 , of a set of linear partial differential equations is defined as @xmath28 where @xmath29 is a positive definite symmetrizer for the system . this energy is not unique , reflecting the freedom in choosing the symmetrizer , @xmath29 . the energy method provides an estimate for the growth rate of the energy in the form of an inequality , say , @xmath30 and thus provides a bound to the energy growth . this bound may be much larger than the actual growth rate for the continuum solution , in which case we say that the energy estimate is _ coarse . _ when the estimate is close to the actual energy growth rate of a solution , the estimate is _ sharp . _ while the behavior of the continuum solution is independent of a particular bound , as we discuss in this paper , numerical solutions sometimes do have a large growth if this is allowed . thus , the sharper the bound is , the smaller the possible spurious growth in the numerical solution . finding a sharp estimate is problem - dependent , and the definition of a convenient energy has to be analyzed on each case . one usually attempts to find a sharp energy bound by exploiting the freedom in choosing the symmetrizer . for example , a well - defined local physical energy exists in many physically motivated problems , which can be used to select a preferred symmetrizer . the evolution equations must be written such that they are symmetrizable with respect to this energy , but this is in general possible . although a locally positive definite energy for the einstein equations does not exist , relatively sharp estimates exist for some problems @xcite . the energy method can be used to derive discrete estimates . using difference operators that satisfy sbp and representing the boundary conditions through orthogonal projections , the semi - discrete calculation mirrors the continuum analysis . when a discrete energy estimate is present , the discrete system is numerically stable by construction . a discrete system with the same energy estimate as the continuum problem may be defined as strictly stable . however , we are particularly interested in systems for which the energy estimate is sharp , and we therefore reserve the term _ strictly stable _ for discrete systems that match a sharp continuum energy estimate . assuming a sharp energy estimate exists for the continuum problem , in this paper we consider how the discrete problem might be implemented to preserve this sharp estimate at the numerical level . this can be exploited to achieve long - term stability in an otherwise numerically stable but long - term unstable scheme . throughout this paper we only use numerically stable schemes , even when not explicitly mentioned . therefore , we concentrate on their long - term stability . finally , we note that the sharp energy estimates discussed here are for semi - discrete systems , where time is continuous but space has been discretized . a sharp estimate may , in principle , be lost owing to error introduced by the temporal integration . the experiments presented in this paper and in @xcite suggest that ( at least in the cases considered ) if strict stability is lost , the effects are not serious . the semi - discrete analysis appears to adequately represent the fully discrete simulation . consider the initial - boundary value problem defined by the equation @xmath31 , \qquad t\ge 0 \ ; , \label{1d}\ ] ] together with some initial data and maximally dissipative boundary conditions . let us consider the numerically stable discretization defined by @xmath32 with the discrete derivative , @xmath9 , given by eq . ( [ simpled ] ) , third - order runge kutta for the time integration [ cf . ( [ eq : rk3 ] ) ] , and maximally dissipative boundary conditions imposed through orthogonal projections . figure [ conv_simple_eq ] shows results from simulations using this numerical scheme and @xmath33 as initial data and @xmath34 as boundary conditions , which corresponds to the time - independent solution @xmath35 . the figure plots the @xmath36 norm of the errors with respect to the exact solution as a function of time , at different resolutions ( obtained on uniform grids , where the number of points ranges from @xmath37 to @xmath38 ) . at a fixed time , the errors decrease with resolution , verifying that , as expected , the scheme is numerically stable . at fixed resolution , however , the error grows as a function of time , i.e. , the solution has long - term error growth , or alti . norm of the solution error for eq . ( [ 1da ] ) as a function of time at many different discretization scales , with initial data @xmath39 and boundary conditions @xmath34 . the solution is calculated with a numerically stable discretization that is not strictly stable . although convergent , the error in the solution clearly displays an exponential growth . the cfl factor used is 0.8 . , height=226 ] . [ conv_simple_eq ] to understand this behavior we examine an energy norm of the solution . let the continuum energy is used to denote that this is the first energy defined for this system . ] be ( using @xmath40 ) @xmath41 we obtain an energy estimate by taking a time derivative in both sides of eq . ( [ cenergy1 ] ) and using the evolution equation ( [ 1d ] ) , obtaining @xmath42 that is , @xmath43 only integration by parts was used to derive this estimate , between eqs . ( [ before ] ) and ( [ after ] ) . now consider the discrete problem . we define a semi - discrete energy analogous to that one defined in eq . ( [ cenergy1 ] ) , namely @xmath44 the derivation of the above continuum energy estimate can be repeated at the semidiscrete level , provided the difference operator satisfies sbp , since only integration by parts was used at the continuum . therefore , the same estimate holds for the semi - discrete energy defined in eq . ( [ denergy1 ] ) . that is , @xmath45 the continuum estimate , eq . ( [ bad_estimate ] ) , and the semi - discrete counterpart , eq . ( [ bad_destimate ] ) , guarantee that the norm of the continuum and numerical solutions will grow with a rate bounded by @xmath46 . as these are estimates , this does not mean that the exponential growth actually occurs . indeed , the continuum solution does not grow , since eq . ( [ 1d ] ) is the advection equation with the time - independent change of variables , @xmath47 . with this transformation , eq . ( [ 1d ] ) becomes @xmath48 . however , the results shown in fig . [ conv_simple_eq ] indicate that an exponential growth does appear in the numerical solution . we now define a new energy , . ] that allows for a sharper energy estimate and , consequently , a better discrete system . let this new energy ( defined with @xmath49 ) be @xmath50 as before , an estimate is obtained by differentiating in time , substituting in the evolution equations , and using integration by parts to get @xmath51 in deriving the estimate ( [ good_estimate ] ) we have used the leibnitz rule , in the third equality . the analogous semi - discrete energy is now @xmath52 however this semi - discrete energy will not , in general , satisfy a discrete version of the continuum estimate given in ( [ good_estimate ] ) . the reason is that although @xmath9 satisfies sbp , it does _ not _ generally satisfy the leibnitz rule , which was necessary in deriving the estimate ( [ good_estimate ] ) . it can actually be checked that , because of this , discretizing the system as in eq . ( [ 1da ] ) will _ not _ reproduce the continuum estimate , i.e. @xmath53 an alternative is to rewrite the evolution equation ( [ 1d ] ) such that the leibnitz rule is unnecessary to obtain the continuum estimate , eq . ( [ good_estimate ] ) . then a semi - discrete version of this estimate does hold if the difference operator satisfies sbp . to do so we write eq . ( [ 1d ] ) as @xmath54 this gives the same estimate as before : @xmath55 ^ 2 \le 0 . \label{est1}\ ] ] however , in deriving the estimate ( [ est1 ] ) we have only used integration by parts , and not the leibnitz rule . therefore , the semi - discrete energy estimate @xmath56 follows immediately if the semidiscrete equation is written as @xmath57 and @xmath9 satisfies sbp . figure [ comparison_discretizations ] shows numerical results obtained by evolving eq . ( [ 1d ] ) using two discretizations , given by eqs . ( [ 1da],[1dc ] ) . ( [ 1d ] ) and ( [ rearrange1 ] ) are equivalent , they have the same set of solutions . however , their discrete counterparts , eqs . ( [ 1da ] ) and ( [ 1dc ] ) , _ do not _ have the same solutions at fixed resolution . in one case the solution error grows as a function of time , in the other case it does not . it is experimentally found that in this example ( and others discussed below ) the growing numerical error in the non - strictly stable discretization is manifest in high frequency modes , which can be eliminated by adding a small amount of dissipation , as is also shown in figure [ comparison_discretizations ] . ) . the dotted line shows the one of the lines of fig.1 , i.e. , the @xmath36 norm of the error for a non - strictly stable discretization , given by eq . ( [ 1da ] ) . the solid line shows the error for the same simulation with some numerical dissipation ( @xmath58 ) . the dashed line , in turn , shows the error for the ( non - dissipative ) strictly stable discretization given by eq . ( [ 1dc ] ) . in all these runs @xmath59 and cfl=0.8 . even though both the strictly stable and the non - strictly but dissipative schemes give long term stability , the former discretization is much more accurate.,height=226 ] consider now an equation with variable - coefficients in the principal part and a lower order term that in principle allows for energy growth , @xmath60 defined on a compact domain @xmath0 $ ] . for simplicity we assume that @xmath61 , and define the energy @xmath62 . homogeneous boundary conditions are given at @xmath63 . rewriting the equation as @xmath64 , one can easily show , using only integration by parts , that the estimate @xmath65 holds . therefore , writing the semi - discrete equation as @xmath66 gives the same semi - discrete energy estimate , @xmath67 if @xmath68 then @xmath69 and the continuum and discrete solutions can not grow in time . equations with features similar to those of eq.([shift_eq ] ) with @xmath68 appear in the klein gordon and maxwell fields on fixed black hole backgrounds , as discussed later in this paper , where the shift plays the role of @xmath70 . if @xmath71 the energy may grow , already at the continuum ; for example , if @xmath72 . in that case one may want to discretize in a way such that artificial numerical growth is avoided as much as possible . for example , if @xmath72 , discretizing as in eq . ( [ discr ] ) gives @xmath73 which is the same equation satisfied by the continuum energy . fast growing numerical errors can arise even in numerically stable discretizations . in some cases they can be traced to the failure of the leibnitz rule for discrete derivatives , and failure of sharp continuum energy estimates to hold in the discrete problem . the use of difference operators satisfying sbp , appropriate numerical boundary conditions , and rearrangement of terms , give raise to strictly stable schemes which satisfy sharp semidiscrete energy estimates , if such estimates are available at the continuum . the goal of these rearrangement of terms is actually not only to avoid the leibnitz rule . clearly , since the growth rate for the energy usually involves derivatives , in general there would still be a discrepancy between the analytical and numerical estimates even if the leibnitz rule was satisfied , arising from the difference between numerical and analytical derivatives . these differences , however , can also be controlled by appropriately rearranging terms ( see , for example , sections [ wave_eqs]-[max_eqs ] , and refs . @xcite ) . in the following we discuss more complex examples : the propagation of scalar and electromagnetic fields on the schwarzschild spacetime in three dimensions , and one - dimensional linear perturbations of gauge waves . in the former cases , we do have readily available sharp energy estimates derived from physical principles . in the latter a sharp estimate is available for the subsidiary system describing the constraint evolution . in this section we describe some model problems commonly used in relativity , while corresponding numerical results are presented in the following section . the first two problems are klein gordon and maxwell fields on a schwarzschild black hole background , and linearization perturbations of the gauge wave is the third one . black holes contain physical curvature singularities that can not be included on the computational domain . black hole excision eliminates this problem by removing a region around the singularity from the computational domain @xcite . this region must lie within the event horizon , and be _ causally disconnected _ from the spacetime outside the black hole so that the excised region does not affect the physics outside the hole . as discussed in the appendix , this region must be carefully chosen for these conditions to hold . numerical implementations of excision have traditionally presented a challenge in deciding how points near the inner boundary are updated . especially in higher dimensions , many combinations of one - sided derivative stencils , interpolation , and extrapolation have been investigated , mostly through experimentation @xcite ( for an alternative excision strategy using pseudospectral methods see @xcite ) . however , as explained in paper i , derivative operators that satisfy sbp together with a diagonal inner product uniquely specify the excision algorithm . the procedure is unambiguous , and for linear problems guarantees a stable implementation . we follow this procedure , detailed in paper i , to excise the singularity . we consider three - dimensional evolutions on a schwarzschild background , in either painlev ' e gullstrand ( pg ) and kerr - schild ( ks ) coordinates . the four - dimensional metric , written in terms of adm quantities , is @xmath74 in pg coordinates @xmath75 while in ks coordinates @xmath76 where @xmath77 and @xmath78 is the black hole s mass . we excise a cubical excision region centered on the black hole , and require the length of the cube , @xmath79 , to satisfy @xmath80 , or @xmath81 . larger cubes , as discussed in the appendix , result in excision boundaries that are _ not _ purely outflow . the massless klein gordon scalar wave equation , @xmath82 , where @xmath83 is the inverse four - dimensional metric , can be be written as @xmath84 where we have introduced new variables @xmath85 and @xmath86 to write the system in first order form . additionally , we define @xmath87 , where @xmath88 is the inverse of the three - metric , @xmath89 , and @xmath90 . this formulation has a number of desirable features when the background metric is stationary @xcite . for example , eqs . ( [ eq : wave1 ] ) are symmetric with respect to the physical energy , and a straightforward estimate shows that this energy does not grow when maximally dissipative boundary conditions , which are automatically constraint - preserving for this formulation , are given . moreover , eqs . ( [ eq : wave1 ] ) have been written such that the leibnitz rule is not needed to guarantee that the energy will not grow . for this reason , we refer to the discretization obtained by replacing @xmath91 by @xmath92 in eqs . ( [ eq : wave1 ] ) as _ energy preserving _ or _ strictly stable_. the physical energy , however , is not positive definite inside the black hole , since the killing vector becomes space - like there . we nevertheless choose this energy - preserving discretization , since it guarantees long - term stability if the computational domain does not include the black hole . more sophisticated formulations and discretizations , which are globally symmetric hyperbolic and conserve the physical energy in the exterior region , up to a distance arbitrarily close to the event horizon , can also be constructed @xcite . as a second example of classical fields on a black hole background we choose the vacuum maxwell equations . we write the maxwell equations in terms of the faraday tensor , @xmath93 , and its dual @xmath94 } & = & 0 \nonumber \\ \nabla_{[a } { } ^{\star}f_{bc ] } & = & 0 . \end{aligned}\ ] ] the dual of the faraday tensor is @xmath95 , where @xmath96 is the completely anti - symmetric symbol . contracting the above equations with a time - like vector field @xmath97 we get the evolution equations : @xmath98c}u^c),\nonumber \\ { \cal l}_u { } ^{\star}f_{ab } & = & - 2 \nabla_{[a}({}^{\star}f_{b]c}u^c ) . \end{aligned}\ ] ] defining a foliation by dragging an initial space - like hypersurface along the integral curves of the vector @xmath97 ( each labeled with the proper time of the integral lines of @xmath97 , @xmath99 ) , and pulling back the above equations onto each of these hypersurfaces we get @xmath100c}u^c),\nonumber \\ { \cal l}_u \phi_{\star}({}^{\star}f_{ab } ) & = & - 2 \partial_{[a}\phi_{\star}({}^{\star}f_{b]c}u^c ) , \end{aligned}\ ] ] where @xmath101 is the pull - back map into the hypersurface . notice that only exterior derivatives appear in the equations , so the geometry only enters through the star operation . the electric and magnetic fields are typically defined as @xmath102 , and @xmath103 , allowing one to decompose @xmath93 and @xmath104 into ( @xmath105 defined by @xmath106 ) @xmath107 } + { \varepsilon}_{ab}{}^{cd } b_c n_d , \nonumber\\ \ , { } ^{\star}f_{ab } & = & 2 b_{[a } n_{b ] } + { \varepsilon}_{ab}{}^{cd } e_c n_d . \end{aligned}\ ] ] contracting eqs . ( [ eq : max_evol_2 ] ) with @xmath108 , we get @xmath109 as with the klein - gordon equation , this system is symmetric hyperbolic outside the event horizon , where @xmath97 is time - like , and only strongly hyperbolic inside the event horizon , where @xmath97 is _ space - like _ , since the symmetrizer is not positive - definite there . we supplement the above equations with a no - incoming radiation boundary condition . no condition is needed to preserve the constraints , for they propagate along @xmath97 , as are the outer boundaries . thus , we set the incoming modes to zero at a boundary with normal @xmath110 , via the projector defined as @xmath111 where @xmath112 is the eigenvector basis , @xmath113 is the co - basis , we define @xmath114 and where the sum extends only over eigenvectors with positive eigenvalues . as with the scalar field , we ask whether the maxwell equations can be written such that the discrete problem has a sharp energy estimate , giving numerical solutions without alti . the background spacetime is invariant under time translations , implying a conserved quantity or energy . this energy is the integral of @xmath115 , where @xmath97 is the killing vector field ( timelike outside the black hole ) over the hypersurface . for the pg foliation this energy is @xmath116\ , dv.\ ] ] just as with the klein gordon equation , this expression is not positive - definite inside the horizon , reflecting a property of the geometry , not of the equation or coordinates used . to discretize the above system we use the method of lines and substitute in eqs . ( [ max1],[max2 ] ) the partial derivatives with difference operators satisfying sbp . in this way strict stability with respect to the discretized version of the above energy is attained . one simple and common test for numerical implementations of the einstein equations is a gauge wave defined by @xmath117 which describes flat spacetime with a sinusoidal coordinate dependence , of amplitude @xmath118 , along the @xmath119 direction . the non - trivial adm variables for this metric are @xmath120 together with the gauge condition @xmath121 we evolve the einstein equations using the symmetric hyperbolic formulation presented in @xcite with dynamical lapse and the time - harmonic gauge . the formulation is cast in first order form by introducing the variables @xmath122 and @xmath123 . rather than solving the full einstein equations here , we study linear perturbations of a background metric given by eq . ( [ gaugewavemetric ] ) . furthermore , we simplify the system by allowing perturbations depending only on @xmath124 , in the form @xmath125 this restriction can be justified by noting that full non - linear evolutions of data defined by eq . ( [ gaugewavemetric ] ) show that only these variables vary dynamically @xcite , and variations in all other variables are consistent with round - off error . the resulting linearized equations are ( henceforth dropping the @xmath126 notation ) @xmath127 where we have defined @xmath128 . this system is symmetric hyperbolic ; three characteristic speeds are @xmath129 and the other two @xmath130 . we consider here an initial boundary value problem on the domain @xmath131 $ ] . we implement boundary conditions via the orthogonal projector method described in appendix d of paper i. the positive / negative eigenmode decomposition of the matrix @xmath132 is @xmath133 ( see paper i ) . here , the matrix @xmath118 is the principal part , and @xmath29 is its symmetrizer . for one - dimensional problems , @xmath134 , and @xmath135 is the matrix that diagonalizes @xmath118 , i.e. , @xmath136 , with @xmath137 a diagonal matrix . the system must satisfy two constraint equations , corresponding to the definition of the variables @xmath138 and @xmath139 . when linearized , these constraints are @xmath140 the other constraints are automatically satisfied owing to the restricted form of the allowed perturbations . despite the simplicity of this system , the numerical integration of these equations reveals some interesting features . indeed , straightforward discretizations of this system have lead to exponentially growing solutions with quickly growing constraint violations . that this instability has been observed by different groups in a variety of situations , led to the speculation that constraint violations _ drive _ the instability . we offer here a different interpretation of the results by examining the evolution of the constraints . evolution equations for the constraints can be obtained by taking the time derivative of @xmath141 and @xmath142 , and substituting in the einstein equations , giving @xmath143 these equations can be integrated in closed form to yield @xmath144 where @xmath145 and @xmath146 are free functions determined by the initial data : @xmath147 clearly the constraint behavior is _ purely oscillatory_. hence , any growth in the constraints observed during evolution must be purely numerical . we will show in the next section that this is indeed the case , and that the error growth can be eliminated . we now present some numerical simulations of the model problems discussed in the previous section . for the black hole tests , we examine configurations where the computational domain is outside the black hole and others that do contain an excised black hole . the former configuration allows us to investigate the evolution of a symmetric hyperbolic system with boundary interactions without the difficulties introduced by excision . we then examine the more demanding case in which excision is required , and the evolution systems here considered are only strongly hyperbolic inside the black hole . this section presents numerical results for the massless klein gordon scalar field on a schwarzschild black hole background . we present some tests of our code , and then discuss differences between the strictly stable and non - strictly stable discretizations . consider an uniform computational domain that _ does not _ contain the black hole , @xmath148 ^ 3 $ ] . initial data with compact support are given to @xmath149 , the time derivative of the scalar field , as @xmath150 ^ 4 \label{eq : initial_data_1 } \end{aligned}\ ] ] when @xmath151 ^ 3 $ ] , and zero otherwise . all other fields are initially set to zero . is of order one despite the large factor in its definition , @xmath152 . ] this data is then evolved using two possible discretizations . the first one is that one obtained by direct substitution of partial derivatives by discrete ones in eq . ( [ eq : wave1 ] ) , this is the strictly stable discretization . the second one is obtained by expanding all derivatives of products with the product rule and then replacing partial derivatives of field variables by their discrete counterparts . we refer to this as the `` na " i ve '' discretization . both discretizations were used to evolve for around 250 crossing times , under different choices of courant factor ( @xmath153 ) . in all cases we found no sign of long term growth when monitoring the variables , though a closer inspection reveals several salient features of the obtained solutions . namely , a self - convergence analysis show results consistent with second order accuracy for a few crossing times ( with better values for the smallest @xmath154 ) . however , at later times , the observed convergence rate deteriorates considerably and achieve artificially large values . the time at which this behavior starts coincides with the fields mostly leaving the computational domain . for instance , figure [ conserve_straight_outside ] displays the @xmath36 norm of @xmath149 obtained with the strictly stable and `` na " i ve '' discretizations . the figure shows the @xmath36 norms of @xmath149 calculated at two resolutions , for 250 crossing times . the large decrease in the norm is a sign that basically most of the fields leave the domain after a few crossing times . both discretizations give a long - term stable algorithm for this case . but , as we will discuss later , the nave discretization without dissipation becomes long - term unstable when the black hole is included on the domain . norm of @xmath149 in a scalar field evolution outside the black hole for about 250 crossing times . the figure compares results for a nave discretization ( see text ) , and a strictly stable discretization at two resolutions , @xmath155 and @xmath156 . the domain is @xmath157 $ ] , @xmath158 , and no dissipation is added to the solution . , width=321 ] to verify that we get the expected self - convergence factor in long term simulations when the fields do not decay to very small values after having left the domain , we now show results of a complementary set of tests on the domain @xmath159 $ ] , @xmath160 $ ] . here we concentrate in what corresponded to the worst case scenario considered , given by @xmath161 . we set the incoming modes on all outer faces to zero , except on the @xmath162 face , where the time derivative of the incoming fields is set to @xmath163 $ } & \nonumber \\ 0\qquad\qquad\qquad\mbox{otherwise } & { } \nonumber \end{array } \right . \label{eq : incoming_modes } \end{aligned}\ ] ] this boundary condition is imposed through an orthogonal projection , as explained in paper i. initial data is given by @xmath164 ^ 4 \ ; , \label{eq : initial_data_1b } \end{aligned}\ ] ] and both discretizations ( strictly and non - strictly stable ) are evolved . figure [ long_conv ] shows the self - convergence factor ( as before , in the @xmath36 norm ) of @xmath165 and @xmath149 versus time for the strictly stable discretization ( similar factors are obtained for the non - strictly stable one ) at the resolutions @xmath166 , @xmath167 and @xmath168 . convergence factors close to the expected value of two are observed , with no qualitative difference in the convergence factor as time progresses . ( solid line ) and @xmath165 ( dashed line ) on a domain outside the black hole , using the strictly stable scheme ( similar factors are obtained for the naive one ) , for about @xmath169 crossing times . these factors are calculated using resolutions of @xmath170 , @xmath167 , and @xmath168 . the obtained factors are close to the expected value of two . the courant factor is @xmath171 , and no dissipation is used . , height=226 ] in this section , the computational domain contains the black hole , which is excised with an inner cubic boundary centered on the black hole . the domain is defined on @xmath172 ^ 3 $ ] , and the excision cube has a total length of @xmath173 . thus , the faces of the inner boundary are at @xmath174 . we first verify that the excision algorithm does not display alti when employing the strictly stable scheme . as mentioned earlier , very little freedom is available in constructing difference operators that satisfy sbp . indeed , for our second - order code with a diagonal scalar product , the stencils are uniquely specified . figure [ long_excision ] shows the @xmath36 norms of @xmath149 and @xmath165 for a long up to @xmath175 evolution ( with no dissipation ) , designed to detect slowly growing modes . as shown in this figure , the algorithm specified by requiring sbp and the equations arranged so as to preserve the energy is long - term stable even with excision boundaries . for this run non - trivial initial data are given , as before , only to @xmath149 . @xmath176 ^ 3 ( r0-r ) & \mbox{if $ |r - r_0|\le |\delta |$ } \\ 0 & \mbox{otherwise \ , . } \end{array } \right . \end{aligned}\ ] ] with @xmath177 and @xmath178 . when studying the convergence of the code one again faces the fact that after a few crossing times the convergence value become meaningless . the behavior observed is analogous to the case where the black hole was outside the computational domain . namely , close to second order convergence is seen during the first few crossing times , and unrealistically high values are obtained afterwards . in fig . [ convergence_excision ] we show a short - term convergence test ( about two crossing times ) with excision . fig . [ norm_excision ] shows the corresponding @xmath36 norms for @xmath165 and @xmath149 . norms of @xmath149 ( solid line ) and @xmath165 ( dashed line ) are shown for about 2500 crossing times . no slowly growing unstable modes are detected in these runs . this run was performed with @xmath179 grid points , @xmath180 , and no dissipation is used . , height=226 ] ( solid line ) and @xmath165 ( dashed line ) using the same configuration as in fig . [ long_excision ] . the self - convergence test is performed on uniform grids with @xmath181 , and @xmath182 grid - points . the convergence factor is close to the expected value of two . the measured convergence factor drops slightly from near two after one crossing time , when most of the field has left the domain and afterwards increases to unrealistic high values . , height=226 ] norms for the runs shown in fig.[convergence_excision ] . the relative difference in the norms for @xmath165 at any two different resolutions , is at most about one percent . similarly for @xmath149 , except at the last point shown , where this field has mostly left the grid . , title="fig:",height=226 ] norms for the runs shown in fig.[convergence_excision ] . the relative difference in the norms for @xmath165 at any two different resolutions , is at most about one percent . similarly for @xmath149 , except at the last point shown , where this field has mostly left the grid . , title="fig:",height=226 ] as in the previous case , we also compare results obtained with the strictly stable and the nave discretizations . long term stability was found for both cases when the computational domain was outside the black hole . with excision , however , the solution obtained with the nave discretization quickly becomes corrupted , as shown in fig . [ conserve_straight_inside ] . as shown in this figure , the amplitude decreases with resolution , indicating that the growth is spurious and that the nave discretization suffers from alti . the strictly stable discretization , however , with a sharp semi - discrete energy estimate , remains well - behaved . the alti that contaminate the navely discretized solution are found to be of high - frequency type , which can be managed by adding explicit dissipation . figure [ conserve_straight_dissip ] shows results using the same initial configurations , except that now dissipation is added to the nave discretization . now all of the runs produce long term stable simulations . norm of @xmath149 in a scalar field evolution on a domain containing an excised schwarzschild black hole . the solid and dot - dashed lines show the @xmath36 norm of @xmath149 obtained using the nave discretization of the wave equation . although the amplitude decrease with resolution , this solution quickly becomes long - term unstable , indicating an artificial instability . the strictly stable discretization ( shown with dashed and dotted lines ) , however , remains long term stable and well - behaved . the results are calculated with the resolutions @xmath183 and @xmath184 , @xmath158 , and no dissipation is added . , height=226 ] norm of @xmath149 in a scalar field evolution on a domain containing an excised schwarzschild black hole . the solid and dot - dashed lines show the solutions obtained using the nave discretization with dissipation , the latter eliminating the artificial instability noted in fig . [ conserve_straight_inside ] . the strictly stable discretization ( shown with dashed and dotted lines ) , without dissipation , is shown for comparison . the results are calculated with the resolutions @xmath183 and @xmath184 , @xmath158 , and the dissipation parameter is @xmath185 . , height=226 ] as with the scalar field tests above , we perform tests with the maxwell equations on two kind of domains . specifically , we choose 1 . computational domain outside the black hole : @xmath186 ^ 2 $ ] and @xmath187 $ ] ; 2 . computational domain including part of the black hole : @xmath188 ^ 3 $ ] . the singularity is excised by means of a cubical region defined by @xmath189 ^ 3 $ ] . the initial data describe a `` pulse '' of electromagnetic field given by @xmath190 ^ 8 \epsilon_{ij3 } ( x^j - x_o^j ) & \mbox{if $ |r - r_0| < \delta$ } \\ 0 & \mbox{otherwise \ , , } \end{array } \right . \end{aligned}\ ] ] where @xmath191 stands for @xmath192 and @xmath193 , @xmath194 is a strength parameter , @xmath195 is the location of the pulse s center , and @xmath126 its width . we discretize in a strictly stable way , as explained in section [ max_eqs ] . let @xmath196 , @xmath197 and @xmath198 , and we set the dissipation to zero . figure [ maxwell12 ] shows @xmath199 vs. time for four different resolutions ( @xmath200 and @xmath182 grid - points , corresponding to @xmath201 and @xmath168 , respectively ) , while the inset shows the decay in time for a long - term run with the coarsest resolution . figure [ maxwell11 ] shows the self - convergence factor of @xmath202 for a short time , computed with @xmath179 , @xmath203 and @xmath182 grid - points , and @xmath203 , @xmath204 , and @xmath182 grid - points . as expected , the factor gets closer to two when resolution is increased . norm of @xmath202 vs. time in an evolution of the maxwell fields at four different resolutions for a strictly stable discretization . the computational domain is outside the black hole . the inset shows a long - term evolution , @xmath205 , using the coarsest resolution . the runs were performed on uniform grids with resolutions @xmath170 , @xmath167 , @xmath206 and @xmath168 , and the courant factor is @xmath207 . , height=226 ] for the runs of fig . [ maxwell12 ] . the factors improve with resolution , in the sense that they closer to the expected value of two . the labels @xmath37-@xmath208-@xmath209 and @xmath208-@xmath210-@xmath209 indicate the number of grid points in each direction used to perform the self - convergence test . , height=226 ] now let @xmath196 , @xmath211 and @xmath212 . in contrast to the scalar field runs , solutions of the maxwell equations without explicit dissipation diverge at long times , even with a strictly stable discretization . the reason for this appears to be that the energy inside the black hole is not positive definite and boundedness ( either at the continuum or numerical level ) is not guaranteed ( it is not clear why this divergence does not appear in the scalar field case , when the singularity is excised , since the energy is also non - positive definite in that case . ) . high frequencies are seen in the solution , and we therefore add a small amount of dissipation , @xmath213 . figures [ maxwell22 ] and [ maxwell21 ] show the norms and convergence data similar to the figures for the domain outside the black hole . as before , the simulations are long - term stable for long times . the self - convergence factors for the coarsest resolutions drop quickly from their initial value , indicating that the coarsest solutions are not in the convergent regime . in addition to effects caused by the fields leaving the computational domain , using a cubical excision boundary forces us to excise very close to the black hole , where the spacetime and field gradients are very large , effectively requiring higher resolution . norm of @xmath202 vs. time in an evolution of the maxwell fields on a computational domain containing an excised black hole for a strictly stable discretization . the norms are shown for four different resolutions , and the inset shows a long - term evolution , @xmath205 , at the coarsest resolution . the runs were performed on uniform grids with resolutions @xmath170 , @xmath167 , @xmath206 and @xmath168 . the courant factor is @xmath180 , and dissipation is added to the evolution equations , with @xmath213 . , height=226 ] . the self - convergence factors are calculated from @xmath202 . the excision boundary is very close to the black hole , and the large gradients encountered in the fields and spacetime near the excision boundary require very high resolution to resolve them adequately . the coarsest solutions are not resolved well enough to be in the convergence regime . however , it can be seen from the figure that the convergence improves with resolution . the labels @xmath37-@xmath208-@xmath209 and @xmath208-@xmath210-@xmath209 indicate the number of grid points in each direction used to perform the self - convergence test . , height=226 ] the gauge wave is a useful test problem for examining some of the difficulties found in the simulation of the einstein equations ( see @xcite ) , and to illustrate the usefulness of the techniques presented in paper i. to this end , we present results obtained using a straight - forward discretization of the right hand side of eqs . ( [ gauge1])([gauge5 ] ) , and show results obtained for long term evolutions with and without dissipation . moreover , we consider cases both where the constraints are initially satisfied , and where they are not . the latter illustrates that even when large constraint violations are introduced in the initial data , the solutions to the linearized equations remain well behaved . this , of course , may not necessarily be the case for the non - linear einstein equations , as for these relatively large amplitudes the linearized system need not be a close approximation to the non - linear ones . or even if the amplitudes were small , there could be purely non - linear effects in the solution which would naturally be missed in a linearized analysis @xcite . for these tests we adopt trivial initial data for all fields , except for @xmath214 , @xmath138 and @xmath215 , which have compact support in @xmath216 $ ] and are given by @xmath217 ^ 4 , \\ d_{xxx } & = & \mu \partial_x g_{xx } , \\ k_{xx } & = & -\frac{\xi}{2 } \partial_x g_{xx } \ , , \end{aligned}\ ] ] where @xmath218 , @xmath219 , and @xmath220 are constants . we set @xmath221 in the background solution and the computational domain is taken to be @xmath222 $ ] . boundary conditions are applied via olsson s projection operators . here , as we are interested in the long term behavior of the solution , we couple the incoming modes the outgoing ones such that @xmath223 for cases where the constraints are initially satisfied , we choose @xmath224 , @xmath225 ; otherwise we set @xmath226 , @xmath225 and @xmath227 . in all these simulations a courant factor of @xmath180 is used , to compare with similar simulations @xcite . figures [ gaugewave_c_straight]-[gaugewave_g_straight ] show results for constraint - satisfying initial data , without dissipation . figure [ gaugewave_c_straight ] shows the @xmath36 norm of the constraint , @xmath228 and its convergence to zero with increasing resolution . initially the norm @xmath228 is almost constant . however , at later times a clear exponential mode is observed . while the solution is well - behaved for longer times with increasing resolution , this approach is impractical for long runs . figure [ gaugewave_con_straight ] shows the convergence ( to zero ) factors in the @xmath36 norm obtained for @xmath228 . for the highest resolutions , the factor is very close to its expected value . an exponential behavior for @xmath214 is also seen in figure [ gaugewave_g_straight ] . even though in principle this quantity could already grow at the continuum , the fact that at fixed time the norm of @xmath214 decreases with increasing resolution suggests that , as discussed below , this is an artifact of an alti . norm of @xmath228 vs. number of crossing times for four different resolutions . after an initial stage where the constraints are well behaved a clear exponential mode dominates the solution.,height=226 ] . [ gaugewave_c_straight ] as a function of number of crossing times , comparing runs with @xmath37 and @xmath208 points ( dotted line ) , @xmath208 and @xmath209 points ( dashed line ) and @xmath209 and @xmath229 points ( solid line ) . notice that even when the solutions are growing exponentially ( see figs.[gaugewave_c_straight],[gaugewave_g_straight ] ) , the constraints converge to zero.,height=226 ] . [ gaugewave_con_straight ] norm of @xmath214 vs. number of crossing times for four different resolutions with no dissipation . at early times , the norm remains fairly constant but at later ones an exponential growth dominates the solution.,height=226 ] . [ gaugewave_g_straight ] the alti for @xmath228 are , as in the previous examples , found to be of high frequency type and , therefore , the addition of dissipation eliminates these growing modes . figures [ gaugewavediss_g_straight][gaugewavediss_con_straight ] illustrate the behavior of simulations with the same initial data and boundary conditions as those of figures [ gaugewave_c_straight]-[gaugewave_g_straight ] , except that now some dissipation is added , @xmath58 . the norm of @xmath142 is illustrated in fig . [ gaugewavediss_c_straight ] , where now only a slow growth is observed at late times . [ gaugewavediss_con_straight ] shows , as before , the convergence factors ( to zero ) for @xmath142 , in the @xmath36 norm . finally , figure [ gaugewavediss_g_straight ] shows the @xmath36 norm of @xmath214 for different resolutions . only a linear growth is observed in the solution is found ; furthermore , this growth decreases with resolution . norm of @xmath214 vs. number of crossing times for four different resolutions . only a small growth is observed in the solution.,height=226 ] . [ gaugewavediss_g_straight ] norm of @xmath228 vs. number of crossing times for four different resolutions . convergence to zero is observed and only slow ( linear ) growth is observed.,height=226 ] . [ gaugewavediss_c_straight ] vs. number of crossing times comparing runs with @xmath37 and @xmath208 points ( dotted line ) , @xmath208 and @xmath209 points ( solid line).,height=226 ] as a last example we consider initial data violating the constraints . here , similarly good behavior is observed when dissipation is employed , in the sense that , as predicted by the continuum analysis of section [ eqs_gauge ] , the constraints oscillate in time . for instance , figure [ gaugewavediss_con_straight ] shows the @xmath36 norm of the constraint @xmath141 vs. time . as expected , these converge to a roughly constant behavior as resolution is increased , and no growth is observed . norm of @xmath230 vs. number of crossing times for three different resolutions , corresponding to the case where the initial data does not satisfy @xmath231 . as expected from the continuum analysis , even in this case the norm does not grow in time . , height=226 ] the construction of stable numerical methods for ibvps often involves a mixture of intuition and experimentation . the choice of numerical scheme and boundary conditions are often considered as necessarily separate ingredients . for certain problems , however , the energy method for discrete systems is an integrated , systematic method for creating numerically stable schemes and analyzing their behavior . while this method has not yet been widely used in general relativity , many current outstanding problems in the field vitally depend on understanding the subtle interplay of boundaries with the interior computational domain . in this paper we have applied the discrete energy method to test problems commonly found in numerical relativity and analyzed the numerical and long - term stability of the resulting schemes , with and without black hole excision . we have also investigated the numerical origin of fast growing modes which may artificially appear in numerical solutions , and used the energy method to discuss possible remedies . in particular , we have presented explicit examples of the techniques described in paper i and discussed what leads to alti and a way to remove or alleviate them . a message from this work is that it is not sufficient to construct numerically stable implementations , but also special care must be taken to devise suitable _ numerical _ methods to rule out , or alleviate alti . ( for a somewhat related approach , see the recent work @xcite ) . to illustrate how this is achieved we have adopted a series of examples , which although considerably simpler than einstein equations , are relevant in the context of numerical relativity . in particular , in efforts to deal with black hole spacetimes and to understand the source of problems encountered with the gauge wave test . as shown throughout this work , even though there are examples in which low - frequency alti are present and for which no sharp estimate is known , a careful analysis and grouping of the variables and/or the addition of ( controllable ) artificial dissipation may in many cases be crucial in handling otherwise artificial long term instabilities . it is expected that these techniques will be important in a number of fronts , in particular in the simulation of einstein s equations . we wish to thank g. calabrese , o. sarbach , j. pullin , e. tadmor , p. olsson , and h .- o . kreiss for helpful discussions , and e. schnetter for helpful comments on the manuscript . this work was supported in part by nsf grants phy0244699 , phy0312049 ; nasa - nag5 - 1430 ; the horace hearne jr . lab for theoretical physics , conicet , and secyt - unc . computations were done at lsu s center for computation and technology and parallelized with the cactus toolkit @xcite . is an alfred p. sloan fellow . black hole excision removes the singularity by placing an inner boundary on the computational domain . while the boundary removes the singularity , the question now becomes finding physically and mathematically consistent inner boundary conditions . clearly , a general solution to this problem for an arbitrary boundary is unknown . the time - development of this boundary presents a further complication . namely , if the boundary becomes time - like , it will encounter the singularity in a finite amount of time . these two issues can be resolved simultaneously by adopting an inner boundary with a space - like world tube . this ensures it will not hit the singularity and that no boundary conditions are necessary if all the characteristics speeds are within the light cone , as the past domain of dependence of boundary points is wholly contained on the previous time slice . some freedom remains in choosing the particular region to be excised , and clearly the largest possible excision boundary is preferable to avoid the strongest gradients of the spacetime . to illustrate some of these issues we now examine cubical excision for both the schwarzschild and kerr black holes , looking for the _ largest possible _ space - like excision cube . as we discuss below , the _ shape _ of the boundary plays a crucial role in satisfying the spacelike condition . we examine one face of the cube , a @xmath232 face . the condition that this surface is space - like implies that @xmath233 , as @xmath234 is the normal to the surface . since the coordinates are fixed _ a - priori , _ we can only vary the position of the boundary to meet this condition . we find that this condition is quite difficult to satisfy , and possible for only very slowly spinning black holes . to see this in detail , consider the kerr metric in kerr schild form @xcite . the quantity @xmath235 is @xmath236 where @xmath237 is related to the cartesian coordinates by @xmath238 the space - like character requirement on the surface implies that @xmath239 for the schwarzschild black hole , @xmath240 , @xmath241 , and the above condition becomes @xmath242 the corner of the cube is the limiting point ( @xmath243 ) for the largest cube , giving the bound @xmath244 , or @xmath245 . for the spinning black hole , the analysis is slightly more involved . the limiting point for the largest cube is given by @xmath246 , @xmath247 , giving the condition @xmath248 examining this condition shows that cubical excision can not be used when @xmath249 , as shown in fig . [ kerr_excise ] . furthermore , the singularity is not at a point , but is ring - like , implying that the cube may not be arbitrarily small . finding the minimum sized cube is more involved , as the faces of the cube need to be examined in addition to the corners @xcite . this limitation of cubical excision for the schwarzschild spacetime was noted before @xcite , by requiring that the excision boundary has no in - coming modes in a particular formulation of einstein equations . a similar analyzis to that of @xcite for the kerr spacetime and cubical excision regions has also recently presented in @xcite . here we would like to point out the following . first , when the eigenmodes are complete and the speeds are physical ( they lie within the null cone ) those results and the one presented here are equivalent , as it is rooted in the physical nature of the excision boundary . second , when dealing with a formulation admitting eigenspeeds outside the light cone , or a non - complete set of eigenmodes , the current straightforward analysis provides necessary conditions that have to be met . these necessary conditions are useful when dealing with any formulation , in particular one whose eigenspeeds might not be known . finally , we emphasize that these results are coordinate dependent ; nevertheless , for the families often used in numerical relativity tests , the same result applies . in particular for the family of slicings reported by martel and poisson @xcite ( which contains both the kerr schild and painlev ' e gullstrand slicings of schwarzschild ) . h. friedrich and a. rendall , in _ einstein s field equations and their physical implications _ , lecture notes in physics , edited by b. g. schmidt ( springer - verlag , berlin , 2000 ) , pp . 127 - 223 ; o. reula , living rev . * 1 * , 3 ( 1998 ) . m. scheel , talk at `` miniprogram on colliding black holes : mathematical issues in numerical relativity '' , institute for theoretical physics , university of california at santa barbara , january 10 - 28,2000 .

particle gels form in suspensions if short - range attractions act between contacting surfaces of particles . depending on the strength of the attraction , one can distinguish them as two types : strongly and weakly aggregated particle gels @xcite . in strongly aggregated particle gels , on which we focus in this work , particles once joined hardly separate at room temperature @xcite . the attractions also prevent tangential displacements of contacting particles @xcite , thus particles form sample - spanning loose networks . their remarkable and complex mechanical properties under external stress are essentially due to the response of a deformable disordered network formed with all the particles . under shear stress , one can sketch out deformation of a particle gel as follows @xcite : * the system is elastic under very low stresses . if the stress is released , the system comes back to the original shape ; * for larger stress , a particle gel may undergo finite deformations , though it does not flow . this is the plastic regime , and it gives malleability ; i.e. , the system mostly keeps the deformed shape if stress is released ; * beyond a yield stress , the particles network is ruptured and the particle gel flows . fluids exhibiting these behavior are called _ yield stress _ ( or _ bingham plastic _ ) fluids @xcite . since the yield stress is the highest stress that can prevent flow , it can be considered as the strength of particle gels . naturally , this strength relates to the strength of the particle connections . in a particle gel under shear stress , transmitting stresses break a portion of particle connections and the sample spanning network is torn apart . the yield stress of particle gels is higher when the packing fraction of particles is larger . in many cases , this variation appears to follow a power law with exponents between 4 and 5 @xcite . under compressive stress , the behavior of a particle gel looks different , due to confinement @xcite . compressive stress appears naturally in a number of applications , for example paste drying @xcite or ceramic processing @xcite . osmotic compression is a generic important case . in this kind of compression process , liquid removal tends to pack the particles into a smaller volume . as long as the applied stress is lower than the strength of the gel , it will retain the volume . if the stress exceeds a compressive yield stress @xmath2 , network restructuring leads to the loss of a part of the liquid phase , and particles reorganize into a smaller volume @xcite . the deformation stops when the mechanical resistance of the network equals the stress . thus , unlike the shear case , deformation does not continue with a constant applied stress @xcite . the growth of mechanical resistance , i.e. compressive consolidation , also shows a highly nonlinear dependence on the packing fraction , typically power - law relations @xmath3 with exponents between 3.2 and 4.5 @xcite . the highly nonlinear dependency of the compressive yield stress @xmath2 on the packing fraction @xmath4 relates to the heterogeneity of the network structure of particles @xcite . the fractal model @xcite successfully captures such packing - fraction dependence of rheological proprieties by considering the characteristic structure of strongly aggregated particle gels . in the model , a particle gel is assumed to be an assembly of fractal flocssuch structural correlation is evidenced by scattering observations @xcite . the entire sample is filled up with flocs of a fractal dimension @xmath0 , and the correlation length @xmath1 is equivalent or proportional to the average size of the flocs . thus , the correlation length @xmath1 can be associated with the packing fraction @xmath4 by using the fractal dimension @xmath0 : @xmath5 , where @xmath6 and @xmath7 are the spatial dimension and proportional coefficient . under this assumption , a gel of a higher packing fraction consists of a higher number of smaller flocs of the same fractal dimension @xmath0 . by deducing average mechanical responses of the flocs , the fractal model predicts a power - law dependence of the shear modulus on @xmath4 with an exponent @xmath8 , where @xmath9 is the fractal dimension of the backbone structure . although this theoretical prediction has often been cited in the literature , we should pay attention to the scope of the model . since fractal correlation yields in the formation process of particle gels , there is no clear reason for @xmath0 and @xmath9 to be kept constant during the compression process . actually , some scattering observations indicate that the compression process changes the local structure ( fractal dimension ) @xcite . to make progress in macroscopic modeling of aggregated colloidal or particle gels @xcite , we need to settle such microstructural issues . there is a recent two - dimensional(2d ) simulation work to support the application of the fractal model to compression processes @xcite . gels which were ( i ) directly prepared to the packing fraction @xmath4 , or ( ii ) prepared by compression from a lower packing - fraction gel @xmath10 showed the same structural characteristics . however , the investigated range of packing fractions was relatively high ( @xmath11 ) . since the fractal model targets on strongly aggregated gels , we should compare simulations at lower packing fractions . indeed , our 2d simulation investigates a wider range of packing fractions ( @xmath12 ) . this article reports results of numerical simulation for compression of particulate networks to discuss the consolidation behavior with restructuring . experimentally observed time evolution of compressed gels should be affected by both the hydrodynamic and network resistances @xcite . but , we here focus on studying the latter ; our simulation work investigates the quasi - static structural evolution determined by the balance between external stresses and network strength at long time . this limitation to quasi - static aspects is common with the above - mentioned fractal model . so , compression considered in this work is sufficiently slow , such as osmotic compression . a contact model is employed to simulate sticky and breakable connections between contacting particles . in granular physics , there are many prior studies to investigate cohesive granular system @xcite . their simulation methods are an extension of the discrete element method @xcite ; the rolling friction has been additionally included into the contact model . though our target is a suspension system , the spirit of the modeling is almost the same . in granular physics , there are also some studies on compression of cohesive powders @xcite , which obtained many common insights with our study . 2d simulation results are reported here . one obvious benefit of studying 2d systems is easy recognition of the structure changes . though the density - density correlation is an effective tool to capture special sorts of heterogeneous structures , it might not capture all features of random structures . 2d experimental system has also been studied to observe the microstructure change under shear @xcite . though we do not expect qualitative difference between 2d and 3d systems , this point should be checked by 3d simulation of the same modeling , which we hope to report in future work . we use fractal aggregation models to generate the initial configurations . networks of particles are formed in a rectangle box of sides @xmath13 and @xmath14 with periodic boundary conditions . we combine two processes . the first step models a formation process of fractal clusters in the dilute limit ; the reaction - limited hierarchical cluster - cluster aggregation model was employed @xcite . each cluster consists of @xmath15 particles . the second step is to build a network with them . @xmath16 of the generated clusters are randomly placed in the box while avoiding overlap with each other , and then they are moved with a random walk . collisions of clusters always lead to coalescence , so a single cluster is formed in the periodic boundary space after a while ( see , e.g. , [ fig_snapshots ] ( a ) , below ) . thus , we can generate random networks of a packing fraction @xmath17 , where @xmath18 is the radius of a particle . one can deduce structural features of the formed networks by evaluating the density - density correlation function @xmath19 ( see appendix [ appendix_density - correlation ] ) . we prepared 10 samples of @xmath20 particles in a box @xmath21 . the packing fraction is @xmath22 . [ 223812_16oct12 ] shows the averages and standard deviations of their correlation functions @xmath19 . a power - law decay of correlation is seen in the finite range @xmath23 . the slope indicates fractal dimension @xmath24 , which is almost the same fractal dimension as the clusters generated by the aggregation model . though a crossover behavior is seen in the range @xmath25 , the correlation decays out after that . prepared in the box @xmath26 . the averages and standard deviations were taken over 10 samples . the red line indicates the slope corresponding to @xmath27 . ] our contact model is designed to simulate strongly aggregated particle gels . the contact model used in the original discrete element method ( dem ) @xcite takes into account friction forces for the sliding mode , which is essential to simulate dense granular systems . for cohesive granular systems , the rolling friction has been added @xcite . this extension is also required to reproduce the behavior of particle gels having chain - like local structures . our implementation is similar to the model used for cohesive granular systems . since the dynamic friction , which is modeled in typical contact models , is unclear for colloidal systems , this part is different in our model . though a brief summary will be given below , the detailed descriptions of our implementation were given in another paper @xcite . we assume that interactions acting among particles are only short range forces . no interparticle interaction acts between two particles before getting into contact . once the gap between two particles becomes zero , a cohesive force starts to act between them . for simplicity , our simulation model assumes immediate bond generation , which is expected to give the upper limit of the consolidation behavior . when compression - rate dependence is problem , the condition of bond generation can be crucial . finite compression rates can suppress bond generation , leading to a weaker consolidation . the forces and moments are assumed to be hookian : let us consider the contact interactions between particles @xmath28 and @xmath29 . the center - to - center distance between them is @xmath30 , and the gap @xmath31 . the normal ( separation / overlap ) displacement is related to the normal element of the contact force , @xmath32 where @xmath33 for @xmath34 and @xmath35 otherwise . @xmath36 is the normal unit vector defined by @xmath37 . the second term of eq . is added to avoid significant overlap between particles . the sliding displacement @xmath38 relates to the tangential part of the contact force , @xmath39 the sliding displacement @xmath38 is the tangential projection of the deviation vector between the original contact points @xmath40 and @xmath41 , @xmath42 where the vectors @xmath43 and @xmath44 indicate the original contact points from the respective centers of the particles and are fixed with them . the bending angle @xmath45 ( @xmath46 is related to the moment acting on the rolling mode , @xmath47 these elastic relations are broken when the force or moment exceed threshold values . we assume simple criteria given with three threshold values : the critical normal force @xmath48 , critical sliding force @xmath49 , and critical rolling moment @xmath50 . when the pulling force exceeds the value ; @xmath51 , the bond is broken and the two particles separate . when the tangential force or rolling torque exceeds the threshold value ; @xmath52 or @xmath53 , both the sliding displacement @xmath54 and rolling angle @xmath55 are reset to zero , which releases the tangential parts of the stored energy . the maximum strains are related to the threshold forces and moment as follows : @xmath56 , @xmath57 , and @xmath58 . it must be noted that in our simple model the bond strength of the tangential mode does not depend on the normal load . however , the enlarged contact area due to the normal load may change the tangential strengths @xcite . our neglect of this effect can affect simulations at higher packing fractions . a particle @xmath28 may be in contact with other particles @xmath29 . the total contact force and torque acting on the particle @xmath28 can be written as follows : @xmath59 particles are immersed in a viscous fluid , so that hydrodynamic interactions also act on them . for micro or nano - sized particles , the particle reynolds number is extremely small , so the stokes equations should be used to evaluate the hydrodynamic interactions among many particles @xcite . however , we do not go into the detailed hydrodynamics here . hydrodynamic interactions in the stokes regime are scaled with the velocities of particles . if the slow compression limit is considered , flows induced by particle movements are negligible in comparison with the contact forces . they just work as energy dissipators . indeed , some experimental observations suggest that a solvent flow does not contribute the rearrangements of particles in a slow sedimentation of colloidal gels @xcite . as an approximation , the stokes formula is used in this work ; the force and torque acting on a particle @xmath28 are proportional to the translational velocity @xmath60 and angular velocity @xmath61 , respectively : @xmath62 where @xmath63 is the viscosity of the solvent . since the interparticle interactions are assumed to be sufficiently stronger than thermal agitations , brownian forces are also neglected in this work . consequently , newton s equations of motions are given as the sum of the contact forces and hydrodynamic interactions , @xmath64 since particle inertia is very small for colloidal systems , it is a good approximation to neglect the inertia terms ( the left hand sides ) and solve the force balance equations @xcite . as explained later ( sect . [ 141932_2nov12 ] ) , however , we took a different strategy to find the static equilibriums . the compression is externally controlled . for simplicity , we focus on examining the uni - axial compression along the @xmath65 axis in this work . the periodic boundaries at @xmath66 and @xmath67 of the simulation box introduced in section [ sec_preparation_fractal_gel ] are replaced by two flat walls . these walls can move along the @xmath65-axis like pistons , and they are assumed to be semipermeable ; liquid passes through them , while particles do not . when particles get in contact with the walls , they are rigidly fixed there ; particles contacting with the wall move together as one object . the periodic boundaries along the @xmath68-axis are still used during the compression simulation . the stress - controlled compression , in which an external stress is applied to cause the compression , is similar to typical experimental situations , such as osmotic compression . however , due to the higher efficiency of the simulation , we use strain - controlled compression . this also allows us to have better statistics in our data . this work investigates the slow compression limit ; the time scale of compression is much longer than the relaxation time of particle motions . in this case , the results obtained with the strain - controlled compression are essentially equivalent to the ones with the stress - controlled compression . the compressive stress balancing the network resistance is evaluated from the total forces acting on the walls from particles . since particles directly connecting to the walls ( called wall particles ) no longer have any mechanical freedom , the forces between wall particles @xmath28 and particles @xmath29 contacting with them are summed up as follows : @xmath69 when we refer to `` compressive stress '' in this 2d simulation work , our simulation box is assumed to have a thickness of particle diameter @xmath70 . therefore , the compressive stresses are given as @xmath71 and @xmath72 . our compression simulations are performed in a stepwise manner in order to obtain numbers of equilibriums by a single simulation run . compression steps are given by two increment ratios @xmath73 and @xmath74 ; the initial compression from @xmath75 to @xmath76 is given as @xmath77 , and the @xmath78-th packing fraction as @xmath79 . the increment ratio @xmath74 is determined with the number of steps @xmath80 from @xmath81 to @xmath82 : @xmath83 . the strain - controlled compression is simply realized by translations of the walls with the velocity @xmath84 , so that the packing fraction increases continuously . when it reaches a target packing fraction @xmath85 , the translations of the walls are suspended to await the equilibrium . usually , the excess stress due to the quickness of the compression is relaxed by spreading the deformations into the entire system . the following criteria are examined every @xmath86 time steps of the simulation : * the relative difference between the top and bottom stresses is smaller than a threshold value @xmath87 : @xmath88 , where @xmath89 is the average , @xmath90 . * the relative change of the compressive stress for the last @xmath91 steps is smaller than a threshold value @xmath92 : @xmath93 , where @xmath94 is the one of @xmath86 steps before . * no bonds were broken for the last @xmath86 steps . thus , the compressive stress @xmath95 balancing the network of packing fraction @xmath85 can be determined . once @xmath95 is determined , the compression continues again for the next target @xmath96 . in order to reduce finite size effects , the simulation box needs to involve a sufficient number of typical structures . the compressive stress to balance sparse networks is very low , and the most of particles are slightly stressed . in this case , the relaxation time is significantly long . usually , the overdamped equations of motion , which are given by dropping the right hand sides of eqs . due to the negligible inertia of particles , are used to simulate colloidal suspensions . however , we have taken a different approach for efficiency . as restricting our scope in the slow compression , we may be allowed to optimize some parameters for a faster simulation . in the slow compression limit , velocities of particles are also slow . in this case , the contact forces dominate the displacements of particles . this is why a reduced viscosity @xmath97 still leads to the same equilibrium configurations with a shorter time . the following viscosity gives the critical damped motion in terms of the rolling mode . @xmath98 though it is not simple problem to find the best value , we used @xmath99 in the equation of motion . this arbitrary selection of the viscosity is justified only if the compression is considered slow enough . this point will be checked in sect . [ 141944_21nov12 ] . the introduced 2d contact model ( see sect . [ 112400_22dec12 ] ) is relatively simple , but it still has 7 parameters to be given . traditional experimental measurements to characterize the particle - particle adhesion determine the pull - off force @xcite , which corresponds to @xmath48 in this work . in recent years the atomic force microscope became a reliable method , which provides more precise data for normal force - displacement relationships @xcite . however , there are fewer experimental measurements to determine the parameters of the other modes @xcite . the experimental measurement of the three - point bending test by using optical tweezers is remarkable @xcite , which provides parameters for the rolling mode , @xmath50 and @xmath100 ( or @xmath101 ) . in any case , all parameters for a single system have never been determined so far . thus our knowledge of the contact forces is still limited , so that we need to set the parameters by hand . but , we may have a certain level of confidence by reproducing a similar bending behavior with the three - point bending measurements . our selected parameters are intended to imitate the observed mechanical behavior : a small extent of elastic deformation and irreversible reorganization due to a critical moment . we have selected two sets of the parameters called bond 1 and bond 2 , which are shown in [ bond_parameters ] . bond 1 has the same strengths for the three modes , while bond 2 is 5 times stronger for the normal and sliding mode than the rolling mode . both have the same maximum strains ; elastic deformations give 5% of the radius . 12cmlccc & & bond 1 & bond 2 + normal strength & @xmath102 & 1 & 5 + sliding strength & @xmath103 & 1 & 5 + rolling strength & @xmath104 & 1 & 1 + normal strain & @xmath105 & 0.05 & 0.05 + sliding strain & @xmath106 & 0.05 & 0.05 + rolling strain & @xmath107 & 0.05 & 0.05 + parameter in eq . & @xmath108 & @xmath109 & @xmath109 + the simulation for the three - point bending test consists of the following three steps : 1 . prepare a linear chain under no stress . apply the forces @xmath110 at the central particle and @xmath111 at both the ends , and wait for equilibrium . 3 . stop applying the forces , and wait for equilibrium . the three snapshots i iii in [ fig_three - point - bending ] ( a ) and ( b ) show the configurations of the respective equilibriums of the bond 1 simulation . though the data is not shown here , the results of bond 2 are similar . the two results were obtained with two slightly different forces : ( a ) @xmath112 , ( b ) @xmath113 . if the external force did not reach the threshold , the initial linear shape remained after the test ( see ( a)iii ) . on the other hand , if the bond stress exceeded the threshold value , some irreversible rearrangement took place . as a result , the initial shape was not recovered ( see ( b)iii ) . in the latter case , the bending deformation proceeds until the moment acting on the broken bond becomes smaller than the threshold due to the increase of the bending angle . this is why the small increment of the applied forces leads to the visible difference of the deformation ( see ii of ( a ) and ( b ) in [ fig_three - point - bending ] ) . the slight bending deformation seen by ( a)ii is close to the limit of elastic deformation of this cluster . thus , the parameters were chosen to simulate some brittle behavior of colloidal aggregates ; the local structures remain unless bond breakups take place . in this test , the rolling mode of the contact model is responsible for the cluster s deformation . ) is applied , the elastic deformation is seen ( a ) , and when it is slightly larger ( @xmath114 ) , the bond breakups cause the plastic deformation ( b ) . ] the initial configurations ( sect . [ sec_preparation_fractal_gel ] ) are fractal - like networks with a packing fraction @xmath115 , where @xmath26 and @xmath116 for bond 1 simulation . since the calculation cost of the bond 2 simulation is higher , smaller system size , @xmath117 and @xmath118 , is simulated . the first target packing fraction is set to @xmath119 . the increment ratio and the number of compression steps are @xmath120 and @xmath121 , respectively . the compression finally reaches to @xmath122 . the relaxation time depends on the parameters of the bonds . to account for this , the wall velocities are set to @xmath123 and @xmath124 for the bond 1 and bond 2 simulations , respectively , where @xmath125 . in the equilibrium checks ( sect . [ sec_uniaxial_compression ] ) after every @xmath126 time integration steps , @xmath127 is used for both the simulations , and @xmath128 and @xmath129 for bond 1 and for bond 2 simulations , respectively . as explained in sect . [ 141932_2nov12 ] , it is useful to introduce artificial parameters to speed up the compression simulation . therefore we started to check that our method reproduced the quasi - static compression as expected . if compression is too fast , the regions near the wall are easily crushed . so , we examined the uniformity of the deformations along the compression axis . the uniformity of compression can be seen via the @xmath65 dependence of the packing fraction @xmath130 , which is defined as the packing fraction in the sliced ranges : @xmath131 ( where @xmath132 was chosen ) . due to the heterogeneous structure and finite system size , the uniformity can not be judged with a single compression simulation . this is why we took the averages over 10 simulation runs with different initial configurations . [ 232801_24oct12 ] shows the results for the bond 1 simulations , where every second compression step is plotted . although we can see an undesirable drop at the middle of the first compression ( @xmath133 ) , the compression is satisfactorily uniform along @xmath65-axis after that . the steep drops at both the edges relate to the integrity of the walls ; particles connecting to the walls are not allowed to move ( see section [ sec_uniaxial_compression ] ) . equivalent results were also obtained with the bond 2 simulation with a slower compaction speed ( see sect . [ 100129_27dec12 ] ) . dependence of packing fractions @xmath130 are shown for every second compression step ( @xmath134 , @xmath135 ) . the solid line indicates the initial configuration . ] once the system reaches a static equilibrium under a compressive stress @xmath89 , further compression requires a finite increment : @xmath136 . therefore , the balancing stress @xmath89 can be considered as the strength of the gel network , i.e. , the _ compressive yield stress _ @xmath2 @xcite . a simulation run of the stepwise compression ( sect . [ 141932_2nov12 ] ) finds a series of @xmath2 _ vs. _ @xmath4 relations . [ fig_compression_curves ] plots the results of the bond 1 and bond 2 simulations , where the unit of the stress is @xmath137 . the compression from @xmath138 to @xmath139 shows wide ranges of @xmath2 ( about four decades ) . these results indicate significantly rapid compressive - strain hardening . is evaluated by the simulations of stepwise compression . its averages and standard deviations taken over 10 runs are shown . the red squares represent the result of the bond 1 simulations , and the blue circles of the bond 2 simulations . the dotted curve shows the fitting function of eq . , and ( solid and dashed ) straight lines show the power - law functions . the arrows indicate ranges of the three regimes : ( i ) elastic - dominant regime , ( ii ) single - mode plastic regime , and ( iii ) multi - mode plastic regime . the average exponents @xmath140 are written on the figure . ] simulation provides various information , which allow us to understand the compression . some of data indicate that there are three regimes in the compression process , which will be seen in the following section . [ [ bond - creation ] ] bond creation + + + + + + + + + + + + + bond creation is essential to understand the irreversible consolidation of the compression . due to heterogeneous structure of the networks , the compressive deformations are not affine . consequently , two separated particles may come in contact and newly bond with each other . the mean contact number per particle reflects new bond creations . the result of bond 1 simulation is shown in [ fig_contact_number_energy ] ( a ) . the compression continuously changes the value from 2 to about 4 with increasing rate , so we do not see distinct regimes in this . . the rates are obtained by normalizing with the compressive strain @xmath141 . the averages and standard deviations are taken over 10 runs of the bond 1 simulation . ] [ [ stored - and - dissipated - energies ] ] stored and dissipated energies + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the bonds have finite strength , and they can be broken up or irreversibly reorganized by certain applied stress . such events cause the plastic deformation of the networks , which contribute the irreversibility . the translations of the wall by external stress inject energy into the system . part of the energy are stored in contact bonds first and dissipated as bond ruptures . we can see this energy flow in the successive increments of the total stored and dissipated energies for each compression step ( [ fig_contact_number_energy ] ( b ) ) . the total dissipated energy exceeds the total stored energy at the early stage , @xmath142 . accordingly , we can identify the boundary of two regimes : elastic dominant regime ( i ) and plastic regime ( ii ) . [ [ bond - rupture ] ] bond rupture + + + + + + + + + + + + the accumulated numbers of the bond ruptures @xmath143 were recorded as distinguishing the stress modes causing the breakup , where the @xmath144 indicates ` connection break ' , ` sliding ' , or ` rolling ' . the rupture rates @xmath145 can be defined as the number of breakups normalized with the total particle number and relative compressive strain : @xmath146 [ bond_ruptures ] shows the results of both the bond 1 and bond 2 simulations . at lower packing fractions , only rupture events occur due to rolling . this indicates that compression of loose networks is achieved by bending deformations of particulate chains . when the compaction reaches a certain point , further compression requires sliding or separation of contacting particles . the onset of this mixed - rupture mode depends on the parameters of the contact forces . the connection breaks start between @xmath147 and @xmath148 in bond 1 simulations , and between @xmath149 and @xmath150 in bond 2 simulations . they split the plastic deformation into two parts : single - mode plastic regime ( ii ) below and multi - mode plastic regime ( iii ) above . are compared for two types of bond : ( a ) bond 1 and ( b ) bond 2 . the bond ruptures are recorded by the causing stress . the rates of the separation due to normal forces are plotted by squares , the ones of the regeneration due to sliding forces or rolling moments by triangles and circles , respectively . ] the date shown in the previous section indicate that three distinct behaviors can be recognized in the compression process . let us discuss them one by one here . [ [ i - elastic - dominant - regime ] ] ( i ) elastic dominant regime + + + + + + + + + + + + + + + + + + + + + + + + + + + there is no percolation path in the 10 initial configurations of @xmath151 ( see [ fig_snapshots ] ( a ) ) . hence , no compressive stress is required to compress them in the slow compression limit . after some compaction , the first spanning path appears , and the network begins to resist the compressive stress . at the gel point , a single - linking chain ( percolation path ) resists the external stress between top and bottom . this system - spanning cluster is deformed as a whole , so that strains of single bonds can be very small . therefore , the dominant response to the external stress is elastic deformation ; the work by compression is stored in the contact bonds ( see [ fig_contact_number_energy ] ( b ) ) . in this range , the power - law form expanding at the gel point @xmath152 is suitable to capture the @xmath2-@xmath4 curve @xcite : @xmath153 in [ fig_compression_curves ] , the dotted curve shows the fitting result with the function for the bond 1 simulation data of @xmath154 . the fitting gives the extrapolated gel point as @xmath155 . the evaluated exponent @xmath156 represents some nonlinearity . the compression induces some new contacts , so that the numbers of percolation paths are increased . the rise of resistance due to such increasing paths can be a plausible interpretation of the nonlinearity in this elastic - dominant regime . it must be noted that our simulation results in this range might have large statistical fluctuations . the typical length scale of the structure is comparable to the system size ( compare the drawn circle and the box size in [ fig_snapshots ] ( a ) ) , and consequently only a few paths enter the averaging . therefore , the gel point @xmath152 obtained in the above fitting has only limited significance . [ [ ii - single - mode - plastic - regime ] ] ( ii ) single - mode plastic regime + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + by considering two data ( [ fig_contact_number_energy ] ( b ) and [ bond_ruptures ] ) , we identified the single - mode plastic regime in the ranges : @xmath157 for bond 1 simulations and @xmath158 for bond 2 simulations . this @xmath4-dependence of the consolidation behavior has been discussed with a scaling concept @xcite . the power - law relationships of @xmath2 and @xmath4 have been also observed in experiments @xcite : @xmath159 a similar power - law behavior can be seen in our simulation results ( [ fig_compression_curves ] ) . the least squares fitting gives the almost same exponents : @xmath160 and @xmath161 , for the bond 1 and bond 2 results , respectively . let us look more closely at this regime . [ fig_snapshots ] ( b ) shows the configuration of particles at the beginning of this range ( @xmath162 ) in a bond 1 simulation . at the gel point , a single linking path spanned the entire box . in contrast , multi - linking structures span the system here . since the maximum bending moment of a random - walk chain is proportional to the inverse of the chain size , the supportable loads of larger loops are smaller . the variety of the mesh sizes indicates a heterogeneity of the local strengths . though large empty spaces remain , the relatively dense meshes fill the space from the bottom to the top ( [ fig_snapshots ] ( b ) ) . further compression requires destroying a part of them . as shown in [ bond_ruptures ] , the rolling - mode ruptures are the only rupture events seen in this range . therefore , the size of the responsible mesh structures determines the compressive yield stress @xmath2 . thus , in the single - mode plastic regime , the length scale is essential ; the compression breaks larger and weaker structures into smaller and more robust structures . fractal model was employed to give the length scale from the packing fraction @xcite ( see sect . [ sec_structure_compression ] ) . according to the nature of a fractal , a portion of a fractal cluster is also a small fractal cluster . if the compression of fractal networks causes only fragmentation of fractal clusters into smaller pieces , it can be expected that the correlation length is decreased while keeping the same fractal dimension . this optimistic expectation leads us to employ the fractal model for the compression problem . we will check this point in the next section . , whose packing fraction is @xmath22 . the configurations ( b ) and ( c ) show the beginning of the _ single - mode plastic regime _ ( @xmath162 ) and the _ multi - mode plastic regime _ ( @xmath163 ) , respectively , and ( d ) shows @xmath164 . the red circles express the correlation lengths ; their diameters are @xmath165 , which will be evaluated in the later section ( sect . [ sec_structure_compression ] ) . ] [ [ iii - multi - mode - plastic - regime ] ] ( iii ) multi - mode plastic regime + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the onset of connection breaks indicates that the compression enters a new regime . this transition point depends on the parameters of the bonds . we consider the data above @xmath163 and @xmath166 in bond 1 and 2 simulations as the multi - mode plastic regime . in these ranges , the compression curves look to follow a power law . however , the evaluated exponents are totally different with each other : @xmath167 and @xmath168 . it is by chance that the exponent of bond 2 simulations is similar with the one in the single - mode plastic regime . the exponent in this regime is easily affected by the parameters of the contact model , that implies absence of a simple scaling law . [ fig_snapshots ] ( c ) depicts the beginning of this regime of a bond 1 simulation ( @xmath163 ) . there are still large loops remaining , but they look like voids due to the denser surroundings . here , we can easily find dense meshes and trimers ( a trimer can be considered as the building block of a close packing ) . if they are next to each other , the entire object can be considered as a lump . the same scenario of the _ single - mode plastic regime _ seems to work ; denser mesh domains are piled up over the entire system . however , due to the finite size of particles , the mechanical response of denser meshes is not the same as looser ones . the crush of them involves sliding ruptures and connection breaks , which are seen in [ bond_ruptures ] . since the criteria of destruction includes the all three modes , the compression curve depends on the details of the contact force . this regime terminates due to the close packing , where rigidity of particles prevents reorganizations of particles . since our simplified contact model is not accurate for that , we can not specify the upper end of the range . during compression , separated particles can newly come into contact . consequently , mesh size decreases as @xmath4 is increased by compression . if we employ the fractal model @xcite to explain the @xmath4-dependence of @xmath2 , the structure change needs to follow a special manner ; the fractal dimension @xmath0 characterizing the local structures does not depend on the packing fraction @xmath4 , and the correlation length @xmath1 is a decreasing function of the packing fraction @xmath4 , @xmath169 . the fractal dimension @xmath0 can be estimated from the density - density correlation function @xmath19 ( see [ appendix_density - correlation ] ) . [ fig_ddc ] shows averages of the correlation functions @xmath19 of every second compression step , @xmath134 ( @xmath170 ) , of the bond 1 simulation results . the results of the bond 2 simulations were almost the same . the correlation of the density is mostly decreasing functions of the two - point distance @xmath171 . the correlation of the configuration before compression is the largest and reaches the farthest distance ( solid line ) . the compression reduces this correlation in a continuous manner . of compressed networks are shown . the solid line shows the initial configuration @xmath172 , and the dashed or dotted lines the every second compression step @xmath134 ( @xmath173 ) . they are averaged values over 10 runs of simulations . ] the fractal correlation appears as a straight slope in the log - log plot of @xmath19 , i.e. , @xmath174 . though the straight lines can be recognized in the initial configuration , the compressed networks show slightly curved lines . in order to make them easier to see , we introduce the fractal dimension profile @xmath175 from the local slope of the correlation function at each distance @xmath171 : @xmath176 the results are shown in [ 143513_26nov12 ] ( a ) . as mentioned in sect . [ sec_preparation_fractal_gel ] , the initial configurations ( solid line ) consist of three parts : a fractal plateau ( @xmath177 ) indicating @xmath178 , crossover range ( @xmath179 ) , and 2d - like uniform network ( @xmath180 ) . thus , though the recognizable fractal plateau is limited , the @xmath175 profiles of the initial configurations agree with the expected behavior as fractal networks . evaluated by eq . are plotted . the solid line shows the initial configuration @xmath172 , and the dashed or dotted solid lines the every second compression step @xmath134 ( @xmath173 ) . the numerical values indicate the packing fractions . ( b ) the @xmath4-dependence of the correlation length @xmath181 ( see the definition in the text ) . ] once the networks are compressed , the fractal correlations are affected immediately . indeed , the first compression step of our simulation ( @xmath182 ) shows the entire change of @xmath175 : a narrower plateau indicating a rise of the fractal dimension , and shifts of the crossover and uniform ranges to shorter lengths . these results suggest that not only long - distance correlations , but short - distance correlations can be affected by the compressive deformation at the early stage . this tendency continues for the further compression . though the profiles @xmath175 for the compressed networks indicate that mono - fractal assumption does not work , it may be worth determining the @xmath4 dependence of the correlation length @xmath1 . since the slopes of @xmath19 in [ fig_ddc ] are not constant , it is ambiguous to determine @xmath1 from them . in our purpose , we may consider the correlation length as the length scale where the network looks uniform . the typical length scale for the uniformity can be determined by evaluating fluctuation of local densities @xcite . the local densities are defined by boxes of size @xmath183 , which are denoted by @xmath184 . if the size of box is larger than the correlation length @xmath1 , the local densities show little fluctuation ; the system looks uniform . thus , the standard deviation of the local densities normalized by the global density , @xmath185 , measures the uniformity . due to the finite system size , we compare here the lengths showing a common finite fluctuation , @xmath186 . the actual values of the correlation lengths @xmath1 should be larger than the evaluated lengths @xmath187 , so a common arbitrary factor may be given to adjust . [ 143513_26nov12 ] ( b ) shows the result of the lengths @xmath187 satisfying @xmath188 . we find a power - law relation between @xmath1 and @xmath4 with an exponent @xmath189 . the obtained exponent is a little bit smaller than the fractal assumption : @xmath190 . the evaluated correlation lengths , @xmath191 , are also shown as the diameter of the circles drawn in [ fig_snapshots ] . we focused on investigating compression of particle gels . inspired by scaling theory for polymer gels , the fractal model has been introduced to explain the packing fraction dependence of particle gels rheology @xcite , and referred for the compression process @xcite . the model assumes that the structure of a particle gel is fractal network . however , the simple question whether a particle gel under compression is fractal network or not , has left unanswered . we have aimed to answer this question by introducing a particle - scale simulation model in the limit of strongly aggregated particle gels . in this case , the cohesive force among contacting particles is assumed to be so strong that brownian forces do not affect the structural evolution of gels . our simulation reproduced the reasonable consolidation behaviors as @xmath4 _ vs_. @xmath2 relationships in the case of quasi - static compression . by monitoring various particle - scale information , i.e. , mean contact number , stored / dissipated energy , and number of bond ruptures , we identified the three distinct regimes : ( i ) elastic dominant regime , ( ii ) single - mode plastic regime , and ( iii ) multi - mode plastic regime . at the gel point , there are many `` dead end '' branches , where the particles in such branches do not contribute to the strength of the network @xmath2 . compression makes separated particles come into contact , so that the mesh size becomes smaller and smaller . this explanation is already known and qualitatively true for the most part . such shrinkage of the typical length scale causes a transition of the local mechanical response from ( ii ) to ( iii ) in the plastic regimes . while the length scale is large enough , the rolling mode is the responsible for the deformation . at a certain length scale , the sliding and normal modes become more important for further compression . thus , we confirmed that a single regime could never cover the wide range of packing fractions in spite of subtle appearance in the @xmath4 _ vs_. @xmath2 relationships . the simulation results also provided an insight that heterogeneity in the local mesh sizes plays some roles in the way of reorganization under compression and overall network strength . mechanical correlations ( connectivity ) determine which parts of the network are destroyed under an applied stress . if a percolation path consisting of robust domains is built , they support a large fraction of the applied stress . in this case , the weakest domain in the path is most probable to be destroyed . this restructuring yields a smaller and robuster domain , which results in the reinforcement of the percolation path . however , such mechanism is subtle because compression is not continuing deformation and tends to reduce the heterogeneity . in order to see whether the fractal assumption remains or not in particle gels under compression , we evaluated the density - density correlation @xmath19 for the simulation results . the evaluated fractal dimension profile @xmath175 says that compression affects both the local structure and correlation length in the early stage . thus , our results are contradictory to the application of the fractal model to compression processes . however , the fractal dimension profiles @xmath175 show a systematic change in the _ single - mode plastic regime_. this result seems to indicate required ingredients to upgrade the fractal model for compression processes . the simulation model of this work is simple ; we neglect various factors in real particle gels to make clear the focus of study and to optimize the simulation speed in order to achieve large system sizes . for example , we employed a simple contact model , which depends on only configuration of particles and has no history dependence . however , cohesion between particles in real system may be more complex , e.g. the bonding or breakup may require certain time periods . in order to deal with such systems , our quasi - static simulation framework needs to be extended . this extension is relevant for the future work because there are many experimental works investigating on the time / rate - depending aspects of colloidal gels and yield stress fluids @xcite . the authors would like to thank prof . morton denn for valuable suggestions on the manuscript , and rs would like to express his gratitude to prof . richard buscall for fruitful discussions , valuable suggestions , and warm encouragement . in order to study colloidal gels , we need to have some tools to characterize random structured systems . the density - density correlation function @xmath192 is used for this purpose . the local density function @xmath193 is defined as follows : @xmath194 by taking the average over the entire space of the system , the density - density correlation function @xmath192 is defined as follows : @xmath195 when the system is isotropic , the correlation is given in terms of the distance , @xmath196 , which can be obtained by taking averages over all directions : @xmath197 the correlation function @xmath19 can be regarded as a conditional probability to find particles . since it is normalized by the entire probability , the value 1 means no correlation . it is known that strongly aggregated gels have structures involving fractal correlations . the local structures are equivalent to fractal flocs . for fractal flocs , the relation between the number of particles @xmath198 and the radius of gyration @xmath199 follows a power - law : @xmath200 the exponent @xmath0 is called fractal dimension . therefore , the average packing - fraction profile of the clusters is given as follows : @xmath201 where @xmath171 is the distance from the center of mass . since fractal gels can be considered as cluster - filling networks , the local structure should coincide with this density profile . consequently , the density - density correlation function @xmath19 has the same functional form : @xmath202 . in a fractal gel , the correlation reaches a finite distance , which is called the correlation length @xmath1 . for @xmath203 , the correlation does not remain : @xmath204 if the local densities are evaluated in boxes of size @xmath1 , the system should look uniform . 52ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.106.118302 [ * * , ( ) ] @noop * * , ( ) @noop , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ , ( , ) chap . @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.99.028303 [ * * , ( ) ] @noop _ _ ( , , ) in @noop _ _ ( ) pp . @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) in @noop _ _ , ( )

galactic cosmic rays ( crs ) are believed to be accelerated via first order fermi mechanism at the expanding shock waves of supernova remnants ( snrs ) ( for a review see * ? ? ? * ; * ? ? ? * ; * ? ? ? two of the strongest points supporting this scenario are : _ i ) _ the fact that snrs alone are capable to maintain the galactic cr flux at the observed level , provided that a fraction of about 10% of their explosion energy is somehow converted into crs @xcite , and _ ii ) _ the fact that some shell - type snrs have been detected at tev energies ( see * ? ? ? * for a review ) , as expected if they indeed are the sources of crs @xcite . in this context , tev gamma rays originate from the decay of neutral pions produced in interactions between the accelerated crs and the interstellar gas swept up by the snr shock . however , the detection of snrs in gamma rays , though encouraging , can not provide by itself the final proof that crs are indeed accelerated at snr shocks . this is because competing leptonic processes , namely inverse compton scattering from accelerated electrons , can also explain the observed tev emission , provided that the magnetic field does not significantly exceed @xmath0 g ( e.g. @xcite ) . evidence for strong @xmath1 g magnetic field , and thus indirect support to the hadronic scenario for the gamma ray emission , comes from the observation of thin x - ray synchrotron filaments surrounding some snrs @xcite and of the rapid variability time scale of the synchrotron x rays @xcite . on the other hand , the absence of thermal x - ray emission from the tev snrs rxj1713.7 - 3946 and rxj0852.0 - 4622 has been interpreted by katz and waxman ( 2008 ) as an argument against the hadronic nature of the gamma - ray emission ( and thus supporting the leptonic scenario ) . their argument can be summarized as follows : in order to fit the tev data with an hadronic model , the gas density can not be too low . the high density would then imply a strong x - ray thermal bremsstrahlung emission which is not observed . however , this argument is not robust given the uncertainties related to the heating of electrons at shocks , and it has been strongly questioned @xcite . moreover , it has been shown that the downstream proton temperature can be effectively reduced if the shock is effectively accelerating cosmic rays . this would also suppress the electron temperature and the thermal x - ray emission @xcite . a conclusive proof of the hadronic nature of the gamma - ray emission will possibly come from the detection of neutrinos from snrs , which are expected to be produced during the same hadronic interactions responsible for the production of gamma rays @xcite . however , since the detection of neutrinos appears challenging even for km@xmath2-scale neutrino telescopes , the search for evidence of cr proton acceleration coming from gamma ray or multi wavelength observations is mandatory . the presence of a massive molecular cloud close to a snr can provide a dense target for cr hadronic interactions and thus enhance the expected gamma - ray emission . a correlation between snrs and molecular clouds is expected , especially in star forming regions @xcite . in general , a spatial correlation between tev gamma rays and dense gas , though not conclusive , would favor an hadronic scenario , where gamma rays are the result of the interactions of cr protons in the dense gas . such correlation has been observed from a number of snr / molecular clouds associations @xcite . if the molecular cloud is overtaken by the snr shock , the enhanced gamma - ray emission is expected to be cospatial with the snr shell , or with a portion of it @xcite . if the cloud is located at some distance from the snr , it can still be illuminated by crs that escape from the snr and produce gamma rays there @xcite . for this scenario , it has been shown that , for typical snr parameters and for a distance @xmath3 kpc , a molecular cloud of mass @xmath4 can emit tev gamma rays at a detectable level if it is located within few hundred parsecs from the snr @xcite . this implies that the angular displacement between the snr shell and the gamma ray emission is of the order of @xmath5 , where @xmath6 is the distance between the snr and the cloud . this translates in the fact that sometimes the association between snrs and molecular clouds can be not so obvious , given that the separation between the two objects can be even larger than the detector s field of view . following this rationale , gabici and aharonian ( 2007 ) proposed that some of the unidentified tev sources detected by hegra @xcite , h.e.s.s . @xcite and milagro @xcite might in fact be such clouds illuminated by a nearby snr . in this paper , we calculate the expected non - thermal emission , from radio to multi - tev photons , from a molecular cloud illuminated by crs coming from a nearby snr . we generalize the model presented by gabici and aharonian ( 2007 ) , which was limited to the hadronic tev photons only , to include the generation of secondary electrons in the cloud and the related synchrotron and bremsstrahlung emission . we found that the total radiation energy output from a cloud is dominated by the gamma ray emission , which can exceed the emission in other bands by an order of magnitude or more . this suggests that some of the unidentified tev sources detected so far , with no obvious or very weak counterparts in other wavelengths ( the so called `` dark sources '' ) , might be in fact associated with massive molecular clouds illuminated by crs . moreover , under certain conditions , the gamma - ray spectrum from the cloud exhibit a concave shape , being steep at low ( @xmath7 gev ) energies and hard at high ( @xmath7 tev ) energies . this fact might have important implications for the studies of the spectral compatibility of gev and tev gamma ray sources . consider a supernova of total energy @xmath8 exploding in a medium of density @xmath9 . the initial shock velocity is @xmath10 cm / s and remains roughly constant until the mass of the swept up material equals the mass of the ejecta . this happens at a time @xmath11^{1/3 } { \rm yr}$ ] , when the shock radius is @xmath12^{1/3 } { \rm pc}$ ] . then the snr enters the sedov phase and the shock radius and velocity scales with time as @xmath13 and @xmath14 . the spectrum of particles accelerated at the snr shock is determined by the transport equation ( e.g. , * ? ? ? * ) : @xmath15 where @xmath16 is the momentum dependent diffusion coefficient and @xmath17 the flow velocity . for a strong shock with compression factor @xmath18 , the test particle theory predicts an universal shape for the cr spectrum at the shock @xmath19 @xcite . if the shock is an efficient accelerator ( as snr shocks are believed to be ) , the cr pressure modifies the flow structure , making the shock more compressible and the spectrum of the accelerated particles harder , @xmath20 with @xmath21 @xcite . detailed calculations compared with multiwavelength observations of snrs suggest the values @xmath22 and @xmath23 @xcite , which we adopt in the following . the maximum momentum of the accelerated particles is determined by a confinement condition , namely that the diffusion length @xmath24 of the particles can not exceed the characteristic size of the system @xmath25 : @xmath26 the maximum possible energies are achieved when the acceleration proceeds in the bohm diffusion limit , @xmath27 , with @xmath28 the magnetic field strength at the shock . in this case the maximum momentum decreases with time as @xmath29 . in fact , the drop of @xmath30 is even faster , given that the magnetic field is also expected to decrease with time . this implies that at any time , particles with momentum above @xmath31 quickly escape the remnant , generating a cutoff in the spectrum . the spectrum of the runaway particles can be approximated as a @xmath32function ( see * ? ? ? * ) : @xmath33 @xmath34 where the integration has to be performed where the integrand is negative . thus , to calculate the flux of the runaway particles one has to know : _ ( i ) _ the cr particle distribution function at @xmath30 at any location in the snr , _ ( ii ) _ the flow velocity both inside the shock and outside it , where the cr precursor forms and _ ( iii ) _ how the maximum momentum varies during the snr evolution . ptuskin and zirakashvili ( 2005 ) showed that it is straigthforward to derive _ ( i ) _ and _ ( ii ) _ using an approximate ( but still reasonably accurate ) linear velocity profile inside the snr : @xmath35 and assuming that the cr pressure at the shock @xmath36 is a fraction @xmath37 of the incoming ram pressure @xmath38 and that @xmath39 . the determination of @xmath30 and its evolution with time requires the knowledge of the diffusion coefficient ( see eq . [ eq : pmax ] ) , which is in turn determined by the level of magnetic turbulence generated by the accelerated particles themselves . this makes the problem nonlinear and very difficult to be solved . the value of @xmath30 depends on a few crucial but poorly known aspects of the problem , including the nature of cr driven instability operating in the shock precursor and the level of wave damping @xcite . because of these uncertainties , we adopt here a phenomenological approach , namely we parametrize the maximum momentum as @xmath40 . we further assume @xmath41 pev and @xmath42 gev at the early ( @xmath43 yr ) and late ( @xmath44 yr ) epochs of the sedov phase respectively . this requires @xmath45 . remarkably , if the maximum momentum is a power law function of time , the spectrum of the escaping particles integrated over the whole sedov phase is also a power law of the form @xmath46 @xcite , which is close ( sligthly harder ) to what needed to fit the cr data below the knee @xcite . following ptuskin and zirakashvili ( 2003 ) , the approximate spectrum of the crs inside the snr @xmath47 can be obtained from eq . [ eq : transport ] by dropping the diffusion term , while the spectrum of the runaway crs at a given distance @xmath48 from the snr and at a given time @xmath49 can be obtained by solving the diffusion equation : @xmath50 the diffusion coefficient @xmath51 describes the propagation of crs in the galactic disk . the available cr data require a power law energy dependence , @xmath52 , with @xmath53 at @xmath54 gev and @xmath55 @xcite . the constraints on the diffusion coefficient are obtained from the comparison between diffusion models and cr data and have to be considered as average galactic values . however , the conditions might be rather different in regions close to cr sources , in particular due to the presence of strong gradients in the cr distribution , which may enhance the generation of plasma waves and thus suppress the diffusion coefficient @xcite . the change in @xmath56 within the allowed range or the choice of a different normalization for @xmath57 does not alter qualitatively the results , the main effect being that the characteristic time scales of the problem change proportional to @xmath58 . remarkably , if @xmath31 scales as a power law of time , eq . [ eq : diffusionism ] can be solved analytically and the distribution function of escaping cosmic rays at any given distance @xmath48 from the snr and at any given time @xmath49 reads , for energies @xmath59 : @xmath60 where @xmath61 is the total supernova explosion energy , @xmath62 is the fraction of such energy converted into crs and @xmath63 and @xmath64 are the maximum and minimum energies of crs accelerated during the sedov phase . the diffusion distance for a cr of energy @xmath65 is : @xmath66 where @xmath67 represents the time after the supernova explosion at which crs with energy @xmath65 are released in the interstellar medium . the solution derived by ptuskin and zirakashvili ( 2005 ) for the total cr spectrum injected by a snr in the interstellar medium during the whole sedov phase can be easily derived by integrating eq . [ eq : crout ] over space . finally , it has to be noticed that the total cr spectrum at a given time and at a given distance from the snr is the sum of two contributions : _ i ) _ a time dependent contribution from crs coming from the snr , whose spectrum is described by eq . [ eq : crout ] and _ ii ) _ a steady contribution from the galactic cr background . following the procedure described above , it is possible to evaluate , for any given time , the cr spectrum in proximity of a molecular cloud located at a given distance from the snr . if the diffusion coefficient inside the cloud is not significantly smaller than the galactic one , cr freely penetrate the cloud and the cr spectrum inside the cloud is not affected by propagation effects . conversely , if the diffusion coefficient is significantly ( more than one order of magnitude ) reduced , low energy crs are excluded from the cloud and a low energy cutoff appears in the cr spectrum , at an energy that depends on the value of the diffusion coefficient ( see * ? ? ? * for details ) . here , we assume free penetration of crs and we refer the reader to gabici et al . ( 2007 ) ( and references therein ) for a detailed discussion on cr exclusions from molecular clouds . we do not consider here any contribution from cr electrons coming from the snr , since they do not escape the remnant due to diffusive confinement ( for low energy electrons ) and severe synchrotron losses in the strong magnetic field ( for high energy electrons ) . cr protons propagating inside a molecular cloud produce secondary electron - positron pairs during inelastic interactions in the dense intercloud medium . we calculate the spectrum of the injected secondary electrons @xmath68 by using the parameterization given by kelner et al . ( 2006 ) and we follow the time evolution of the electron distribution function @xmath69 by solving the equation : @xmath70 + q_e(t , e ) - \frac{f_e(t , e)}{\tau_{esc}^e}\ ] ] where @xmath71 is the energy loss rate for electrons , @xmath72 the energy loss time , and @xmath73 the diffusive escape time from the cloud . all these time scales will be discussed and estimated in section 4 . once the proton and electron cr spectra have been derived , the non - thermal radiation from the molecular cloud can be calculated . we consider here a molecular cloud located at a given distance @xmath6 from a snr and we calculate the cr spectrum at the cloud location . as mentioned above , the spectrum is the sum of two disctinct components : _ i ) _ the crs coming from the nearby snr , described by eq . [ eq : crout ] , and _ ii ) _ the galactic cr background , which results from the superposition of all the cr sources in the galaxy . while the latter contribution is constant in time , the first one changes , since the flux of crs escaping from the snr evolves in time as described in sec . [ sec : model ] . [ fig : crs ] shows the spectrum of crs at the location of the molecular cloud . the galactic cr background is plotted as a thin dot dashed line labeled as _ cr sea _ , while the spectrum of the crs coming from the snr is plotted as a thin line for different times after the supernova explosion : 500 yr ( solid ) , 2000 yr ( dotted ) , 8000 yr ( short dashed ) and 32000 yr ( long dashed ) . thick lines represent the sum of the two contributions . the distance between the snr and the molecular cloud is 50 pc ( left panel ) and 100 pc ( right panel ) . we assume a total supernova explosion energy of @xmath74 ergs and a high cr acceleration efficiency at the snr shock equal to @xmath75 . the normalization of the cr spectrum at the cloud scales linearly with these two quantities . , since the estimated flux of particles escaping the snr comes from a nonlinear theory of shock acceleration ] the diffuse galactic cosmic ray spectrum is assumed to be identical to the one observed near the sun ( see e.g. * ? ? ? * ) : @xmath76 while the diffusion coefficient , needed to evaluate eq . [ eq : crout ] , is taken equal to : @xmath77 compatible with cr propagation models ( e.g. * ? ? ? the evolution with time of the cr spectrum at the position of the molecular cloud can be understood as follows . according to the model described in sec . [ sec : model ] , crs with different energies leave the snr at different times . the highest energy ( @xmath78 ) crs leave the snr first , while crs with lower and lower energy are released at later times . moreover , higher energy crs diffuse faster , thus the spectrum of crs at the cloud exhibit a sharp low energy cutoff at an energy @xmath79 , which moves to lower and lower energies as time passes . the position of the cutoff represents the energy of the least energetic particles that had enough time to reach the cloud . from fig . [ fig : crs ] it is clear that the influence of the presence of a nearby snr close to the cloud is reflected in the cr spectrum at the cloud position , but this influence depends on many parameters , such as the distance between the snr and the cloud , the time since the supernova explosion , and the cr energy . the crs coming from the snr dominate the total cr spectrum at high energy , while at lower ( @xmath7 gev ) energies the galactic cr background is always the dominant component , unless the molecular cloud is located at distances significantly smaller than @xmath80 50 pc from the snr . however , such small distances are comparable to the size of the snr itself and thus in this case an interaction between the snr shock and the molecular cloud is expected ( see e.g. * ? ? ? the investigation of this scenario goes beyond the scope of this paper and thus we limit ourselves to the case in which the distance between the snr and the cloud is significantly bigger than the size of both objects . aharonian and atoyan ( 1996 ) also evaluated the cr spectrum in the vicinity of a cr accelerator by using an approach similar to the one developed here . they made no specific assumption about the nature of the accelerators and solved the cr transport equation by assuming that a power law spectrum of crs is injected in the interstellar medium . they considered both the case of a continuous injection of particles during the whole lifetime of the accelerator and the case of an impulsive source that releases all the crs at the same time . the approach we adopt here is different , because it is specific for a given class of sources , namely snrs . in this specific case , particles having different energies are released at different times in the interstellar medium . molecular clouds are characterized by a wide range of masses , going from @xmath81 to @xmath82 or even more and have typical sizes ranging from few to few tens of parsecs . the typical density of a cloud is of about few hundred atoms per cubic centimeter , but much higher densities can be found in less massive and smaller ( parsec or sub - parsec scale ) molecular cloud cores , dark clouds or bok globules ( see * ? ? ? * for a review ) . the typical magnetic field of the intercloud medium is @xmath83 g @xcite , and it scales roughly as the square root of the gas density , thus reaching the mg level in the densest regions with density @xmath84 @xmath85 @xcite . the propagation of high energy crs inside molecular clouds has been studied in gabici et al ( 2007 ) ( see also dogel and sharov , 1990 ) , where an extensive discussion can be found . here we summarize the most relevant aspects . once the crs from the snr reach the molecular cloud , they diffusively penetrate with typical time scale : @xmath86 where @xmath87 is the cloud radius and @xmath88 is the diffusion coefficient which we assume here to depend on energy and on magnetic field as : @xmath89 here @xmath90 is a factor that takes into account deviations from the average galactic diffusion coefficient described by eq . [ eq : diff ] ( see * ? ? ? * ) and @xmath91 g is the average magnetic field in the galactic disk . at very high energies , when the larmor radius of the particle becomes comparable with or even larger than the size of the cloud , particles propagate almost rectilinearly , and the propagation time reduces to the crossing time @xmath92 . in the following we assume the characteristic propagation time for a cr in a molecular cloud to be : @xmath93 which is a rough approximation which still describes with sufficient accuracy the two different regimes of propagation . crs can freely penetrate the molecular cloud if the diffusion time is shorter than the energy loss time which , for cr protons with energy above @xmath42gev , is dominated by inelastic proton proton interactions in the dense gas and reads ( see e.g. * ? ? ? * ) : @xmath94 where @xmath95 is the gas density , @xmath96 is the speed of light and @xmath97 and @xmath98 are the cross section and inelasticity of the process , respectively . for nonrelativistic protons ( energies below 1 gev ) ionization losses become relevant , with time scale ( e.g. * ? ? ? * ) : @xmath99 @xmath100^{-1 } ~ yr .\ ] ] here @xmath101 is the proton kinetic energy and @xmath102 is the proton mass . the total energy loss time for cr protons due to both processes is given by : @xmath103 secondary electron - positron pairs are produced inside the cloud during inelastic collisions between cr protons in the dense gas . the typical diffusion time for such electrons is given by eq . [ eq : difftime ] , while the relevant channels for electron energy losses are ionization losses , bremsstrahlung emission and synchrotron emission with characteristic time scales ( e.g. * ? ? ? * ) : @xmath104 @xmath105 @xmath106 respectively . here , @xmath107 is the electron lorenz factor and @xmath65 the total electron energy . the ionization losses are computed for ultra - relativistic electrons . the total energy loss time for cr electrons due to the three processes listed above is given by : @xmath108 in writing the equation above , we neglected the possible role of inverse compton losses off soft background photons . such losses are expected to be relevant when the energy density in the radiation field @xmath109 is greater than the magnetic field energy density . in the following we will consider a cloud with radius @xmath110 pc and magnetic field @xmath111 g , and in this case the condition reads : @xmath112 . such condition can be realized inside molecular clouds only close to star forming regions . in particular , if a star forming region or a stellar ob association with total photon output equal to @xmath113 erg / s is located within a molecular cloud , inverse compton losses will dominate over synchrotron losses around the star forming region up to a distance of @xmath114 pc , which becomes comparable with or larger than the molecular cloud radius @xmath87 when @xmath115 . since this is a quite high luminosity , only a small fraction of molecular clouds are expected to host such luminous ob associations ( williams and mckee 1997 ) . we can thus safely neglect the role of inverse compton losses in most of the situations . inverse compton losses can certainly play a role if one considers small regions within the cloud which surrounds star forming regions . such regions constitute a small fraction of the cloud volume and thus are not expected to affect significantly the non thermal emission from the whole cloud . in fig . [ fig : timegmc ] ( left panel ) the characteristic time scales listed above have been plotted as a function of particle energy for a giant molecular cloud with total mass @xmath116 and radius @xmath110 pc . assuming a flat density profile the density is @xmath117 @xmath85 . the magnetic field is assumed to be @xmath118 g . the dotted line refers to proton energy losses , which are dominated by ionization losses at energies below @xmath42 gev and by inelastic proton proton interactions at higher energies . the solid line represents the electron energy loss time . the three different power law behaviors reflect the dominance of ionization , bremsstrahlung and synchrotron losses at low , intermediate and high energies , respectively . finally , the dashed line represents the propagation time over a distance equal to the cloud radius . the propagation time has been evaluated by using the diffusion coefficient from eq . [ eq : diffcoeff ] with @xmath119 ( no suppression with respect to the average galactic value ) . for proton energies above the threshold for pion production ( @xmath120 mev ) , the propagation time is always shorter than the energy loss time . this means that cr protons which produce gamma rays and secondary electrons can freely penetrate the cloud and their flux is not attenuated due to energy losses . the propagation time for cr electrons is also shorter than the energy loss time for particle energies between @xmath121 mev and a few hundreds of tev . this implies that , within this energy range , the secondary electrons produced inside the cloud quickly escape , and have little effect on the non - thermal emission from the cloud . on the other hand , extremely energetic electrons with energies above a few hundreds tevs radiate all their energy in form of synchrotron photons before leaving the cloud . in a typical magnetic field of a few tens of microgauss , these electrons emit synchrotron photons with energy : @xmath122 thus , the most relevant contribution from secondary electrons to the cloud non thermal emission falls in the hard x - ray band . the middle panel of fig . [ fig : timegmc ] shows the time scales for a compact cloud with radius @xmath123 pc and mass @xmath124 , which implies , in case of a flat density profile , a density of @xmath125 @xmath85 . a strong magnetic field of @xmath126 g is assumed , as suggested by observations of very dense clouds @xcite . these parameters are typical for molecular cloud cores or for dark clouds @xcite . the left and middle panels in fig . [ fig : timegmc ] looks very similar , except for a scaling factor . this implies that the same conclusions can be drawn as for the case of a giant molecular cloud , and thus a similar behavior is expected from giant molecular clouds and compact clouds , the only difference being that all the time scales are much shorter in the latter case , due to the high gas density and to the reduced size of the system . the properties of giant molecular clouds located in the galactic centre region can differ significantly from the average figures reported above . as an example , we plot in the right panel of fig . [ fig : timegmc ] the typical time scales for the sgrb2 cloud . this is a very massive cloud located at 100 pc ( projected distance ) from the galactic centre . the cloud virial mass is @xmath127 @xcite and a magnetic field at the milligauss level has been measured in the outer envelope of the cloud complex @xcite . the mass distribution can be fitted with a radial gaussian density profile with @xmath128 pc ( * ? ? ? * and references therein ) . to compute the curves plotted in fig . [ fig : timegmc ] ( right panel ) , we assumed a cloud radius of @xmath129 pc , which encloses @xmath130 of the total cloud mass . this gives an average density of @xmath131@xmath85 ( roughly a factor of 2 below the central density ) . it is evident from the right panel of fig . [ fig : timegmc ] that the sgrb2 cloud is remarkably different from a typical giant molecular cloud . in particular , the very high values of the magnetic field and of the gas density make the energy loss time of cr protons significantly shorter than the propagation time for energies below a few hundred gevs . moreover , for cr electrons the energy loss time is always shorter than the propagation time . this means that cr protons with energies up to few hundred gevs can not penetrate the molecular cloud , as they do in the cases considered in the left and middle panel of fig . [ fig : timegmc ] . primary cr electrons can not penetrate the cloud , while secondary cr electrons produced inside the cloud in hadronic interactions can not leave the cloud and radiate all their energy close to their production site . these characteristics make sgrb2 a very peculiar objects whose modelling needs a specific treatment . a detailed study of the cr penetration in the sgrb2 cloud has been performed by protheroe et al . ( 2008 ) , who also computed the synchrotron radio emission from secondary electrons . the effects of crs exclusion from giant molecular clouds on their gamma ray emission have been discussed in detail by gabici et al . ( 2007 ) . in the following we will not focus onto any specific object but rather investigate the case of the typical molecular clouds as the ones considered in the left and middle panels of fig . [ fig : timegmc ] . to conclude the discussion on characteristic time scales , it is interesting to note that under certain conditions all the energy loss times , for both protons and electrons , scales in the same way with gas density , namely as @xmath132 . this is evident from eqns . [ eq : pptime ] and [ eq : iontimep ] , that describe inelastic proton - proton scattering and ionization losses for protons respectively , and from eqns . [ eq : iontimee ] and [ eq : bremstime ] that describe ionization and bremsstrahlung losses for electrons respectively . moreover , also for synchrotron losses [ eq : syntime ] ) it is possible to derive the same scaling with gas density by recalling that observations suggest that the magnetic field in molecular clouds scales as the square root of gas density @xcite . this has the important consequence that , in typical molecular clouds , particles with a given energy ( protons or electrons ) lose their energy always through the same channel , independently on the gas density . as said above , this conclusion may not hold for peculiar obkects such as sgrb2 . molecular clouds are now estabilished gamma ray sources @xcite . their gamma ray emission is believed to be the result of the decay of neutral pions produced during the inelastic collisions of crs with the dense gas which constitutes the cloud @xcite . during the same interactions , also electrons and positrons are produced via the decay of charged pions . these electrons and positrons produce a broad spectrum of radiation from radio waves to gamma rays due to synchrotron emission and non thermal bremsstrahlung . in this section we compute the expected non - thermal emission from a molecular cloud located in the proximity of a snr . the emission is the result of cr interactions with the dense gas and magnetic field in the cloud and is made up of two contributions : a steady state contribution from the interactions of background crs that penetrate the cloud and a time dependent contribution from the interactions of crs coming from the nearby snr . we consider a giant molecular cloud of mass @xmath116 , radius @xmath110 pc and we assume an uniform density of @xmath117 @xmath85 . the magnetic field is @xmath133 g . the relevant time scales for cr propagation and energy losses in such an environment have been plotted in the left panel of fig . [ fig : timegmc ] . in order to show all the different contributions to the total non - thermal emission , in fig . [ fig : gmc2000 ] we plot the broad band spectrum from the cloud at a time @xmath134 years after the supernova explosion . the snr is located 100 pc away from the molecular cloud and the distance to the observer is 1 kpc . the dotted line ( curve 3 ) represents the emission from neutral pion decay ( from both background crs and crs from the snr ) , the dot dashed lines represent the synchrotron ( curve 2 ) and bremsstrahlung ( curve 4 ) emission from background cr electrons that penetrate the molecular cloud and the dashed lines represent the synchrotron ( curve 1 ) and bremsstrahlung ( curve 5 ) emission from secondary electrons produced during inelastic cr interactions in the dense gas . for the spectrum of background cr electrons we use a fit to the measured spectrum at earth ( see kobayashi et al . 2004 for a recent compilation of data ) . electrons can freely penetrate the cloud except for the highest ( @xmath135 ) and lowest ( @xmath136 ) part of the energy spectrum , where the energy loss time scale is significantly shorter than the propagation time ( see fig . [ fig : timegmc ] ) . for this energies the cr electron flux inside the cloud is suppressed and we estimated the suppression by assuming that cr electrons can penetrate undisturbed the cloud only up to a given depth , which is estimated as @xmath137 and @xmath138 for the high and low energy end of the spectrum respectively . this approximation is satisfactory for the purposes of this paper , given that the suppression becomes crucial only for particles which emit negligible non - thermal emission . the decay of neutral pions dominates the total emission for energies above @xmath139 mev . the two peaks in the emission reflects the shape of the underlying cr spectrum , which , as illustrated in fig . [ fig : crs ] , is the sum of the steep background cr spectrum , which produces the @xmath140bump at a photon energy of @xmath141 mev ( in the photon flux @xmath142 ) , and an hard cr component coming from the snr that produces the bump at higher energies . the flux level at 1 tev is approximatively @xmath143erg/@xmath144/s , detectable by currently operating imaging atmospheric cherenkov telescopes , even taking into account the quite extended ( @xmath145 ) nature of the source . it is remarkable that such a cloud would be detectable even if it were located at the distance of the galactic centre , as can be easily estimated by taking into account that the sensitivity of a cherenkov telescope like h.e.s.s . after 50 hours of exposure , is @xmath146tev/@xmath144/s , where @xmath147 is the source extension . this means that very massive clouds can be used to reveal the presence of enhancements of the cr density in different locations throughout the whole galaxy . similar conclusions can be drawn for the expected emission in the gev range , which is currently probed by the agile and fermi satellites . in particular , fermi , with a point source sensitivity of @xmath148gev/@xmath149/s at energies above 1 gev ( www-glast.slac.stanford.edu ) , will be able to detect such giant molecular clouds as extended sources if they are located within @xmath150 kpc from the earth , or as point sources if they are at larger distances . such a use of molecular cloud as cr _ barometers _ has been discussed in several papers for both gev @xcite and tev gamma rays @xcite . here we demonstrated that snrs can provide enhancements in the cr density that can generate gamma ray fluxes well within the capabilities of currently operating instruments . the spectral shape in the gamma ray energy range deserves further discussion . for the situation considered in fig . [ fig : gmc2000 ] , the gev gamma rays are the result of the decay of neutral pions produced by background crs that penetrate the cloud . thus , the gamma ray spectrum above gev energies simply mimics the underlying cr spectrum , which is a steep power law of the form @xmath151 . on the other hand , the neutral pion decay spectrum at tev energies is , in this case , dominated by the contribution from crs coming from the nearby snr . after 2000 years from the supernova explosion , only crs with energies above several tens of tev had enough time to leave the snr and reach the cloud and thus the cr spectrum at the cloud exhibits an abrupt low energy cutoff at that energy , that we call here @xmath152 ( see fig . [ fig : crs ] ) . as a consequence , the gamma ray spectrum is expected to be peaked at an energy @xmath153 of several tev . the slope of the gamma ray spectrum below the peak is determined by the physics of the interaction only , and not by the shape of the underlying cr spectrum , and is roughly of the form @xmath154 ( e.g. * ? ? ? thus , a loose association between a snr and a massive molecular clouds as the one studied here , is expected to be characterized , at least at some stage of the snr evolution , by a very peculiar spectrum which is steep at low ( gev ) energies and hard at high ( tev ) energies . the possibility of detecting sources with such a distinct spectrum is also relevant for the issue of identifying gev and tev unidentified sources . one of the criteria suggested to support an association between a gev and tev source is , beside the positional coincidence , the spectral compatibility . in a recent paper , funk et al . ( 2008 ) , investigated the spectral compatibility of egret and h.e.s.s . unidentified sources located in the inner galactic region . for sources showing positional coincidence , they found generally a good spectral compatibility , but due to the small number of sources , it was not possible to claim an association between the two populations of sources at a statistically significant level . for sources detected by only one of the two instruments , funk et al . ( 2008 ) investigated the consequences of extrapolating the measured spectrumat higher or lower energies . they considered only single power laws extending over the whole ( gev tev ) energy range and discussed the implications of adding a high or low energy cutoff . the scenario presented here adds the new interesting possibility of a single source showing a dramatically different behavior at gev and tev energies , namely a spectrum which shows a significant hardening at higher energies . the evolution with time of the emission from the cloud is shown in fig . [ fig : gmctot ] , where the solid , dotted , short dashed and long dashed lines show the spectrum at a time equal to 500 , 2000 , 8000 and 32000 years after the supernova explosion respectively . for comparison , the emission from a molecular cloud , with no snr located in its proximity , is plotted as a dot dashed line . in this case , only background crs that penetrate the molecular cloud contribute to the emission . it is evident from fig . [ fig : gmctot ] that the radio ( @xmath155 mm ) and the soft gamma ray ( @xmath156 mev @xmath157 gev ) emission from the cloud is constant in time . this reflects the fact that such emission is produced by background crs that enter the cloud . the emission in the other energy bands is variable in time , being produced by the crs coming from the snr . the flux of this latter component changes with time as indicated in fig . [ fig : crs ] . the two most prominent features in the variable emission are two peaks , in x and gamma rays respectively . the peak observable at gamma ray energies is the result of the decay of neutral pions produced when crs of different energies coming from the snr reach the cloud at different times . the peak moves at lower and lower energies with time , reflecting the fact that crs with lower and lower energies progressively reach the cloud as time flows . at early times , the emission can extend up to @xmath7 100 tev ( solid line ) , revealing the presence of pev crs and thus indirectly the fact that the nearby snr is acting as a cr pevatron ( see * ? ? ? * for a discussion of this issue ) . moreover , the gamma ray emission in the tev range is enhanced with respect to the one expected from an isolated molecular cloud ( dashed dotted line in fig . [ fig : gmctot ] ) for a period of several @xmath158 yr . this is much longer than the period during which snrs are effectively accelerating the multi - tev crs responsible for the tev emission , which lasts few thousands years . this is because the duration of the gamma ray emission from the cloud is determined by the time of propagation of crs from the snr to the cloud and not by the much shorter cr confinement time in the snr . therefore , the gamma ray emission from the cloud lasts much longer than the emission from the snr , making the detection of clouds more probable @xcite . the peak in the x - ray spectrum is due to synchrotron emission from secondary electrons produced in cr interactions in the cloud . the peak is moving to lower energies with time but , unlike the gamma ray peak , it is also becoming less and less pronounced . this fact can be understood with the help of fig . [ fig : timegmc ] ( left panel ) . after 500 yr from the supernova explosion ( solid line in fig . [ fig : gmctot ] ) , pev crs from the snr reach the cloud and produce there secondary electrons with energy in the @xmath139 tev range . for these electrons , the synchrotron cooling time is comparable with the escaping time from the cloud . thus , they release a considerable fraction of their energy in form of x - ray synchrotron photons before leaving the cloud . as time passes , lower energies crs reach the cloud and secondary electrons with lower energies are produced . for these electrons the cooling time becomes progressively longer than the escape time and this explain the suppression of the synchrotron emission . the x - ray synchrotron emission is weaker than the tev emission for any time and for times @xmath159 yr the ratio between tev and kev emission can reach extreme values of a few tens or more . these values are observed from some of the unidentified tev sources such as hess j1616 - 508 @xcite and more in general unidentified tev sources are characterized by the absence of any clear counterpart at other wavelengths . due to this peculiar spectral properties , such sources have been labeled as _ dark _ , since they seemed to emit gamma rays only . however , in the scenario presented in this paper , spectra showing a high tev / kev flux ratio can be produced very naturally if a cloud is illuminated by crs coming from a nearby snr . this suggestion is also supported by the fact that most of the unidentified tev sources are spatially extended , as molecular clouds are expected to be . in the hard x - ray / soft gamma ray region of the spectrum ( from tens of kevs to hundreds of mevs ) , partially covered by the integral satellite , extremely hard spectra ( @xmath160 ) may result due to the bremsstrahlung emission from primary and secondary cr electrons . finally , the radio emission from the cloud is the result of the synchrotron emission of background cr electrons that penetrate the cloud . the contribution from secondary electrons is subdominant , due to the fact that the @xmath7 gev secondary electrons that might emit synchrotron radio waves escape from the cloud before losing energy ( see fig . [ fig : timegmc ] ) . moreover , in the gev energy range , ionization and bremsstrahlung losses dominate , and this would further reduce the expected synchrotron emission . recently , protheroe et al . ( 2008 ) estimated the expected synchrotron radio emission from secondary electrons produced by cr interactions in the sgr b2 giant molecular cloud . they did not consider the contribution from background cr electrons and neglected the effects of diffusive transport of secondary electrons ( namely , they assumed no penetration of electrons from outside the cloud and instantaneous cooling of secondary electrons produced inside the cloud ) . these assumptions are valid for the very strong magnetic field of the order of one milligauss measured for sgrb2 . such field is much stronger than the one assumed here . thus , due to these somewhat extreme assumptions , possibly justified in the particular case of sgr b2 cloud , a direct comparison between their findings and the results presented in this paper is not straightforward . protheroe et al . ( 2008 ) also noticed that the radio emission from a cloud at frequencies above @xmath42 ghz can be dominated by free free thermal emission . being focused on the non - thermal emission from clouds , we did not attempt here to model their thermal emision . however , our findings can still be tested and constrained by observations in the ghz range . this can be done by requiring the predicted synchrotron emission not to overproduce the observed free free thermal emission . a similar approach has been adopted by jones et al . ( 2008 ) . as noticed in the previous section , concave gamma ray spectra may be produced in a molecular cloud located in proximity of a snr , as the result of the decay of neutral pions produced in cr interactions . such concavity reflects the shape of the underlying cr spectrum , which consist of the superposition of two components : the galactic cr background , characterized by a steep spectrum , and the crs coming from the nearby snr , which exhibit a hard spectrum . with this respect , the distance between the snr and the molecular cloud @xmath6 plays a crucial role . this is because , the larger the distance between the snr and the cloud , the lower the level of the cr flux coming from the snr . moreover , also the time evolution of the emission from a cloud changes with @xmath6 since the time it takes a particle with given energy to cover such a distance scales as @xmath161 , where @xmath88 is the diffusion coefficient . in fig . [ fig : vshaped ] the total gamma ray spectrum from a molecular cloud is shown as a function of the distance between the snr and the cloud . the cloud mass is @xmath82 and the distance from the snr is 50 , 100 and 200 pc for the left , central and right panel , respectively . similarly to fig . [ fig : gmctot ] , the solid , dotted , short - dashed and long - dashed lines refer to the emission for 500 , 2000 , 8000 and 32000 years after the supernova explosion . it is evident from fig . [ fig : vshaped ] that a great variety of gamma ray spectra can be produced . in almost the entirety of the cases considered , the gamma ray emission is characterized by the presence of two pronounced peaks . the low energy peak , located in the gev domain is steady in time and it is the result of the decay of neutral pions produced in hadronic interactions of background crs in the dense intercloud gas . the high energy peak is the result of hadronic interactions of crs coming from the nearby snr , and thus it is moving in time to lower and lower energies ( see previous section for a discussion ) . both the relative intensity and position of the two peaks depend on the distance between the snr and the cloud . interestingly , the gev emission from the cloud is affected by the presence of the nearby snr only at late times after the explosion and only if the distance from the snr is comparable or smaller than @xmath80 50 pc ( see fig . [ fig : vshaped ] , left panel ) . in all the other cases the gev emission is always the result of the interactions of background crs and thus , at least in this case , observations of molecular clouds in the gev gamma ray domain can not be used to infer the presence of a cr accelerator located at a distance greater that @xmath80 50 pc from the cloud . similar concave or `` v shaped '' spectra have been recently obtained by rodriguez marrero et al . ( 2008 ) in a different context in which two molecular clouds are assumed to be located in the proximity of a cr accelerator . if the two clouds happen to be located at different distances from the cr accelerator but within an angular separation smaller than the fermi angular resolution , then they would appear as a single gev source , and the superposition of their emission might result in concave spectra . however , rodriguez marrero et al . ( 2008 ) did not include in their calculations the contribution to the emission coming from the ubiquitous galactic cr background , which in most cases dominates over the contribution of crs from the nearby snr , at least for what concerns the gev emission from the cloud ( see e.g. fig . [ fig : vshaped ] ) . moreover , they also considered very short distances between the cloud and the cr accelerator ( down to 5 pc ) , but in this case an accurate modeling of the cr accelerator itself is needed , especially for what concerns its own gamma ray emission that might add up to the total emission . in addition to that , distances as short as @xmath162 pc are , under many circumstances , smaller than both the source size and the radius of the molecular cloud itself , and this changes significantly the problem since the interactions between the accelerator and the mc are likely to play an important role . the prediction of v - shaped spectra that we make in this paper is more general than the one by rodriguez marrero et al . ( 2008 ) , since it represents an intrinsic feature of a single molecular cloud which is located close to a cr source . the v - shaped gamma ray spectrum reflects the shape of the underlying cr spectrum which is the superposition of the steep spectrum of the background crs and the hard spectrum of crs coming from the nearby snr . the value of the magnetic field in the cloud is an important parameter since it regulates the synchrotron energy losses of high energy electrons , and the diffusive escape time of relativistic particles from the cloud . observations suggest that the value of the magnetic field in a molecular cloud scales with the square root of the gas density , reaching very high values ( 1 mg or more ) in the very dense sub - parsec scale cloud cores @xcite . however , dense cores constitute a very small fraction of the total volume of molecular clouds . since here we are interested in calculating the emission from the whole cloud , the relevant parameter is the volume averaged value of the cloud magnetic field . it seems reasonable to assume that the total magnetic energy in the cloud @xmath163 , where @xmath164 is the cloud volume , does not exceed the total gravitational energy of the cloud @xmath165 . this leads to a maximum value of the average magnetic field of : @xmath166 g . this is in general agreement with observations , from which a value of @xmath83 g can be inferred for typical cloud densities of @xmath139 @xmath85 . however , since the dispersion around this mean value is considerable ( see e.g. * ? ? ? * ) , it is worth to investigate how the non thermal emission from a molecular cloud depends on the actual value of the magnetic field . [ fig : bfield ] shows the broad band spectrum from a molecular cloud of mass @xmath82 , radius @xmath167 pc , density @xmath168 @xmath85 . the distance between the cloud and the snr is 50 pc . in the top and bottom panels is plotted the cloud emission after 500 and 2000 years from the supernova explosion , respectively . the solid line refers to a value of the magnetic field of 30 @xmath169 g , while the dashed line represents the emission for a smaller value of the field equal to 10 @xmath169 g . the cloud emission from radio frequencies up to the hard x - ray band strongly depends on the value of the magnetic field , while the gamma ray emission is unaffected , being the result of hadronic interactions of cr protons . the strong dependence of the radio and x - ray emission on the magnetic fields is evident , and this demonstrate that this effect has to be taken into account in multi wavelength studies of molecular clouds . the non thermal emission from a molecular cloud located in proximity of a snr is the result of the interactions of crs that penetrate the cloud . both crs in the galactic background and runaway crs from the snr contribute to the emission . in this paper , we calculated the expected non thermal emission from the molecular cloud as a function of several parameters , such as the time after the supernova explosion , the distance between snr and cloud , and the cloud magnetic field . in all the cases , the gamma ray emission from the cloud is by far exceeding the energy output in other energy bands . the gamma ray emission from the cloud consists of two distinct components . the first one is due to the interactions of crs from the galactic background , which are characterized by a steep spectrum . this component is steady in time and peaks in the gev energy region . the second component is the result of the interactions of runaway crs that escape the snr and diffusively reach the cloud . the spectrum of these runaway crs is hard and variable in time and both this characteristics are reflected in the related gamma ray emission . in particular , this component is producing a second peak in the gamma ray spectrum which moves to lower and lower energies , reflecting the fact that high energy crs are released from the snr earlier and diffuse faster than low energy ones , and as a consequence they reach the cloud earlier . the superposition of these two components of the gamma ray emission may produce concave spectra , which are steep at low ( @xmath7 gev ) energies and hard at high ( @xmath7 tev ) energies . this peculiar spectra might be revealed by means of joint observations of fermi and ground based cherenkov telescopes . the detection of such emission would provide a strong hint for the presence of a cr accelerator in the neighbourhood of the molecular cloud . it has been suggested that some of the _ dark _ tev unidentified sources , with no obvious counterparts at any other wavelength , might be indeed associated with molecular clouds illuminated by crs from nearby accelerators such as snrs @xcite . we showed here that , in this scenario , the _ `` darkness '' _ of a source , often defined as the radio @xmath170 between the tev and x - ray observed flux , depends on several parameters such as the distance between the snr and the cloud , the time after the supernova explosion and the value of the magnetic field in the molecular cloud . however , in most cases it is very natural to obtain very high values for @xmath170 , compatible with the estimates on @xmath170 claimed for some of the unidentified tev sources ( see e.g. * ? ? ? * ; * ? ? ? this further supports an association between unidentified tev sources and molecular clouds and suggests that the unknown nature of these objects might be soon revealed .

the search for the exotic states ( tetraquarks , hybrids , hadronic molecules , etc ) in the observed hadron spectrum has been a subject of both theoretical and experimental investigations for decades . the exact pattern , how these states emerge , should be strictly determined by the underlying theory and should therefore contain important information about the behavior of qcd at low energies . in practice , however , extracting such information from the data encounters certain challenges , which are in part of a conceptual nature . in the present paper we wish to focus exactly on this issue . in general , a state is called `` exotic '' if its quark content does not correspond to the `` standard '' constellation given by the non - relativistic quark model ( @xmath2 for mesons and @xmath3 for baryons ) . consequently , one needs to use a particular model as a reference point to define how the exotic states are meant ( note that the very notion of constituent quarks is , strictly speaking , model - dependent ) . putting it differently , one has to _ agree _ on certain _ criteria _ formulated in terms of certain hadronic _ observables _ : if these observables are measured , or calculated on the lattice , and the results do not follow the pattern predicted by the quark model , this then should be interpreted as a signature for exotica . a standard example for the exotic state candidates is given by the scalar nonet with the masses around 1 gev . as it is well known , the observed mass hierarchy in this nonet is reversed as compared to , e.g. , the pseudoscalar or vector multiplets . such a mass ordering is counter - intuitive from the point of view of the naive quark model , but can be easily understood , if the scalar mesons were interpreted as tetraquark states ( see , e.g. , @xcite ) . this is , however , not the only possible interpretation . in refs . @xcite , the @xmath4 and @xmath5 were considered as hadronic molecules , whereas in refs . @xcite these states were described as a combination of a bare pole and the rescattering contribution . in the jlich meson - exchange model , the @xmath5 appears to be a bound @xmath6 state , whereas the @xmath4 is a dynamically generated threshold effect @xcite . similar conclusions were inferred in ref . @xcite from the calculations in the unitarized chpt with explicit resonance states . finally , the investigations carried out within the framework of qcd sum rules are also indicative of the non-@xmath2 nature of @xmath4 @xcite . given these multiple interpretations , it is natural to look for the clear - cut criteria based on the _ observables _ in order to minimize the _ model - dependence _ of the statements about the nature of the hadronic states in question . in fact , such criteria are known for quite some time already . the `` pole counting '' method , considered in refs . @xcite , relates the number of the @xmath7-matrix poles near threshold to the molecular nature of the states corresponding to these poles . namely , it has been argued that the loosely bound states of hadrons ( hadronic molecules ) correspond to a single pole , whereas the poles corresponding to the tightly bound quark states ( of standard or exotic nature ) always come in pairs . a closely related criterion goes under the name of weinberg s compositeness condition @xcite , which uses the quantity called the wave function renormalization constant @xmath0 , where @xmath8 , to differentiate between the loosely bound states and tight qcd composites , the values @xmath9 corresponding to the molecular states and _ vice versa . _ the application of these methods for the analysis of the data on scalar mesons are considered in refs . @xcite , and the recent review on the subject may be found in ref . moreover , _ theoretically , _ one may study the dependence of the pole positions on the number of the colors @xmath10 ( see refs . @xcite ) or the quark masses ( refs . @xcite ) . from the above studies , one can judge about the precise structure of these states beyond the simple alternative between a molecule and a tight quark composite . recent years have seen a renewed interest in the field , which is partly related to the progress in the lattice calculations of the qcd spectrum at the quark masses close to the physical values . it should be realized that the lattice studies have powerful tools at their disposal to analyze the nature of the states that emerge in qcd . apart from the information about the dependence of the spectrum on quark masses , a valuable information comes from the volume dependence of the calculated spectrum as well as its dependence on the twisting angle in case of twisted boundary conditions , see refs . note that all this information is obtained from the first - principle calculations on the lattice and is thus in principle devoid of any model - dependent input . in this paper we investigate the nature of the scalar states in the sector with one charm quark that is a natural generalization of our treatment of the light scalar mesons . we mainly focus on the case of the @xmath11 meson @xcite , albeit the formalism , which we develop here , can be straightforwardly applied to the other cases where a bound state close to the elastic threshold emerges ( note that , in this paper , we do not consider the generalization of the approach to the inelastic case . this forms a subject of a separate investigation . ) . the @xmath11 does not fit very nicely to the quark - model picture , and its structure is still debated , see , e.g. , ref . @xcite for a recent review . the molecular picture , due to the closeness of the @xmath1 threshold and a large coupling to the @xmath1 channel looks most promising among other alternatives . it would be highly desirable to verify this conjecture in a model - independent manner , on the basis of the lattice calculations . to this end , one may use the fact that the dependence of the bound - state energy on the kaon mass is very different for a molecule and a standard quark - model state , see ref . another possible method to address this issue has been described , e.g. , in refs . @xcite , where the authors propose to study the volume - dependence of the spectrum in order to apply the weinberg s compositeness criterion on the lattice . the exploratory study of light pseudoscalar mesons @xmath12 of @xmath13 in full lattice qcd has been carried out in refs . @xcite . in some isospin channels the study is plagued by the presence of disconnected contributions . the implementation of the method from refs . @xcite , which implies carrying out calculations at different volumes , could be therefore quite expensive . in this paper we propose an alternative , which requires calculations at one volume , albeit with twisted boundary conditions . moreover , we show that , in the study of @xmath11 , one may use _ partially twisted _ boundary conditions , despite the fact that the quark annihilation diagrams are present . the method used in the proof is the same as in ref . generally , one may expect that the simulations with partially twisted boundary conditions could be less expensive than working at different volumes , while they provide us the same information about the nature of the bound states in question . this article is organized as follows . in sect . [ compositeness ] , we briefly review weinberg s argument for the compositeness of particles . in sect . [ sec : mass - shift ] we describe the procedure of extraction of the parameter @xmath0 from the data with twisted boundary condition . further , in sect . [ models ] we use some models and produce synthetic lattice data in order to check the procedure of the extraction in practice . the error analysis has also been carried out . separately , in sect . [ partially ] , we discuss the use of the partially twisted boundary conditions and show that they are equivalent to the full twisting in our case . [ conclusions ] contains our conclusions . as mentioned before , in view of the plethora of candidates of exotic hadrons , it is very important to make model - independent statements on the nature of these states . model - independence requires that we can only study the physical observables which can be defined in terms of the matrix elements between asymptotic states . in particular , we would like to ask a question , whether a given particle , corresponding to the @xmath7-matrix pole , can be regarded as `` elementary '' or rather as a bound state ( molecule ) of other hadrons . the central place in this identification belongs to the so - called wave function renormalization constant @xmath0 , which has been used to distinguish composite particles from elementary ones since the early 1960 s @xcite . to see its role , we will first discuss a non - relativistic quantum mechanical system , following the discussion of ref . @xcite . in this section , we will restrict our discussion to the infinite volume . let us consider a two - body system with a hamiltonian @xmath14 , where @xmath15 is the free hamiltonian , and @xmath16 specifies the interaction . both @xmath17 and @xmath15 have a continuum spectrum . let us assume that there is a bound state solution of the schrdinger equation with a binding energy @xmath18 , @xmath19 and @xmath15 also has a discrete spectrum which are the bare elementary particles . for simplicity , we will assume that there is only one such state , denoted by @xmath20 . in the hilbert space spanned by the eigenstates of the free hamiltonian , the completeness relation is thus given by @xmath21 where @xmath22 is the reduced mass . thus , the probability for the physical state @xmath23 overlapping with the elementary state @xmath20 which , by definition , equals to @xmath0 , is given by @xmath24 ^ 2 } , \label{eq : bb0}\end{aligned}\ ] ] where eq . is used . the quantity @xmath25 then describes the probability of the physical state not being the elementary state or finding the physical state in the two - particle state . in other words , @xmath26 corresponds to a mostly elementary state whereas a state with @xmath9 can be interpreted as a predominately molecular one . in general , the above integral depends on the matrix element @xmath27 , which is not directly measurable . however , for loosely bound states , the quantity @xmath0 can be related to the observables . consider , for instance , an s - wave bound state with a small binding energy . the binding energy should be much smaller than the inverse of the range of forces so that the matrix element @xmath27 can be approximated by a constant @xmath28 . we get from eq . @xmath29 note that , in the past , this equation has been often applied to distinguish composite particles from elementary ones , see e.g. @xcite . the non - relativistic coupling constant @xmath30 coincides with the residue of the non - relativistic scattering matrix at the bound state pole . this can be immediately seen , considering the low equation @xmath31 in the vicinity of the pole @xcite . here , @xmath32 . finally , we would like to relate the quantity @xmath0 to the physical observables , namely , to the scattering length @xmath33 and effective range @xmath34 . here , we are closely following the path of ref . it is important to note that these relations can be derived when the binding energy is much smaller than the inverse of the range of forces . we start with the twice - subtracted dispersion relation for the inverse of @xmath35 @xmath36 where the two subtraction constants have been determined from eq . . the s - wave transition matrix element is related to the non - relativistic s - wave scattering amplitude @xmath37 $ ] as @xmath38 with @xmath39 and @xmath40 being the s - wave phase shift . thus , one gets @xmath41 . inserting this into eq . , we obtain @xmath42 where @xmath43 denotes the characteristic distance between the constituents in the two - body bound system . comparing the above expression with the effective range expansion @xmath44 , and using eq . , one can express the scattering length and effective range in terms of the binding energy and compositeness @xcite @xmath45 therefore , for an s - wave shallow two - body bound state , the compositeness can be measured by measuring the low - energy scattering parameters . next , we turn to the compositeness condition within the framework of the quantum field theory . for simplicity , let us first consider the situation when a scalar particle described by a field @xmath46 with the bare mass @xmath47 couples with two scalars @xmath48 with the masses @xmath49 . the interaction lagrangian takes the form @xmath50 . consider now the two - point function of the field @xmath46 @xmath51 summing up one - loop bubble diagrams to the two - point function , one arrives at the expression ( see fig . [ fig : bubble ] ) @xmath52 where the one - loop self - energy is given by @xmath53 the relativistic scattering amplitude for the process @xmath54 in the same approximation is given by ( see fig . [ fig : bubble ] ) is used in the definition of the @xmath55-matrix . ] @xmath56 the relativistic and the non - relativistic scattering matrices are the same up to an overall normalization . in the rest frame of the bound system , the relation takes the form @xmath57 where @xmath58 . now , let us consider the behavior of the scattering amplitude in the vicinity of the bound - state pole . the two - point function has the following behavior @xmath59 where @xmath60 is the physical mass . ( a ) and the two - point function of the field @xmath61 ( b ) . only one - loop bubbles are summed up . solid ( dashed ) lines denote @xmath62 ( @xmath61 ) fields , respectively.,width=377 ] the residue of the propagator determines the wave function renormalization constant for the particle @xmath61 : @xmath63 where @xmath64 is the renormalized coupling constant , and @xmath65 . in order to establish the relation of the quantity @xmath0 , defined by eq . ( [ eq : zrel ] ) , with its non - relativistic counterpart , we perform the contour integration over @xmath66 of the loop integral in eq . : @xmath67 where @xmath68 and @xmath69 . in the rest frame of the bound state , one has @xmath70 . taking derivative with respect to @xmath71 , and then taking the non - relativistic approximation which amounts to @xmath72 and @xmath73 , we get @xmath74 ^ 2 } , \ ] ] where we have used @xmath75 . taking into account the difference between relativistic and non - relativistic normalizations , we finally arrive at the relation @xmath76 , cf . with eq . ( [ eq : low ] ) . comparing now this relation with eq . , one immediately sees that the wave function renormalization constant @xmath0 is the same as its non - relativistic counterpart and thus the compositeness condition for an s - wave bound state can be written as @xmath77 one might treat the above argumentation with a grain of salt , since it is based on certain approximations . namely , the amplitude is given as a sum of one - loop diagrams only . it is , however , clear that the result is valid beyond this approximation , if bound states close to an elastic threshold are considered . the justification is provided by the statement that such bound states can be consistently described within a non - relativistic effective field theory , which is perturbatively matched to the underlying relativistic theory ( see , e.g. , ref . @xcite for a review on the subject ) . such an effective theory is equivalent to the non - relativistic quantum mechanics ( the number of particles is conserved ) and hence the compositeness can be rigorously defined along the lines discussed above . finally , we would like to mention that the quantity @xmath0 , which is defined in eq . ( [ eq : zrel ] ) , is ultraviolet finite , since the quantity @xmath78 is defined through the residue of the _ renormalized _ scattering amplitude . as stated above , the wave function renormalization constant , @xmath0 , gives an overlap of the physical state with the elementary state and hence could be used as a parameter that describes the compositeness of a given state . lattice calculations provide a model - independent way to determine @xmath0 from the volume dependence of the spectrum @xcite , or as we propose in this paper from the dependence on the twisting angle . in this section we set up a finite - volume formalism , which describes the dependence of the bound - state mass on the volume or twisting angle . we consider elastic scattering of particles with the masses @xmath79 and @xmath80 in the s - wave . then , generally , a unitary partial - wave amplitude in infinite volume is given by @xmath81 where @xmath82[s-(m_1-m_2)^2]$ ] is the relative momentum squared in the center of mass ( c.m . ) frame . further , the function @xmath83 ( `` the inverse potential '' ) is a regular function in the vicinity of the threshold . the notation used here is reminiscent of that of unitarized chiral perturbation theory , but eq . may in fact describe any elastic unitary amplitude , with the particular dynamics encoded in the function @xmath84 . the loop function @xmath85 is given by eqs . and . this function contains a unitarity cut . across this cut , we have @xmath86 . other ( distant ) cuts that may be also present are included in @xmath84 . the loop function @xmath85 is divergent and has to be renormalized . here we do the renormalization with a subtraction constant . as it will be seen below , the extension to the finite volume is independent of any regulator . when the particles are put in a finite box of size @xmath87 , their momenta become discretized due to boundary conditions . so , the continuum spectrum , which gives rise to the cut in the infinite volume , becomes a discrete set of two - particle levels . in order to obtain the spectrum in a finite volume , one should replace the momentum integrals by the sums over the discretized momenta in the expression of the scattering amplitude . then , the `` finite volume scattering amplitude '' @xmath88 contains poles on the real axis that correspond to the discrete two - particle levels . it should be noted that the finite - volume effects in @xmath84 are exponentially suppressed ( see , e.g. , @xcite ) , so the the finite volume scattering amplitude can be obtained just by changing the loop function by its finite volume counterpart @xmath89 @xcite , where @xmath90\ , .\ ] ] here @xmath91 denotes the integrand in eq . , and @xmath92 the allowed momenta in a finite volume , whose value depends on the box size @xmath87 and the boundary conditions used . for the periodic boundary conditions we have @xmath93 , @xmath94 . in case of twisted boundary conditions , the momenta also depend on the twisting angle @xmath95 according to @xmath96 . using the methods of ref . @xcite , it can be shown that @xmath97 can be related to the modified lscher function @xmath98 , see appendix [ appendix - deltag ] , @xmath99 where @xmath100 and the dots stand for terms that are exponentially suppressed with the volume size @xmath87 @xcite . in this paper , we are going to apply lscher formalism to study shallow bound states , where the finite - volume effects are exponentially suppressed . since , for such states , the binding momentum @xmath101 is presumed to be much smaller than the lightest mass in the system , the exponentially suppressed corrections emerging , e.g. , from the potential @xmath84 could be consistently neglected as compared to the corrections @xmath102 that arise from @xmath103 . note however that , if masses of the constituents increase for a fixed binding energy , then the magnitude of the binding momentum also increases and , for the bound states of heavy mesons , may become comparable to the pion mass . in this case , further study of the problem is necessary . a recent example of such a study ( albeit in the light quark sector ) is given in ref . @xcite . in the present paper this issue is not addressed . finally , note that the divergences arising at @xmath104 in eq . cancel between the sum and the integral , so we can safely send the cutoff to infinity . thus , @xmath97 does not depend on any regulator . in appendix [ appendix - deltag ] we show in detail , how @xmath105 could be calculated below threshold for different types of boundary conditions . bound states show up in the scattering amplitude as poles on the real axis below threshold . namely , if we have a bound state with the mass @xmath60 in the infinite volume , the scattering amplitude should have a pole at @xmath106 , with the corresponding binding momentum @xmath107 , @xmath108 . from eq . , it is clear that @xmath60 and @xmath109 satisfy the equation @xmath110 = 0,\ ] ] where @xmath111 is the analytic continuation of @xmath112 for arbitrary complex values of @xmath113 , which is needed since the bound state is located below threshold , @xmath114 . on the other hand , the discrete levels in a finite volume are obtained as the poles of the finite - volume scattering amplitude @xmath88 and , in particular , the bound state pole gets shifted to @xmath115 , with binding momentum @xmath116 , given by @xmath117 note that , below threshold , both @xmath118 and @xmath97 are real , so the pole position is real . the discrete scattering levels above threshold are real as well ( as they should be ) , since the imaginary part of @xmath97 cancels exactly with that of @xmath118 . next , we relate the finite - volume pole position with the infinite - volume quantities as the bound state mass , @xmath60 , and the coupling , @xmath119 ( defined as the residue of the scattering amplitude at the pole @xmath106 ) . to this end , we expand @xmath120 around the infinite - volume pole position , @xmath121 , @xmath122 where the prime denotes a derivative respect to @xmath113 . then , evaluating the residue at @xmath123 in eq . we obtain @xmath124 where the derivative @xmath125 is to be evaluated at @xmath106 . finally , using eqs . and , we obtain for the pole position shift @xmath126\ ] ] this equation gives the bound state pole position , @xmath127 ( or , equivalently , @xmath128 ) as a function of the infinite - volume parameters @xmath119 and @xmath101 . it is worth noting that , within the approximation , the position of the bound state pole in a finite volume depends only on these two parameters . this approximation works remarkably well in all cases considered in this paper . if the difference @xmath129 is small enough , eq . can be solved iteratively . for periodic boundary conditions , with the use of eq . , it can be shown that the lowest - order iterative solution reads @xmath130 which coincides with the result given in refs . however , it will be shown below that , for shallow bound states , where @xmath101 is very small , one should take more than just the first term in the sum . moreover , in some cases , the iterations converge very slowly , if at all . therefore , in our opinion , it is safer to consider solving eq . numerically , without further approximations , in order to obtain the finite volume pole position @xmath127 this is the way we proceed . using eq . , it is possible to fit the infinite - volume parameters @xmath60 and @xmath119 from the bound state levels @xmath127 , obtained through lattice simulations at different @xmath87 or @xmath95 . this , in turn , allows one to determine the compositeness parameter from eq . . however , in actual lattice simulations , the measured energy levels have some uncertainty , and the number of different volumes or different twisting angles might be not very large . therefore , it is important to know in advance , at which accuracy should be the lattice measurements carried out , in order to render the extraction of the parameter @xmath0 reliable . we address this question in some exactly solvable models with a given @xmath84 , producing `` synthetic lattice data , '' adding random errors and trying to extract back the infinite volume parameters @xmath131 and @xmath0 from data . the potential in this model is given by a `` bare state pole '' , @xmath132 which depends on two parameters : a bare pole position @xmath133 and a bare coupling constant @xmath134 . by appropriately choosing the value of the bare parameters , we can reproduce a bound state with any given mass @xmath60 and coupling @xmath78 . if our model describes the interaction of two particles , where a bound state with the mass @xmath60 is present , the scattering partial wave amplitude should have a pole at @xmath106 , @xmath135 the physical coupling of the bound state , @xmath78 , is given by the residue of the scattering partial - wave amplitude at the bound state pole @xmath136g_0 ^ 2=zg_0 ^ 2\ , .\ ] ] one can use above equations to trade the bare parameters for the physical ones in the expression of the scattering amplitude and write the latter in terms of @xmath60 and @xmath0 : @xmath137}.\ ] ] note that the above amplitude does not depend on the subtraction constant that renders @xmath85 finite . this model can describe a bound state with any given value of the wave function renormalization constant . next , we study the finite volume effects in the bound - state mass . in the actual calculations , we take @xmath138 , @xmath139 and choose the mass of the bound state to be @xmath140 mev . this is a shallow bound state at 20 mev below threshold , which corresponds to a binding momentum @xmath141 mev . for the mainly molecular state we take @xmath142 , and @xmath143 is chosen for the mainly elementary one . for each of these two states , we calculate their finite - volume mass @xmath115 as the subthreshold pole position in the finite - volume scattering amplitude . in the left panel of fig . [ toy - dependence ] , we show the mass of the two states with @xmath142 and @xmath143 as a function of @xmath87 for periodic boundary conditions and do not discuss the pion mass dependence . ] . these are obtained from the solution of the exact equation . it is easy to see that the finite volume effects are much bigger in the case of the molecular state with @xmath142 than in the case of an elementary state with @xmath143 . this was of course expected in advance , since small finite - volume effects point on a compact nature of the state in question . here we also plot the solutions of eq . , using the known values of @xmath60 and @xmath78 , taken from the infinite volume model . in this way we can test the validity of the approximation in eq . , used to derive eq . from eq . , which basically states that all relevant dynamics is encoded only in the two parameters @xmath60 and @xmath78 . as can be seen in fig . [ toy - dependence ] , eq . is able to reproduce the synthetic lattice results very accurately . on the other hand , note that for shallow bound states the binding momentum @xmath101 is small , so no wonder that the expansion in @xmath97 converges rather slowly . consequently , retaining only the leading - order term and constructing iterative solution , see eq . , might not be sufficient in all cases . in the right panel of the same figure we show the dependence of the bound - state mass on the twisting angle @xmath144 for the fixed value of @xmath145 . we see that , for such a choice of twisting , the size of the effect of twisting for a fixed @xmath87 is almost the double of the maximal effect caused by the variation of @xmath87 from the same value to infinity ( periodic boundary conditions ) . thus , using ( partially ) twisted boundary conditions to determine @xmath0 , besides being cheaper , could give more accurate results than a method based on the study of the volume - dependence of the energy level . note also that , for the above choice of the twisting angle , the twisting effect is maximal . other choices , e.g. , @xmath146 lead to a smaller effect . now we turn our attention to the realistic case of the hadronic bound state @xmath11 in the @xmath1 scattering channel with isospin @xmath147 and strangeness @xmath148 . when isospin symmetry is exact , this state is stable under strong interactions , since it does not couple to the lighter hadronic channels ( the observed decay @xmath149 breaks isospin symmetry ) . thus , the formalism above , tailored for stable bound states , does apply in this case . the case of quasi - bound states , which are coupled to inelastic channels , requires special treatment and is not addressed here . a popular view on the @xmath11 meson is that this state is dynamically generated as a pole through the s - wave interactions between the @xmath150-meson and the kaon in the isoscalar channel @xcite . we shall study this system , using the model used from ref . @xcite , which is based on the leading - order heavy flavor chiral lagrangian @xcite and unitarizes the amplitude @xcite . namely , the infinite - volume amplitude is obtained from eq . with the s - wave - projected potential @xmath151,\ ] ] where @xmath152 is the cosine of the scattering angle , @xmath153 mev is the pion decay constant , and @xmath71 and @xmath154 are usual mandelstam variables . we regularize the loop function with a subtraction constant @xmath155 , as done in refs . its value at the scale @xmath156 is taken to be @xmath157 . with this value of the subtraction constant , we find a bound state pole , associated with the @xmath11 , at @xmath158 , and the coupling to @xmath1 , which is given by the residue of the pole , @xmath159 takes the value @xmath160 . one can easily calculate the compositeness parameter of the bound state as well , using eq . . the calculation yields @xmath161 . hence , in this model , the @xmath162 is predominately a molecular state . next , we study this model in a finite volume and consider twisting of different quarks , from which the @xmath150 and @xmath163 mesons consist . the net effect is that these mesons get different momenta as a result of such twisting , so the expression for @xmath164 changes . note that this issue is important in view of the fact that partial twisting is allowed only for certain quarks ( see section [ partially ] for more details ) . in fig . [ dk - l - dependence ] , we display the volume dependence of the bound state mass for different twisting angles which are again chosen as @xmath144 . in the left panel , we plot the @xmath87-dependence for three different values of the twisting angle , when twisted boundary conditions are applied to the @xmath154-quark . in the right panel , twisted boundary conditions are applied to the @xmath71-quark . as we shall see later , in the latter case the use of partial twisting gives the same results as using fully twisted boundary conditions . the size of the finite volume effects , using twisted boundary conditions for the @xmath165-quark , is very small , so we do not discuss this case . in this model , we test again that the predictions obtained from eq . , using the values of @xmath60 and @xmath78 from the infinite - volume model , reproduce very well the exact solution . consequently , all relevant dynamics of the model near threshold is encoded in just two parameters @xmath78 and @xmath60 . on the other hand , we see that retaining only the leading exponential in the expansion of @xmath164 will have a large impact on the accuracy . consequently , the first few terms should be retained . we see that the convergence is satisfactory : e.g. , taking @xmath166 , where @xmath167 denotes the number of terms retained in the expansion , we see that the largest difference between the synthetic data and the prediction from eq . is less than 0.1 mev . analyzing fig . [ dk - l - dependence ] , we again come to the conclusion that the use of ( partially ) twisted boundary conditions can provide a better way to extract the compositeness parameter @xmath0 from lattice results . this can already be seen by comparing the curves for @xmath168 and @xmath169 . one namely observes that the size of the effect due to twisting at a fixed volume is almost twice as big as due to changing the volume for periodic boundary conditions . in fig . [ dk - theta - dependence ] , for three different volumes , we show the dependence of the bound - state mass on the twisting angle both for @xmath154- and @xmath71-quark twisting . on the other hand , taking the results of the @xmath170-dependence ( like in fig . [ dk - theta - dependence ] ) at a fixed volume for granted , one could fit the value of the infinite - volume mass and coupling constant to these data , using eq . . after this , it is straightforward to obtain the value of @xmath0 . in fact , producing four synthetic lattice data points at a fixed @xmath171 and @xmath172 ( either for @xmath154- or @xmath71- quark twisting ) , we were able to obtain values for @xmath60 and @xmath78 that differ less than 1% from those calculated from the infinite volume model by fitting the solution to eq . ( with @xmath173 ) to the synthetic data . real lattice simulations , however , produce results which carry uncertainties . hence , the question arises , how big these errors could be in order to be still able to determine @xmath0 with a desired accuracy . since , as seen from the figures , the finite volume effects ( for reasonable volume sizes , say , above @xmath171 ) are at most around 10 mev , one expects that a relatively high accuracy will be needed in the measurement of the bound - state energy . in order to determine , how high this accuracy should actually be , we assign an uncertainty to the synthetic data that we generate from our model . in particular , using the von neumann rejection method , from the `` exact '' data points we generate a new , `` randomized '' data set , where the central values of each data point are shifted randomly , following the gaussian distribution centered at exact data values and with a standard deviation , given by the lattice data error . repeating this process several times , we obtain several sets of synthetic lattice data with errors and central values shifted accordingly . we then fit each of the randomized data sets and obtain a corresponding value for @xmath60 and @xmath78 ( and therefore , for @xmath0 ) , one for each set , ending up with as many values for the parameters , as many randomized data sets we have generated . we can obtain then the mean and standard deviation of the distributions for @xmath60 , @xmath78 and @xmath0 . thus , for a given data error , we can estimate the accuracy of the parameter extraction . .the accuracy of the extraction of the parameter @xmath0 from the fits to the synthetic lattice data for different input error @xmath174 . four or eight data points and two different volumes @xmath171 and @xmath175 were used , see main text for details . [ cols="^,^,^,^,^ " , ] here , we address the same problem in the context of the @xmath1 scattering . the method is described in ref . @xcite , to which the reader is referred for further details . consider first the scattering in the infinite volume . we start from building the channel space by tracking the quarks of different species following through the quark diagrams describing the @xmath1 scattering . it is clear that , since only light quarks may annihilate , the possible final states contain _ valence _ , _ sea _ or _ ghost _ light quarks with equal masses , as given in table [ tab : channels ] . omitting channel indices , the resulting _ algebraic _ lippmann - schwinger equation couples 3 different channels [ eq : ls - full ] t = v+vg_dkt , where @xmath55 , @xmath16 and @xmath176 are given by @xmath177 matrices . the free green function is given by [ eq : g ] g_dk(s ) = g(s ) ( 1,1,-1 ) where @xmath85 is defined in eqs . and , supplemented by the prescription that the integral is performed in dimensional regularization _ after _ expanding the integrand in powers of 3-momenta ( see refs . @xcite for details ) . the minus sign on the diagonal of the matrix @xmath176 arises due to fermionic nature of @xmath150 and @xmath163 mesons composed of valence and ( commuting ) ghost quarks . the crucial point now is that there exist linear symmetry relations between various elements of @xmath55 due to equal valence , sea and ghost quark masses . note that scattering matrix elements are given by residues of the 4-point green functions @xmath178 of the bilinear quark operators at the poles , corresponding to the external mesonic legs . decomposing @xmath178 into _ connected _ @xmath179 and _ disconnected _ @xmath180 pieces through wick contractions ( see fig . [ fig : cd ] ) and noting that quark propagators are the same for all light quark species , we get [ eq : cd ] _ 11=_22=t_c - t_d , _ 33=-t_c - t_d , _ 12=_13=_23=_21=_31=_32=-t_d , ) and disconnected ( @xmath180 ) diagrams , emerging in @xmath181 scattering amplitudes with various quark species ; @xmath182=@xmath183 . ] since in our case there are no neutral states and thus no mixing occurs , following the argumentation given in ref . @xcite , it is easy to show that @xmath55-matrix obeys the same symmetry relations as @xmath184 [ eq : ex1 ] t_11=t_22=t , t_33=-t+2y , t_12=t_13=t_23=t_21=t_31=t_32=y . here @xmath185 corresponds to the _ physical _ elastic @xmath1 scattering amplitude , i.e scattering in the sector with valence quarks only . other diagonal entries are _ unphysical _ in the sense that they correspond to scattering of particles , composed of sea and ghost light quarks . non - diagonal elements of @xmath55-matrix describe coupling between valence and sea / ghost sectors through _ disconnected _ diagrams . furthermore , it is straightforward to check from eq . ( [ eq : ls - full ] ) that the elements of potential matrix @xmath16 satisfy the same symmetry relations as @xmath55 and can be expressed in the following form [ eq : v - matrix ] v= ( c c c & & + & & + & & -+2 + ) , let us now turn to the case of a finite volume and derive the lscher equation for a couple of particular choices of partially twisted boundary conditions . note that the potential @xmath16 remains the same ( up to exponentially suppressed in terms @xmath87 ) , while in the loop functions the integration is substituted by summation over lattice momenta . 1 . twist the @xmath71/@xmath165-quark , leaving @xmath154 and @xmath186-quarks to obey periodic boundary condition . in this case , the matrix of the green functions is @xmath187 . the solution of the lippmann - schwinger equation in a finite volume for the _ physical _ amplitude @xmath188 is given by t= , where @xmath189 is the loop function @xmath85 in a finite volume . we see that the finite - volume spectrum in case of the partial twisting is determined from the lscher equation 1-g_l^(s)=0 , in the same way as in the full - twisting case . thus , the results obtained by using of the partially twisted boundary conditions on the @xmath165- or @xmath71-quark are equivalent to those using full twisting . 2 . twist the valence @xmath154- and @xmath186-quarks simultaneously , leaving @xmath71- and @xmath165-quarks obey periodic boundary condition . in this case , the ghost light quarks also need to be twisted , and the matrix of the green functions is @xmath190 . + the lscher equation determining the finite volume spectrum now takes the form ^2=0 . vanishing of the first bracket on the r.h.s gives the lscher equation with _ no twisting_. note also that the quantity @xmath191 is in fact the connected part of the scattering potential for the isoscalar @xmath1 system , which is identical to the @xmath1 scattering potential in the isovector channel . hence , vanishing of the second bracket is equivalent to the fully twisted lscher equation for the isovector @xmath1 scattering scattering , partial twisting is always equivalent to the full twisting in this case . ] . * lattice qcd does not only determine the hadron spectrum . under certain circumstances , it may provide information about the nature of hadrons , which renders lattice simulations extremely useful for the search and the identification of exotic states . note that the lattice qcd possesses unique tools at its disposal ( e.g. , the study of the volume and quark mass dependence of the measured quantities ) , which are not available to experiment . * in the present paper , we concentrate on the identification of hadronic molecules on the lattice . experimentally , one may apply weinberg s compositeness condition to the near - threshold bound states , in order to distinguish the molecular states from the elementary ones . to this end , one may use the value of the wave function renormalization constant @xmath0 which obeys the inequalities @xmath8 . the vanishing value of the parameter @xmath0 corresponds to the purely molecular state . in this paper we consider the lattice version of the weinberg s condition . * it is known that the quantity @xmath0 can be extracted from lattice data by studying the volume dependence of the measured energy spectrum . we have shown that the same result can be achieved by measuring the dependence of the spectrum on the twisting angle in case of twisted boundary conditions . moreover , within the method proposed , the expected effect is approximately twice as large in magnitude and comes at a lower computational cost . further , we have analyzed synthetic data to estimate the accuracy of the energy level measurement which is required for a reliable extraction of the value of @xmath0 on the lattice . * as an illustration of the method , we consider the @xmath11 meson , which is a candidate of a @xmath1 molecular state . it is proven that , despite the presence of the so - called annihilation diagrams , one may still use the partially twisted boundary conditions for the extraction of @xmath0 from data if the charm or strange quark is twisted . the effects which emerge due to partial twisting , are suppressed at large volumes . the authors thank m. dring , l. liu , u .- g . meiner and s. sasaki for interesting discussions . this work is partly supported by the eu integrated infrastructure initiative hadronphysics3 project under grant agreement no . we also acknowledge the support by the dfg ( crc 16 , `` subnuclear structure of matter '' ) , by the dfg and nsfc ( crc 110 , `` symmetries and the emergence of structure in qcd '' ) , by the shota rustaveli national science foundation ( project di/13/02 ) , by the bonn - cologne graduate school of physics and astronomy , by the nsfc ( grant no . 11165005 ) , and by volkswagenstiftung under contract no . 86260 . we compute the scattering amplitude in a finite volume by replacing the loop function @xmath176 by its finite volume counterpart @xmath192 and obtain synthetic data from the poles of the finite volume scattering amplitude . in particular , the pole below threshold gives the mass of the bound state in a finite volume . for the case of a level below threshold , there exists a fairly simple way to calculate @xmath97 defined by eq . , so that the equation for @xmath127 can be easily solved . here , we consider three different cases , one with periodic boundary conditions , and two with twisted boundary conditions . depending on which quarks are twisted , the momenta of the mesons are modified accordingly . in the case of periodic boundary conditions , the meson momenta in a box are given by @xmath193 we can evaluate the sum in eq . , using the poisson summation formula @xmath194 . transforming the sum into the integral gives @xmath195 next , we note that the integrand @xmath91 can be approximated by @xmath196 , since the difference is exponentially suppressed @xcite . here , @xmath113 is the three - momentum squared of the particles in the center of mass ( c.m . ) then , for @xmath197 , @xmath97 reads @xmath198 the function @xmath199 can be expressed in terms of the lscher zeta - function @xmath200 , as follows @xcite : @xmath201 z_{00}(1;\hat k^2)&=&\frac{1}{\sqrt{4\pi}}\,\sum_{{\vec n}\in \mathbb{z}^3 } \frac{1}{\vec n^2-\hat k^2}\ , , \end{aligned}\ ] ] where @xmath100 . in the case of twisted boundary conditions , when the momenta of both particles are shifted but the particles still are in the c.m . frame , the allowed momenta in a box are : @xmath202 where @xmath95 is the twisting angle . now , acting in the same way , we can evaluate the sum in eq . @xmath203 and @xmath97 becomes @xmath204 again , we can express @xmath97 in terms of the lscher zeta - function with twisted boundary conditions , @xmath103 , as follows , @xmath205 z_{00}^{\vec\theta}(1;\hat k^2)&=&\frac{1}{\sqrt{4\pi}}\,\sum_{{\vec n}\in \mathbb{z}^3 } \frac{1}{\bigl(\vec n + \vec\theta/2\pi\bigr)^2-\hat k^2}\ , .\end{aligned}\ ] ] for the particular case of @xmath144 , the first few terms of the above expansion are given by @xmath206\end{aligned}\ ] ] with @xmath207 . finally , in the case of twisted boundary conditions , when only the momentum of one of the particles ( say , particle 1 ) is shifted , the allowed momenta in a box are @xmath208 the particles are not in the c.m . frame any more : the c.m . momentum is equal to @xmath209 . hence , we have to evaluate @xmath97 in a moving frame with momentum @xmath210 , @xmath211 \omega_1 ^ 2=&(\vec p-\vec q)^2+m_1 ^ 2,\qquad \omega_2 ^ 2=\vec q^2+m_2 ^ 2,\quad\quad \vec q_{n}=\frac{2\pi\vec n}{l},\qquad p^2=p_0 ^ 2-\vec p^2=s.\end{aligned}\ ] ] again , we can approximate the integrand by @xcite @xmath212 where @xmath213 , @xmath214 is the momentum of the particles in the c.m . frame , and the dots denote exponentially suppressed terms . using the poisson summation formula , we arrive at @xmath215 where @xmath216 and @xmath217 are the components parallel and perpendicular to @xmath210 of @xmath218 , and @xmath219 is the relativistic gamma - factor . once again , we can relate @xmath97 in this case with the lscher zeta function in the moving frame @xmath220 @xcite , see also refs . @xcite : @xmath221 p_d&=&\{{\vec r}=\mathbb{r}^3~|~r_\parallel=\gamma^{-1}(n_\parallel-\mu|{\vec d}| ) , ~{\vec r}_\perp={\vec n}_\perp,~{\vec n}\in\mathbb{z}^3\}\ , , \end{aligned}\ ] ] where @xmath222 . for the case of @xmath144 , the first few terms in the above expansion are @xmath223.\end{aligned}\ ] ] in the case of shallow bound states , the exponential factor @xmath101 will be usually quite small , so in order to reproduce accurately the full function , one should take several terms in the expansion for @xmath97 above . r. l. jaffe , phys . d * 15 * ( 1977 ) 267 . d. black _ et al _ , phys . d * 59 * ( 1999 ) 074026 [ arxiv : hep - ph/9808415 ] . n. n. achasov and a. v. kiselev , phys . d * 68 * ( 2003 ) 014006 [ arxiv : hep - ph/0212153 ] . j. d. weinstein and n. isgur , phys . * 48 * ( 1982 ) 659 . j. a. oller , e. oset and j. r. pelaez , phys . * 80 * ( 1998 ) 3452 [ hep - ph/9803242 ] . j. a. oller , e. oset and j. r. pelaez , phys . d * 59 * ( 1999 ) 074001 [ erratum - ibid . d * 60 * ( 1999 ) 099906 ] [ erratum - ibid . d * 75 * ( 2007 ) 099903 ] [ hep - ph/9804209 ] . j. a. oller and e. oset , phys . d * 60 * ( 1999 ) 074023 [ arxiv : hep - ph/9809337 ] ; t. hyodo , int . j. mod . a * 28 * ( 2013 ) 1330045 [ arxiv:1310.1176 [ hep - ph ] ] . j. nebreda , j. r. pelaez and g. rios , phys . d * 84 * ( 2011 ) 074003 [ arxiv:1107.4200 [ hep - ph ] ] . j. r. pelaez and g. rios , phys . * 97 * ( 2006 ) 242002 [ hep - ph/0610397 ] . c. hanhart , j. r. pelaez and g. rios , phys . * 100 * ( 2008 ) 152001 [ arxiv:0801.2871 [ hep - ph ] ] . j. r. pelaez and g. rios , phys . d * 82 * ( 2010 ) 114002 [ arxiv:1010.6008 [ hep - ph ] ] . v. bernard , m. lage , u .- g . meiner and a. rusetsky , jhep * 1101 * ( 2011 ) 019 [ arxiv:1010.6018 [ hep - lat ] ] . m. albaladejo and j. a. oller , phys . d * 86 * ( 2012 ) 034003 [ arxiv:1205.6606 [ hep - ph ] ] . a. martinez torres , l. r. dai , c. koren , d. jido and e. oset , phys . d * 85 * ( 2012 ) 014027 [ arxiv:1109.0396 [ hep - lat ] ] . t. sekihara and t. hyodo , phys . c * 87 * ( 2013 ) 045202 [ arxiv:1209.0577 [ nucl - th ] ] . s. ozaki and s. sasaki , phys . d * 87 * ( 2013 ) 014506 [ arxiv:1211.5512 [ hep - lat ] ] . m. albaladejo , c. hidalgo - duque , j. nieves and e. oset , phys . rev . d * 88 * ( 2013 ) 014510 [ arxiv:1304.1439 [ hep - lat ] ] . b. aubert _ et al . _ [ babar collaboration ] , phys . * 90 * ( 2003 ) 242001 [ hep - ex/0304021 ] ; d. besson _ et al . _ [ cleo collaboration ] , phys . d * 68 * ( 2003 ) 032002 [ erratum - ibid . d * 75 * ( 2007 ) 119908 ] [ hep - ex/0305100 ] . l. liu , k. orginos , f .- k . guo , c. hanhart and u .- g . meiner , phys . d * 87 * ( 2013 ) 014508 [ arxiv:1208.4535 [ hep - lat ] ] . g. moir , m. peardon , s. m. ryan , c. e. thomas and l. liu , arxiv:1312.1361 [ hep - lat ] . d. mohler , c. b. lang , l. leskovec , s. prelovsek and r. m. woloshyn , phys . * 111 * ( 2013 ) 22 , 222001 [ arxiv:1308.3175 [ hep - lat ] ] ; c. b. lang , l. leskovec , d. mohler , s. prelovsek and r. m. woloshyn , phys . rev . d * 90 * ( 2014 ) 034510 [ arxiv:1403.8103 [ hep - lat ] ] . m. lscher , commun . math . * 104 * ( 1986 ) 177 . s. r. beane , p. f. bedaque , a. parreno and m. j. savage , phys . b * 585 * ( 2004 ) 106 [ hep - lat/0312004 ] . y. koma and m. koma , nucl . b * 713 * ( 2005 ) 575 [ hep - lat/0406034 ] . s. sasaki and t. yamazaki , phys . d * 74 * ( 2006 ) 114507 [ hep - lat/0610081 ] . z. davoudi and m. j. savage , phys . d * 84 * ( 2011 ) 114502 [ arxiv:1108.5371 [ hep - lat ] ] . m. lscher , nucl . b * 354 * ( 1991 ) 531 . m. dring , u .- g . meiner , e. oset and a. rusetsky , eur . j. a * 47 * ( 2011 ) 139 [ arxiv:1107.3988 [ hep - lat ] ] . m. albaladejo , g. rios , j. a. oller and l. roca , arxiv:1307.5169 [ hep - lat ] . e. e. kolomeitsev and m. f. m. lutz , phys . b * 582 * ( 2004 ) 39 [ hep - ph/0307133 ] . g. burdman and j. f. donoghue , phys . b * 280 * ( 1992 ) 287 . m. b. wise , phys . d * 45 * ( 1992 ) 2188 . t. -m . yan , h. -y . cheng , c. -y . cheung , g. -l . lin , y. c. lin and h. -l . yu , phys . d * 46 * ( 1992 ) 1148 [ erratum - ibid . d * 55 * ( 1997 ) 5851 ] . j. a. oller and e. oset , nucl . a * 620 * ( 1997 ) 438 [ erratum - ibid . a * 652 * ( 1999 ) 407 ] [ hep - ph/9702314 ] . k. rummukainen and s. a. gottlieb , nucl . b * 450 * ( 1995 ) 397 [ hep - lat/9503028 ] . m. gckeler , r. horsley , m. lage , u .- g . meiner , p. e. l. rakow , a. rusetsky , g. schierholz and j. m. zanotti , phys . d * 86 * ( 2012 ) 094513 [ arxiv:1206.4141 [ hep - lat ] ] .

nuclear resonances and excited states can be very complicated many - body structures with a number of different decay modes . the simplest decay , perhaps beside @xmath2-emission , is breakup into two particles as exemplified by nucleon- and @xmath3-emission , and binary fission @xcite . the deceivingly simple breakup into three particles is much less studied and far from understood . this is in contrast to bound three - body cluster structures where a variety of techniques are available and able to predict the properties , even of the exotic quantum halo states @xcite . the three - body continuum properties are less established although rather well studied over many years @xcite . experimental information is obtained by measuring the properties of the particles in the final state . the experimental techniques are now advanced to a level where accurate and kinematically complete measurements are available on a number of different systems @xcite , and many more are expected to follow . reliable theoretical descriptions are needed to interpret existing data , to predict unknown decay results and to help in the design of new interesting experiments . both the structure of the initial state and the decay mechanism are essential and both must therefore be properly described simultaneously . clearly for a genuine many - body state only the intermediate and large - distance structure is decisive . the small distance behavior is artificial and only serving to provide the proper continuous boundary conditions . for two - body decay , like @xmath3-emission , the relative potential determines all properties . in the example of @xmath3-emission , the two - body potential can be divided into short- , intermediate- and long - distance . the short - distance part is artificially adjusted to allow the correct resonance energy and the barrier region then determines the width . at large distances where the potential has vanished the energy of the @xmath3-particle is determined by energy conservation . in two - body decay the energy distribution then only reflects the width of the initial resonance . these properties are very different for decays with more than two particles in the final state . for three - body bound states and resonances `` large distance '' is less well defined . however , the corresponding structure can efficiently be computed by use of the hyperspherical adiabatic expansion method @xcite . the wave functions are in this technique expanded on basis states related to adiabatic potentials calculated as functions of the hyperradius @xmath4 . then @xmath4 provides a measure of distances for the three - body problem . the wave function is usually dominated by the component related to the lowest adiabatic potential . the small - distance part ( @xmath4 small ) of both wave function and potential are only directly meaningful if the particles appearing in the final state form a genuine three - body system . otherwise this part of the effective potential is constructed to produce the correct resonance energy and provide an appropriate boundary condition for the wave function . at intermediate distances the potential has a barrier which is decisive for the width of the resonance . at large distances the resonance wave function is characterized by outgoing waves which contain information about distributions of relative energies and possibly other quantities like spin distributions . for three - body decay the two most obvious decay mechanisms , direct and sequential , were recently studied in a schematic model in @xcite . another schematic model is also formulated in the limit where only the coulomb interaction is important at intermediate and large distances @xcite . the detailed resonance structure at small and intermediate distances were investigated in realistic models in @xcite . the large - distance properties are much more difficult to calculate accurately , because either the correct continuum three - body coulomb wave functions are unknown , or the short - range potentials may produce an almost long - range ( inverse square ) effective potential at large distance . in the latter case the origin is precisely as for the efimov effect @xcite . the corresponding potential is most likely the lowest at large distance but not necessarily also at small distances . thus , the resonance wave function may change structure from small to large distance . the relative energy distribution arises as the result appearing at large distance . the numerical computations are then sometimes rather tricky . the purpose of the present paper is to establish a general method to compute relative energy distributions after decay into three particles . the short- and intermediate - distance resonance structure from @xcite is a good starting point but we need in addition to calculate the asymptotic behavior in momentum space . the asymptotics vary for the different decay mechanisms , and the related numerical treatment is difficult when all the possibilities simultaneously have to be accounted for . we assume that formation and decay of the resonances are independent processes . the resonance could be formed by beta - decay from a neighbouring nucleus , or a window with the relevant energies can be selected in an experimental setup where contributions from other processes also are eliminated . in section 2 we develop the theoretical formalism for resonance decay . this was previously sketched by use of the saddle point approximation @xcite , while we here shall instead use the zeldovich regularization of the divergent fourier integrals @xcite . in section 3 we discuss some of the important features arising from calculation of resonance wave functions by use of the hyperspherical adiabatic expansion method combined with the complex scaling method . in section 4 we illustrate in details with realistic computations for the @xmath0-resonance in @xmath1he . finally , section 5 contains a brief summary and the conclusions . we assume that the system of particles has been generated in a meta - stable quantum state ( a resonance ) that is a generalized eigen - state of the hamiltonian with complex energy . this is a decaying state it describes a constant flux of particles towards infinity . suppose we have a system of detectors at large distances which measure the momenta of the particles emerging from this decaying state . clearly these detectors will measure the probability distribution of particle momenta in the meta - stable state , that is the absolute square of the momentum space wave - function . the theoretical derivation is most easily understood if we first explain the idea for simple resonance decay into a two - body system . we need resonance inventions in coordinate and momentum - space and transformations between these non - square integrable functions . the momentum space wave - function of a resonant state with the complex energy @xmath5 has the form @xcite @xmath6 where @xmath7 is the relative momentum and @xmath8 indicates the two angles defining the direction of the vector @xmath7 . we assume that the wave function @xmath9 only has the pole at @xmath10 and @xmath11 is then a continuous function of the momentum @xmath7 with no poles . the distribution @xmath12 of the relative momentum @xmath7 of the two particles in the resonant state @xmath13 is given by the absolute square of the momentum - space wave function , i.e. @xmath14 where the real observable energy @xmath15 is @xmath16 . the approximation that the system is generated in a pure resonant state @xmath13 is most likely only valid in the neighborhood of the resonant energy , i.e. @xmath17 . furthermore , the function @xmath18 is smooth and varies by definition much less than the denominator in eq.([psi_k ] ) . in any case we shall only consider energies where @xmath19 is less than a few times @xmath20 . we can then confidently substitute the momentum @xmath21 with the resonant momentum @xmath22 in @xmath18 and thus arrive at the expression of the famous breit - wigner type @xmath23 where the energy dependence is given by the factor @xmath24 ^{-1}$ ] while the angular dependence is given by the ( absolute square of the ) function @xmath25 . thus , the momentum - space wave - function eq.([psi_k ] ) of the resonance allows direct calculation of the momentum distributions of the decay fragments through eq.([p_k ] ) . clearly improvements are possible by use of different approximations of @xmath18 . instead of momentum - space , the wave - function of the resonance may be available in coordinate - space where the large - distance asymptotic form is given by @xmath26 where @xmath27 denotes the two angles defining the direction of the relative coordinate @xmath28 . the structure is an outgoing spherical wave potentially modified by an angular dependence contained in @xmath29 . generally the resonance wave function can be written as a partial - wave expansion in the spherical harmonics @xmath30 , i.e. @xmath31 where @xmath32 are constants depending on angular momentum and projection quantum numbers @xmath33 and @xmath34 . the radial functions @xmath35 are those solutions of the radial schrdinger equation that asymptotically approach the outgoing spherical wave in eq.([e60 ] ) , i.e. @xmath36 this defines the asymptotic behavior of the decaying resonance wave function which in turn determines the energy distribution in the observable final state . the coordinate- , @xmath37 , and momentum - space , @xmath38 , wave - functions are connected via a fourier transform @xmath39 expansion of the plane - wave in terms of spherical harmonics @xmath40 reduces the fourier integral in eq.([e70 ] ) to a one - dimensional radial integral @xmath41 because of the asymptotics in eq.([chi- ] ) the radial integral is seen to diverge . the large - distance behavior is responsible for the divergence . the physics content , expressed by a finite value , then has to be extracted by a suitable regularization . we use the prescription proposed by zeldovich @xcite , i.e. multiplication of the integrand by a gaussian where the range after integration is increased to infinity . for an exponential this gives @xmath42 where @xmath43 can be complex and @xmath44 is the error function of complex argument . in the present context the radial integral in eq.([e75 ] ) is first rewritten by adding and subtracting the asymptotic expression of the diverging integrand . the difference between the true and the asymptotic expression then remains finite even without multiplication by the gaussian function . only the asymptotic expression then diverges when the gaussian smoothly converges to an overall factor of one . the physics content is extracted by dividing with a similarly diverging normalization integral of the square of the wave function @xmath45 . also this integral , now in the denominator , is rewritten by addition and subtraction of its asymptotic expression . again only the asymptotic expression diverges . the zeldovich prescription now leaves the ratio of these two diverging integrals of the asymptotic expressions . however , this ratio does not diverge but converge towards the physically meaningful result . apart from a normalization we therefore have to regularize only the asymptotic expression obtained by use of eq.([chi- ] ) and the asymptotic approximation of @xmath46 , i.e. @xmath47 the radial integral is then by use of eq.([e77 ] ) evaluated to be @xmath48 \nonumber\\ = \frac{i^{l}}{2}\frac{k+k_{0}-(-1)^{l}(k - k_{0})}{k^{2}-k_{0}^{2 } } = i^{l } \frac{k_{0}}{k^{2}-k_{0}^{2 } } \;\ ; \;\ ; { \rm or } \;\;\;\ ; i^{l } \frac{k}{k^{2}-k_{0}^{2 } } \ ; \end{aligned}\ ] ] for even or odd @xmath33 , respectively . the summation in eq.([e85 ] ) is proportional to the angular amplitude @xmath49 from eq.([chi- ] ) , but now as a function of the momentum @xmath7 . in any case we assumed earlier that @xmath50 in the smooth functions . we therefore arrive at the final expression for the fourier - transform of the resonance wave - function , i.e. @xmath51 thus the function @xmath52 from eq.([psi_k ] ) is then related to @xmath49 from eq.([e60 ] ) by @xmath53 the convenient fact that only the asymptotic limit of the resonance wave function enters after the regularization procedure is perhaps more surprising in mathematics than in physics where the observable energy distributions always are obtained from the properties at large distances . the observable distribution in momentum - space is determined by the angular wave function in coordinate - space evaluated for angles describing the direction of the momentum . this peculiar fact can intuitively be understood by the geometry of particles moving towards the detectors at infinitely large distances . coordinates and momenta then must point in the same direction . a mathematical formulation is available from ionization cross sections calculated for atomic physics processes @xcite . the generalization to three particles first requires a convenient set of coordinates . we choose the scaled jacobi coordinates @xcite @xmath54 where @xmath34 is a mass scale , and @xmath55 and @xmath56 are the coordinate and mass of the particle number @xmath57 . the hyper - spherical coordinates are then the hyper - radius @xmath58 , the hyper - angle @xmath59 , and the directional angles @xmath60 and @xmath61 of the vectors @xmath62 and @xmath63 * * d**efined by @xmath64 the jacobi coordinates depend on the sequence chosen for the particles , and the three different pairs of @xmath62 and @xmath63 could be labeled to distinguish . we omit these labels when the meaning is clear . the corresponding momentum - space variables are @xmath65 where @xmath66 and @xmath67 are the conjugate momenta related to the coordinates @xmath62 and @xmath63 . the generalization of the two - body spherical harmonics are the so - called hyper - spherical harmonics @xcite @xmath68 where @xmath69 , @xmath70 , and @xmath71 are the angular quantum numbers related to coordinates @xmath72 and @xmath73 , and @xmath74 is a normalization factor . these functions are the eigen - functions of the angular part @xmath75 of the three - body kinetic energy operator @xmath76 @xmath77\ ] ] with the eigenvalues @xmath78 , i.e. @xmath79 where @xmath80 is a non - negative integer . without coulomb and without bound two - body subsystems the three - body resonance wave - function @xmath81 with the complex energy @xmath82 can be expanded in terms of the hyper - spherical harmonics @xmath83 where the hyper - radial functions @xmath84 have the usual resonance asymptotic behavior of an out - going hyper - spherical wave @xmath85 the three - body wave - function asymptotically has the form of the hyper - spherical wave with an angular amplitude @xmath86 determined by the expansion coefficients @xmath87 , i.e. @xmath88 the momentum - space wave - function is the fourier transform @xmath89 the three - body plane - wave can be expanded in hyper - spherical harmonics as @xmath90 due to orthogonality of the hyper - spherical harmonics the angular part of the integral in eq.([e90 ] ) is trivial and we are only left with the hyper - radial integral @xmath91 precisely as in the two - body case , in the vicinity of the resonance the integrand can be made ready for regularization by substitution of eq.([e115 ] ) and the asymptotic form @xmath92 this results in a diverging integral similar to that of the two - body case @xmath93 using the zeldovich regularization leads to @xmath94 that is , in the vicinity of the resonance , the angular wave function in momentum - space is proportional to that in coordinate - space but evaluated for the momentum variables . the energy distribution is determined by the breit - wigner factor where the width is obtained from the three - body resonance . the function @xmath95 now contains information about the non - trivial energy distribution between the three particles . this is in contrast to the two - body decay where all the energy is in the only existing relative degree of freedom . sometimes a bound two - body subsystem can be emitted from a three - body resonance . such a final state configuration can not be described by hyper - spherical harmonics . however if this is the only open channel , the description of such a decay reduces to the two - body case . indeed the asymptotics of the resonance wave - function is then @xmath96 where @xmath97 describes a bound system of particles 2 and 3 with binding energy @xmath98 , @xmath99 , @xmath100 is the angular amplitude and @xmath101 is the complex three - body energy . if both three - body and two - body decays are possible the wave - function contains asymptotics of both two- and three - body types , @xmath102 where @xmath103 and @xmath104 are the asymptotic coefficients determining the relative weights of the two decay channels . both @xmath105 and @xmath49 are dimensionless and the dimension ( length to @xmath106 ) of @xmath107 compensate for the one length in the denominator of the last term . the fourier transform and the corresponding regularization then give the momentum - space wave function , i.e. @xmath108 where @xmath109 , @xmath110 is the momentum - space wave function of the two - body bound state @xmath107 . these two channels correspond , respectively , to two close - lying particles in a bound state far away from the third one , and three particles all far away from each other . thus , in the of large distances they do not interfere , and the resulting momentum distribution is simply a weighted sum of the corresponding distributions . the relative contributions of the two channels are given by @xmath111 and @xmath112 correspondingly . generalization to describe decays into more than one two - body bound state is formally straightforward , i.e. the corresponding non - interfering asymptotic terms should simply be added . this holds for more than one bound state in the same two - body system as well as for bound states in different two - body systems . instead of a bound state the decay via a two - body resonance is often considered in interpretation and analysis of experiments . clearly the narrower the resonance the more similarity to the case of two - body bound states . in any case the hyper - spherical expansion must eventually converge , although the convergence can be too slow for a reliable extraction of the asymptotic coefficients in eq.([asy ] ) from a numerical solution of the three - body problem . however , in this case the ( slowly convergent ) two - body resonance configuration can then be explicitly included into the asymptotics while only the remaining ( hopefully fast convergent ) part is expanded , i.e. @xmath113 where @xmath114 is the complex three - body energy , @xmath115 is the ( complex ) energy of the two - body resonance and the remaining part is described by the complex momentum @xmath116 . the precise definition of @xmath117 arises from a constraint to be seen below . the corresponding momentum - space wave - function is again given by the regularized fourier transform , i.e. @xmath118 the momentum distribution is given by the absolute square of this momentum - space wave - function . in the center of mass system we can directly find the distribution of particle 3 arising from the sequential emission via the two - body resonance . absolute square of the last term in eq.([e95 ] ) and use of the energy conservation @xmath119 ( or @xmath120 ) immediately gives the energy distribution for particle @xmath121 @xmath122 } \frac{1}{[(e_{23}+e_1 - e_{0})^2 + \gamma_{0}^2/4 ] } \nonumber \\ \propto \frac{1}{(e_1 - ( e_{0}-e_{23}^{(0)})^2 + ( \gamma_{0}+\gamma_{23})^2/4 } \ ; , \label{e100}\end{aligned}\ ] ] which states that the most probable energy of particle @xmath121 is @xmath123 and the width of the distribution is the sum of the two and the three - body widths . precisely this breit - wigner distribution only arises when the @xmath124 in eq.([e95 ] ) is proportional to @xmath125 , i.e. given by the probability distribution in the initial three - body state , which also is of breit - wigner form . thus the definition of @xmath117 must involve @xmath126 and not @xmath127 . the same integration could of course be performed on the energy of particle 1 , but this would only give the two - body breit - wigner distribution for @xmath128 whereas the measurements provide individual energies , @xmath129 and @xmath130 , for particles 2 and 3 in the center of mass system . to get the distributions of @xmath129 and @xmath130 involve trivial but tedious kinematical transformations where also energies and directions of particle 1 are required . the relative energy distributions can be obtained from the angular resonance wave function calculated without complex scaling . we assume that the system of particles is produced in an initial state for example by a beta - decay process . we can then imagine the subsequent decay as the time evolution of the initial non - stationary state . this can be formulated as a time dependent coupled channels problem . it can also be viewed intuitively as a particle described by time dependent coordinates determined by classical equations of motion . this should be done with the appropriate initial amplitudes for all parts of the initial wave function . the hyperradius must vary from being very small to infinitely large . we increase @xmath4 until all particles are outside the interaction ranges of all other particles . from then on all distances scale as @xmath4 and all other coordinates remain unchanged with time until @xmath131 . energy conservation is maintained at large distances by converting the potential energy into kinetic energy in the scaling degree of freedom , i.e. by increasing the velocity @xmath132 of the @xmath4-coordinate . the wave function then evolves with all angular degrees of freedom frozen eventually reaching the detectors placed far away . the absolute square of the angular wave function as function of @xmath133 then provide the energy distributions simply because the kinetic energy of particle 1 is given by the velocity of @xmath134 , i.e. @xmath135 . then the energy distribution as function of the kinetic energy , proportional to @xmath136 , is the probability coordinate - space distribution as function of @xmath133 , apart from the phase space conversion from @xmath3 to energy , i.e. division by a factor proportional to @xmath137 . this procedure is tempting since we only need to increase the maximum value of @xmath4 in all the numerical implementations and plot the wave function at that large distance . for this to be accurate the asymptotics has to be well described by the hyperspherical expansion or the basis has to be very large and able to describe the necessary large @xmath4-behavior however , when the basis functions have asymptotics different from one of the intermediate structures the size of the basis needed to reach convergenge can easily become huge , making the procedure impractical . this is not necessarily easy to see in the numerical results where an increase of basis size usually is rather expensive while the convergence could be extremely slow . the procedure is probably only directly useful for direct decay or for sequential decay via broad resonances . otherwise the different intermediate structure should be computed somehow and extrapolations to large distances applied to each component individually . an example to illustrate the present alternative formulation is available in the schematic model discussed in details in @xcite where the widths but not the energy distribution were computed . assume that only the coulomb potential is important and the most probable path ( ridge in the wave function ) from small to large distances can be described by scaling the hyperradius . the corresponding optimum path is defined by minimizing the wkb - tunneling expression as function of different relative scaling parameters @xmath138 , where @xmath139 is the distance between particles @xmath57 and @xmath21 . the path is given by @xmath140 ( see also @xcite ) , where @xmath141 is the charge of particle @xmath57 . we then arrive at the most probable value for the energy division , i.e. @xmath142 where @xmath143 is the total energy distributed among all the three particles . this expression is simple but not very accurate . it also only provides an estimate of the peak value . to compute the distribution other paths must also be considered . this is possible but we shall leave this for a later discussion in connection with a detailed treatment of the coulomb interaction . the resonance wave function contains all information including that of the relative energy distribution after the decay . the calculations must then first provide the corresponding three - body resonance states . second the large - distance behavior must be accurately extracted . due to the different structures , these steps are not trivially completed by use of only one method . we briefly describe first the main ingredients in our computations and the features of the wave function . second we explain how the asymptotic behavior is obtained in practice . we use the hyperspherical adiabatic expansion method combined with complex scaling to obtain resonance wave functions . the coordinates are defined in section 2 along with our basis functions in angular space , i.e. the hyperspherical harmonics in each of the three jacobi systems . we solve the complex scaled faddeev equations as function of hyperradius @xcite . the complex scaled coupled set of radial equations are subsequently solved with the appropriate boundary conditions , i.e. exponentially vanishing with increasing @xmath4 for both bound states and resonances . thus here we assume that we do not need to treat the coulomb interaction explicitly at asymptotic large distances . as shown in @xcite , the results obtained with this method agree well with some of the most common procedures , as for instance the complex energy method . the energies are usually accurately determined in this method . the same applies to the wave functions at small and intermediate distances where the exponential fall - off still is not too restrictive . however , we need the information at distances where the asymptotic limit is reached , i.e. possibly at very large @xmath4 where the complex scaled wave functions are very small . furthermore , more than one geometric structure can be important at the same time , e.g. two different spatially confined two - body configurations with the ( different ) third particle far away . this happens frequently with two identical particles like neutrons and protons as constituent particles , because the nucleon - core interaction must be sufficiently attractive to produce a bound or resonating three - body system , and this implies that such two - body configurations are favored . to account simultaneously for different two - body substructures , it is essential to use three components as in the faddeev decomposition adopted by us . the same efficiency can be achieved in a variational approach by allowing faddeev - like components in the trial wave function @xcite . it is much more difficult , if not impossible , to reach convergence with only one component as in the hyperharmonic expansion method @xcite . even with three faddeev components a large basis has to be employed . to describe substructures inside one of the two - body potentials ( range @xmath144 ) for large @xmath4 we need values of the hyperpsherical quantum number @xmath80 up to a few times @xmath145 where @xmath146 is the reduced mass of the two particles . this is because @xmath147 is the number of nodes in the basis , and details can only be described if a few nodes can be placed inside the structure in question . thus , for nuclear systems , where @xmath148 fm , we need @xmath149 of at least @xmath150 to describe such structures for @xmath151 fm . employing complex scaling transforms resonances into states obeying the numerically easier bound state boundary conditions . the required basis size is larger , essentially because the exponential fall - off moves to larger distances with increasing scaling angle . these estimates provide necessary conditions for a reasonable description of two - body substructures which in turn are necessary to describe the sequential decay mechanism . the requirement of large @xmath80 to describe substructures must be reconciled with the fact that an increase of @xmath4 towards infinity results in convergence of the angular eigenvalues to the hyperharmonic spectrum for free particles . we show an example in fig . [ fig1 ] where the imaginary parts are omitted as they both oscillate around zero and approach zero at large @xmath4 . the real parts of the angular eigenvalues approach @xmath78 as @xmath4 increases while the corresponding potentials all approach zero faster than @xmath152 . the attractive pockets in the eigenvalues at short distance disappear in the potentials except for the two lowest where the negative values remain as a prominent feature . the approach to the asymptotic values is very fast except for the levels where @xmath153-waves contribute significantly . the low energies favor these levels at large distances . the related adiabatic wave functions approach the hyperspherical harmonics . the reason is that the regions in space , where the short - range interactions are significant , are shrinking in size with increasing @xmath4 relative to the total space available . the radial extension of these regions , responsible for two - body correlations , decrease with @xmath4 as @xmath154 . the interactions are non - vanishing in smaller and smaller regions . consequently they become less and less important for both energies and wave functions . thus , the basis size has to increase with @xmath4 in order to allow a description of the two - body substructures or equivalently of sequential decay , but the lowest adiabatic potentials approach the free solutions . the basis size in practice always has to remain finite and the substructures eventually become impossible to describe in this way . the interactions have to be chosen to reproduce the pairwise low - energy scattering properties of the three particles appearing in the final state . clearly we must accurately include all the partial waves necessary to describe the quantum numbers of the decaying resonance . however , even with a sufficiently large basis the three - body system does not necessarily appear with the correct energy and width . in fact , there may not even be an attractive region at small distances as required to produce a resonance of finite width . this could occur when we are dealing with a many - body resonance without traces of any three - body cluster structure . nevertheless , a meaningful computation can be carried out of the energy distributions emerging after a three - body decay . the philosophy is the same as for @xmath3-emission where the inner part of the effective potential is replaced by an attractive square well with a depth adjusted to reproduce the resonance energy . the resulting barrier is then used to derive the width , usually in the wkb approximation . we generalize this concept to the adiabatic potentials , i.e. we add a three - body potential of short range in the hyperradius . it is intended to describe interactions beyond the two - body level such that the three - body system has a resonance at the desired energy . by doing this we have substituted the possibly complicated many - body structure at small distance with the three - body cluster structure resulting from an effective potential , which in turn also provides the correct boundary conditions for a three - body decaying resonance . this principle was introduced for fine - tuning in the first calculation with the correct boundary conditions of the three-@xmath3 decay of the second @xmath155-state in @xmath156c @xcite . it has later become the standard procedure to adjust three - body energies without significant changes of the underlying substructure @xcite . the radial solution is often strongly dominated by one or two of the lowest adiabatic components at small distance where the relative probability is large . this is because all three short - range two - body interactions contribute simultaneously and the result is the energetically most favored three - body resonance structure consistent with the boundary conditions . as @xmath4 increases , at least one particle has to move away from the other leaving at most one non - vanishing two - body interaction . coherent contributions from different of these configurations are possible and sometimes even favored . at large distance , where the energy distribution is determined , several more adiabatic potentials are often needed . the couplings due to the coulomb interaction would normally increase the necessary number of potentials . it is established @xcite that the adiabatic potentials of lowest energy at large @xmath4 are related to configurations with relative @xmath153-waves between the two closest particles . this is the basis for the efimov effect @xcite . if these large - distance configurations differ from the resonance structure at small @xmath4 , possibly with higher partial waves , the lowest angular wave function must change its structure accordingly as @xmath4 increases . in @xcite we showed the structure for @xmath1he(2@xmath157 ) for the lowest eigenvalue which approach the @xmath158 value at large @xmath4 . in fig . [ fig2 ] we show the results for the similar eigenvalue approaching the @xmath159 level for large @xmath4 . the pronounced and rapidly changing structure is qualitatively similar to the lower - lying @xmath158 level , i.e. dominated by @xmath160 neutron - core structure at small @xmath4 and by @xmath153-waves between the two neutrons at large @xmath4 . essentially all other allowed components contribute with equally small amounts . these rather dramatic changes imply that it is crucial to include all adiabatic potentials with significant couplings to those dominating the structure at small hyperradii . this is simply because the couplings are responsible for changing the radial weights of the different adiabatic components as function of @xmath4 , e.g. no couplings imply the same occupation independent of @xmath4 . on the other hand , each of the angular wave functions related to the adiabatic potentials are themselves functions of @xmath4 , sometimes rapidly varying as seen in fig . [ fig2 ] . in principle the non - diagonal couplings could be vanishingly small and all change of structure would be described by the lowest adiabatic wave function . however , this is rather unlikely because the couplings are defined as matrix elements of first and second radial derivatives of the angular wave functions . thus , radial couplings between rapidly changing angular structures are inevitable . in fig . [ fig3 ] we show how the strongly varying coupling potentials can be related to the changing angular structure seen in fig . [ fig2 ] . the peaks are most pronounced when a crossing is avoided and the two levels switch characteristics @xcite . the rather confusing coupling picture is crucial for the asymptotic behavior of the wave function at large distance where the energy distribution is determined . the second order terms are substantially larger than the first order couplings but all vanish at large @xmath4 . thus the numerical computations in the present case must extend at least beyond @xmath161 fm where the couplings have reached very small values . this is about two times the largest scattering length which defines the distance of convergence towards asymptotic values . we have now established two important but competing effects , which determine the three - body resonance structure from small to large values of @xmath4 . in the extreme , the structure can either remain unchanged by climbing correspondingly up on the adiabatic potentials , or the structure can change to follow that of the lowest - lying adiabatic wave function . a compromise between following the energetically most favored configuration and the resistance to a change of structure therefore must be reached . the combination of these effects determine the relative population of the different components in the radial solution , which in turn determines the observable energy distribution . the couplings are more important here than for widths , energies and small - distance wave functions . they must be accurately computed to provide the energy distribution . the structure of the resonance wave function at large @xmath4 could remain unchanged and only exhibit a simple scaling behavior proportional to @xmath4 . this is typical of direct decay . the wave function could also have large probability for finding two close - lying particles where the hyperradius mainly changes by moving the third particle as @xmath4 increases . this is typical of sequential decay via more or less stable two - body configuration , e.g. sequential decay through two - body resonances . the intermediate configuration does not necessarily need a confining barrier , but could be provided by low - lying two - body virtual @xmath153-states @xcite . mixtures of all types can occur giving rise to the description of decay properties as fractions proceeding via individual two - body configurations . all these structures can be computed by use of our method , although convergence for the coulomb interaction is more difficult . the best strategy to get reliable results is not obvious , because the brute force method of increasing basis size and hyperradius until convergence is reached may be beyond any reasonable computer effort . the indecision is related to the requirement of an increasing basis with increasing @xmath4 , which means that a smaller @xmath4 and a smaller basis could provide a better description with much less effort . in other words a convergence may be reached in a region of @xmath4-values for a moderate basis size . this convergence would be destroyed as @xmath4 is allowed to increase because the basis size can not follow . the convergence can possibly be reached faster by extrapolation of the observable distribution by use of a known or anticipated dependence of @xmath4 and basis size @xcite . different parts of the wave function may extrapolate differently . the most efficient choice depends on the ( mixtures of ) decay mechanisms which therefore has to be determined first . therefore the first step is to compute the structure of the resonance wave functions as discussed in @xcite . nuclear three - body decay without complications of the coulomb interaction must involve emission of two neutrons . the decaying states do not have to be three - body structures although such two - neutron halos are available and rather well studied . the most obvious case is the established @xmath0 resonance in @xmath1he which is formed by the same neutron - core components as in the ground state . without coulomb interactions the computations should quickly lead to the desired energy distributions . however , even short - range interactions can present difficulties as highlighted by the intricate description needed for the efimov effect @xcite . both the @xmath3-neutron and the neutron - neutron interactions are previously employed in ground state computations @xcite . we follow the procedure outlined in the preceding sections . different prescriptions are possible to implement the pauli principle @xcite , all of them providing indistiguishable angular wave functions at large distances . we adjust the three - body potential to give the correct resonance energy . this only requires marginal fine - tuning . in total 1132 hyper - spherical harmonics are used in the expansion ( [ eq21 ] ) , and the maximum value of @xmath80 is 200 for the most relevant partial wave components and never smaller than 40 . the resonance wave function is then available as function of the hyperspherical coordinates . a complex scaling angle of 0.10 rads is enough to produce an exponentially vanishing with increasing @xmath4 wave function for the resonance . a different scaling angle , where the numerical calculations have converged , produce the same results . we already showed the angular eigenvalues and the adiabatic potentials in fig . the probability distribution arising from only the lowest potential was shown in @xcite . the structure changes from peaks at small @xmath4 corresponding to @xmath3-neutron @xmath162-structure to a probability with one broad peak corresponding to comparable distances between all three particles . this reflects the change of structure of this angular wave function from small to large @xmath4 as seen in details in fig . eventually the lowest hyperharmonic function with @xmath158 is approached . this indicates in itself a direct decay mechanism . however , in this case the lowest potential provides rather misleading results . the rapidly changing structure seen in fig . [ fig2 ] at around @xmath163 fm could easily lead to occupation of higher - lying levels . these occupation probabilities are functions of @xmath4 and simply found as squares of the radial wave function obtained by solving the coupled set of radial equations . the results are shown in fig . [ fig4 ] for the lowest adiabatic components . at small @xmath4 the lowest potential is totally dominating but as @xmath4 increases the lowest three components contribute with comparable amplitudes . all radial wave functions vanish by oscillating around zero with decreasing amplitudes . the relative sizes are more clearly seen in the right hand side of fig . [ fig4 ] . after the transition around 20 fm the individually very small radial amplitudes stabilize on relatively constant finite ratios . the square of these give the relative weights , i.e. reduced compared to the first component by about 0.6 , 0.25 , 0.01 for the second , third and fourth potential , respectively . the transition to stable ratios is consistent with the disappearance of the coupling terms shown in fig . [ fig3 ] . the total probability distribution in fig . [ fig5 ] are quite different from that of the lowest eigenvalue . at large @xmath4 the probability now peaks at a smaller distance between the two neutrons and correspondingly the @xmath3-neutron distance is increased . still fairly broad distributions remain . the decay mechanism indicated by this structure is now instead of direct rather a mixture between the preferred sequential decay via a neutron - neutron intermediate configuration and a smaller direct component . reliable computation of the energy distribution requires numerically converged results in an appropriate region of @xmath4-values . the energy distributions are shown in fig . [ fig6 ] as functions of @xmath4 for a sufficiently large number of adiabatic potentials . the resemblance with the probability distribution is not surprising since only the volume element has been changed . the observable distribution is the cut for constant , and sufficiently large , @xmath4 where convergence has been reached as function of basis size . the neutron energy distribution has two peaks for small @xmath4 corresponding to the geometric configurations of one neutron close to the @xmath3-particle and the other neutron further away . this is reflected in the peak in the @xmath3-spectrum at intermediate energies corresponding to the same geometric configurations . the structure changes with @xmath4 into a broad peak at intermediate energies for the neutron spectrum , and one peak very close to the maximum energy for the @xmath3-spectrum . this is the fingerprint of sequential decay via emission of the @xmath3-particle followed by decay of an intermediate two - neutron structure . this is easily visualized as the two - body decay process where the @xmath3-particle receives maximum energy when the two neutrons move together in the opposite direction . in the subsequent decay each neutron then must share the remaining energy which leads to an intermediate energy between zero and the maximum value . this inferred decay mechanism is perhaps counter - intuitive because stable intermediate configuration of two neutrons do not exist neither as bound states nor as resonances . it would be much more acceptable with the @xmath3-neutron @xmath162-resonance as the intermediate structure . however , one characteristic feature of the neutron - neutron interaction is the low - lying virtual @xmath153-state which simply means that there is a substantial @xmath153-wave attraction . apparently this is decisive for the decay process where the two neutrons end up by moving essentially in the same direction , and then necessarily guided by the attraction . the interesting point is maybe that this is not the way they started out at small distance in the spatially confined part of the wave function . this change of structure with hyperradius is in a sense reflecting the dynamic character of the decay process . at distances larger than the scattering length the short - range interactions are negligibly small , the wave function changes are completed and the asymptotics are established . the microscopic structure of the energy distributions can be studied by dividing into contributions from the individual adiabatic potentials as seen in fig . [ fig7 ] for the emitted @xmath3-particle . the total distribution remains essentially unchanged if more than the four dominating potentials are included . each contribution has its own characteristic feature . the first has a peak close to the maximum energy , i.e. resembling @xmath3-emission from a neutron - neutron @xmath155-state . the second has a peak at intermediate energy , i.e. resembling sequential decay by the @xmath3-neutron resonance . the third has a peak at small energy , i.e. resembling @xmath3-emission from an excited neutron - neutron @xmath0-state . in addition the fourth potential also gives a small contribution with maxima at intermediate and maximum energy . the size of about 1% can not be seen in the total distribution . however , this eigenvalue has the same angular momentum quantum numbers as the first level . therefore the non - diagonal interference term would be about 10% of the total contribution . it turns out that the interference essentially is destructive and responsible for the almost flat region at intermediate energies . the decay mechanism is then not simple although understandable in terms of our formulation . the main contribution is decay via the virtual @xmath153-state and the second is from direct decay . the third mechanism is produced by the coupling to the higher - lying state taking place at relatively small @xmath4 . this populates the level eventually approaching the @xmath159 hyperspherical level at large @xmath4 . the interference with the dominating contribution then leads to the total distribution . the division into these different contributions of direct and sequential is to some extent artificial but perhaps useful in connection with the experimental analysis . the mixture of all these contributions leads to the total distribution which has the right features but without precise reproduction of the high - energy peak , see fig . [ fig7 ] . experimental resolution would not improve very much because either the peak gets broader and lower , or it gets higher and narrower . the discrepancies can originate from the presence of the target and the reaction mechanism itself as well as from contributions to the experimental points from other than resonance decays . the experiment selects the window of energies around the @xmath0 resonance position in the reaction @xmath164li(@xmath165h,@xmath166he)@xmath167he@xmath168 . this necessarily includes some background which perhaps has a different energy distribution than the @xmath0 resonance we investigated in the present calculations . we find a distribution where the two - neutron virtual @xmath153-state dominates whereas the measurements are broader as expected from non - resonance decays . in this work we focus on the decay of `` populated '' resonances , and an appropriate description of this reaction goes beyond the scope of the paper . an attempt to understand the distribution was published soon after the experiment in @xcite . the measured distribution was fitted by a linear combination of the lowest hyperharmonic functions of @xmath158 and 4 . the conclusion was that a substantial @xmath159 component is needed to reproduce the experiment . the decay mechanism dominated by the neutron - neutron virtual @xmath153-state was abandoned in favor of the @xmath159 component . this phenomenological analysis provides a good fit even for energies above the maximum allowed by the resonance energy . the decaying resonance wave function does not enter anywhere . the significance is not easy to interpret in terms of decay mechanisms as attempted in the present work . the neutron energy distribution is not measured but for future comparison we show our prediction in fig . the division into different adiabatic components show that the broad total distribution centered around an intermediate energy is obtained by adding several qualitatively similar contributions . the different mechanisms would all produce most likely energies around half the maximum value . to distinguish it is therefore necessary to measure both @xmath3-particles and neutrons after the decay . it is instructive to investigate the dependence of the distributions on the choices of @xmath4 , @xmath169 and basis size . the @xmath4-dependence is already indicated in fig . [ fig6 ] where the distributions are very stable as soon as @xmath4 is larger than about @xmath150 fm . however , this stability does require a sufficiently large basis which at least up to @xmath170 fm still can be handled in modest - size computers . it is also clear that a finite basis can not accurately describe the solutions when @xmath4 increases towards @xmath171 . then the angular solutions approach the hyperharmonics but a large basis is still required to reproduce the structures at small distances between pairs of particles . eventually this becomes impossible . the many basis functions cancel each other at larger distances . at intermediate distances , where the basis is sufficiently large , the resonance wave functions are independent of @xmath4 . for the radial solution this is seen in the right part of fig . [ fig4 ] , where for @xmath4 larger than about 50 fm , the ratio between the different radial components is approximately constant . the energy distributions are mainly dominated by the absolute squares of these ratios , although when different adiabatic components interfere also the real parts of these complex ratios may contribute individually . this behavior of the radial ratios is responsible for the stable behaviour of the energy distributions for sufficiently large values of rho . the constant behavior of the radial ratios also implies that the radial wave functions have already reached asymptotics as given in eq.([eq22 ] ) for all the channels , and therefore the distributions are independent of the scaling angle . however , the latter conclusion is based on an assumption of analyticity of the angular solutions as function of @xmath169 . when this scaling angle is changed corresponding to a rotation across a singularity like a two- or three - body resonance the continuity is broken and the solutions change as well . this is especially clear when we compare two solutions with @xmath169 smaller and larger than the angle corresponding to a two - body resonance . for the large @xmath169 one angular eigenvalue changes character and increases towards infinity as @xmath172 , see @xcite . this qualitative change of behavior necessarily causes a change of the angular wave functions because the upgoing eigenvalue at large distances fully describes the properties of the two - body resonance . these features were distributed over several wave functions for the small @xmath169-value . in between singularities the individual angular solutions are independent of both @xmath4 and @xmath169 . this may not be an apparent feature of the numerical solutions because the basis has to be sufficiently large for a complete description . as @xmath169 increases the effective ranges of the two - body interactions also increase and the stable region is pushed to larger @xmath4-values . this means that the minimum basis size has to increase with @xmath169 . we formulate a method to compute the energy distribution of three particles emerging after three - body decay of a many - body resonance . the complex energy of a resonance corresponds to a pole in the momentum - space wave function which has an absolute square of the form as breit - wigner shape multiplied by a smoothly varying function . in coordinate - space this form corresponds to a large - distance asymptotic wave function consisting of only outgoing waves . we show formally by fourier transformation that the coordinate - space asymptotic angular dependence determines the energy distribution by substituting momentum directions for the conjugate coordinate directions . for this the divergent fourier integral is regularized by the zeldovich prescription . for two - body decay the energy distribution is trivially given by the breit - wigner distribution of the initial resonance . energy conservation is taking care of everything else . for three - body decay the total energy can be distributed continuously among the three particles . we show that the resonance decay results in distributions obtained from the large - distance angular behavior of the coordinate wave function . the asymptotic behavior can correspond to either genuine three - body structures or two - body substructures for example corresponding to two - body resonances or configurations favored by substantial attraction as for virtual states . also virtual population of two - body intermediate substructures is allowed as an appropriate asymptotic behavior with a resulting special energy distribution . the different asymptotics characterize the different decay modes used in analyses of experimental data . different modes can co - exist . we illustrate by the decay of the @xmath0-state in @xmath1he . the practical computations employ the hyperspherical adiabatic expansion combined with the complex scaling method . we discuss how a large hyperradius necessarily must be accompanied by a large basis . convergent results may then be obtained with less efforts at moderate hyperradii and moderate basis sizes . for convergence it is crucial to have all three faddeev components , and especially if all decay mechanisms simultaneously should be included in the theoretical formulation . the wave function undergoes dramatic changes from small distances , where the resonance properties usually are determined , and large distances from which the energy distributions emerge . the reasons for these structural changes are that the small distance behavior is determined by the two - body resonance substructures , whereas the large - distance behavior is determined as a competition between two effects , i.e. the energetically favored configuration of smallest two - body angular momentum with attractive two - body potentials and maintaining the same structure as at small distances in higher - lying levels . in conclusion , theoretical interpretation of the simplest nuclear three - body decay without coulomb interactions is already rather complicated . it is then advisable to test any given method on these systems . the accuracy of computations of the more complicated decaying charged systems can then be judged . this is important since almost all nuclear three - body resonance decays involve charged systems . the goal is to interpret the soon - to - come accurate experimental correlation data for three - body decays of charged systems .

recently , interests have been generated on cooling techniques for mechanical systems at nano- and micron - scales @xcite . among them , the typically employed is the feedback cooling technique where an external feedback circuit is responsible for detecting the motion of the target and feeding a counteracting force against this motion ; through a general decrease of magnitude in the density noise spectrum , it was shown that the feedback can effectively reduce the fluctuation of the target and provide a cooling mechanism @xcite . some experiments based on the models that contain the feedback loops have been carried out in the past few years . a few are directed towards the cooling of micron- to nanometer - size mechanical resonators , aiming to reach a macroscopic quantum mechanical ground state and serving as a powerful manifestation of quantum mechanical effects @xcite . other experiments succeed in slowing down the motion of micron - size mirrors through the radiation pressure of an optical field in a fabry - prot ( fp ) cavity , aiming to reach a noise level and equivalently an effective temperature that are pertinent to the employment of high - precision detection of gravity waves @xcite . the forementioned implementations of feedback cooling through reduction of noise fluctuations invariably rely on an electrical circuitry external to the target system to be cooled . the controller here is usually fixed and attracts or repels the resonator through either electrostatic coulomb force or lorentzian force . if such an external detection - control unit could be eliminated in favor of a mechanism with self - detection of and self - adjustment to the target s thermal motion , we call the mechanism `` self - cooling '' @xcite . devices implementing this self - cooling use less components and are free from the reliance on an external circuitry and hence prone to less noise sources . in one case @xcite , an augmented cavity along with an extra optical cavity field is established on the other side of the mirror , in addition to the regular fp cavity , so as to counteract the radiation pressure from the original cavity field . this extra field cushions the motion of the pressure mirror and plays the role of feedback . in another case @xcite , a set of josephson junctions behaving as a qubit , serially connected to a mechanical resonating beam , serves delayed supercurrent into the circuit according to the magnetic flux through the circuit loop . the magnetic flux is controlled by the vibrating motion of the beam , which in turn is controlled by the magnetic field generated by the current feed . such a mutual dependence furnishes a self - feedback mechanism . it should be pointed out that both of the self - feedback setups require delayed feedback , which assumed _ a priori _ a non - markovian approximation that explicitly depends on the history of the target s motion . in these phenomenological treatments , the cooling target either couples itself to a static controller and makes itself prone to the noise stemmed from the feedback , or couples to a mechanically static detection construct and receives manually delayed feedback . hereby , we present a dynamic model based on an intrinsic mechanism with non - markovian feedback , which is obviously free from an external feedback loop and does not rely on a presupposition of historical dependence . this mechanism is illustrated by a simple mechanical system in which the target is modeled by a harmonic oscillator and attached to a dynamic controller , which is a relatively heavier resonator , through a spring . the target is controlled by an intrinsic feedback through the dynamics of coupling : earlier positions and velocities of the target affects the motion of the controller and this influence is subsequently fed back to the target . consequently , the accumulation of earlier states of the target will affect the state of itself later . with proper parameter setup , the target essentially experiences a resistance and decelerates its motion ; its displacement variance is shrunk , noise suppressed and effective temperature cooled down . the lack of a specific detection device for the motion of the target resonator and an external feedback circuit characterizes the intrinsic nature of the mechanical feedback . our numerical analysis shows the existence of a noise suppression capability of our scheme , e.g. the variance of displacement can be reduced to @xmath0 , and therefrom a cooling capability under a practical setting accessible in current experiments . we note that the scheme is theoretically illustrative through its simple model setup yet widely applicable because the general oscillator systems can be extended to quantum bosonic systems and other cases . in fact , a similar model and mechanism has been proposed to actively cool down the torsional vibration of a nanomechanical resonator through spin - orbit interactions @xcite . the model will be explained in sec.[sec : model ] and its delay function then derived _ a posteriori _ to examine its non - markovian dependence . the complete solution of the system dynamics is given in sec.[sec : exact_soln ] , with which we will derive the density noise spectrum and calculate the theoretical noise suppression rate . the associated numerical results will be presented in sec.[sec : num_analysis ] , given various parameter setups . the analysis is extended to the domain of momentum noise in sec.[sec : momentum - noise ] . , the target mass @xmath1 , the spring of constant @xmath2 , the ancillary mass @xmath3 , and the spring of constant @xmath4 . @xmath5 is the harmonic driving force.,width=288 ] our model setup ( see fig.[fig : model ] ) comprises two masses and three springs . the two masses are denoted by @xmath1 and @xmath3 , respectively . the mass @xmath1 is the target and typically lighter whereas the mass @xmath3 serves as an ancillary controller and is relatively heavier . the three springs are denoted by their hooke s constants @xmath6 , @xmath2 and @xmath4 , respectively . the spring of constant @xmath6 attaches the lighter mass @xmath1 to the fixed wall on the left and the spring of constant @xmath4 attaches the heavier mass @xmath3 to the fixed wall on the right . the spring of constant @xmath2 strings the two masses together . such a setup , intuitively , grants the heavier mass @xmath3 the function of a suspension system and a medium for the feedback . the symbol @xmath5 represents an external driving force which is necessary for the discussion of cooling but can be deemed zero for the present . the connected springs will give rise to mechanical vibrations of the masses . we let @xmath7 denote the effective mechanical resonance frequency for the mass @xmath1 , assuming the other mass @xmath3 is fixed , and @xmath8 the equivalent for the mass @xmath3 , assuming the mass @xmath1 is fixed . besides these mechanical vibrations , we assume each of the masses experience a frictional damping and we let @xmath9 denote the damping coefficient for the mass @xmath1 and @xmath10 that for mass @xmath3 . then according to the setup above , the coordinates of the two masses obey a coupled system of classical langevin equations @xmath11 where @xmath12 is the coordinate of the mass @xmath1 and @xmath13 that of the mass @xmath3 . @xmath14 and @xmath15 on the right hand side of the equations denote the random thermal noise generated by the mass @xmath1 and @xmath3 , respectively , due to their frictional damping . the frictional damping terms as dissipation and the random noise terms as fluctuations constitute the total brownian thermal force in the classical langevin formalism . this thermal force induces an thermal environment , the effect of which is divided among the two terms @xmath16 and @xmath17 according to the fluctuation - dissipation relations @xcite @xmath18 for now we assume the two masses are independently interacting with two thermal environments . that is , there are no correlations between the fluctuations of the two masses actually , we can realize from eq.([eq : langevin_x ] ) and eq.([eq : fd - relation ] ) that the motion of the mass @xmath1 is resisted by a frictional force @xmath9 , which in turn is transduced into thermal energy and heats up itself . this process , however , is mediated by the mass @xmath3 that stands between @xmath1 and the fixed wall through the term @xmath19 . if there were not the mass @xmath3 , the kinetic oscillation of @xmath1 would be instantly reacted by the surrounding springs of hooke s constant @xmath6 and @xmath4 . with the presence of the mass @xmath3 and the extra spring of constant @xmath2 , the oscillation of the mass @xmath1 will first squeeze the spring of constant @xmath2 , and then the squeezed spring will release and push the mass @xmath3 to the right . it follows that the spring of constant @xmath4 will be squeezed successively . the mediating mass @xmath3 breaks the original single spring into two and permits these two springs to stay in different states , squeezed and released , and hence essentially delays the reaction of the springs . this cascaded process is then reflected by the wall and executed in a reversed order to the left ; the delayed reaction of the springs of constant @xmath2 and @xmath4 acts back onto the oscillating mass @xmath1 . the delayed reaction depends on the oscillating velocity of the target mass @xmath1 and can thus be considered a feedback onto itself . the entire process can be regarded as a feedback loop from the view of control theory through the flow diagram shown in fig.[fig : feedback_flow ] . the target resonator @xmath1 is the system to be controlled and the ancillary resonator @xmath3 becomes the controller which detects the signal @xmath20 as its input and feeds the signal @xmath21 as its output . the cycle time of the loop corresponds to the delay of the controller - to - system reaction . the dependence of this delayed reaction on the tunable parameters , mainly those spring constants , allows us to control this self - feedback precisely to counteract the oscillating motion of the target mass . receives signal @xmath20 and outputs signal @xmath21 according to the motion of the target @xmath1.,width=302 ] the weakened motion is converted to a reduction of effective temperature of the target mass through an equivalent relation between the autocorrelation of the target s displacement and temperature , derived from the fluctuation - dissipation relation eq.([eq : fd - relation ] ) . under normal circumstances , that is , when the coupling mass @xmath3 were not present , this equivalent relation can be written as @xmath22 where @xmath9 is defined as in eq.([eq : langevin_x ] ) and @xmath23 is a function to be given explicitly . experimentally , the statistical variance of the coordinate is measured and the above relation is used to compute an effective temperature . increasing the damping coefficient @xmath9 will certainly increase the enveloping rate of the oscillating motion of @xmath20 . however , this reduction of motion can not lead to an equivalent noise suppression effect upon the target mass @xmath1 because of the constraint imposed by the friction - induced fluctuation relation described above . in fact , according to this relation , increasing damping rate leads to more thermal dissipation of the system . that is to say , simply enlarging frictional force merely results in a heating effect upon the target system . our aim , therefore , is to reduce the temperature of the target mass by reducing its coordinate variance through the feedback that does not simultaneously increase the damping coefficient . after we derive the explicit feedback response below , we will show the feedback is actually shrinking the variance in sec.[sec : exact_soln ] . generally , a feedback external to an oscillating system has the effect of adding an extra driving force term on the right hand side of the equation of motion , i.e. @xmath24 where we have let @xmath25 denote the coordinate of a general dynamic system , @xmath26 the feedback force , and @xmath9 , @xmath27 , @xmath14 terms of similar meanings to those defined in eq.([eq : langevin_x ] ) . if the target were to be cooled , the motion of the system should be slowed down . in other words , behaving as a function of the target s velocity , the force @xmath26 should have an equivalent effect of increasing the damping coefficient , but do not increase the fluctuations . besides , the feedback force should be dependent on the target s position and the noise source . summing up these requirements , we expect @xmath26 to be a function of @xmath28 , @xmath29 and @xmath14 . to overcome the generic fluctuation - dissipation relations and hence accomplish an efficient noise suppression , we shall use a non - markovian type feedback : @xmath26 not only depends on the current value of the velocity and the position of the target , but also their historical values at past times . if we represent the historical dependence by a time - derivative @xmath30 and weigh the contribution of the histories by a delay function @xmath31 , where @xmath32 stands for the current time and @xmath33 for the time in the past , the feedback force can be written as an integral with respect to @xmath33 , @xmath34 the above formula shows the mathematical character of a general feedback force . inversely , any function that can express the same character should be considered a feedback force . therefore , we can verify the dynamic response of the coupling mass @xmath3 in our model as an effective feedback mechanism by finding the corresponding specific expression for @xmath35 and @xmath31 . to do so , we solve eq.([eq : langevin_q ] ) by using fourier transforms and integration by parts ( the detailed derivation is given in appendix [ sec : derv_delay ] ) @xmath36 where @xmath37 denotes the inhomogeneous part of the equation , i.e. the external driving force to @xmath3 . in the solution , @xmath38\left\ { \sqrt{4\bar{\omega}^{2}-\gamma^{2}}\times\right.\label{eq : delay}\\ & & \left.\cos\left[\frac{\xi}{2}\sqrt{4\bar{\omega}^{2}-\gamma^{2}}\right]+\gamma\sin\left[\frac{\xi}{2}\sqrt{4\bar{\omega}^{2}-\gamma^{2}}\right]\right\ } \nonumber \end{aligned}\ ] ] is the delay function that we look for and @xmath39 correspondingly shows the historical dependence . if we plug the solution ( [ eq : soln_q ] ) into eq.([eq : langevin_x ] ) , i.e. reduce the degree of freedom of the variable @xmath21 in the equation by substituting with its formal solution , we arrive at an integro - differential equation of only a single variable @xmath20 , which is comparable to the feedback - containing equation of motion that appears in previous literature @xcite . however , in the latter case the delay function is assume to be a non - markovian approximation @xmath40 and @xmath41 is left unknown . here , we have explicitly implemented a self - feedback mechanism that exerts force back onto the mass @xmath1 after certain delay through the use of the coupling mass @xmath3 . the sinusoidal factor in the delay eq.([eq : delay ] ) illustrates the damped oscillating motion of the mass @xmath3 . the implicit time derivative of @xmath20 within the integrand implies up to an equivalent effect an additional friction to the motion of the mass @xmath1 , which damps the oscillation without increasing the target s fluctuation . the non - integral term is markovian and has the same order as @xmath42 . though it does not appear in eq.([eq : feedback_form ] ) , this term effectively reduces the oscillating frequency and shall not counteract the feedback effect . therefore , both terms impose noise suppression effect to our target . the only limitation , however , is the thermal fluctuation @xmath17 from the coupling mass @xmath3 itself and it will result in a limit for the suppression because of the competition between this fluctuation and the effective feedback . in order to examine the validity of the above proposed mechanism and verify the efficacy of the cooling rate , we find the analytic solutions of the langevin equations ( [ eq : langevin_x ] ) and ( [ eq : langevin_q ] ) to reflect the displacement of the target @xmath1 as a response to its own thermal noise , the motion of the ancillary mass @xmath3 and the thermal noise of mass @xmath3 . from the response function , we shall derive the effective damping coefficient and vibrating frequency of the target mass @xmath1 as well as the autocorrelation function of its coordinate . the noise spectrum and the total noise fluctuation are then defined upon these derived quantities . the displacement spectrum can be written as the sum of responses of the noise terms ( the derivation is given in appendix [ sec : derv_soln ] ) @xmath43 note that the two different susceptibilities @xmath44 and @xmath45\left[\bar{\omega}_{e}^{2}-\omega^{2}+i\omega\gamma_{e}\right]}\label{eq : response_f}\ ] ] reflect the system s different responses to thermal fluctuations from the mass @xmath1 and the mass @xmath3 . here , we have defined the effective vibrating frequency of the mass @xmath1 to be @xmath46 and the effective damping coefficient of the mass @xmath1 to be @xmath47 eq.([eq : fourier_x ] ) means that the mechanical susceptibility or response function of the target @xmath1 is adjusted because of the dynamic coupling of the target mass @xmath1 to the ancillary mass @xmath3 . the first term represents the direct effect of the thermal bath acting on the target mass @xmath1 ; whereas the second term represents the indirect effect of the thermal bath onto the mass @xmath1 through the mediating mass @xmath3 and the coupling between the two masses . the positivity of the second term in eq.([eq : damp_eff ] ) has asserted our expectation of increasing the damping rate without increasing thermal force . we shall also note that an additional noise source @xmath48 is imposed onto the adjusted susceptibility eq.([eq : response ] ) due to the dynamic coupling . but seeing that it is divided by a frequency - squared term , we expect it to be negligible when the target mass @xmath1 is not resonating at a frequency close to that of the mass @xmath3 . the density noise spectrum ( dns ) , which is defined by the equation@xmath49 can thus be computed from eq.([eq : fourier_x ] ) @xmath50 comparing the above equation with the case when @xmath3 is absent , i.e. when @xmath51 is degenerated to @xmath9 , we observe a general suppression at the noise peak and spreading of the noise spectrum . the observable effect of the noise sources on the motion of the mass @xmath1 is equivalent to the variance of the displacement of the mass @xmath1 in time domain , which is defined as the average of the entire noise spectrum , i.e. the integral of the dns of @xmath12 , @xmath52 the integral is computable after we approximate the effective damping rate and vibrating frequency by truncating their expansions shown in appendix [ sec : derv_var ] . as a result , the variance is a variable of temperature and the spring constants @xmath53 the noise fluctuation depends on the three spring constants @xmath6 , @xmath4 and @xmath2 as its parameters , and is independent of the masses @xmath1 and @xmath3 and the damping rate @xmath9 and @xmath10 of the resonators . we shall notice that the noise suppression effect for the target resonator is always present for all values the spring constant @xmath2 takes . this ideal result is due to the non - markovian feedback we derived in sec.[sec : model ] , which always increases the effective damping while retaining the same fluctuations . eq.([eq : supp_noise ] ) shows a complex relation between itself and its three parameters of spring constants . to illustrate its behavior , we focus on its relation with the spring constant @xmath2 . it is a monotonic decreasing function of @xmath2 and its two limiting values are @xmath54 which coincides with our expectation that enlarging the constant @xmath2 will render the feedback more effective due to the enlarged feedback amplitude eq.([eq : phi(x , t ) ] ) . the symmetry between the target mass and the ancillary mass in the model setup allows us to compute the variance of the ancillary mass following the same methodology @xmath55 the limiting value of @xmath56 at @xmath57 is identical to that of @xmath58 , which shows that under the extremal case where the two oscillating masses combines into one by a rigid body the limiting behaviors of the two bodies become the same . to show completely the noise suppression behavior of the two oscillating masses , the numerical analysis is separated into two parts with each part for each extremal end of the values of the spring constants whereas the other parameters are set to laboratory accessible values for common micromechanical resonators . the first case , the identical case , is where the springs attaching the two masses to the walls share the same hooke s constant , i.e. @xmath59 . we examine how the noise suppression behavior is affected by varying the value of the constant @xmath2 of the middle spring . the second case , the large detuning case , occurs when the springs attaching the two masses to the walls take vastly different values of their hooke s constant . again we examine the noise suppression limit for different values taken for the constant @xmath2 of the middle spring . we assume the springs with one end fixed to walls have @xmath60 . the lighter target mass has @xmath61 and its frictional damping rate @xmath62 . the heavier ancillary mass has @xmath63 and its frictional damping rate @xmath64 . when free from the stringing spring @xmath2 , the target mass will oscillate at a natural frequency of @xmath65 and the ancillary mass at @xmath66 . the system s initial temperature is set to room temperature @xmath67 . we first look at the density noise spectrum of the ancillary resonator when the middle spring is set to have its constant @xmath68 , @xmath69 , @xmath70 , @xmath71 and @xmath72 as shown in fig.[fig : dns_of_q_case1 ] . the noise peaks at the frequencies @xmath73 , @xmath74 , @xmath75 , @xmath76 and @xmath77 . with parameters : @xmath60 , @xmath63 , @xmath61 , @xmath64 , @xmath62 , @xmath67 . curves from top to bottom are plotted from @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72.,width=288 ] we notice that when tuning the spring constant @xmath2 , not only the peaking frequency is shifted to the right , the peak amplitude is reduced along with the increased value of @xmath2 . this proves a general suppression in noise and an equivalent cooling effect to the system . the details of the spectrum and the spread width can be shown more apparently when we cluster the peaks together with a common frequency @xmath78 for their corresponding peaking frequencies , which is shown in fig.[fig : dns_of_q_case1_shift ] . the total noise fluctuation reached after suppression can be computed from the area under each curve in the figure , using eq.([eq : var_x ] ) . the attenuated noise levels corresponding to the 5 values of the spring @xmath2 are , respectively , @xmath79 , @xmath80 , @xmath81 , @xmath82 and @xmath83 times a common factor of @xmath84 . with parameters : @xmath60 , @xmath63 , @xmath61 , @xmath64 , @xmath62 , @xmath67 and peaking frequencies shifted to the center of the plot . curves from top to bottom are plotted from @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72.,width=268 ] the cooling effect for the target mass @xmath1 is more obvious if we examine the clustered peak plot of the target mass density noise spectrum shown in fig.[fig : dns_of_x_case1 ] . however , differing from the behavior of the ancillary mass shown in fig.[fig : dns_of_q_case1 ] , the target mass is resonant at two peak frequencies for each of the 5 values of the spring constant @xmath2 . among the pairs of peaking frequencies , one group clusters in the low frequencies and the other spreads out in the high frequencies . the low frequency group , the rescaled along peak center plot shown on the left of fig.[fig : dns_of_x_case1 ] , shares exactly the same peaking frequencies as those of the ancillary mass and we expect this behavior takes place when the target mass is resonating with the ancillary mass . this harmonic driven noise associates with the noise source @xmath85 in eq.([eq : fourier_x ] ) and , as we argue before , does not contribute much to the overall thermal noise . the high frequency group spreads out to peak frequencies @xmath86 , @xmath87 , @xmath88 , @xmath89 and @xmath90 for the varying spring constant @xmath2 . the anharmonic noise with respect to the ancillary mass associates with the noise source @xmath91 in eq.([eq : fourier_x ] ) . our cooling mechanism in this identical case is fairly effective , with the noise level reached down to @xmath79 , @xmath80 , @xmath81 , @xmath82 and @xmath83 times a common factor of @xmath84 , respectively , identical to the values of the ancillary mass . with parameters : @xmath60 , @xmath63 , @xmath61 , @xmath64 , @xmath62 , @xmath67 and peaking frequencies shifted to the center of the plot . the left and the right subplots corresponds to the low and the high peak frequency groups . curves from top to bottom are plotted from @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72.,width=288 ] we assume , in this case , the spring constants @xmath92 and @xmath93 ; the masses @xmath94 and @xmath61 ; thus the natural oscillating frequencies for the two masses are retained . the damping coefficients and the initial temperature are left unchanged . varying the spring constant of the middle spring @xmath2 over the same five values gives the noise spectrum plot of the target mass @xmath1 shown in fig.[fig : dns_of_x_case2 ] . the plot is again rescaled to the center along the peaking frequencies @xmath86 , @xmath87 , @xmath95 , @xmath96 and @xmath97 . we note that the noise peaks at one frequency for each value of the spring @xmath2 . these peaking frequencies are close to those in the identical case above but the noise suppression rate , as we have expected , is much better . that means the large detuning not only helps suppress the noise source stemmed from the coupling mass @xmath3 to negligible amplitude but also makes the feedback more effective for countering the target s noise . the 5 values of the spring constant @xmath2 corresponds to noise fluctuations of @xmath79 , @xmath98 , @xmath99 , @xmath100 and @xmath101 times a common factor of @xmath84 . with parameters : @xmath92 , @xmath93 , @xmath94 , @xmath61 , @xmath64 , @xmath62 , @xmath67 and peaking frequencies shifted to the center of the plot . curves from top to bottom are plotted from @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72.,width=288 ] fig.[fig : dns_of_q_case2 ] shows the plot of the ancillary resonator s noise spectrum , again rescaled to the center along the peaking frequencies . these peaking frequencies are very close to those of the target mass . we predict that the large detuning between the springs @xmath4 and @xmath6 puts the ancillary mass @xmath3 into a particularly passive role that reflects the minute motion of the target . this also helps explain why the noise suppression is especially effective in this case by the fact that the mass @xmath3 has its speed comparable to the target @xmath1 but in an opposite direction such that the force it exerts through the spring @xmath2 can favorably counteract the motion of the target @xmath1 . nonetheless , the reinforced role that the ancillary mass @xmath3 plays means that itself does not belong to the target system . as shown in the figure , the ancillary mass almost retains its original noise level throughout the varying values of the spring @xmath2 . with parameters : @xmath92 , @xmath93 , @xmath94 , @xmath61 , @xmath64 , @xmath62 , @xmath67 and peaking frequencies shifted to the center of the plot . curves from left to right are plotted from @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72.,width=288 ] the effective temperature is usually defined according to the equipartition theorem through the equality between the thermal energy and the mechanical potential energy . the intrinsic nature of our model renders the stored potential energy of the target and of the environment in their shared link , the middle spring @xmath2 , inseparable and thus forbids our discussion of the effective temperature in the coordinate space . to make our model relevant to the discussion of cooling , we examine the system dynamics in its momentum space and define the effective temperature according to the kinetic energy , i.e. the momentum fluctuation , of the target . in addition , to follow the convention along previous literature , a harmonic driving force term @xmath102 where @xmath103 is the driving frequency , should be added to the feedback loop ( cf . fig.[fig : model ] ) . in our case , this harmonic force appears as an extra term in the equation of motion of the ancillary resonator @xmath3 . following the above arguments , we rewrite the coupled system of classical langevin equations eq.([eq : langevin_x ] ) and eq.([eq : langevin_q ] ) as@xmath104 where @xmath105 and @xmath106 denote , respectively , the momenta of the target resonator and the ancillary resonator . the thermal fluctuations @xmath14 and @xmath15 still obey the same set of relations eq.([eq : fd - relation ] ) and eq.([eq : fd - relation_f ] ) . consequently , we have a modified linear response ( cf . eq.([eq : fourier_x ] ) , eq.([eq : response ] ) and eq.([eq : response_f ] ) ) after combining the system of equations in the frequency domain ( see appendix [ sec : derv_eff_temp])@xmath107\bigr]\label{eq : fourier_p}\end{gathered}\ ] ] where @xmath108 and @xmath109 are the susceptibilities eq.([eq : response ] ) and eq.([eq : response_f ] ) . we can thus arrive at the noise spectrum in momentum , after following the same routine of computations,@xmath110 the square of momentum of the target is defined similarly as in eq.([eq : var_x])@xmath111 which can be separated into two parts : the variance of momentum , i.e. the fluctuation or noise in momentum space,@xmath112 and the mean squared , i.e. , the square of the steady state value of momentum , @xmath113 note that the harmonic force term only contributes to the first moment of the momentum . since thermal dissipation only induces variance of the target s momentum , the effective temperature of the target can inversely be determine by the variance of the kinetic energy through the energy equipartition theorem @xmath114 eq.([eq : dns_mo ] ) is independent of the two parameters @xmath115 and @xmath103 of the harmonic driving force . we conclude that the harmonic force bears no effect in reducing the system s effective temperature and hence in cooling . eq.([eq : dns_mo ] ) is also not analytically integrable , numerical integration shows the integral takes value similar to the case when the spring constant @xmath116 . that means the feedback is only effective in the displacement domain but not in the momentum domain . in summary , we have proposed a theoretical model to demonstrate the general self - feedback type noise suppression technique through the coupling between the target resonator and an adjuvant system . in particular , such a self - feedback is achieved through the dynamics of the adjuvant system . the explicit delay function has been given and its efficacy in reducing the noise spectrum verified . we have also used numeral results to confirm our observations . before concluding this paper , we add some remarks as follows : first , the theory of thermal equilibrium state can be used to explain why the system comprising two resonating masses has its noise reduced to a limit when the spring constant @xmath2 tends from zero to infinity . if @xmath2 were zero , the stringing spring between the two masses would disappear and the system would comprise two independent resonators , each of which will attain its thermal equilibrium over time . in this case , the system has two degrees of freedom . according to the energy equipartition theorem , the system must contain twice as much energy as @xmath117 . for the other limit , when @xmath2 tends to infinity , the two resonators become a rigid body and possess only one degree of freedom , which means the system would contain only @xmath117 of energy . this explicit picture clarifies how the combined system reduces its energy over the increasing value of @xmath2 . however , how each resonator in the combined system reduces its own energy is an open question though we can intuitively think the heat bath may play crucial role in the asymmetric thermalization of the two resonators as tow open system . another remark is that the energy equipartition theorem is applicable only after the system enters thermal equilibrium state . when interaction occurs between the two degrees of freedom contributed by the two masses , such an illustration is not appropriate . this is the reason why the nonlinear character of the noise suppression rate can not be explained by the energy equipartition theorem alone . in addition , we have assumed there does not exist correlation between the two noise sources for the two masses , i.e. @xmath118 . however , when the two masses oscillate very closely with each other , the above assumption based on independent thermal environments will not hold and we need to consider the case where @xmath119 another setting we shall consider is when the motions of the two masses are quantized : we need to change the fluctuation - dissipation relations eq.([eq : fd - relation ] ) and eq.([eq : fd - relation_f ] ) to @xmath120 where @xmath121 and @xmath122 are not simply @xmath123-functions and depends on the planck constant @xmath124 . we will present the general investigation on the case with quantum fluctuations and thermal bath correlations , but here we have concentrated on the simple case . the authors thank yong li of university of basel and ying dan wang of ntt basic research laboratories for helpful discussions . this work is supported by the nsfc with grants no.90203018 , no.10474104 , and no.60433050 . it is also funded by the national fundamental research program of china with grants no.2001cb309310 and no.2005cb724508 . fourier transforming eq.([eq : langevin_q ] ) gives @xmath125 where we use @xmath126 , @xmath127 and @xmath128 to denote the fourier transforms of @xmath12 , @xmath13 and @xmath15 , respectively , @xmath129 substituting the first two transform of the above into eq.([eq : q_tfm ] ) and taking the inverse fourier transform of @xmath127 , we get @xmath130 the linear integral with respect to @xmath131 can be viewed as a contour integral along a semicircle in the upper half - plane with two poles at @xmath132/2 $ ] . assuming @xmath133 , the corresponding residues are , @xmath134\nonumber \end{aligned}\ ] ] where @xmath135 denotes the unit step function originated from the positive locus of the integration path . using cauchy s theorem , @xmath21 is reduced to a single - integral form @xmath136\sin\left[\frac{1}{2}\sqrt{4\bar{\omega}^{2}-\gamma^{2}}(t-\tau)\right]\nonumber \end{aligned}\ ] ] where we have used the shorthand @xmath137 as defined in eq.([eq : phi(x , t ) ] ) . to write @xmath21 in our desired form , we further integrate by parts with respect to @xmath138 @xmath139\sin\left[\frac{t-\tau}{2}\sqrt{4\bar{\omega}^{2}-\gamma^{2}}\right]\nonumber \\ & = & \frac{1}{\bar{\omega}^{2}}\phi(x , t)+\frac{2}{\sqrt{4\bar{\omega}^{2}-\gamma^{2}}}\times\\ & & \int_{0}^{\infty}\mathrm{d}(t-\tau)\frac{\mathrm{d}\phi(x,\tau)}{\mathrm{d}(t-\tau)}2\exp\left[-\frac{1}{2}\gamma(t-\tau)\right]\times\nonumber \\ & & \left\ { \frac{\exp\left[\frac{1}{2}i\sqrt{4\bar{\omega}^{2}-\gamma^{2}}(t-\tau)\right]}{\gamma - i\sqrt{4\bar{\omega}^{2}-\gamma^{2}}}\right.+\nonumber \\ & & \left.\frac{\exp\left[-\frac{1}{2}i\sqrt{4\bar{\omega}^{2}-\gamma^{2}}(t-\tau)\right]}{\gamma+i\sqrt{4\bar{\omega}^{2}-\gamma^{2}}}\right\ } \nonumber \end{aligned}\ ] ] where the factor in the last two lines constitute the function @xmath31 as defined in eq.([eq : delay ] ) . to find the response function , we first take the fourier transform of eq.([eq : langevin_q ] ) and eq.([eq : langevin_x ] ) to get , respectively , eq.([eq : q_tfm ] ) and @xmath140 where we let @xmath126 , @xmath127 and @xmath141 denote the fourier transforms of @xmath12 , @xmath13 and @xmath14 as we did in appendix [ sec : derv_delay ] . substituting eq.([eq : q_tfm ] ) into eq.([eq : x_tfm ] ) , we have@xmath142 reshuffling the terms and putting again @xmath143 on one hand side , the equation becomes @xmath144}{\bar{\omega}^{2}-\omega^{2}+i\omega\gamma - g^{2}/[mm(\bar{\omega}^{2}-\omega^{2}+i\omega\gamma)]}.\ ] ] to find explicitly the response function , we group together the real terms and the imaginary terms in the denominator to write @xmath143 in a familiar form similar to the case where the coupling mass @xmath3 were not present ; whence the denominator becomes@xmath145-\omega^{2}\\ + i\omega\left[\gamma+\frac{g^{2}}{mm}\frac{\gamma}{(\bar{\omega}^{2}-\omega^{2})^{2}+\omega^{2}\gamma^{2}}\right]\end{gathered}\ ] ] and we can define the terms in the two brackets as in eq.([eq : freq_eff ] ) and eq.([eq : damp_eff ] ) . using eq.([eq : fourier_x ] ) , the associated fluctuation - dissipation relations of eq.([eq : fd - relation ] ) and eq.([eq : fd - relation_f ] ) in the frequency domain @xmath146 and the independence between the two noise sources , we find by using definition eq.([eq : dns_defn ] ) @xmath147\label{eq : dns_ll}\ ] ] where @xmath108 and @xmath109 are defined in eq.([eq : response ] ) and eq.([eq : response_f ] ) . seeing that @xmath148 we can factor out @xmath149 from the bracket in eq.([eq : dns_ll ] ) @xmath150\nonumber \\ & = & \frac{2k_{b}t}{m}\left[\frac{1}{(\bar{\omega}_{e}^{2}-\omega^{2})^{2}+\omega^{2}\gamma_{e}^{2}}\right]\times\\ & & \left[\gamma+\frac{g^{2}}{mm}\frac{\gamma}{(\bar{\omega}^{2}-\omega^{2})^{2}+\omega^{2}\gamma^{2}}\right]\nonumber \end{aligned}\ ] ] which equals to eq.([eq : dns ] ) . the integrand of eq.([eq : var_x ] ) has highest order of @xmath151 in the denominator and can not be integrated analytically . to make the integrand integrable , we introduce a parameter @xmath152 as a variable of the spring constant @xmath4 and approximate the integrand by expanding @xmath153 with respect to @xmath154 up to 2nd order in the vicinity of @xmath155 , in other words , when @xmath156 . that is , we let @xmath157 to calculate the three terms in the expansion , we first substitute @xmath4 with @xmath158 in eq.([eq : freq_eff ] ) , @xmath159 since the lowest - order term in the numerator and the denominator is , respectively , @xmath160 and @xmath161 , the second term of the above formula goes to 0 when @xmath162 . thus we find the first term in the expansion @xmath163 . the first - order derivative of @xmath164 reads @xmath165}{\left[(gm^{-1}\mu^{-1}-\omega^{2})^{2}+\omega^{2}\gamma^{2}\right]^{2}}.\ ] ] again by comparing the coefficients of the lowest - order terms in the numerator and the denominator , we shall see the second term in the expansion @xmath166 . the third term in the expansion can be obtained similarly by observing the limiting behavior of the numerator and the denominator of the second - order derivative of @xmath164 @xmath167^{3}},\ ] ] which reads @xmath168 after taking the limit . combining the results , we have @xmath169 the truncation error rate introduced in this approximation , by comparing eq . ( [ eq : exact_omega ] ) and eq.([eq : appr_omega ] ) , is @xmath170}{\bar{\omega}_{e}^{2}}.\ ] ] expanding @xmath153 and @xmath154 gives @xmath171(f^{2}+m^{2}\omega^{2}\gamma^{2})-g^{2}f\right\ } \nonumber \\ & & \times\left[(g+k)(f^{2}+m^{2}\omega^{2}\gamma^{2})-g^{2}f\right]^{-1}\end{aligned}\ ] ] where we have used the shorthand @xmath172 . since we only consider the usual cases with @xmath173 , the terms containing the damping coefficient can be omitted@xmath174f - g^{2}}{(g+k)f - g^{2}}.\ ] ] expanding @xmath15 and multiplying the numerator and denominator by @xmath175 , we find @xmath176}\ ] ] and at the other extreme of the expansion @xmath177 whence @xmath178 and @xmath2 can take any value @xmath179 therefore , the error can be sufficiently suppressed if we let @xmath57 and the expansion of @xmath153around @xmath177 is validated . following the same reasoning and procedure , we can approximate the effective damping coefficient eq.([eq : damp_eff ] ) , @xmath180 which becomes @xmath131-independent . substituting eq.([eq : appr_omega ] ) and eq.([eq : appr_gamma ] ) into eq.([eq : var_x ] ) , we get , after minor algebra , @xmath181 where @xmath182 can be considered the effective damping coefficient and vibrating frequency of the mass @xmath1 after the approximation . we can apply a limiting process to eq.([eq : appr_var ] ) @xmath183 and compute the integral using the theorem of residues as we did in appendix [ sec : derv_delay ] @xmath184 fourier transforming the langevin equations in the momentum space eq.([eq : langevin_mo_x ] ) and eq.([eq : langevin_mo_q ] ) and recognizing @xmath185 we have@xmath186 we can follow the lines in appendix [ sec : derv_soln ] at this point to get the fourier transform of the momentum of the target resonator@xmath187\\ \times\left[\bar{\omega}^{2}-\omega^{2}+i\omega\gamma-\frac{g}{mm}\frac{1}{\bar{\omega}-\omega^{2}+i\omega\gamma}\right]^{-1}\end{gathered}\ ] ] which can be written in a compact form as in eq.([eq : fourier_p ] ) after grouping the terms . the density noise spectrum in momentum can then be computed@xmath188\\ & + 2\pi f_{0}^{2}\omega^{2}m^{2}\delta(\omega+\omega_{0})l_{f}(\omega)l_{f}(-\omega_{0})\nonumber \\ & = 2k_{b}tm\frac{\omega^{2}\gamma_{e}}{(\bar{\omega}_{e}^{2}-\omega^{2})^{2}+\omega^{2}\gamma_{e}^{2}}+\\ & + 2\pi f_{0}^{2}\omega_{0}^{2}m^{2}\delta(\omega+\omega_{0})l_{f}(\omega_{0})l_{f}(-\omega_{0})\nonumber \end{aligned}\ ] ]

the study of the formation of population iii stars in a cosmological context via high - resolution simulations is becoming a mature discipline , with multiple groups finding similar results ( bromm , coppi , & larson 2002 ; bromm & loeb 2004 ; abel , bryan , & norman 2002 ; oshea & norman 2007 ; yoshida et al . 2006b ; gao et al . however , all of these calculations operate under a common fundamental assumption namely , the absence of an ultraviolet background . given the abundance of halos in which population iii stars would form and the general consensus that these stars must be massive , one expects that , though only a small fraction of the volume of the universe would be ionized , a significant background of ultraviolet radiation in the lyman - werner ( lw ) band ( 11.18 - 13.6 ev ) , which is capable of photodissociating molecular hydrogen , would be present for the formation of the bulk of population iii stars ( omukai 2001 ; yoshida et al . 2003 ; machacek , bryan , & abel 2001 ; wise & abel 2005 ) . given that atomic hydrogen is optically thin to this radiation and that it easily destroys molecular hydrogen , this could have a significant impact on the formation of primordial stars . ( 2001 ) used cosmological amr simulations to study the formation of primordial stars in the presence of a soft uv background , and later including an x - ray background ( machacek , bryan , & abel 2003 ) . they found that the lw background delays the formation of population iii stars and shifts halo formation to higher masses . however , their calculations were not of sufficiently high resolution to give any significant information about the formation of the protostellar cloud at the core of each collapsing cosmological halo . ( 2003 ) performed a suite of large sph simulations of high redshift structure formation , including two calculations with a soft uv background with @xmath6 and @xmath7 , where @xmath8 is the mean intensity of the uv background in the lw band in units of @xmath9 erg s@xmath10 @xmath11 hz@xmath10 sr@xmath10 . they found that gas cooling is suppressed in the higher lw background case , and predicted that star formation would not occur . on the other hand , & haiman ( 2002 ) showed analytically that primordial gas in halos with t@xmath12 k can still collapse to high densities in the presence of a strong molecular hydrogen - dissociating background via atomic hydrogen line radiation , and can eventually form h@xmath13 . they referred to these objects as second generation objects " in the sense that although still of primordial composition , their formation pathway is different from the first stars . an issue that needs to be addressed concerns the mode of primordial star formation in halos with t@xmath5 k and above . it is apparent from recent studies using large numbers of simulations that smaller halos , with masses of @xmath14 m@xmath15 , appear to always form primordial stars in isolation ( gao et al . 2006 ; oshea & norman 2007 ) . is this still the case in halos whose masses are an order of magnitude or more larger ? does a single primordial star form , or several ? this is particularly relevant given that many semi - analytic models of structure formation that follow the chemical evolution of structures at high redshift ( scannapieco et al . 2005 ; schneider et al . 2006 ) suggest that halos of this scale which have not been enriched by metals ( and thus will form primordial stars ) exist up to at least @xmath16 . if this is true , there is expected to be a strong photodissociating background at that time ( yoshida et al . 2003 ; wise & abel 2005 ) . this paper is the second in a series . in & norman ( 2007 ) ( hereafter referred to as paper i ) , we examined several aspects of population iii star formation in a @xmath17cdm universe , in the absence of an ultraviolet background which photodissociates molecular hydrogen . in this paper , we study population iii star formation in a single cosmological realization varying the strength of the lw background , with the goal of investigating the effect that this background has on the evolution and properties of gas in the halo core ( at radii @xmath18 pc ) and the ultimate fate of the gas at the center of t@xmath5 k halos . we use a single cosmological realization , varying the strength of the photodissociating background over the range suggested by & abel ( 2005 ) , and look for the cutoff in population iii star formation suggested by ( 2003 ) . in agreement with the findings of ( machacek et al . 2001 ) , we find that the addition of a photodissociating background results in a delay of the collapse of high density gas at the center of the most massive halo in the simulation and , as a result , an increase in the virial mass of this halo at the onset of baryon collapse . we find that , contrary to the results of ( 2003 ) , star formation is not suppressed for j@xmath0 , but occurs even with backgrounds as high as j@xmath1 . we find that 2 cooling leads to collapse despite the depressed halo core molecular hydrogen fractions f@xmath19 by two multiplicative effects : ( 1 ) the elevated 2 cooling rates per molecule at @xmath2 k , and ( 2 ) time . we find that halo core collapse occurs in the usual way once the gas in the halo core has become dense enough that the cooling time becomes much less than the hubble time . we also observe a relationship between the strength of the photodissociating background and the rate of accretion onto the evolving protostellar cloud core , with higher lw background fluxes resulting in higher accretion rates . this is a simple consequence of the suppression of molecular hydrogen formation ( and thus suppression of cooling ) by the photodissociating background , as well as the higher virial temperatures of the more massive halos at the epoch of collapse . this may have implications for the range of population iii stellar masses . finally , we find that the collapsing halo cores in our simulations do not fragment at densities below @xmath3 @xmath4 regardless of the strength of the soft uv background , suggesting that population iii stars forming in halos with t@xmath5 k may still form in isolation . the organization of this paper is as follows . in section [ methodology ] we provide a description of enzo , the code used to perform the calculations in this paper and of the simulation setup . the results from our simulations are presented in sections [ results.meanprop ] through [ results.rephalos ] : section [ results.meanprop ] discusses some of the mean halo properties observed in the calculations , section [ results.evolution ] discusses the evolution of the halo core prior to collapse , section [ results.collapse ] compares spherically - averaged halo properties for all simulations at the epoch of collapse , section [ results.fixred ] discusses a variety of halo properties at a fixed redshift , and section [ results.rephalos ] compares the evolution of two representative simulations taken from our ensemble . in section [ sect.issues ] we discuss neglected physics and possible numerical issues , and in section [ discuss ] we discuss some of the results presented in this work and their implications . finally , we present a summary of the main results in section [ summary ] . ` enzo ' is a publicly available , extensively tested adaptive mesh refinement ( amr ) cosmology code developed by greg bryan and others ( bryan & norman 1997a , 1997b ; norman & bryan 1999 ; oshea et al . 2004 , 2005b ) . the specifics of the enzo code are described in detail in these papers ( and references therein ) , but we present a brief description here for clarity . the enzo code couples an n - body particle - mesh ( pm ) solver ( efstathiou et al . 1985 ; hockney & eastwood 1988 ) used to follow the evolution of a collisionless dark matter component with an eulerian amr method for ideal gas dynamics by & colella ( 1989 ) , which allows high dynamic range in gravitational physics and hydrodynamics in an expanding universe . this amr method ( referred to as _ structured _ amr ) utilizes an adaptive hierarchy of grid patches at varying levels of resolution . each rectangular grid patch ( referred to as a `` grid '' ) covers some region of space in its _ parent grid _ which requires higher resolution , and can itself become the parent grid to an even more highly resolved _ child grid_. enzo s implementation of structured amr places no fundamental restrictions on the number of grids at a given level of refinement , or on the number of levels of refinement . however , owing to limited computational resources it is practical to institute a maximum level of refinement , @xmath20 . additionally , the enzo amr implementation allows arbitrary integer ratios of parent and child grid resolution , though in general for cosmological simulations ( including the work described in this paper ) a refinement ratio of 2 is used . since the addition of more highly refined grids is adaptive , the conditions for refinement must be specified . in enzo , the criteria for refinement can be set by the user to be a combination of any or all of the following : baryon or dark matter overdensity threshold , minimum resolution of the local jeans length , local density gradients , local pressure gradients , local energy gradients , shocks , and cooling time . a cell reaching any or all of the user - specified criteria will then be flagged for refinement . once all cells of a given level have been examined , rectangular solid boundaries are determined which minimally encompass the flagged cells on that level . a refined grid patch is then introduced within each such bounding volume , and the results are interpolated to a higher level of resolution . in enzo , resolution of the equations being solved is adaptive in time as well as in space . the timestep in enzo is satisfied on a level - by - level basis by finding the largest timestep such that the courant condition ( and an analogous condition for the dark matter particles ) is satisfied by every cell on that level . all cells on a given level are advanced using the same timestep . once a level @xmath21 has been advanced in time @xmath22 , all grids at level @xmath23 are advanced , using the same criteria for timestep calculations described above , until they reach the same physical time as the grids at level @xmath21 . at this point grids at level @xmath23 exchange baryon flux information with their parent grids , providing a more accurate solution on level @xmath21 . cells at level @xmath23 are then examined to see if they should be refined or de - refined , and the entire grid hierarchy is rebuilt at that level ( including all more highly refined levels ) . the timestepping and hierarchy rebuilding processes are repeated recursively on every level to the maximum existing grid level in the simulation . two different hydrodynamic methods are implemented in enzo : the piecewise parabolic method ( ppm ) ( woodward & colella 1984 ) , which was extended to cosmology by ( 1995 ) , and the hydrodynamic method used in the zeus magnetohydrodynamics code ( stone & norman 1992a , 1992b ) . we direct the interested reader to the papers describing both of these methods for more information , and note that ppm is the preferred choice of hydro method since it is higher - order - accurate and is based on a technique that does not require artificial viscosity , which smoothes shocks and can smear out features in the hydrodynamic flow . the chemical and cooling properties of primordial ( metal - free ) gas are followed using the method of ( 1997 ) and ( 1997 ) . this method follows the non - equilibrium evolution of a gas of primordial composition with 9 total species : @xmath24 , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , and @xmath32 . the code also calculates radiative heating and cooling following atomic line excitation , recombination , collisional excitation , free - free transitions , molecular line cooling , and compton scattering of the cosmic microwave background , as well as any of approximately a dozen different models for a metagalactic ultraviolet background that heat the gas via photoionization and/or photodissociation . we model the cooling processes detailed in ( 1997 ) , but use the & palla ( 1998 ) molecular hydrogen cooling function . the multispecies rate equations are solved out of equilibrium to properly model situations where , e.g. , the cooling time of the gas is much shorter than the hydrogen recombination time . a total of 9 kinetic equations are solved , including 29 kinetic and radiative processes , for the 9 species mentioned above . the chemical reaction equation network is technically challenging to solve due to the huge range of reaction time scales involved ; the characteristic creation and destruction time scales of the various species and reactions can differ by many orders of magnitude . as a result , the set of rate equations is extremely stiff , and an explicit scheme for integration of the rate equations can be costly if small enough timestep are taken to keep the network stable . this makes an implicit scheme preferable for such a set of equations , and enzo solves the rate equations using a method based on a backwards differencing formula ( bdf ) in order to provide a stable and accurate solution . it is important to note the regime in which this chemistry model is valid . according to ( 1997 ) and ( 1997 ) , the reaction network is valid for temperatures between @xmath33 k. the original model discussed in these two references is only valid up to n@xmath34 @xmath4 . however , addition of the 3-body h@xmath13 formation process allows correct modeling of the gas chemistry up until the point where collisionally induced emission from molecular hydrogen becomes an important cooling processes , which occurs at @xmath35 @xmath4 . we do not include heating by molecular hydrogen formation , which will be significant at densities of @xmath36 @xmath4 and above , and may affect temperature evolution at these high densities . a further concern is that the optically thin approximation for radiative cooling breaks down beginning at n@xmath37 @xmath4 . beyond this point , modifications to the cooling function that take into account the non - negligible opacity of the gas to line radiation from molecular hydrogen must be made , as discussed by & abel ( 2004 ) . even with these modifications , a more correct description of the cooling of gas of primordial composition at high densities will require some form of radiation transport , which will greatly increase the cost of the simulations . the simulations discussed in this paper are set up in a similar way to those in paper i. a dark matter - only calculation with @xmath38 particles in a three - dimensional simulation volume which is @xmath39 h@xmath10 mpc ( comoving ) on a side is set up at @xmath40 assuming a `` concordance '' cosmological model with no baryons : @xmath41 , @xmath42 , @xmath43 , @xmath44 ( in units of 100 km / s / mpc ) , @xmath45 , and using an eisenstein & hu power spectrum ( eisenstein & hu 1999 ) with a spectral index of @xmath46 . the cold dark matter cosmological model is assumed . this calculation is then evolved to @xmath47 using a maximum of four levels of adaptive mesh refinement , refining on a dark matter overdensity of 8.0 . at @xmath47 , the hop halo finding algorithm ( eisenstein & hut 1998 ) is used to find the most massive halo in the simulation . at this point , we generate a new set of initial conditions which contain the same large - scale power as the dark matter - only calculation , but include both dark matter and baryons such that the lagrangian volume in which the halo in the dark matter - only calculation formed is resolved at high spatial and mass resolution using a series of static nested grids , with a @xmath38 root grid and three static nested grids , for an overall effective root grid size of @xmath48 cells . the highest resolution grid is @xmath49 grid cells , and corresponds to a volume @xmath50 h@xmath10 comoving kpc on a side . the dark matter particles in the highest resolution grid are 14.48 h@xmath10 m@xmath15 and the spatial resolution of the highest resolution grid is 586 h@xmath10 parsecs ( comoving ) . previous work shows that this particle mass resolution is adequate to fully resolve the collapse of the halo ( abel et al . 2002 ; oshea & norman 2007 ) . all simulations are performed using the adaptive mesh cosmology code enzo , which is described in section [ enzocode ] . the simulations are started at @xmath40 and allowed to evolve until the collapse of the gas within the center of the most massive halo , assuming the presence of an unevolving soft uv background with intensities in the lyman - werner band of j@xmath51 and @xmath9 erg s@xmath10 @xmath11 hz@xmath10 sr@xmath10 . this range covers a much wider range of parameter space than the results described in ( 2001 ) , and completely encompasses the range of possible j@xmath52 values suggested by & abel ( 2005 ) , for a wide range of mean population iii stellar masses . note that many publications use f@xmath52 instead of j@xmath52 : f@xmath52 has units of erg s@xmath10 @xmath11 hz@xmath10 , and thus f@xmath53 j@xmath52 . we will use j@xmath52 throughout this paper , and for convenience express values in units of j@xmath54 , where @xmath8 is the mean intensity of the uv background in the lw band in units of @xmath9 erg s@xmath10 @xmath11 hz@xmath10 sr@xmath10 . the equations of hydrodynamics are solved using the ppm method with a dual energy formulation , which is required to adequately resolve the thermal properties of gas in high - mach flows . the nonequilibrium chemical evolution and optically thin radiative cooling of the primordial gas is modeled as described in section [ enzocode ] , following 9 separate species including molecular hydrogen ( but excluding deuterium ) , with an initial electron fraction of @xmath55 ( which is roughly consistent with ( 1968 ) for an @xmath56 , @xmath57 universe ) . note that the initial electron fraction in the calculation is relatively unimportant to the molecular hydrogen formation rates in halo cores , as the electron fraction at the center of a given halo is controlled primarily by mergers and the shock formed by accretion of gas onto the halo . adaptive mesh refinement is used throughout the innermost high resolution region such that cells are refined by factors of two along each axis , with a maximum of 22 total levels of refinement . this corresponds to a maximum spatial resolution of 115 h@xmath10 au ( comoving ) at the finest level of resolution , with an overall spatial dynamical range of @xmath58 . to avoid effects due to the finite size of the dark matter particles , the dark matter density is smoothed on a comoving scale of @xmath59 pc ( which corresponds to @xmath60 0.03 proper pc at @xmath61 ) . this is reasonable because at that scale in all of our calculations the gravitational potential in the halo of interest is completely dominated by the baryons . grid cells are adaptively refined based upon several criteria . we refine on baryon and dark matter overdensities in cells of 4.0 and 8.0 , respectively . this corresponds to a maximum mass of gas or dark matter per cell ( on the most highly refined static grid ) of m@xmath62 and @xmath63 m@xmath15 , respectively . in addition , the _ minimummassforrefinementlevelexponent _ parameter is set to @xmath64 for both the dark matter and baryon overdensity refinement criteria , meaning that the mass required to refine to a higher level decreases as a function of increasing level , as : @xmath65 where l is the current level of refinement . the negative exponent means that the mass resolution in the calculations is super - lagrangian for example , m@xmath66 m@xmath67 . in addition to refining on baryon and dark matter overdensity , these simulations include refinement criteria which ensure that shocks are always well - resolved , that the cooling time in a given cell is always longer than the sound crossing time of that cell , and that the jeans length is always resolved by at least 16 cells . this last criterion guarantees that the truelove criterion ( truelove et al . 1997 ) is always resolved by a factor of four more cells in each dimension than is strictly necessary , ensuring that no artificial fragmentation will take place . figures [ fig.image_jlw_0 ] and [ fig.image_jlw_1em21 ] show projections of baryon density and temperature for the j@xmath68 and j@xmath1 calculations , respectively , at the epoch at which each calculation collapses , defined as the redshift at which the baryon number density reaches @xmath69 @xmath4 . note that due to the rapid evolution of gas at high density , the `` collapse redshift '' depends very weakly on the exact choice of density threshold . the calculations are started from the same set of initial conditions , and the j@xmath1 calculation is clearly a later stage in the evolution of the j@xmath68 run . the satellite halos surrounding the primary halo which collapses at z=24.12 in the j@xmath68 run have merged by z=17.32 when the halo in the j@xmath1 case collapses ( cf . [ fig.merger ] ) . with virial masses of @xmath70 and @xmath71 , respectively , the latter halo is more than 20 times as massive as the former . though quite different in mass , both halos exhibit similar morphologies they are extremely centrally - concentrated , and only a single condensed object ( that is to say , a primordial protostellar cloud core ) is visible in the highest - resolution panel in each image . the existence of a single cloud core is common to all calculations discussed in this paper . simulation ( @xmath72 ; m@xmath73 m@xmath15 ) . top row : projected log baryon density . bottom row : projected , mass - weighted log baryon temperature . left column : region 2.13 kpc ( proper ) across and deep . middle column : region 133.3 pc ( proper ) across and deep . right column : region 0.130 pc ( proper ) across and deep . the middle and right column are zoomed images focusing on the high baryon density core which forms in the center of the halo . , scaledwidth=90.0% ] simulation ( @xmath74 ; m@xmath75 m@xmath15 ) . top row : projected log baryon density . bottom row : projected , mass - weighted log baryon temperature . left column : region 2.13 kpc ( proper ) across and deep . middle column : region 533 pc ( proper ) across and deep . right column : region 0.260 pc ( proper ) across and deep . the middle and right column are zoomed images focusing on the high baryon density core which forms in the center of the halo . , scaledwidth=90.0% ] figure [ fig.halovals1 ] shows several mean halo quantities for all simulations discussed in this paper , including halo collapse redshift as a function of j@xmath52 , virial mass as a function of j@xmath52 , virial temperature as a function of j@xmath52 , and the virial mass as a function of collapse redshift . simulations where the soft uv background is turned on are shown by solid squares , and the `` control '' j@xmath68 calculation is shown by an open square . in plots where j@xmath52 is shown on the x - axis , the j@xmath68 simulation is placed at log@xmath76 . figure [ fig.halovals1 ] shows that there is a clear relationship between the lw intensity and the collapse redshift and virial mass of the halo . a larger lw intensity results in a later collapse time and larger virial mass because the halo must be hotter in order to have a cooling time which is less than a hubble time despite the depressed 2 abundance in the high - density gas at the halo s center . this is discussed further in section [ results.evolution ] the final mass of the halos in the simulations with j@xmath1 is approximately a factor of 20 higher than that in the `` control '' simulation . this is in qualitative agreement with ( 2001 ) , who suggest that there is a `` minimum halo mass '' as a function of the strength of the lw background of the form @xmath77 this threshold mass is plotted in panel ( c ) of figure [ fig.halovals1 ] . [ eqn - mach - mthresh ] is only strictly valid over the range @xmath78 , because this was the range simulated by ( 2001 ) . we see that our points for @xmath79 parallel the threshold curve but at a mass approximately four times higher . this difference can be ascribed to two factors . first , the threshold mass is the lowest possible halo mass that can collapse , derived from a statistical sample . since we study only one realization which focuses on the most massive halo in the box , its mass is bound to be higher than the statistical minimum . second , ( 2001 ) s criterion for cooling catches halos at an earlier stage of evolution compared to our data points , which give the halos virial masses at the time of central baryon collapse ( the `` collapse redshift '' ) . examination of panels ( b ) and ( c ) of figure [ fig.halovals1 ] show that there is some sort of `` phase change '' between j@xmath80 and @xmath81 causing the halo mass at the time of collapse to increase steeply and non - monotonically . the collapse redshift steadily decreases as a function of uv increasing background strength ( i.e. collapse of gas at the center of the halo is delayed ) . however , panel ( c ) shows that the virial mass steadily increases with increasing uv background strength until j@xmath80 , at which point there is a jump in mass by more than a factor of four , above which the halo mass increases only slightly with increasing uv background strength . that something interesting should happen at these lw intensities is consistent with the findings of ( 2003 ) who found that 2 cooling is strongly suppressed at j@xmath82 . they predicted that star formation would be inhibited since the equilibrium 2 fraction is below the critical fraction for cooling derived by ( 1997 ) . contrary to these predictions , we find collapse not only at j@xmath82 , but also at values as large as j@xmath83 . we note that highest value considered by ( 2001 ) was j@xmath84 , and the highest simulated by ( 2003 ) was j@xmath82 . therefore our cases j@xmath85 have not been examined before , and certainly not at the resolution of our simulations . how is the gas in the center of these halos able to cool and collapse in such high radiation backgrounds ? this is analyzed in some detail in the next section . a hint is provided in ( d ) of figure [ fig.halovals1 ] , which plots the halo virial temperature as a function of j@xmath52 . at background strengths above j@xmath80 , the virial temperature is consistently approximately @xmath86 k , which is roughly the temperature at which atomic hydrogen cooling is effective . however , an examination of the radial temperature profiles shows that high - density gas in the center of the halo never reaches these temperatures , but is more typically 2000 k. atomic line cooling is unimportant at such low temperatures ; however the 2 cooling rate per particle is roughly 100 times as large at 2000k as at 500k . 2 cooling still operates in the halo centers despite low 2 abundance due to the higher cooling rates and long evolutionary timescales . in the rest of this section we merely present additional mean properties of the halo and its central region at collapse , and defer discussion of the relevant timescales to section [ results.evolution ] . vs. halo collapse redshift . panel ( b ) : halo virial mass vs. halo collapse redshift panel ( c ) : halo virial mass vs. j@xmath52 panel ( d ) : halo virial temperature vs. j@xmath52 the j@xmath68 `` control '' result are shown as an open square ( and is at log j@xmath87 in the panels which are a function of j@xmath52 ) . in the bottom left panel , the dashed line corresponds to the fitting function for threshold mass from ( 2001 ) , eqtn . 8.,scaledwidth=90.0% ] figure [ fig.halovals2 ] shows several properties of the halo core at the epoch of collapse , including the core temperature , molecular hydrogen fraction , and spherically - averaged accretion rate as a function of ultraviolet background strength , and the spherically - averaged accretion rate as a function of the molecular hydrogen fraction . all `` core '' values are spherically - averaged and measured at the mass shell where @xmath88 m@xmath15 of gas is enclosed . panel ( a ) shows that the core h@xmath13 fraction decreases significantly with increasing fuv flux , with a corresponding increase in the core temperature ( panel ( b ) ) . this relationship is similar to that noted in & norman ( 2007 ) , where the amount of molecular hydrogen at densities of @xmath89 @xmath4 varies between simulations , and correlates strongly with baryon temperature . given that the accretion of gas onto the protostellar cloud is subsonic , this results in a strong relationship between the soft uv background flux and accretion rate onto the protostellar cloud , as shown by panel ( c ) , with accretion rates varying by more than a factor of 30 between the j@xmath68 and @xmath90 cases . panel ( d ) shows the strength of the correlation between core h@xmath13 fraction and accretion rate . note that the values discussed above are insensitive to the exact definition of the halo `` core . '' core for several simulations with the same cosmic realization but a range of lw background intensities . panel ( a ) : j@xmath52 vs. baryon core temperature . panel ( b ) : j@xmath52 vs. baryon core @xmath31 fraction . panel ( c ) : j@xmath52 vs. instantaneous accretion rate . panel ( d ) : baryon core @xmath31 fraction vs. accretion rate . the j@xmath68 `` control '' result are shown as an open square ( and is at log j@xmath87 in the panels which are a function of j@xmath52 ) . all `` core '' values are spherically - averaged and measured at the mass shell where @xmath88 m@xmath15 of gas is enclosed . , scaledwidth=90.0% ] figure [ fig.halovals3 ] shows the halo core mass and mass fraction at the epoch of collapse ( note that the definition of `` core '' is somewhat different than in the previous figure ) . the halo `` core mass '' is defined as being all gas at a density of n@xmath91 @xmath4 or above , and the core mass fraction is defined as the core mass divided by the virial mass of the halo at the epoch of collapse . this choice of minimum density ensures that we have captured the entirety of the quasistatically contracting analog of a galactic molecular cloud core , and for practical purposes is comparable to the gas fraction of cold , dense gas discussed in ( 2001 ) . the size of the halo core increases in absolute terms as the strength of the photodissociating background is increased , from @xmath92 m@xmath15 until j@xmath93 , where it reaches @xmath94 m@xmath15 . above this value of j@xmath54 , the core mass then begins to decrease , reaching @xmath95 m@xmath15 at j@xmath1 . the core mass fraction generally decreases ( though with some noise in the relationship ) from @xmath96 at j@xmath68 to @xmath97 at j@xmath1 , though the total overall change is less than a factor of 2 . @xmath4 or above . the `` core mass fraction '' is defined as the `` core mass '' divided by the virial mass of the halo at the epoch of collapse . the j@xmath98 `` control '' result are shown as an open square ( and is at log j@xmath87 in the panels which are a function of j@xmath52 ) . , scaledwidth=90.0% ] we now examine in more detail the evolution of the conditions in the cloud center leading to collapse . in figure [ fig.evolution ] we plot the central density , temperature , entropy @xmath99 , and 2 fraction versus time for the calculations with @xmath100 ( solid line ) , @xmath101 ( short - dashed line ) , @xmath81 ( long - dashed line ) and @xmath102 ( dot - dashed line ) . in these plots , central values of temperature and density are defined as being the values of these quantities in the cell with the highest baryon density . it is most instructive to compare the two extreme cases , j@xmath103 , which very closely resembles the j@xmath104 case , and the j@xmath83 case , which according to ( 2003 ) should not collapse at all . in panel ( a ) we see the central density increase modestly from @xmath105 2 @xmath4 to @xmath105 20 @xmath4 over 40 myr , and then increase rapidly thereafter , reaching @xmath106 @xmath4 in a scant additional 15 myr . this evolution is driven by 2 cooling , as can be seen from the entropy evolution in panel ( c ) . as expected , runaway cooling occurs when the 2 fraction reaches @xmath107 ( panel d ) , consistent with the ( 1997 ) analysis . by contrast the j@xmath83 case requires 120 myr for the central density to increases from @xmath105 2 @xmath4 to @xmath105 20 @xmath4 . as shown in panel ( c ) , the first 50 myr of this contraction is adiabatic , followed by an increase in entropy due to some heating event ( mergers ) . the fact that central entropy is not decreasing for 90 myr is indicative of the fact that 2 cooling is unimportant over this interval due to extremely low equilibrium 2 fractions ( f@xmath108 ; ( panel d ) ) . however , something interesting happens in the j@xmath1 calculation at @xmath109 myr . prior to that time the central temperature has crept up to 5000 k due to the increased virial mass . the cooling rate per 2 molecule is about 3 orders of magnitude higher at 5000k than at 500k a typical temperature at the halo center in low uv background evolutions . we believe this high temperature is due to a combination of adiabatic heating as the potential well deepens , and merger - induced shock heating . as shown in figure [ fig.merger ] , three large halos merge between @xmath110 and @xmath111 , quadrupling the halo mass ( figure [ fig.halovals1 ] , panel ( b ) ) . the elevated cooling rates cool the gas at the halo center to @xmath112 , allowing the gas density to increase slightly . the two effects reduce the central entropy from @xmath113 700 k @xmath114 to @xmath115 250 k @xmath114 . this is followed by another heating event at t=190 myr , presumably due to another merger , followed by a second more catastrophic cooling event . at t=200 myr , the halo core collapses , driving the central density higher , temperature and entropy lower , and 2 fraction higher . why does this collapse occur ? this is analyzed in figure [ fig.timescales ] , where we plot the evolution of the important timescales in the center of the halo . in the j@xmath103 case , the 2-body 2 formation timescale is always roughly an order of magnitude smaller than the photodissociation timescale . consequently , the 2 abundance is out of equilibrium and grows steadily with time ( figure [ fig.evolution ] panel d ) . the 2 cooling time in the central zone ( @xmath116 myr ) is far less than the hubble time ( @xmath117 myr ) , and before too long runaway collapse occurs . at higher values of j@xmath52 , the 2-body 2 formation time always hovers around the photodissociation timescale , regardless of how long that timescale is . this tells us that 2 is in equilibrium : the 2-body formations balance the photo - destructions by the lw background . the equilibrium 2 fraction is given by @xmath118 where k@xmath119 is the rate coefficient for the formation of h@xmath120 ( the limiting reaction for h@xmath13 formation at the temperature and density range considered here ) , @xmath121 is the electron density , and k@xmath122 is the 2 photodissociation rate due to the solomon process ( abel et al . 1997 ; yoshida et al . 2003 ) . using t @xmath123 k and @xmath124 , we get @xmath125 , which is in good agreement with the 2 fractions we see in figure [ fig.evolution ] . since @xmath126/ n@xmath127 , the smaller @xmath128 is , the shorter its formation time . now let us consider the 2 cooling time evolution ( short - long - dashed line ) . it is far lass than a hubble time for j@xmath129 , ensuring that these halo cores will eventually cool and collapse . in the j@xmath83 case , the cooling time drops below the hubble time for the first time at t=170 myr , and again at t=190 myr where it begins a steady decrease toward @xmath130 yr . despite the low equilibrium 2 fraction , if we wait long enough the inexorable press of time eventually establishes the condition @xmath131 and we get runaway collapse , even for j@xmath83 . note that we get catastrophic cooling because we are in the density regime ( @xmath132 @xmath4 ) where 2 cooling is proportional to the square of the gas density , not linear in the gas density ( abel , bryan , & norman 2000 ; abel et al . 2002 ) . ( solid ) , @xmath101 ( short - dashed ) , @xmath81 ( long - dashed ) and @xmath90 ( dot - dashed ) . , scaledwidth=90.0% ] and @xmath90 ( panels ( a)(d ) , respectively ) : 2-body 2 formation time ( solid ) ; 3-body 2 formation time ( short - dashed ) ; 2 collisional dissociate time ( long - dashed ) ; 2 photodissociation time ( dot - dashed ) ; 2 cooling time ( short - long - dashed ) . , scaledwidth=90.0% ] , t@xmath133 k halo that collapses at z=17.32 with a lyman - werner background flux @xmath134 . in comic strip order , the redshifts are z=22 , 21 , 20 , 19 , 18 , and 17.32 . field of view is 53.571 comoving kpc ( 2.678 proper kpc at z = 19 ) . logarithm of the projected baryon density ( column density ) is displayed , with color table scaled to the maximum and minimum values in each image . , scaledwidth=90.0% ] figure [ fig.radprof_zcoll.1 ] shows radial profiles of baryon number density , temperature , enclosed mass , circular velocity , and rms mach number as a function of radius , as well as specific angular momentum as a function of enclosed mass , for all simulations discussed in this paper . to facilitate comparison , results from the output of each simulation where the peak baryon number density at the center of the halo is approximately @xmath135 @xmath4 are shown , in order to capture each halo at a similar evolutionary stage , rather than at a fixed point in time . panel ( a ) shows that all of the simulations have similar density profiles . the scatter in number density at a given radius is readily explained by variation in halo mass halos which collapse at later times ( due to higher uv background strength ) are more massive , and thus have higher overall baryon densities . this is shown in another way in panel ( c ) , in a plot of enclosed mass as a function of radius which shows that all halos have very similar profiles at @xmath136 pc , but a significant variation at larger radii which is related to halo mass . the plot of temperature as a function of radius in panel ( b ) shows an interesting trend as the lw background is increased , the overall halo temperature as well as the halo core ( where `` core '' is roughly defined as gas within @xmath137 pc of the halo center ) temperature go up . in the outskirts of the halo , where baryon densities are low and thus cooling times are long , this is due primarily to the increase in halo mass , with the peak temperature corresponding approximately to the halo virial temperature . in the halo core , however , this temperature is correlated more strongly with the h@xmath13 fraction , and thus the uv background , as suggested by figure [ fig.halovals2 ] . the plots of specific angular momentum as a function of enclosed mass and keplerian velocity fraction ( defined as the local circular velocity divided by the keplerian orbital velocity due to mass within that radius ) as a function of radius in panels ( d ) and ( e ) show some mild trends trends . it appears that halos in simulations with stronger uv backgrounds tend to have less specific angular momentum at a given mass shell , and also tend to have a lower keplerian velocity fraction at a given radius than gas in halos which form in the presence of a lower uv background indeed , the halo which forms in the j@xmath1 calculation has the least angular momentum of all of the simulations , and the lowest keplerian velocity fraction out to @xmath137 pc . we speculate that this is due to transport and segregation of angular momentum by turbulence halos whose collapse is delayed have more time for these processes to act within the halo core , resulting in less angular momentum overall when the halo ultimately collapses ( note that this will be discussed more fully in a later paper ) . the plot of rms mach number as a function of radius in panel ( f ) also shows no obvious trend , though at radii below @xmath138 pc , simulations with a smaller uv background generally have higher rms mach numbers . this is due primarily to the gas being colder , and not the gas velocities being higher . . panel ( a ) : @xmath139 as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : enclosed gas mass as a function of radius . panel ( d ) : specific angular momentum as a function of enclosed mass . panel ( e ) : keplerian velocity fraction as a function of radius . panel ( f ) : gas rms mach number as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to simulations , as follows . black solid line : j@xmath68 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : @xmath139 as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : enclosed gas mass as a function of radius . panel ( d ) : specific angular momentum as a function of enclosed mass . panel ( e ) : keplerian velocity fraction as a function of radius . panel ( f ) : gas rms mach number as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to simulations , as follows . black solid line : j@xmath68 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : @xmath139 as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : enclosed gas mass as a function of radius . panel ( d ) : specific angular momentum as a function of enclosed mass . panel ( e ) : keplerian velocity fraction as a function of radius . panel ( f ) : gas rms mach number as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to simulations , as follows . black solid line : j@xmath68 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : @xmath139 as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : enclosed gas mass as a function of radius . panel ( d ) : specific angular momentum as a function of enclosed mass . panel ( e ) : keplerian velocity fraction as a function of radius . panel ( f ) : gas rms mach number as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to simulations , as follows . black solid line : j@xmath68 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : @xmath139 as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : enclosed gas mass as a function of radius . panel ( d ) : specific angular momentum as a function of enclosed mass . panel ( e ) : keplerian velocity fraction as a function of radius . panel ( f ) : gas rms mach number as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to simulations , as follows . black solid line : j@xmath68 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : @xmath139 as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : enclosed gas mass as a function of radius . panel ( d ) : specific angular momentum as a function of enclosed mass . panel ( e ) : keplerian velocity fraction as a function of radius . panel ( f ) : gas rms mach number as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to simulations , as follows . black solid line : j@xmath68 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=30.0% ] figure [ fig.radprof_zcoll.2 ] shows the h@xmath13 fraction , electron fraction , h@xmath120 fraction , ratio of cooling time to sound crossing time , ratio of cooling time to dynamical time , and ratio of sound crossing time to dynamical time as a function of radius , for all simulations . outputs and line types correspond to those in figure [ fig.radprof_zcoll.1 ] . there is a clear relationship at all radii between h@xmath13 fraction and uv background strength in panel ( a ) as the fuv background is increased , the overall h@xmath13 fraction decreases . this difference is most noticeable at radii of @xmath146 pc , but is maintained at larger and smaller radii . there is a `` kink '' at approximately @xmath101 pc where the values are all quite similar . this corresponds to a baryon number density of @xmath36 @xmath4 in all simulations , which is where 3-body h@xmath13 formation begins to occur . panels ( b ) and ( c ) show that e@xmath120 and @xmath29 fractions track each other , which is to be expected the local electron fraction controls the amount of h@xmath120 which can be produced , which is the limiting reaction in the dominant mode of h@xmath13 formation for n @xmath147 @xmath4 and there is a general trend with increasing electron and h@xmath120 fractions with increasing uv background flux . the plots of the ratio of cooling time to sound crossing time , ratio of cooling time to dynamical time , and ratio of sound crossing time to dynamical time as a function of radius shown in panels ( d ) , ( e ) and ( f ) display some interesting trends . the three simulations with the lowest uv background strengths have a much longer cooling time than the other calculations , which are all grouped roughly together with no discernible pattern . this agrees well with the plot of the ratio of cooling time to dynamical time , where a similar trend is observed . the difference between the three simulations with the lowest uv background strengths and the others is due to the somewhat higher h@xmath13 fraction in the halo core of these calculations . the gas temperature is @xmath148 k at the center of the halo cores in these calculations , and the gas can not cool further , which results in a very long cooling time . the sound crossing time is also increased , but not as significantly . all simulations have approximately the same ratio of sound crossing to dynamical times at all radii . given that the cooling time is typically longer than both the sound crossing time and dynamical time , one can infer that the collapse of the halo is occurring quasi - statically for all simulations . . panel ( a ) : molecular hydrogen fraction as a function of radius . panel ( b ) : electron fraction as a function of radius . panel ( c ) : @xmath29 fraction as a function of radius . panel ( d ) : ratio of gas cooling time to sound crossing time as a function of radius . panel ( e ) : ratio of gas cooling time to dynamical time as a function of radius . panel ( f ) : ratio of gas sound crossing time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : molecular hydrogen fraction as a function of radius . panel ( b ) : electron fraction as a function of radius . panel ( c ) : @xmath29 fraction as a function of radius . panel ( d ) : ratio of gas cooling time to sound crossing time as a function of radius . panel ( e ) : ratio of gas cooling time to dynamical time as a function of radius . panel ( f ) : ratio of gas sound crossing time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : molecular hydrogen fraction as a function of radius . panel ( b ) : electron fraction as a function of radius . panel ( c ) : @xmath29 fraction as a function of radius . panel ( d ) : ratio of gas cooling time to sound crossing time as a function of radius . panel ( e ) : ratio of gas cooling time to dynamical time as a function of radius . panel ( f ) : ratio of gas sound crossing time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : molecular hydrogen fraction as a function of radius . panel ( b ) : electron fraction as a function of radius . panel ( c ) : @xmath29 fraction as a function of radius . panel ( d ) : ratio of gas cooling time to sound crossing time as a function of radius . panel ( e ) : ratio of gas cooling time to dynamical time as a function of radius . panel ( f ) : ratio of gas sound crossing time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : molecular hydrogen fraction as a function of radius . panel ( b ) : electron fraction as a function of radius . panel ( c ) : @xmath29 fraction as a function of radius . panel ( d ) : ratio of gas cooling time to sound crossing time as a function of radius . panel ( e ) : ratio of gas cooling time to dynamical time as a function of radius . panel ( f ) : ratio of gas sound crossing time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . , title="fig:",scaledwidth=30.0% ] . panel ( a ) : molecular hydrogen fraction as a function of radius . panel ( b ) : electron fraction as a function of radius . panel ( c ) : @xmath29 fraction as a function of radius . panel ( d ) : ratio of gas cooling time to sound crossing time as a function of radius . panel ( e ) : ratio of gas cooling time to dynamical time as a function of radius . panel ( f ) : ratio of gas sound crossing time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . , title="fig:",scaledwidth=30.0% ] figure [ fig.radprof_zcoll.3 ] shows radial velocity as a function of radius , accretion time ( defined as @xmath149 ) as a function of enclosed baryon mass @xmath150 as a function of enclosed baryon mass and @xmath150 as a function of time for all simulations . output time and lines correspond to those in figures [ fig.radprof_zcoll.1 ] and [ fig.radprof_zcoll.2 ] . panel ( a ) shows a clear relationship between the strength of the soft uv background and the radial velocity , with the calculations that have higher uv backgrounds also having higher infall velocities . this is easily understood by the quasistatic contraction of the gas , which takes place at or below the local sound speed in the gas , as discussed by ( 2002 ) and & norman ( 2007 ) . given that the halo temperatures are systematically higher in calculations where the uv background is stronger , the sound speed is higher , and thus the rate at which the gas at the center of the halo contracts is higher . this leads to lower accretion times ( panel ( b ) ) and higher accretion rates ( panel ( c ) ) , with the accretion rates for the simulations with the strongest uv backgrounds being higher than those with a low ( or no ) uv background . panel ( d ) shows that the time evolution of the accretion rate in all simulations is qualitatively similar , but varies strongly in absolute magnitude and in the time at which the accretion rate onto the evolving protostellar cloud peaks , suggesting that the stars forming in these simulations may have very different evolutionary histories . . panel ( a ) : radial velocity as a function of radius . panel ( b ) : accretion time as a function of enclosed baryon mass . panel ( c ) : instantaneous accretion rate as a function of enclosed baryon mass . panel ( d ) : estimated accretion rate as a function of time . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . the upper and lower light short - long - dashed curves which extend from the upper left corner of panel ( b ) correspond to the main sequence lifetime of a massive population iii star of that mass and the kelvin - helmholtz timescale of a population iii timescale with a given luminosity and radius . all values are taken from ( 2002 ) . the three light diagonal short - dashed lines which extend from bottom left to top right in panel ( b ) correspond to masses accreted using constant accretion rates of ( from top to bottom ) @xmath151 , @xmath101 , and @xmath81 m@xmath152yr . the light horizontal short - dashed line in panels ( c ) and ( d ) correspond to the `` critical '' accretion rate of & palla ( 2003 ) , @xmath153 m@xmath152yr . , title="fig:",scaledwidth=45.0% ] . panel ( a ) : radial velocity as a function of radius . panel ( b ) : accretion time as a function of enclosed baryon mass . panel ( c ) : instantaneous accretion rate as a function of enclosed baryon mass . panel ( d ) : estimated accretion rate as a function of time . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . the upper and lower light short - long - dashed curves which extend from the upper left corner of panel ( b ) correspond to the main sequence lifetime of a massive population iii star of that mass and the kelvin - helmholtz timescale of a population iii timescale with a given luminosity and radius . all values are taken from ( 2002 ) . the three light diagonal short - dashed lines which extend from bottom left to top right in panel ( b ) correspond to masses accreted using constant accretion rates of ( from top to bottom ) @xmath151 , @xmath101 , and @xmath81 m@xmath152yr . the light horizontal short - dashed line in panels ( c ) and ( d ) correspond to the `` critical '' accretion rate of & palla ( 2003 ) , @xmath153 m@xmath152yr . , title="fig:",scaledwidth=45.0% ] . panel ( a ) : radial velocity as a function of radius . panel ( b ) : accretion time as a function of enclosed baryon mass . panel ( c ) : instantaneous accretion rate as a function of enclosed baryon mass . panel ( d ) : estimated accretion rate as a function of time . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . the upper and lower light short - long - dashed curves which extend from the upper left corner of panel ( b ) correspond to the main sequence lifetime of a massive population iii star of that mass and the kelvin - helmholtz timescale of a population iii timescale with a given luminosity and radius . all values are taken from ( 2002 ) . the three light diagonal short - dashed lines which extend from bottom left to top right in panel ( b ) correspond to masses accreted using constant accretion rates of ( from top to bottom ) @xmath151 , @xmath101 , and @xmath81 m@xmath152yr . the light horizontal short - dashed line in panels ( c ) and ( d ) correspond to the `` critical '' accretion rate of & palla ( 2003 ) , @xmath153 m@xmath152yr . , title="fig:",scaledwidth=45.0% ] . panel ( a ) : radial velocity as a function of radius . panel ( b ) : accretion time as a function of enclosed baryon mass . panel ( c ) : instantaneous accretion rate as a function of enclosed baryon mass . panel ( d ) : estimated accretion rate as a function of time . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at the point where the maximum number density is approximately @xmath135 @xmath4 . line types and weights correspond to those in figure [ fig.radprof_zcoll.1 ] . the upper and lower light short - long - dashed curves which extend from the upper left corner of panel ( b ) correspond to the main sequence lifetime of a massive population iii star of that mass and the kelvin - helmholtz timescale of a population iii timescale with a given luminosity and radius . all values are taken from ( 2002 ) . the three light diagonal short - dashed lines which extend from bottom left to top right in panel ( b ) correspond to masses accreted using constant accretion rates of ( from top to bottom ) @xmath151 , @xmath101 , and @xmath81 m@xmath152yr . the light horizontal short - dashed line in panels ( c ) and ( d ) correspond to the `` critical '' accretion rate of & palla ( 2003 ) , @xmath153 m@xmath152yr . , title="fig:",scaledwidth=45.0% ] figure [ fig.radprof_fixred.1 ] shows the status of several spherically - averaged quantities for halos from all simulations at a fixed redshift , @xmath154 , which is shortly before the baryons in the most massive halo in the j@xmath68 simulation collapses to high density . this is at an early point in the evolution of the majority of these halos , and several conclusions can be drawn from the plots of number density , temperature , h@xmath13 fraction , and radial velocity as a function of radius . the molecular hydrogen fraction declines monotonically with increasing uv background strength , in agreement with an estimate for equilibrium h@xmath13 fraction at low ( @xmath155 @xmath4 ) densities : @xmath156 where k@xmath119 is the rate coefficient for the formation of h@xmath120 ( the limiting reaction for h@xmath13 formation at the temperature and density range considered here ) and k@xmath52 is the rate coefficient for h@xmath13 in the lw band ( abel et al . 1997 ; yoshida et al . halo cores at higher central densities are departing from the approximations used in the above estimate and thus have values of the h@xmath13 fraction that are not quite the equilibrium values . halo cores which have reached core h@xmath13 fractions higher than @xmath157 have been able to cool to significantly below the virial temperature ( since the cooling time is much less than the hubble time at that value of the h@xmath13 fraction ) , allowing the baryon density to increase . the effect of cooling can also be seen in the plot of radial velocity as a function of radius , where the core regions of halos that have cooled are contracting ( albeit very slowly ) , while those which are still at the virial temperature are not showing significant signs of contraction within @xmath158 pc . all halos have some inflow at large radii , due to the infall of gas from adjoining filaments . . panel ( a ) : number density as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : @xmath31 fraction as a function of radius . panel ( d ) : radial velocity as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at @xmath154 ( shortly before the j@xmath68 simulation collapses to high density . line types and weights correspond to simulations , as follows . black solid line : j@xmath159 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=45.0% ] . panel ( a ) : number density as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : @xmath31 fraction as a function of radius . panel ( d ) : radial velocity as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at @xmath154 ( shortly before the j@xmath68 simulation collapses to high density . line types and weights correspond to simulations , as follows . black solid line : j@xmath159 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=45.0% ] . panel ( a ) : number density as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : @xmath31 fraction as a function of radius . panel ( d ) : radial velocity as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at @xmath154 ( shortly before the j@xmath68 simulation collapses to high density . line types and weights correspond to simulations , as follows . black solid line : j@xmath159 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=45.0% ] . panel ( a ) : number density as a function of radius . panel ( b ) : temperature as a function of radius . panel ( c ) : @xmath31 fraction as a function of radius . panel ( d ) : radial velocity as a function of radius . all quantities except enclosed gas mass are mass - weighted , and all simulations are shown at @xmath154 ( shortly before the j@xmath68 simulation collapses to high density . line types and weights correspond to simulations , as follows . black solid line : j@xmath159 . black short - dashed line : j@xmath140 . black long - dashed line : j@xmath141 . black dot short - dashed line : j@xmath142 . black dot long - dashed line : j@xmath80 . black short dashed - long dashed line : j@xmath143 . red solid line : j@xmath144 . red short - dashed line : j@xmath145 . red long - dashed line : j@xmath1 . , title="fig:",scaledwidth=45.0% ] in this section we compare the evolution of two representative simulations . we choose the calculations which correspond to j@xmath68 and @xmath90 . these two simulations represent the extremes in halo evolution the j@xmath68 calculation is an example of the evolution of a `` standard '' halo , as discussed in previous literature , and the j@xmath1 simulation is the most extreme example of a population iii protostar forming in a t@xmath5 k halo available from our suite of calculations . as described in section [ sect.simsetup ] , both calculations start from the same initial conditions , with only the strength of the lw background being different . projections of the log baryon density and temperature for the j@xmath68 and @xmath90 simulations are shown in figures [ fig.image_jlw_0 ] and [ fig.image_jlw_1em21 ] , respectively . these projections are at the epoch of collapse , which is @xmath160 @xmath161 for the j@xmath68 @xmath162 simulation . the halo masses are significantly different : @xmath163 and @xmath164 m@xmath15 for the two calculations , reflecting the rapid pace of mergers at that redshift , and the amount of structure evolution that takes place over that relatively short time period ( @xmath165 years ) . this can clearly be seen in the left column of both figures , when the large amount of structure apparent in figure [ fig.image_jlw_0 ] has merged into the main halo by the time of protostellar cloud formation in the j@xmath1 case , as seen in figure [ fig.image_jlw_1em21 ] . despite the differences in evolutionary states , the halos themselves are quite similar in figures [ fig.image_jlw_0 ] and [ fig.image_jlw_1em21 ] . both calculations show halos that are strongly centrally concentrated , as shown by the center column in both figures . this column shows projections of the halos which are centered on the baryon peak and are approximately one virial radius across ( and thus are scaled differently in both images ) . no fragmentation of the halo core is visible . the rightmost panel in both calculations shows the center of the halo , where the protostellar cloud is evolving . there is still no evidence for fragmentation up to a central baryon density of @xmath3 @xmath4 . figures [ fig.radprof_jlw0.1 ] and [ fig.radprof_jlw1em21.1 ] show the temporal evolution of several spherically - averaged baryon quantities for the j@xmath68 and @xmath90 calculations , respectively . these plots show baryon number density , enclosed baryon mass , temperature , h@xmath13 fraction , ratio of cooling time to sound crossing time , and ratio of cooling time to dynamical time , as a function of radius , evolving from a number density of @xmath166 @xmath4 to @xmath69 @xmath4 in both calculations . some qualitative commonalities are obvious between the two calculations . both halos steadily grow in central density ( panel ( a ) ) , which is coupled to the growing molecular hydrogen fraction in the central region of the halo ( panel ( d ) ) . both halos collapse quasi - statically , as shown in panels ( e ) and ( f ) . also , the temperature profiles ( as shown in panel ( c ) ) are broadly similar , in that the gas in the halo central region cools to some minimum value at r @xmath167 pc ( corresponding to n @xmath168 @xmath4 in both simulations ) and then creeps steadily upward as gas collapses to higher densities . though there are qualitative similarities between the evolution of the two calculations , detailed examination shows some significant quantitative differences . the time that the halo in the j@xmath68 simulation takes to evolve from a central density of @xmath169 @xmath4 to @xmath135 @xmath4 is @xmath170 myr , with 21 myr of that being the time required to evolve from @xmath169 to n @xmath171 @xmath4 . the time required for the j@xmath1 calculation to evolve from @xmath169 @xmath4 to @xmath135 @xmath4 is slightly shorter , @xmath60 17 myrs , with 8 myrs required to evolve to n @xmath172 @xmath4 . the temperature evolution is also somewhat different the j@xmath68 simulation has a minimum temperature at r @xmath137 pc of 200 k , with n @xmath171@xmath173 and a molecular hydrogen fraction of @xmath174 , and roughly 1000 m@xmath15 of enclosed gas . as the gas within this radius evolves to higher densities , the temperature increases by a factor of a few , peaking at t @xmath175 k at n @xmath176 @xmath4 , with f@xmath177 . the gas cooling time within r @xmath137 pc in the j@xmath68 simulation is always larger than the sound crossing and dynamical times by at least a factor of two , implying a quasistatic gas contraction at all times . the j@xmath1 simulation evolves somewhat differently . the collapsing gas reaches a minimum temperature of @xmath178 k at r @xmath179 pc , with a baryon density of @xmath180 @xmath4 , an enclosed gas mass of roughly @xmath86 m@xmath15 , and a molecular hydrogen fraction of @xmath181 , two orders of magnitude below that in the j@xmath68 calculation at the equivalent temperature minimum . the low h@xmath13 fraction is due to the high lw background radiation flux . as the halo evolves to higher densities , the gas temperature and h@xmath13 fraction also creep upwards . the temperature of the gas when the peak reaches roughly @xmath135 @xmath4 is @xmath182 k , with a h@xmath13 fraction of @xmath183 . the cooling time at r @xmath179 pc is an order of magnitude higher than the sound crossing or dynamical times , implying an extremely slow contraction of the gas at that radius . as the central density increases , however , the cooling time becomes lower , with t@xmath184/t@xmath185 and t@xmath184/t@xmath186 dipping to @xmath187 at its minimum , but the latter increasing toward the central density peak to a ratio of @xmath188 . this implies a more rapid contraction of gas in the halo center in the simulation with a higher uv background , despite the lack of molecular hydrogen . it is also worth noticing that the j@xmath1 calculation has h@xmath13 fractions at n @xmath168 and @xmath106 @xmath4 of @xmath189 and @xmath190 , respectively , as compared to f@xmath191 and @xmath192 at the same densities in the j@xmath68 calculation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath193 , @xmath194 years . red dot long - dashed line : @xmath195 , @xmath196 years . red dot short - dashed line : @xmath197 , @xmath196 years . red long - dashed line : @xmath198 , @xmath199 years . red short - dashed line : @xmath200 , @xmath201 years . red solid line : @xmath202 , @xmath203 years . black short dashed - long dashed line : @xmath204 , @xmath205 years . black dot long - dashed line : @xmath206 , @xmath207 years . black dot short - dashed line : @xmath208 , @xmath209 years . black long - dashed line : @xmath210 , @xmath211 years . black short - dashed line : @xmath212 , @xmath213 years . black solid line : @xmath214 , @xmath215 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath193 , @xmath194 years . red dot long - dashed line : @xmath195 , @xmath196 years . red dot short - dashed line : @xmath197 , @xmath196 years . red long - dashed line : @xmath198 , @xmath199 years . red short - dashed line : @xmath200 , @xmath201 years . red solid line : @xmath202 , @xmath203 years . black short dashed - long dashed line : @xmath204 , @xmath205 years . black dot long - dashed line : @xmath206 , @xmath207 years . black dot short - dashed line : @xmath208 , @xmath209 years . black long - dashed line : @xmath210 , @xmath211 years . black short - dashed line : @xmath212 , @xmath213 years . black solid line : @xmath214 , @xmath215 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath193 , @xmath194 years . red dot long - dashed line : @xmath195 , @xmath196 years . red dot short - dashed line : @xmath197 , @xmath196 years . red long - dashed line : @xmath198 , @xmath199 years . red short - dashed line : @xmath200 , @xmath201 years . red solid line : @xmath202 , @xmath203 years . black short dashed - long dashed line : @xmath204 , @xmath205 years . black dot long - dashed line : @xmath206 , @xmath207 years . black dot short - dashed line : @xmath208 , @xmath209 years . black long - dashed line : @xmath210 , @xmath211 years . black short - dashed line : @xmath212 , @xmath213 years . black solid line : @xmath214 , @xmath215 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath193 , @xmath194 years . red dot long - dashed line : @xmath195 , @xmath196 years . red dot short - dashed line : @xmath197 , @xmath196 years . red long - dashed line : @xmath198 , @xmath199 years . red short - dashed line : @xmath200 , @xmath201 years . red solid line : @xmath202 , @xmath203 years . black short dashed - long dashed line : @xmath204 , @xmath205 years . black dot long - dashed line : @xmath206 , @xmath207 years . black dot short - dashed line : @xmath208 , @xmath209 years . black long - dashed line : @xmath210 , @xmath211 years . black short - dashed line : @xmath212 , @xmath213 years . black solid line : @xmath214 , @xmath215 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath193 , @xmath194 years . red dot long - dashed line : @xmath195 , @xmath196 years . red dot short - dashed line : @xmath197 , @xmath196 years . red long - dashed line : @xmath198 , @xmath199 years . red short - dashed line : @xmath200 , @xmath201 years . red solid line : @xmath202 , @xmath203 years . black short dashed - long dashed line : @xmath204 , @xmath205 years . black dot long - dashed line : @xmath206 , @xmath207 years . black dot short - dashed line : @xmath208 , @xmath209 years . black long - dashed line : @xmath210 , @xmath211 years . black short - dashed line : @xmath212 , @xmath213 years . black solid line : @xmath214 , @xmath215 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath193 , @xmath194 years . red dot long - dashed line : @xmath195 , @xmath196 years . red dot short - dashed line : @xmath197 , @xmath196 years . red long - dashed line : @xmath198 , @xmath199 years . red short - dashed line : @xmath200 , @xmath201 years . red solid line : @xmath202 , @xmath203 years . black short dashed - long dashed line : @xmath204 , @xmath205 years . black dot long - dashed line : @xmath206 , @xmath207 years . black dot short - dashed line : @xmath208 , @xmath209 years . black long - dashed line : @xmath210 , @xmath211 years . black short - dashed line : @xmath212 , @xmath213 years . black solid line : @xmath214 , @xmath215 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath216 , @xmath217 years . red dot long - dashed line : @xmath218 , @xmath219 years . red dot short - dashed line : @xmath220 , @xmath221 years . red long - dashed line : @xmath222 , @xmath223 years . red short - dashed line : @xmath224 , @xmath225 years . red solid line : @xmath226 , @xmath227 years . black short dashed - long dashed line : @xmath228 , @xmath229 years . black dot long - dashed line : @xmath230 , @xmath231 years . black dot short - dashed line : @xmath232 , @xmath233 years . black long - dashed line : @xmath234 , @xmath235 years . black short - dashed line : @xmath236 , @xmath237 years . black solid line : @xmath238 , @xmath239 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath216 , @xmath217 years . red dot long - dashed line : @xmath218 , @xmath219 years . red dot short - dashed line : @xmath220 , @xmath221 years . red long - dashed line : @xmath222 , @xmath223 years . red short - dashed line : @xmath224 , @xmath225 years . red solid line : @xmath226 , @xmath227 years . black short dashed - long dashed line : @xmath228 , @xmath229 years . black dot long - dashed line : @xmath230 , @xmath231 years . black dot short - dashed line : @xmath232 , @xmath233 years . black long - dashed line : @xmath234 , @xmath235 years . black short - dashed line : @xmath236 , @xmath237 years . black solid line : @xmath238 , @xmath239 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath216 , @xmath217 years . red dot long - dashed line : @xmath218 , @xmath219 years . red dot short - dashed line : @xmath220 , @xmath221 years . red long - dashed line : @xmath222 , @xmath223 years . red short - dashed line : @xmath224 , @xmath225 years . red solid line : @xmath226 , @xmath227 years . black short dashed - long dashed line : @xmath228 , @xmath229 years . black dot long - dashed line : @xmath230 , @xmath231 years . black dot short - dashed line : @xmath232 , @xmath233 years . black long - dashed line : @xmath234 , @xmath235 years . black short - dashed line : @xmath236 , @xmath237 years . black solid line : @xmath238 , @xmath239 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath216 , @xmath217 years . red dot long - dashed line : @xmath218 , @xmath219 years . red dot short - dashed line : @xmath220 , @xmath221 years . red long - dashed line : @xmath222 , @xmath223 years . red short - dashed line : @xmath224 , @xmath225 years . red solid line : @xmath226 , @xmath227 years . black short dashed - long dashed line : @xmath228 , @xmath229 years . black dot long - dashed line : @xmath230 , @xmath231 years . black dot short - dashed line : @xmath232 , @xmath233 years . black long - dashed line : @xmath234 , @xmath235 years . black short - dashed line : @xmath236 , @xmath237 years . black solid line : @xmath238 , @xmath239 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath216 , @xmath217 years . red dot long - dashed line : @xmath218 , @xmath219 years . red dot short - dashed line : @xmath220 , @xmath221 years . red long - dashed line : @xmath222 , @xmath223 years . red short - dashed line : @xmath224 , @xmath225 years . red solid line : @xmath226 , @xmath227 years . black short dashed - long dashed line : @xmath228 , @xmath229 years . black dot long - dashed line : @xmath230 , @xmath231 years . black dot short - dashed line : @xmath232 , @xmath233 years . black long - dashed line : @xmath234 , @xmath235 years . black short - dashed line : @xmath236 , @xmath237 years . black solid line : @xmath238 , @xmath239 years . , title="fig:",scaledwidth=30.0% ] simulation . panel ( a ) : number density as a function of radius . panel ( b ) : enclosed gas mass as a function of radius . panel ( c ) : temperature as a function of radius . panel ( d ) : @xmath31 fraction as a function of radius . panel ( e ) : ratio of cooling time to sound crossing time as a function of radius . panel ( f ) : ratio of cooling time to dynamical time as a function of radius . all quantities except enclosed gas mass are mass - weighted . line types and weights correspond to different times , as follows . red short dashed - long dashed line : @xmath216 , @xmath217 years . red dot long - dashed line : @xmath218 , @xmath219 years . red dot short - dashed line : @xmath220 , @xmath221 years . red long - dashed line : @xmath222 , @xmath223 years . red short - dashed line : @xmath224 , @xmath225 years . red solid line : @xmath226 , @xmath227 years . black short dashed - long dashed line : @xmath228 , @xmath229 years . black dot long - dashed line : @xmath230 , @xmath231 years . black dot short - dashed line : @xmath232 , @xmath233 years . black long - dashed line : @xmath234 , @xmath235 years . black short - dashed line : @xmath236 , @xmath237 years . black solid line : @xmath238 , @xmath239 years . , title="fig:",scaledwidth=30.0% ] in this paper we have examined aspects of the formation of population iii stars in the presence of a soft ultraviolet background , using an adaptive mesh refinement cosmological structure formation code . given the nature of the tool and current limitations in computing power , some physics were neglected , and possible numerical issues may arise . we discuss this here . the primordial chemistry model used in these calculations ignores the effects of deuterium , lithium , and the various molecules that form between these elements and ordinary hydrogen . deuterium and lithium have been shown to be unimportant in the temperature and density regimes that we have examined in this paper ( galli & palla 1998 , 2002 ; lipovka , nez - lpez , & avila - reese 2005 ) . however , it is possible that they may be relevant in other situations of importance to population iii star formation in particular , regions which have been ionized to very high electron fractions may experience significant cooling from the hd molecule , which due to its permanent dipole moment makes it more than 100 times more effective as a cooling agent than molecular hydrogen ( per molecule ) , and has the potential to cool gas down to approximately the temperature of the cosmic microwave background , which scales with redshift as @xmath240 k ( flower et al . 2000 ; galli & palla 2002 ; lipovka et al . this gives a minimum baryon temperature of approximately @xmath241 kelvin at @xmath110 and could reduce the minimum accretion rate onto a primordial protostar significantly . lithium , while in principle an effective coolant as lih , is safely ignored , since only a tiny fraction of lithium is converted into lih ( mizusawa , omukai , & nishi 2005 ) . a final chemical process that is omitted from our calculation is heating caused by molecular hydrogen formation at high densities ( n @xmath242 @xmath4 ) , the inclusion of which may result in differences in the temperature evolution of the gas in the highest density regimes considered here . this does not significantly affect the conclusions of our paper , as the significant differences we observe between simulations occur at densities much lower than @xmath36 @xmath4 . self - shielding of the photodissociating background by molecular hydrogen in the high - density gas is ignored in these calculations . though this effect could in principle be important , the actual column densities of molecular hydrogen are typically far too small to actually block the soft uv flux . according to ( 2001 ) ( and references therein ) , a column density of @xmath243 @xmath11 is enough for shielding to become important . however , this was derived for a static distribution of h@xmath13 , while the lw band consists of hundreds of individual lines whose width in this case is dominated by doppler broadening . it is useful to note in this case that the average line width is @xmath179 km / s and the rms baryon velocity in our calculations are @xmath244 km / s . in order for self - shielding to be important in the case of a turbulent medium , the column density must be much higher . typical maximum h@xmath13 column densities in our calculations are on the order of @xmath245 @xmath11 , but these occur late in the collapse of the halo core , and in the highest density regions the cooling and h@xmath13 production times are much shorter than the photodissociation time scale , at which point self - shielding becomes unimportant . it is worth noting that there are some regimes where self - shileding can be critical . ( 2007 ) finds that self - shielding can strongly affect the evolution of a collapsing primordial halo which is being illuminated with photodissociating flux by a neighboring star , though the situation is somewhat idealized . additionally , we do not consider the more complicated effects relating to population iii stars which form in halos in the cosmic neighborhoods where previous generations of stars have existed . this allows us to ignore complex radiative , chemical and dynamical effects that would vastly complicate our calculations ( oshea et al . 2005a ; johnson , greif , & bromm 2006 ; yoshida et al . 2006a ; greif et al . 2007 ; abel , wise , & bryan 2007 ; ahn & shapiro 2007 ) . a further effect that is ignored in this paper is @xmath29 photodetachment . this could in principle be a significant effect @xmath29 is the catalyst for molecular hydrogen formation in the dominant 2 formation channel at densities @xmath147 @xmath4 , and it can be detached by photons with energies @xmath246 ev . photons with this energy would be produced in great numbers by the same massive population iii stars that we assume are producing the molecular hydrogen photodissociating background . however , as shown by ( 2001 ) , the rate of photodetachment is orders of magnitude smaller than the rate of @xmath29 formation at densities comparable to that found in the centers of the halos examined in this work , and thus the process of @xmath29 photodetachment can be safely ignored . in this paper , we examine the effects of molecular hydrogen dissociating backgrounds which range in strength from j@xmath247 to @xmath9 ergs@xmath10 @xmath11 hz@xmath10 sr@xmath10 in the lw band ( @xmath248 ev ) . this range is in good agreement with the expected range of photodissociating backgrounds predicted by & abel ( 2005 ) , but it is still finite . examination of our simulations show that the case with the lowest uv background ( j@xmath140 ) is almost identical to the `` control '' ( j@xmath68 ) calculation , justifying our choice of minimum value . our upper value is consistent with wise & abel , and examination of figures [ fig.halovals1 ] and [ fig.radprof_zcoll.1][fig.radprof_zcoll.3 ] suggests that increases in the strength of the uv background ( within reasonable values ) will result in a further delay in halo collapse , but no major change in the mode of star formation observed . this is an effect of the halo properties a small amount of molecular hydrogen will always exist in these halos , allowing the gas at the center to cool and contract quasi - statically to higher densities . once the gas has collapsed to very high densities , rapid h@xmath13 formation via the 3-body process will occur , essentially independent of the strength of the soft uv background , and the gas will be able to cool down to @xmath249 k very quickly . the effects of magnetic fields are completely ignored in the simulations discussed in this work . magnetic fields are discussed in detail in paper i , but , to summarize , a fairly high seed magnetic field is needed to be dynamically significant at the relatively low densities we explore in this work . the possible importance of magnetic fields has been explored in analytic and semi - analytic work ( tan & mckee 2004 ; silk & langer 2006 ; maki & susa 2007 ) . we will examine the possible evolution of magnetic fields within the context of cosmological amr simulations in a later paper ( oshea & turk 2007 , in preparation ) . the simulations presented in this paper are generated assuming a cosmology that is somewhat different than the currently - favored wmap year iii `` best - fit '' model ( spergel et al . most importantly , our value of @xmath250 is 0.9 , which is significantly higher than the wmap value of 0.761 . the general effect of a higher @xmath250 is to cause structure formation to take place earlier , and thus at a given redshift one would expect significantly more halos in our chosen cosmological model than in the wmap year iii model . however , the evolution of any single cosmological halo , such as the one examined in this work , is not particularly affected by this parameter in the sense that we are not examining halo stastical properties . in addition to @xmath250 , the ratio of @xmath251 in our simulations is 0.1337 , while it is @xmath252 in the wmap year iii cosmology . this may result in some small quantitative differences in , e.g. , the redshift of halo collapse , but should not significantly affect our results . we direct the reader to paper i for a detailed discussion of other possible numerical issues , such as the generation of cosmological initial conditions , the assumption that the cold dark matter model is correct , the choice of halo in our simulations , and the size of the simulation volumes used . this paper explores the formation of population iii stars in simulations with a constant soft uv background . our results agree well qualitatively with that of ( 2001 ) ; we both find that a soft uv background can delay the formation of molecular hydogen , and thus the cooling and collapse , of small ( @xmath253 m@xmath15 ) cosmological halos in which population iii stars form . we also find that increasing the soft uv background increases the minimum halo mass required for a halo to collapse ( in a way similar to that of smoothing the dark matter power spectrum at small scales see & norman ( 2006 ) for a discussion ) . machacek et al . derived a mass threshold for collapse as a function of the lw background flux that agrees well with our simulations , though the halo masses in our calculation are significantly higher . this is due to the halo that we examine being a `` typical '' halo rather than at the threshold mass for star formation . presumably , if we performed these calculations using many halos in a range of cosmological realizations , we would find a minimum halo mass that is somewhat lower than the masses seen in our calculations . our work is a significant improvement upon that of machacek et al . in some respects , as our simulations are much more highly resolved and we examine the evolution of a single halo over a much wider range of soft uv background fluxes . one drawback of our work compared to machacek et al . is that we examine the evolution of a single halo , albeit with a broad range of uv backgrounds . our work is similar to the calculations in ( 2003 ) which examine population iii star formation in the presence of a photodissociating ultraviolet background . they find that for values of j@xmath52 above @xmath254 erg s@xmath10 @xmath11 hz@xmath10 sr@xmath10 , hydrogen molecules are rapidly dissociated and gas cooling is inefficient , implying that the collapse of gas in the halo center may be delayed until atomic line cooling dominates , in halos with t@xmath12 k. while we see delays in halo collapse , we do see that a tiny amount of 2 can form even at high fuv background strengths , eventually allowing gas to cool and collapse to high densities without ever reaching temperatures at which atomic line cooling would be effective ( t @xmath255 k ) . yoshida et al . make this statement based on a set of simulations using two different values of j@xmath52 ( @xmath254 and @xmath256 ) , which bracket the `` break '' in properties seen in our results . the simulations used to obtain this result had sph particles with m@xmath257 h@xmath10 m@xmath15 and thus an effective mass resolution comparable to the mass of gas within the quasistatically - collapsing central region of a halo . our mass resolution is more than two orders of magnitude higher in the center of the cosmological halos that we examine . the differences between our result and theirs is likely due to the significant differences in resolution and may in fact be consistent once resolution is taken into account . we see a strong relationship between accretion rates onto the protostellar cloud and the strength of the photodissociating background . calculations with higher ultraviolet background strengths typically have a larger spherically - averaged baryon accretion rate onto the evolving protostellar cloud . this is clearly due to variation in halo central temperature relating to the amount of molecular hydrogen existing in the center of the halo at the epoch of collapse . accretion onto the evolving protostellar cloud is subsonic , and thus regulated by the local sound speed ( which scales as t@xmath258 ) . this implies some relationship between the strength of the photodissociating background and the final stellar mass of the primordial star however , the details of this relationship are unclear , and depends on many factors . for example , & palla ( 2003 ) suggest that an increase in accretion rate above @xmath153 m@xmath15/yr will actually result in a _ decrease _ in the final mass of the star due to radiative feedback from the evolving protostar . however , their results use one - dimenstional , fairly idealized models , and geometrical effects may be important . this is explored in & mckee ( 2004 ) , who use a combination of analytic and semi - analytic models of the evolving system . they suggest that accretion onto the protostar is highly non - spherical , and is in fact mediated by an accretion disk . they argue that a larger accretion rate onto the disk will lead to a larger star , though with some limitations . finally , ( 2001 ) and & yoshii ( 2003 ) study the imf of stars that form in t@xmath12 k metal - free protogalaxies in the presence of an h@xmath13 photodissociating background . they conclude that photodissociation actually decreases the jeans mass of the gas at high densities , and thus reduces the fragmentation mass scale of the clouds and presumably the stellar mass . they use one - dimensional , spherically symmetric simulations , however . we can explore this issue in more detail in later calculations , when appropriate models for shielding , cooling via @xmath29 , and more advanced chemistry have been implemented into enzo . this will be examined in a later paper . it is worth noting that the range of accretion rates observed in this situation are within the range of rates seen in the @xmath17cdm simulations discussed in paper i. this implies that the mass ranges of the resulting stars will not be significantly different than in the `` standard '' population iii star formation scenario . we see no fragmentation of the high - density baryon core , up to a baryon number density of @xmath259 @xmath4 , for the entire range of simulations explored in this paper . this is due to a combination of effects , but predominantly the poor cooling properties of molecular hydrogen . the gas is relatively hot ( @xmath260 kelvin ) and thus has a high sound speed , which helps to damp out perturbations in the halo center which would otherwise result in multiple fragments . this effect is exacerbated in simulations with high ultraviolet background strengths , as the temperatures ( and thus the sound speeds ) are generally higher . this result implies that there is no fundamental change in the mode of population iii star formation as halos grow , and that more massive `` proto - galactic '' halos , with t@xmath261 k , will continue to form a single massive star per halo . these halos will have a much larger binding energy than the smaller , m @xmath14 m@xmath15 halos which have traditionally been examined by numerical simulations , and implies that multiple generations of star formation may be able to take place in a single halo . this leads to the possibility of `` self - enrichment , '' where a single population iii star enriches the high - density center of the halo to metallicities high enough to change the cooling properties of the gas . this could cause a strong change in the imf ( santoro & shull 2006 ; smith & sigurdsson 2007 ) , and these objects could make a significant contribution to the reionization history of the universe ( mackey , bromm , & hernquist 2003 ; greif & bromm 2006 ) . we will examine whether the high - density gas in the centers of these more massive halos fragments in a later work . the values of accretion rates onto the evolving protostellar cloud which are observed with increasing photodissociating background flux ( which are always comparable to those observed in paper i ) , in addition to the apparent lack of fragmentation of gas in the halo center up to densities of n@xmath262 @xmath4 , implies that there is no major change in the mass range of population iii stars as halo masses increase to t@xmath5 k. this is in mild disagreement with previous semianalytic results by & haiman ( 2002 ) and ( 2003 ) . this is apparently because previous authors assumed that 2 cooling in `` second generation '' ( t@xmath12 k ) objects is unimportant , while we observe that a small 2 fraction , which forms in equilibrium with the uv background , can cool gas effectively to temperatures of @xmath263 k. we do agree with these authors that the majority of primordial star formation will take place in objects with t@xmath5 k , the so - called `` second generation '' objects ( macintyre , santoro , & thomas 2006 ) our work appears to contradict the results of & bromm ( 2006 ) , who use analytic and semianalytic techniques to follow the collapse of gas within halos of t@xmath264 . they find that , in their model , gas cools to @xmath265 k via atomic line cooling , and then contracts nearly isothermally to high densities , allowing a molecular hydrogen fraction of f@xmath266 to build up independent of local density and temperature . the gas in the center of the halo then cools and allows the gas to fragment on scales of @xmath158 m@xmath15 . we do not see this mode of star formation , possibly because we never generate the significantly ionized halos with t@xmath267 k upon which their scenario depends . rather , we simulate halos that in the cases of j@xmath0 are built up to t@xmath5 k via mergers , and the gas is never ionized to a significant degree . one possible reason for the observed differences is that we ignore the formation of deuterium hydride and its effects on the cooling properties of primordial gas . it is possible that , in some contexts , the inclusion of hd may result in enhanced fragmentation . this will be examined in future studies . the scenario described by oh & haiman may still in fact be able to occur , and deserves further detailed numerical study to determine the fate of the collapsing gas and the implications of the ( presumably ) extremely high accretion rates onto the forming protostellar clouds . our results are similar to those shown by & abel ( 2007 ) , who perform similar simulations examining the collapse of gas in cosmological halos in the presence of a photodissociating background . they also find that , regardless of the strength of the uv background , the collapse of gas in the centers of the cosmological halos in question still occurs via h@xmath13 cooling . this holds even for quite extreme examples , such as when the original electron fraction is set to unphysically low levels ( which should suppress h@xmath13 formation , at least temporarily ) . in this paper we have performed a suite of high dynamical range ( l@xmath268 ) 3d adaptive mesh cosmological simulations of the formation of population iii stars in a @xmath17cdm universe in the presence of a molecular hydrogen photodissocating ( `` lyman - werner '' ) ultraviolet background . the purpose of these calculations is to understand the effect that the soft ultraviolet background has on the evolution of the gas in collapsing cosmological halos and to determine possible effects on the forming protostellar cloud . these simulations are all of a single cosmological realization , but with varied ultraviolet background strengths . our principal results are as follows : \1 . our calculations show that , as the flux of ultraviolet radiation in the lyman - werner band is increased , population iii star formation in a given cosmological halo is delayed to later times and , as a result , an increase in the virial mass of this halo at the onset of baryon collapse . this is in good agreement with previous work by ( 2001 ) and ( 2003 ) . we find that , contrary to previous work , the formation of primordial stars is never completely suppressed , regardless of the strength of the uv background . a small amount of molecular hydrogen always exists in cosmological halos , and allows cooling and halo collapse in gas which is bathed in a strong photodissociating background but has not reached @xmath86 k. the previously - suggested mode of star formation in `` second generation '' halos , where the collapse of gas to high density is completely suppressed until cooling can take place via atomic hydrogen line transitions , is never observed in our calculations , which have photodissociating background strengths up to j@xmath1 . \3 . though the molecular hydrogen fraction in the centers of halos which are bathed in strong ( j@xmath0 ) lyman - werner radiation is very small at early times ( f@xmath269 ) , the gas at the center of the halo can still collapse due to the efficient cooling of molecular hydrogen at @xmath270 k , and to the extended time these halos require to collapse compared to halos in the presence of much smaller uv backgrounds . we observe that the estimated accretion rate onto the forming protostellar cloud varies strongly as a function of j@xmath52 , with simulations that have a stronger ultraviolet background having higher accretion rates . this is a function of the suppression of molecular hydrogen formation ( and thus suppression of effective cooling ) by the photodissociating background and of the higher virial temperatures of these halos when the gas finally collapses . only a single clump is formed at the center of each collapsing cosmological halo , regardless of the strength of the photodissociating background , up to a baryon number density of @xmath3 @xmath4 . this implies that , as in the more commonly - studied mode of star formation , we will find only a single star per halo even in objects which are massive enough that t@xmath12 k.

one of the longstanding questions of elementary particle physics is whether nature makes use of the higgs mechanism of spontaneous symmetry breaking to endow the particles with their masses . the higgs boson , @xmath5 , is the missing link sought to verify this theoretical concept in the standard model ( sm ) . its mass , @xmath6 , is essentially a free parameter of the sm . the failure of experiments at the cern large electron - positron collider ( lep 1 ) and the slac linear collider ( slc ) to observe the decay @xmath7 has ruled out the mass range @xmath8 gev at the 95% confidence level @xcite . the discovery potential of lep 2 , with centre - of - mass ( c.m . ) energy @xmath9 gev and luminosity @xmath10 pb@xmath11 per experiment , will reach up to @xmath12 gev @xcite . we note in passing that @xmath6 lower bounds may also be derived theoretically by demanding that the higgs vacuum be ( meta)stable @xcite . several theoretical arguments bound @xmath6 from above . the requirement that partial - wave unitarity in intermediate - boson scattering at high energies be satisfied at tree level establishes an upper bound on @xmath6 at about @xmath13 tev @xcite . this unitarity bound is significantly lowered by including one- @xcite and two - loop @xcite corrections . however , the improved bound depends on the energy scale , @xmath14 , up to which the sm is assumed to remain valid . the same is true for the triviality bound , which is derived perturbatively @xcite by requiring that the running higgs self - coupling , @xmath15 , stays finite for renormalization scales @xmath16 . lattice computations , which take all orders of the perturbative expansion into account , yield a triviality bound of @xmath17 gev @xcite . the precise value of this bound depends on the specific form of the lattice action and the regularization method used . an alternative way of constraining @xmath6 from above is to require that the higgs sector be weakly interacting , so that perturbation theory is meaningful @xcite . the resulting bounds depend somewhat on the considered process and the precise definition of perturbation - theory breakdown , but they are independent of assumptions concerning the scale @xmath14 beyond which new physics operates . at one and two loops , the leading high-@xmath6 corrections to physical observables related to higgs - boson production or decay are of @xmath18 and @xmath3 , respectively . the one - loop corrections to the partial widths of the higgs - boson decays to pairs of fermions @xcite and intermediate bosons @xcite are relatively modest , below 15% at @xmath19 tev if the on - mass - shell renormalization scheme is employed . the situation is similar for higgs - boson production @xcite . for @xmath6 increasing , the @xmath3 corrections will eventually exceed the @xmath18 ones in size . in the case of the fermionic higgs - boson decays , this happens at @xmath20 tev in the on - shell scheme @xcite . obviously , the perturbative expansion in @xmath21 ceases to usefully converge for larger values of @xmath6 . the study of high - energy weak - boson scattering at two loops ( with neglect of trilinear couplings ) leads to a perturbative upper bound on @xmath6 which depends on the considered c.m . energy . for @xmath22 in the 1-tev range , the @xmath6 upper bound comes out as low as 450 gev @xcite . however , this result may be relaxed using resummation techniques @xcite . since a higgs boson with @xmath23 decays almost exclusively to @xmath24 and @xmath25 pairs , it is desirable , both from the theoretical and phenomenological points of view , to find the two - loop @xmath3 corrections to the partial widths of these decays as well . this is the purpose of the present paper . in contrast to the fermionic case @xcite , this involves the calculation of massive three - point functions at two loops . due to the presence of a large external momentum , heavy - mass expansion techniques are not applicable here . as in a previous analysis @xcite , we take advantage of the goldstone - boson equivalence theorem @xcite and neglect the gauge and yukawa couplings as well as the intermediate - boson masses . in this approximation , the radiative corrections to the @xmath26 and @xmath27 decay widths coincide . in the evaluation of the relevant two - loop feynman diagrams , both infrared ( ir ) and ultraviolet ( uv ) divergences occur . we calculate all divergent diagrams analytically , using dimensional regularization . this guarantees that the cancellation of the ir and uv divergences is manifest . the remaining diagrams are individually devoid of singularities . they are computed numerically in four space - time dimensions after formulating the integrals in parallel and orthogonal space . this has the advantage that potential numerical instabilities in connection with the conventional feynman parameterizations are avoided . we list our results for the individual diagrams , so that they are available to other authors for comparisons and further calculations . we anticipate that we find good agreement with a recent , pioneering work by ghinculov @xcite , which fully resorts to numerical methods . in ref . @xcite , the ir divergences are regularized by introducing a small goldstone - boson mass . the final result is obtained by numerically taking the massless limit . another conceptual difficulty in ref . @xcite is related to the treatment of ( endpoint ) singularities in the feynman - parameter integrals . these are controlled by choosing suitable integration paths in the complex plane with the aid of spline functions . the values of the individual diagrams are not specified in ref . we believe that , in view of the enormous complexity of the problem at hand , it is indispensable to have at least two independent analyses that agree with each other . this paper is organized as follows . in sec . 2 , we catalogue the contributing feynman - graph topologies and explain our renormalization framework . in sec . 3 , we describe our analytical and numerical techniques . in sec . 4 , we present our results for the individual two - loop vertex diagrams and assemble the renormalized @xmath28 ( @xmath29 ) transition amplitude . in sec . 5 , we explore the phenomenological implications of our result and draw the conclusions . the relevant one - loop three - point amplitudes to @xmath30 , where @xmath31 is the dimensionality of space time , are listed in the appendix . in this section , we set the stage for our calculation of the @xmath28 ( @xmath29 ) decay widths to @xmath3 . as mentioned in the introduction , this task may be greatly simplified in the limit of interest , @xmath23 , through the use of the goldstone - boson equivalence theorem @xcite . this theorem states that the leading high-@xmath6 electroweak contribution to a feynman diagram may be calculated by replacing the intermediate bosons @xmath32 , @xmath2 with the would - be goldstone bosons @xmath33 , @xmath34 of the symmetry - breaking sector of the theory . in this limit , the gauge and yukawa couplings may be neglected against the higgs self - coupling . by the same token , the goldstone bosons may be taken to be massless , and the fermion loops may be omitted . adopting the conventions of ref . @xcite , we may write the relevant lagrangian for the symmetry - breaking sector of the sm in terms of bare quantities , marked by the subscript 0 , as @xmath35 where @xmath36 is the higgs quartic coupling , @xmath37 is the higgs vacuum expectation value , @xmath5 is the real scalar higgs field , and the real scalar so(3 ) triplet , @xmath38 , is related to the goldstone bosons , @xmath33 and @xmath34 , by @xmath39 and @xmath40 , respectively . the tadpole counterterm , which cancels all tadpole contributions of @xmath41 order by order , has been omitted in writing eq . ( [ eqsbs ] ) . consequently , all diagrams which include tadpole contributions need to be dropped in calculations @xcite . we work in the on - mass - shell renormalization scheme as formulated in refs . the residual so(3 ) symmetry present in eq . ( [ eqsbs ] ) is preserved in the renormalization procedure , so that the @xmath33 and @xmath34 bosons share one mass counterterm and one wave - function renormalization constant . specifically , the renormalization conditions are as follows : ( i ) the mass counterterms , @xmath42 and @xmath43 , are determined so that the renormalized masses , @xmath44 and @xmath45 , coincide with the physical ( pole ) masses , _ i.e. _ , the pole positions of the radiatively corrected propagators . in the case of the goldstone bosons , one has @xmath46 . ( ii ) the wave - function renormalization constants , @xmath47 and @xmath48 , which enter the relations @xmath49 and @xmath50 between the bare and renormalized fields are adjusted so that the on - shell propagators have unit residues . ( iii ) the physical higgs quartic coupling and vacuum expectation value are fixed by requiring that @xmath51 and @xmath52 , respectively , where @xmath53 is fermi s constant . this leads to the following relations @xcite : @xmath54 where @xmath55 and @xmath56 are the self - energy functions of the bare goldstone and higgs fields , respectively , calculated from lagrangian ( [ eqsbs ] ) . two - loop expressions for @xmath42 , @xmath43 , @xmath47 , @xmath48 , and @xmath57 in terms of @xmath6 and @xmath36 may be found in refs . @xcite . the feynman diagrams that contribute to the @xmath28 transition amplitude through @xmath3 are depicted in fig . [ fig1 ] . dashed and solid lines represent goldstone and higgs bosons , respectively . adjacent propagators with identical four - momenta , which occur in connection with mass - counterterm insertions , are separated by solid circles . the calligraphic labels stand for the values of the diagrams and are used in our formulas ; the nomenclature extents the one introduced in ref . the underlying conventions are explained in sec . 3 and the appendix . there are 6 one - loop diagrams ( @xmath58 , @xmath59 , and @xmath60 ) , 18 reducible two - loop diagrams ( @xmath61 , @xmath62 , @xmath63 , and @xmath64 ) , and 52 irreducible two - loop diagrams ( @xmath65 , @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72 , and @xmath73 ) . in addition , there are 8 one - loop diagrams with one mass - counterterm insertion ( @xmath74 and @xmath75 , where @xmath76 ) . further reducible two - loop diagrams emerge as products of one - loop vertex and wave - function - renormalization diagrams ; these are not counted here . the one - loop diagrams need to be evaluated through @xmath30 . the corresponding results for the tadpole ( @xmath58 ) and two - point topologies ( @xmath59 and @xmath77 ) may be found in ref . @xcite ; those for the three - point topologies ( @xmath60 and @xmath78 ) are listed in the appendix . the irreducible two - loop two - point topologies ( @xmath65 , @xmath66 , and @xmath67 ) may be taken from ref . @xcite . here , we regard two external legs that merge into the same vertex as one point ; solid circles are not counted as points . for example , we view both @xmath77 and @xmath65 as two - point topologies . the main technical difficulty of the present paper is to tackle the irreducible two - loop three - point topologies ( @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72 , and @xmath73 ) . topologies @xmath68 , @xmath69 , and @xmath70 are ultraviolet divergent . topologies @xmath79 and @xmath80 are also plagued by infrared singularities . topologies @xmath71 , @xmath72 , and @xmath73 are uv and ir finite . after describing our computational techniques in sec . 3 , we itemize the results for topologies @xmath81 through @xmath82 in sec . 4 . calculating the combinatorial factors with the help of the software package diagrammar @xcite , we find the following master formula for the two - loop @xmath28 transition amplitude in terms of @xmath57 , @xmath83 , and @xmath6 : @xmath84\delta m_w^2 \nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5mm}}&\left . { } + \left[i\lambda_0(8{\cal t}_2 + 12{\cal t}_3 ) + i(\lambda_0 v_0)^2(8{\cal d}_{2a}+48{\cal d}_{3})\right]\delta m_h^2 \right\}.\end{aligned}\ ] ] here , the tree - level term has been factored out together with the wave - function renormalization constants of the external legs ; in our two - loop approximation , these quantities need to be evaluated through @xmath85 . the second and third terms within the curly bracket arise at one loop ; their prefactors of @xmath57 and @xmath83 need only to be expanded through @xmath86 . the residual terms are of genuine two - loop order ; here , @xmath57 and @xmath83 may immediately be replaced with @xmath36 and @xmath37 , and it is sufficient to insert the one - loop expressions for @xmath42 and @xmath43 . notice that we have already performed mass renormalization in eq . ( [ eqmaster ] ) , so that @xmath6 , which appears inside the expressions for the various diagrams , is the only mass parameter left . as usual , we calculate the feynman diagrams in @xmath31 space - time dimensions and introduce an unphysical t hooft mass scale , @xmath87 , to keep the higgs quartic coupling dimensionless . after coupling renormalization , ( [ eqmaster ] ) will be finite and @xmath87 independent in the physical limit @xmath88 . to fix the notation , we write down the generic one - loop two - point integral ( with two scalar propagators ) : @xmath89}.\ ] ] the one - loop two - point diagrams in fig . [ fig1 ] are then defined as @xmath90 analytic results for these diagrams to @xmath30 may be found in ref . note that , in contrast to ref . @xcite , we exclude the factors of @xmath91 connected with the scalar propagators in the definitions of the diagrams . for the one - loop tadpole diagram , we have @xmath92 . the one - loop three - point diagrams , @xmath60 and @xmath78 , are defined and evaluated to @xmath30 in the appendix . the two - loop diagrams are defined in analogy to eqs . ( [ eqtwo ] ) and ( [ eqthree ] ) . we first consider the ir - finite two - loop diagrams in fig . [ fig1 ] . in the limit @xmath88 , any member , @xmath93 , of this class may be cast into the generic form @xmath94,\ ] ] where @xmath95 is the euler - mascheroni constant , @xmath96 is the number of propagators in @xmath97 , and @xmath98 , @xmath99 , and @xmath100 are finite , complex numbers . the special form of the prefactor is to suppress the appearance of the familiar term @xmath101 in @xmath99 and @xmath100 . ( by contrast , the @xmath95 piece of the prefactor was expanded in ref . coefficients @xmath98 , @xmath99 , and @xmath100 for diagrams @xmath65 , @xmath66 , and @xmath67 are available in analytic form @xcite . diagrams @xmath102 and @xmath103 are ir divergent and will be discussed further below . the irreducible two - loop three - point diagrams with four propagators , @xmath68 , exhibit uv divergences in the form of quadratic and linear poles in @xmath104 , but they are ir safe . they may be evaluated analytically from appropriate feynman - parameter representations , proceeding along the lines of ref . @xcite . in the case of diagram @xmath81 , where all internal lines are massless , we find agreement with ref . @xcite . diagrams @xmath69 and @xmath70 are more involved , and it is advantageous to apply dispersion relations in a way similar to ref . let us first concentrate on diagrams @xmath69 . we observe that they contain a one - loop two - point subdiagram consisting of a bubble and an adjacent propagator . calling the masses inside this bubble @xmath105 and @xmath106 , the mass of the adjacent leg @xmath107 , and the common four - momentum @xmath108 , we may rewrite this subdiagram as @xmath109 where @xmath110[s-(m_1-m_2)^2]$ ] is the klln function . ( [ eqdr ] ) is inserted in the expression for the one - loop seed diagram , the first term turns into a product of the type @xmath62 , which contains all divergences . in the second term , we may interchange the dispersion and loop integrations and are left with a finite dispersion integral , which may be solved analytically . diagrams @xmath70 may be treated in a similar fashion . they contain a one - loop two - point subdiagram consisting of a bubble and two identical adjacent propagators . a useful representation of this subdiagram emerges by differentiating eq . ( [ eqdr ] ) with respect to @xmath111 . the divergences of diagrams @xmath70 are contained in products of the form @xmath112 . in particular , this method allows us to separate the ir divergences of diagrams @xmath79 and @xmath80 . by the way , this method may also be applied to the ir - divergent diagrams @xmath102 and @xmath103 . let us now turn to the cancellation of the ir divergences in eq . ( [ eqmaster ] ) . it is possible to extract the ir divergences along with the uv ones as poles in @xmath104 using dimensional regularization . in fact , this avenue was taken in ref . @xcite . here , we adopt an alternative strategy . we identify a basic set of ir - divergent one - loop diagrams and show that their prefactors vanish . in this way , we do not need to solve any ir - divergent loop integral . it is easy to see that the diagrams which emerge by duplicating any of the goldstone - boson propagators in diagrams @xmath113 , @xmath114 , @xmath115 , and @xmath116 are ir divergent . the resulting diagrams are @xmath117 , @xmath118 , @xmath119 , and @xmath120 , respectively . as for the irreducible two - loop diagrams , @xmath102 , @xmath103 , @xmath79 , and @xmath80 are ir divergent . hence there are twelve ir - divergent terms in eq . ( [ eqmaster ] ) . by means of the dispersion - relation method described above , we find that @xmath121 , @xmath122 , @xmath123 , and @xmath124 are finite . exploiting the one - loop identity @xcite @xmath125 it follows on that the twelve ir - divergent terms in eq . ( [ eqmaster ] ) may be grouped in four finite sets . specifically , we have @xmath126,\nonumber\\ ( \lambda{\cal q}-i\delta m_w^2){\cal t}_1 + 4(\lambda v)^2{\cal e}_2&{\hspace*{-2.5mm}}={\hspace*{-2.5 mm } } & \lambda\left[-2+o(\epsilon)\right],\nonumber\\ ( \lambda{\cal q}-i\delta m_w^2){\cal d}_1 + 4(\lambda v)^2{\cal h}_3&{\hspace*{-2.5mm}}={\hspace*{-2.5 mm } } & \frac{1}{v^2}\left[\frac{3}{8}\zeta(2)-\frac{1}{2 } + i\frac{\pi}{2}\ln2+o(\epsilon)\right],\nonumber\\ ( \lambda{\cal q}-i\delta m_w^2){\cal d}_{2b}+4(\lambda v)^2{\cal h}_4&{\hspace*{-2.5mm}}={\hspace*{-2.5 mm } } & \frac{1}{v^2}\left[-\frac{\zeta(2)}{12 } -\frac{\sqrt3}{2}{\mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{\pi}{3}\right)+\frac{\pi}{2\sqrt3 } + \frac{1}{2}+o(\epsilon)\right],\end{aligned}\ ] ] where @xmath127 is riemann s zeta function , with values @xmath128 and @xmath129 , and @xmath130 is clausen s integral , with value @xmath131 . in the remainder of this section , we describe the evaluation of the finite diagrams @xmath132 though @xmath82 . this is achieved with the help of the methods developed in refs . these methods suggest a separation of the integration space in an orthogonal and a parallel space . the parallel space is defined to be the linear span of the exterior momenta involved in the process under consideration . for a three - point function , this is a two - dimensional space . each loop momentum is a sum of two vectors , one being its projection onto the two - dimensional plane spanned by two non - degenerate exterior momenta , the other one being orthogonal to this plane . this splitting does not break lorentz invariance @xcite . in the following , we define orthogonal- and parallel - space variables for our purposes . we assume two independent exterior momenta , @xmath133 and @xmath134 , and two loop momenta , @xmath135 and @xmath136 . without loss of generality , we assume @xmath133 and @xmath134 to be timelike . then , @xmath137 and @xmath138 are orthogonal unit vectors , with @xmath139 and @xmath140 . we further define @xmath141 we then have @xmath142 futhermore , we call the angle between the two - loop momenta in orthogonal space @xmath143 in its most straightforward application , this approach delivers a two - fold integral representation for the two - loop two - point function @xcite . typically , the remaining integrations are with respect to the parallel - space component of each of the two loop momenta . for two - loop three - point functions @xcite , and for four - point functions as well @xcite , this approach opens a route to three - fold integral representations , regardless of the internal topology of the function . in ref . @xcite , this was improved to give a two - fold integral representation for the planar three - point case . remarkably , the so - obtained integral representations are solely defined on finite domains of integration @xcite . these results are sufficient to treat topologies @xmath72 and , with slight modifications , topologies @xmath71 as well . a similar result for the three - point non - planar case was obtained recently @xcite , and was used here for the first time to handle topologies @xmath73 . we now wish to describe some features of the method . following separation into parallel - space and orthogonal - space variables , the angular integration in the orthogonal space is complicated due to the presence of propagators involving both loop momenta , which results in a non - trivial integration over the variable @xmath34 . in fact , the @xmath34 integration delivers a result of the form @xmath144 , where @xmath145 is a function having a single branch - cut and @xmath146 is such a propagator involving both loop momenta . in @xmath145 , this propagator is evaluated at nullified orthogonal loop momenta . typically , in four - dimensional field theory , @xmath145 turns out to be a logarithm in the two - loop two - point case , and a square root in the two - loop three - point case . by the very definition of a scalar field propagator , @xmath146 is a quadratic form in parallel - space variables . parallel - space integrations over the real line are most conveniently performed with the help of the residue theorem . to this end , we wish to close the contour in either the upper or the lower halfplane . this is obscured by the cut of @xmath144 , which is present in both halfplanes . the problem is readily solved by using translations @xmath147 and @xmath148 . using the @xmath149 signature of space - time , we see that these translations render the propagators linear in the variables @xmath150 and @xmath151 . accordingly , the cut defined by the equation @xmath152 with @xmath153 , is now in either the lower or upper complex @xmath150 ( resp . @xmath151 ) plane . this allows us to close the contour in the opposite halfplane . the precise location of the cut becomes a function of the other integration variables . this results in constraint equations for the contributing residues , and thus in constraints on the domains of integration for the remaining variables . these domains turn out to be restricted to triangles in the @xmath154 plane in the case of two - loop three - point functions @xcite . at this stage , one is left with a four - fold integral representation . there are two integrations for the remaining parallel - space variables @xmath155 and @xmath156 , and the moduli @xmath157 and @xmath158 still have to be integrated over the positive real axis . these last two integrations can be carried out next . these @xmath159 and @xmath160 integrals have the form @xcite @xmath161 with the cut of the square root chosen to be along the positive real axis . the coefficients are functions of the parallel space variables @xmath155 and @xmath156 . with due care to the analytic structure of the integrand , one achieves the first integration by using either landen @xcite or euler transformations @xcite . the results allow one to interpret the @xmath160 integration as an integral representation of dilogarithms and associated functions , and thus one obtains a two - fold integral representation as the final result . we once more stress that the final two integrations only cover finite triangular domains . they are suitable for numerical integrations using gaussian quadrature . this programme was first successfully applied to two - loop three - point functions in ref . @xcite , and then extended to the crossed topology in ref . @xcite . for the crossed topology , the unavoidable presence of two propagators involving both loop momenta results in a more complicated cut structure . there are many more cases to be considered , but the calculation still follows the scheme outlined above . we now present our results for diagrams @xmath81 through @xmath82 in the form of eq . ( [ eqtem ] ) . the number of propagators is @xmath162 for @xmath68 , @xmath163 for @xmath69 and @xmath71 , and @xmath164 for @xmath70 , @xmath72 , and @xmath73 . coefficients @xmath98 , @xmath99 , and @xmath100 are listed in tables [ tab1 ] and [ tab2 ] . the formulas for coefficients @xmath100 of @xmath165 , @xmath166 , and @xmath167 are somewhat lengthy , and we have introduced the following constants : @xmath168\nonumber\\ & { \hspace*{-2.5mm}}\approx{\hspace*{-2.5mm}}&-0.647\,466\,172,\nonumber\\ { \rm g}_5&{\hspace*{-2.5mm}}={\hspace*{-2.5mm}}&-\int_0^{1/4}dx\,\frac{\sqrt{1 - 4x}}{x(1-x)}\,\ln(1+x ) \nonumber\\ & { \hspace*{-2.5mm}}={\hspace*{-2.5mm}}&\sqrt3\left[{\mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{2}{3}\pi + \arctan\sqrt{15}\right ) + { \mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{2}{3}\pi - \arctan\sqrt{15}\right)\right ] -\frac{4}{\sqrt3}{\mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{\pi}{3}\right)\nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5mm}}&-2\ln^2\frac{\sqrt5 - 1}{2 } + \frac{\pi}{\sqrt3}\ln2 \nonumber\\ & { \hspace*{-2.5mm}}\approx{\hspace*{-2.5mm}}&-0.177\,283\,264,\nonumber\\ { \rm h}_5&{\hspace*{-2.5mm}}={\hspace*{-2.5mm}}&\int_0^{1/4}dx\,\frac{\sqrt{1 - 4x}}{(1-x)^2 } \left[{\mathop{{\mbox{li}}_2}\nolimits}(-x)+\ln x\ln(1+x)\right]\nonumber\\ & { \hspace*{-2.5mm}}\approx{\hspace*{-2.5mm}}&-0.065\,325\,931,\nonumber\\ { \rm h}_5&{\hspace*{-2.5mm}}={\hspace*{-2.5mm}}&-\int_0^{1/4}dx\,\frac{\sqrt{1 - 4x}}{(1-x)^2}\ln(1+x)\nonumber\\ & { \hspace*{-2.5mm}}={\hspace*{-2.5mm}}&-\frac{2}{\sqrt3 } \left[{\mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{2}{3}\pi + \arctan\sqrt{15}\right ) + { \mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{2}{3}\pi - \arctan\sqrt{15}\right)\right ] + \frac{8}{3\sqrt3}{\mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{\pi}{3}\right)\nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5 mm } } & -\frac{\sqrt5}{2}\ln\frac{3-\sqrt5}{2 } -\frac{\pi}{\sqrt3}\left(\frac{2}{3}\ln2+\frac{1}{2}\right ) \nonumber\\ & { \hspace*{-2.5mm}}\approx{\hspace*{-2.5mm}}&-0.021\,441\,581.\end{aligned}\ ] ] the analytic results for @xmath169 and @xmath170 are rather messy , and we refrain from writing them down here . in sec . 3 , we discussed the elimination of the ir - divergent pieces in eq . ( [ eqmaster ] ) . substituting in the remainder of eq . ( [ eqmaster ] ) the expressions for @xmath43 , @xmath47 , @xmath48 , @xmath57 , and @xmath83 in terms of @xmath6 and @xmath36 @xcite as well as the results for the various one- and two - loop diagrams , we may set @xmath171 and so obtain the physical result @xmath172\right . \nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5 mm } } & + \hat\lambda^2\left [ - 2\zeta(3 ) + \zeta(2)\left(- 45\pi\sqrt3 + 144\ln2 + \frac{1173}{2}\right ) + 4\left(33\sqrt3 + 4\pi\right){\mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{\pi}{3}\right ) \right . \nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5 mm } } & \phantom{+\hat\lambda^2\left[\right . } - 181\pi\sqrt3 - \frac{749}{8 } + 189{\rm l}_5 - 2{\rm f}_2 - 12{\rm g}_5 - 18{\rm h}_5 \nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5 mm } } & \phantom{+\hat\lambda^2\left[\right . } + 12{\rm i}_1 + 12{\rm i}_2 + 8{\rm i}_3 + 72{\rm i}_4 + 40{\rm i}_5 + 24{\rm i}_6 + 8{\rm i}_7 \nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5 mm } } & \phantom{+\hat\lambda^2\left[\right . } + 108{\rm j}_1 + 12{\rm j}_2 + 12{\rm j}_3 + 4{\rm j}_4 + 72{\rm j}_5 + 24{\rm j}_6 + 24{\rm j}_7 + 8{\rm j}_8 + 36{\rm k}_1 + 4{\rm k}_2 \nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5 mm } } & \phantom{+\hat\lambda^2\left[\right . } + i\left ( \zeta(2)\left ( - 108\sqrt3\ln2 + 174\sqrt3 - 53\pi\right ) + \pi\left(57\ln2 - \frac{41}{2 } - 2{\rm f}_2 - 12{\rm g}_5 - 18{\rm h}_5\right ) \right . \nonumber\\ & { \hspace*{-2.5mm}}{\hspace*{-2.5 mm } } & \phantom{+\hat\lambda^2\left[\right . } + 40{\rm i}_5 + 8{\rm i}_7 + 12{\rm j}_2 + 12{\rm j}_3 + 4{\rm j}_4 + 24{\rm j}_7 + 8{\rm j}_8 + 4{\rm k}_2 \left.\left.\left.\vphantom{\frac{1}{1}}\right)\right]\right\},\end{aligned}\ ] ] where @xmath173 . here , @xmath174 stems from the all - massive lemon diagram , @xmath175+o(\epsilon)$ ] , and is given in terms of a two - dimensional integral in eq . ( a86 ) of ref . @xcite , where it is called @xmath176 . all other constants are listed in tables [ tab1 ] and [ tab2 ] together with eq . ( [ eqcon ] ) . inserting in eq . ( [ eqfin ] ) the numerical values of these constants , we find @xmath177\right\}.\end{aligned}\ ] ] this agrees reasonably well with eq . ( 13 ) of ref . @xcite , where the two - loop coefficient is found to be @xmath178 $ ] . taking the absolute squared of eq . ( [ eqnum ] ) , we obtain the correction factor , @xmath179 , in the relation @xmath180 between the radiatively corrected and born values of the @xmath28 decay width as @xmath181 for comparison , we also list the analogous correction factor , @xmath182 @xmath183 , for the higgs - boson decay to a pair of fermions @xcite : @xmath184 in the derivation of eq . ( [ eqkf ] ) , yukawa couplings have been set to zero . therefore eq . ( [ eqkf ] ) should be a useful approximation to the full electroweak two - loop result as long as @xmath185 . we are now in a position to explore the phenomenological consequences of our result . in fig . [ fig2 ] , we display the @xmath28 correction factor @xmath179 of eq . ( [ eqkv ] ) to @xmath18 and @xmath3 as a function of @xmath6 . for comparison , also the corresponding approximations for the @xmath186 correction factor @xmath187 of eq . ( [ eqkf ] ) are shown . in both cases , the @xmath18 terms enhance the born results . this effect is slightly more pronounced for @xmath179 . in the case of @xmath179 , the @xmath3 term acts in the same direction as the @xmath18 term . for @xmath6 increasing , it rapidly gains importance relative to the @xmath18 term ; at @xmath188 gev , both terms have the same numerical value , 12.6% . in the case of @xmath187 , the @xmath3 term reduces and , for @xmath189 tev , overcompensates the effect of the @xmath18 term . obviously , the perturbation expansions of @xmath179 and @xmath187 in @xmath21 cease to usefully converge for @xmath6 values in excess of about 1 tev . consequently , this value may be considered as an upper bound on the higgs - boson pole mass @xmath6 , if the sm is to be weakly interacting @xcite . we emphasize that this observation is independent of speculations concerning the energy scale up to which the sm is supposed to be valid . the goldstone - boson equivalence theorem @xcite allowed us to extract power corrections in @xmath21 . in the high-@xmath6 limit , these are clearly dominant over corrections in @xmath190 and @xmath191 arising from the gauge and yukawa sectors of the sm , respectively . for the @xmath26 and @xmath27 decays , this was verified for the @xmath18 terms in refs . @xcite , where the respective partial widths were completely calculated to one loop in the sm . there is no obvious reason why the situation should be different at the electroweak two - loop order . qcd corrections enter the stage only at the two - loop level , in connection with quark loops . the dominant two- and three - loop qcd corrections to the @xmath192 and @xmath193 couplings , of @xmath194 @xcite and @xmath195 @xcite , respectively , are well under control and suppressed against the @xmath21 power corrections for @xmath6 large . the dominance of the @xmath21 power corrections in the case of the @xmath196 yukawa couplings was discussed in great detail in refs . @xcite . in this context , we should mention that the @xmath18 two - loop corrections to the partial width of the loop - induced @xmath197 decay have been computed @xcite with the aid of the goldstone - boson equivalence theorem . since a higgs boson with mass @xmath23 decays almost exclusively to pairs of intermediate bosons and top quarks , and we have gained control over the dominant two - loop corrections to the respective partial widths , we are now able to predict the lifetime of a high - mass higgs boson with two - loop precision . for the reader s convenience , we list here the born formulas of the relevant partial decay widths : @xmath198 where @xmath199 and @xmath200 ( @xmath201 ) . we compute the total higgs - boson decay width as @xmath202 + k_f\gamma_b\left(h\to t\bar t\,\right)\,.\ ] ] we display the result as a function of @xmath6 in fig . [ fig3 ] , assuming @xmath203 gev . comparing the born , one - loop , and two - loop approximations , the significance of radiative corrections to the heavy - higgs lifetime , @xmath204 , becomes evident . equations ( [ eqkv ] ) , ( [ eqkf ] ) , and ( [ eqdec ] ) refer to the on - mass - shell renormalization scheme and are thus independent of the unphysical renormalization scale @xmath87 , which enters via dimensional regularization . by contrast , in other renormalization schemes such as the @xmath205 scheme , there remains a weak @xmath87 dependence in the loop - corrected expressions for the various higgs - boson decay widths . it is generally believed that the scheme and scale dependences of a calculation up to a given order indicate the size of the unknown higher - order contributions , _ i.e. _ , they provide us with an estimate of the theoretical uncertainty . applying this empirical rule to the fermionic and bosonic higgs - boson decays , one may conclude that the respective perturbative results are likely to become unreliable already for @xmath206 gev @xcite . in fact , for @xmath206 gev , the @xmath87 dependences of the @xmath205 results are no longer reduced if the two - loop corrections are included @xcite . finally , we should mention that monte - carlo studies on finite lattices provide us with a hint on the all - order relation between the mass and the bosonic decay width of the higgs boson @xcite . in ref . @xcite , elastic @xmath207 scattering was analyzed nonperturbatively in the framework of the four - dimensional o(4)-symmetric nonlinear @xmath208 model in the broken phase , and the @xmath208 resonance was observed . the mass @xmath209 and width @xmath210 of this resonance could be extracted from the measured scattering phases . at the same time , the pion mass @xmath211 and wave - function - renormalization constant @xmath2 as well as the field expectation value @xmath212 in infinite volume were determined . in this way , it is possible to compare the nonperturbative value of @xmath210 with the respective perturbative value obtained by inserting @xmath211 , @xmath209 , @xmath2 , and @xmath212 into the tree - level formula for @xmath210 , given by eqs . ( 9.4 ) and ( 9.5 ) of ref . @xcite . for the central simulation point , the nonperturbative result is found to exceed the tree - level one by approximately 16.7% @xcite . it is interesting to confront this all - order correction with the one- and two - loop corrections of eq . ( [ eqkv ] ) . if we identify @xmath209 , @xmath2 , and @xmath212 of ref . @xcite with our parameters @xmath6 , @xmath47 , and @xmath83 , respectively , we have @xmath213 , where the lattice constant cancels on the right - hand side . we so find that the central simulation point of ref . @xcite corresponds to @xmath214 gev , for which the corrections through @xmath18 and @xmath3 in eq . ( [ eqkv ] ) amount to approximately 7.7% and 12.5% , respectively . the tree - level , one - loop , two - loop , and nonperturbative values of @xmath179 at this value of @xmath6 are indicated by the crosses in fig . [ fig2 ] . in other words , the one- and two - loop terms of @xmath179 are 7.7% and 4.7% , respectively , while the higher - order terms apparently add up to a total of 4.3% . this is a very reasonable result , which gives support to the notion that the perturbative expansion in @xmath21 of the bosonic higgs - decay width is still usefully convergent for this value of @xmath6 , in accordance with the conclusions drawn above on the basis of fig . however , this observation should be taken with a grain of salt . the analysis of ref . @xcite was performed with @xmath211 finite , while we set @xmath215 for the goldstone - boson pole mass . nevertheless , we believe that such finite - mass effects will not drastically change our conclusions as long as we compare the relative corrections to the decay widths rather than the absolute values of the latter . this assumption is substantiated by the detailed inspection of the full one - loop results for the @xmath26 and @xmath27 partial decay widths @xcite . * acknowledgements * we thank a. ghinculov for providing us with some partial results of ref . @xcite , and m. gckeler , m. lscher , p. weisz , and j. westphalen for valuable advice regarding the comparison of our result with ref . the collaboration on this project was initiated during the workshop on _ higher order perturbative corrections in the standard model _ at the _ aspen center for physics_. here , we present the relevant one - loop three - point diagrams to @xmath30 . we write the generic one - loop three - point integral as @xmath216 [ ( q+p+k)^2-m_3 ^ 2+i\varepsilon]}.\end{aligned}\ ] ] the one - loop three - point diagrams in fig . [ fig1 ] are then defined as @xmath217 we find @xmath218 + o(\epsilon^2)\right\ } , \nonumber\\ { \cal c}_2&{\hspace*{-2.5mm}}={\hspace*{-2.5 mm } } & \frac{i}{(4\pi)^2m_h^2 } \left(\frac{4\pi\mu^2}{m_h^2e^\gamma}\right)^\epsilon \left\ { -\frac{2}{3}\zeta(2 ) + \epsilon\left[\frac{2}{9}\zeta(3)-\frac{4\pi}{9}{\mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{\pi}{3}\right)\right ] + o(\epsilon^2)\right\ } , \nonumber\\ { \cal d}_{2a}&{\hspace*{-2.5mm}}={\hspace*{-2.5 mm } } & \frac{i}{(4\pi)^2m_h^4 } \left(\frac{4\pi\mu^2}{m_h^2e^\gamma}\right)^\epsilon \left\ { i\frac{\pi}{2 } + \epsilon\left[-\frac{5}{4}\zeta(2)+i\frac{3\pi}{2}\ln2\right ] + o(\epsilon^2)\right\ } , \nonumber\\ { \cal d}_3&{\hspace*{-2.5mm}}={\hspace*{-2.5 mm } } & \frac{i}{(4\pi)^2m_h^4 } \left(\frac{4\pi\mu^2}{m_h^2e^\gamma}\right)^\epsilon \left\{\frac{\pi}{3\sqrt3 } + \epsilon\left[\frac{2}{3}\zeta(2)+\frac{4}{3\sqrt3}{\mathop{{\mbox{cl}}_2}\nolimits}\left(\frac{\pi}{3}\right ) -\frac{\pi}{3\sqrt3}\ln3\right ] + o(\epsilon^2)\right\}.\nonumber\\\end{aligned}\ ] ] as already stated in sec . 3 , @xmath119 and @xmath120 are ir divergent . we need not calculate them , since their prefactors in eq . ( [ eqmaster ] ) vanish according to eq . ( [ irfinsets ] ) . e. gross , b.a . kniehl , and g. wolf , z. phys . c * 63 * , 417 ( 1994 ) ; * 66 * , 321(e ) ( 1995 ) ; m. carena , p.m. zerwas , _ et al . _ , in _ physics at lep2 _ , edited by g. altarelli , t. sjostrand , and f. zwirner , cern yellow report no . 9601 , vol . 1 ( february 1996 ) p. 351 . linde , pisma zh . teor . fiz . * 23 * , 73 ( 1976 ) [ jetp lett . * 23 * , 64 ( 1976 ) ] ; s. weinberg , phys . 36 * , 294 ( 1976 ) ; m.j . duncan , r. philippe , and m. sher , phys . * 153b * , 165 ( 1985 ) ; m. sher , phys . rep . * 179 * , 273 ( 1989 ) ; phys . b * 317 * , 159 ( 1993 ) ; * 331 * , 448(a ) ( 1994 ) ; m. lindner , m. sher , and h.w . zaglauer , phys . lett . b * 228 * , 139 ( 1989 ) ; c. ford , d.r.t . jones , p.w . stephenson , and m.b . einhorn , nucl . * b395 * , 17 ( 1993 ) ; g. altarelli and g. isidori , phys . b * 337 * , 141 ( 1994 ) ; j.a . casas , j.r . espinosa , and m. quiros , phys . b * 342 * , 171 ( 1995 ) . r. dashen and h. neuberger , phys . lett . * 50 * , 1897 ( 1983 ) ; w. langguth and i. montvay , z. phys . c * 36 * , 725 ( 1987 ) ; a. hasenfratz , t. neuhaus , k. jansen , h. yoneyama , and c.b . lang , phys . b * 199 * , 531 ( 1987 ) ; a. hasenfratz and t. neuhaus , nucl . phys . * b297 * , 205 ( 1988 ) ; p. hasenfratz and j. nager , z. phys . c * 37 * , 477 ( 1988 ) ; m. lscher and p. weisz , phys . b * 212 * , 472 ( 1988 ) ; u.m . heller , h. neuberger , and p. vranas , nucl . phys . * b399 * , 271 ( 1993 ) ; u.m . heller , m. klomfass , h. neuberger , and p. vranas , nucl . * b405 * , 555 ( 1993 ) . cornwall , d.n . levin , and g. tiktopoulos , phys . d * 10 * , 1145 ( 1974 ) ; * 11 * , 972(e ) ( 1975 ) ; c.e . vayonakis , lett . nuovo cim . * 17 * , 383 ( 1976 ) ; m.s . chanowitz and m.k . gaillard , nucl . * b261 * , 379 ( 1985 ) ; g.j . gounaris , r. kgerler , and h. neufeld , phys . d * 34 * , 3257 ( 1986 ) ; y .- p . yao and c .- p . yuan , phys . d * 38 * , 2237 ( 1988 ) ; j. bagger and c. schmidt , phys . d * 41 * , 264 ( 1990 ) ; h. veltman , phys . d * 41 * , 2294 ( 1990 ) ; h .- j . he , y .- p . kuang , and x. li , phys . lett . * 69 * , 2619 ( 1992 ) ; phys . d * 49 * , 4842 ( 1994 ) ; h .- j . he , y .- p . kuang , and c .- p . yuan , phys . d * 51 * , 6463 ( 1995 ) . .coefficients @xmath98 , @xmath99 , and @xmath100 to be inserted in template ( [ eqtem ] ) to obtain the values of the divergent two - loop three - point diagrams of fig . @xmath162 for @xmath68 , @xmath163 for @xmath69 , and @xmath164 for @xmath70 . analytical and numerical results for @xmath219 , @xmath220 , @xmath169 , @xmath221 , @xmath170 , and @xmath222 are given in eq . ( [ eqcon ] ) . [ cols="^,^,^,<",options="header " , ]

i would like to revive , from a new viewpoint , a somewhat old fashioned discussion , which recently seems to be all but extinct . as goes without saying , every contribution , correction , criticism or suggestion is greatly appreciated . the discussion , which lasted for the major part of the early stages of atomic , nuclear and particle physics , as we know them today , dealt primarily with the scientific and logical implications of quantum theory . again revived by david bohm s contribution , the `` hidden variables '' , it gradually began to loose on impact , until today any fundamental question in this respect seems to be considered a lack of scientific soundness . from an epistemological point of view , the fact is not easily understood , since all the fundamental problems are still far from any solution . second quantization , on the contrary , provides an even more problematic conception by its inherent infinity problems @xcite , which , till now , have not been solved in a satisfying way . that all the original problems are still there , can be seen by a survey of late publications of bohm @xcite , de broglie @xcite , dirac @xcite , or schrdinger @xcite . from the viewpoint of common sense the most profound reason of uneasiness with quantum theory , despite its indisputable successes , has been expressed by richard feynman : _ you see , my physics students do nt understand it either . that is , because i do nt understand it . nobody does . _ @xcite that the discussion , mentioned above , is nevertheless non existent , might have its origins therefore not so much in a satisfying status quo , but rather in a lack of alternative concepts . from an epistemological point of view a well working theory , despite all its conflicts with logical reasoning , is better than no theory . the discussion then leads to the question , why so far no promising concept has been put forth . comparing with other sciences , the situation seems very unusual : seventy years of continuous research from schrdinger s first publications @xcite have neither altered the principles of quantum theory , nor have they removed the axiomatic qualities of its fundamental statements . this situation basically leaves two possibilities : either to accept , that we can not understand , i.e. that we _ have to believe _ , or to search for alternatives . the attempts so far , to remove the axiomatic qualities of quantum theory , which can be summarized by the _ copenhagen interpretation _ @xcite , mainly focussed on the statistical qualities attributed to fundamental physical events . it was thought , that a suitable theoretical framework could prove quantum events to be integral results of basically continuous processes . the first of these attempts can be attributed to erwin schrdinger @xcite . his wave mechanics , now a fundamental part of quantum theory , does not allow for an interpretation of wave functions as _ physical _ waves due to the contradiction a spreading wave packet presents to mass conservation . on a fundamental level , the contradiction originates in the difference between phase velocity and group velocity of de broglie waves @xcite . since schrdinger s equation is based on exactly these qualities of de broglie waves , it can not be retained without rejecting the interpretation of wave functions as physical waves . the problem was solved by born s interpretation @xcite of @xmath4 as a _ probability amplitude _ , an interpretation which constitutes an integral part of the copenhagen interpretation . this interpretation of the wave function is not without its inherent problems , as einstein , podolksky and rosen pointed out @xcite . if quantum theory contained all the physical information about a spin system , thus their line of reasoning , then information about a measurement process must be exchanged with a velocity faster than c , if spin conservation is to be retained . this leaves , as they point out , only two alternatives : either quantum theory contradicts the theory of relativity , or it does not contain all the information about the system : in this case , quantum theory is incomplete . the scientific controversy was subject to much discussion , but never conclusively solved . recently , its implications have been extended by the phenomenon of interaction free measurements @xcite , based on a thought experiment by renninger @xcite and realized by a mach zehnder interferometer . interaction free measurements have different implications in a classical or quantum context : while a measurement without interaction in classical physics does not alter the system , it leads to a reduction of the wave function in quantum theory . according to the solution , proposed by heisenberg ( see ref.@xcite ) only the reduction of the wave function constitutes physical reality : in this case the question is legitimate , why a wave function , obviously devoid of any physical meaning , plays such an important role in current theories . david bohm s famous theory of hidden variables @xcite was a far more elaborate concept along the same lines . he tried to prove , that the formalism of quantum theory allows for a sub layer of hidden variables , and that the framework therefore is , as einstein suspected , incomplete . the modification was rejected due to the work of bell , who devised a simple situation , where this assumption could be put to a test @xcite . the bell inequalities proved that , within the framework of quantum theory , such an assumption leads to logical contradictions . the theory of hidden variables was therefore rejected . the combined evidence of all these attempts allows the conclusion , that a modification of quantum theory within its fundamental axioms , yielding a framework of which the quantum aspect will be but a result , is not very promising . at best , it reproduces numerical results by a different mathematical structure , at worst , it not even accounts for basic experiments . there remain , therefore , three fundamental questions to be answered prior to any attempted modification of micro physics : * on what theoretical or experimental basis could an alternative formulation be based ? * which theoretical concept of physics , statistical , field theory , or mechanics , shall be employed ? * what constitutes a successful alternative to current formulations ? experimental evidence exists , that moving electrons exhibit wave like features . these qualities have been established beyond doubt by diffraction experiments of davisson and germer @xcite . quantum theory can not directly access wave like qualities , because , as already mentioned , the periodic wave function needs to be subjected to a complicated evaluation process in order to extract its physical relevance . if the initial problems of wave dispersion and velocity can be overcome , the experimental result of wave characteristics of moving mass and its mathematical formalization may well constitute the foundation of a different theoretical conception . the only other experimental evidence at hand with equivalent significance , the quantum phenomena , do not as easily fit into a physical interpretation , since the basic obstacle , why discrete values should be the only ones measured , has to be removed : quantum phenomena are therefore more remote in terms of basic concepts . concerning the theoretical context , it has to be considered , that quantum theory originally started as a modification of mechanics . most of the concepts of the theory are consequently still disguised mechanical concepts . field theoretical results are therefore only achieved by a statistical superstructure : the mechanical basis then leads to a statistical formulation of field theories , which allows for particle interactions , but not for internal particle structures . modifications of this methodology basically occur along two lines : either by a modification of interactions , which requires additional qualities of particles , or by postulating the existence of sub particles . physical progress then requires an ever refined toolkit of particles , sub particles and interactions . from the viewpoint of epistemology , the problem inherent to this approach is the limited degree of freedom . every modification requires a major change of particle properties , and since every new experiment provides data not yet accounted for , the development of physics becomes an ever increasing inventory of individual particles . the particle zoo , high energy physicists so frequently refer to , can , along these epistemological lines , be seen as a consequence of theoretical shortcomings . since the experimental basis the wave characteristics and thus the internal structure of particles suggest an approach based on field theory rather than mechanics , this essay will be based on field theory . if mechanical concepts are required , they always refer to continuous distributions of physical qualities , which yield discrete mass or charge values only by integration . if one is to try an alternative route of theoretical development , it seems important , not to get carried away and extend a concept beyond its limits of proper application . to this end , a goal has to be defined as well as careful checks of experimental evidence . a theory , which does not allow for falsification , remains useless , the same applies to a new theory , which does not deviate from the current one by at least one significant result . that a new theory can not and should not reproduce all results of an existing one , goes without saying personally , i think that the following results are sufficient to justify serious consideration of a new approach : * the reproduction of fundamental results of early quantum theory * the formalization of internal particle structures and the proof of physical significance of these structures * a model of ( hydrogen ) atoms consistent with the formalism developed and consistent with experimental evidence * the basics of a theory of particle interaction due to intrinsic characteristics * provision of at least one significant deviation which allows for an experimentum crucis the fundamental relations and concepts of this new approach have already been published or are in print in the _ speculations in science and technology _ ( chapman and hall , london ) @xcite . the accounts given in this and follow up papers are somewhat enhanced and thorougher versions , making use of the greater freedom of length . as already emphasized , any feedback is greatly appreciated and readers are welcomed to contact me directly via email at the email address : whofer@eapa04.iap.tuwien.ac.at to allow for internal characteristics of moving particles we base the following section on the experimental evidence of wave features . theoretically , the framework will be based on a suitable adaptation of the fundamental relations found by l. de broglie . originally , de broglie s formalization of waves @xcite was based on the framework of special relativity , the amplitude of a de broglie wave given by : @xmath5 the variables are four dimensional vectors , their components are described by the following relations : @xmath6 in a three dimensional and non relativistic ( @xmath7 = 1 ) adaptation of the original equation ( [ ws001 ] ) the amplitude can be expressed as a plane wave . this adaptation is quite common in quantum theory and therefore no deviation from the original framework : @xmath8 the first major deviation from the traditional framework is the interpretation of @xmath9 as a real function , as a physical wave . this conception is not compatible with quantum theory for the following reasons : as a moving particle can not consist of a specific plane wave ( particle mass in this case would be distributed over an infinite region of space ) , a moving particle in quantum theory _ must _ be formalized as a fourier integral over partial waves . an application of schrdinger s equation then leads to the , above mentioned , spreading of the wave packet , which contradicts experimental evidence . the interpretation of the wave function as a _ physical entity _ , furthermore , contradicts heisenberg s uncertainty relation @xcite , because it implies the possibility of measurements below quantum level . @xmath10 the reason , that this real formalization of de broglie remained so far unconsidered , is not only its inconsistency with quantum theory , but also a consequence of planck s relation : dispersion of de broglie waves is described by two independent relations , which are widely confirmed in experiments . the first is the wavelength of de broglie waves , established by diffraction experiments of davisson and germer @xcite , the second planck s energy relation @xcite : @xmath11 if particle energy is classical _ kinetic _ or relativistic energy , the phase velocity of a de broglie wave is not equivalent to particle velocity : @xmath12 the deduction of phase velocity , leading to the contradiction in eq . ( [ ws006 ] ) is , contrary to a first impression , not self evident . it contains , on thorougher analysis , a conjecture and a subsequent logical circle . firstly , it is not self evident , that on a micro level energy of a particle is restricted to _ kinetic _ energy of inertial mass : assuming , that a particle is , basically , a mechanical entity without any internal energy component , can not be inferred from experimental or theoretical results . it is therefore an _ assumption _ lacking a solid basis . based on this conjecture is the following logical circle : _ if energy is kinetic energy , phase velocity of a de broglie wave is not equal to mechanical velocity of a particle . if phase velocity does not equal mechanical velocity , a free particle can not consist of a single wave of specific frequency . if it does not consist of a single wave , then the wave features of a particle must be formalized as a fourier integral over infinitely many partial waves . in this case any partial wave can not be interpreted as a physical wave . if the interpretation as a physical wave is not justified , then internal wave features can not be related to physical qualities . if they can not be related to physical qualities , then internal processes must remain unconsidered . and if internal processes remain unconsidered , then energy of the particle is kinetic energy _ apart from the question , whether this assumption is necessary , which can only be answered a posteriori , it seems also legitimate to ask , whether it is the best possible assumption , especially in view of complications arising from fundamental particles formalized as fourier integrals over infinitely many partial waves . from the viewpoint of simplicity , the concept is necessary , if no alternative exists , it is unnecessary complicated , if the same theoretical result the wave features of single particles can be obtained by a simpler formalism . to this aim alternatives have to be analyzed . a direct way of formalization would be based on two basic statements : ( 1 ) the experimental evidence of wave features , and ( 2 ) the assumption , that wave features must be reflected by _ physical _ qualities of particles , density for example . mathematically , these statements are expressed by : @xmath13 @xmath14 in view of the dispersion relations eq . [ ws005 ] these relations seem contradictory : the standard evaluation procedure would , for a specific _ physical _ wave , lead to a difference between phase velocity and mechanical velocity in conflict with the energy principle . this contradiction depends , though , on the principal conjecture , that a moving particle is equivalent to inertial mass , and that all its energy is contained in the longitudinal motion of an inertial mass @xmath15 with velocity @xmath16 . as soon as wave features are assumed to describe the physical nature of particle motion , this conjecture can no longer be sustained : energy of the particle will then be contained in its _ mass oscillations_. since mass oscillations determine a periodic change of kinetic energy , the energy principle requires the existence of an _ intrinsic _ potential of particle motion , and averaging over kinetic and potential energy during particle motion then leads to a total energy of double the original kinetic value . using this total energy for planck s relation satisfies the dispersion relation of a monochromatic plane wave : based on the intrinsic properties of physical waves any monochromatic plane particle wave then satisfies the wave equation . averaging density of the particle wave over a full period we get initially : @xmath17 and the kinetic energy of particle motion is therefore ( @xmath18 denotes the volume of the particle ) : @xmath19 since kinetic energy is not constant but periodic , the result requires that an intrinsic potential exists , which yields an equivalent average of potential energy : @xmath20 and total energy of the particle wave is then : @xmath21 phase velocity of this particle wave equals , consistent with quantum theory , mechanical velocity of particle motion , although , contrary to quantum theory , no fourier integral over partial waves is required . therefore any monochromatic plane particle wave , consisting of oscillating mass and a complementary potential , describes motion of a _ free particle _ : @xmath22 from a theoretical point of view the existence of intrinsic potentials is a _ conjecture _ , it is more problematic than the original one based on inertial mass , since it implies the existence of physical variables not accounted for in the current framework of quantum theory . relating @xmath23 to total energy of the particle is , in this respect , not problematic , since experimentally energy of a particle can only be estimated by its state of motion or wavelength : and the relation between particle velocity @xmath24 , wavelength @xmath25 , and frequency @xmath26 is retained . with the dispersion relation given by eq . [ ws013 ] the wave function @xmath27 and density @xmath28 of free particles comply with a wave equation : @xmath29 the wave equations display intrinsic features of particle motion and possess , as will be seen in the following sections , the same dimensional level as electrodynamic formulations . the logical structure of physical statements is therefore modified : while in the standard procedure of quantum field theory the fundamental statements of quantum theory determine events on a micro level , and electromagnetic fields are logical superstructures , in the present context statements about the wave features of particles are different aspects of the same physical phenomena dealt within electrodynamics . this parallel development of electrodynamics and quantum theory from the same root the intrinsic wave features of particles must be understood in a literal sense , because the potentials deduced from mass oscillations will be related to electromagnetic fields . the schrdinger equation , in this context , is an expression of mechanical features i.e. potential and kinetic energy of particle motion . from the viewpoint of physical waves it is not without a certain degree of arbitrariness . to derive the relation we proceed from the periodic wave function eq . [ ws013 ] . then the time differential is given by : @xmath30 kinetic energy @xmath31 of particle motion will therefore be : @xmath32 and substituting eq . [ ws015 ] into [ ws016 ] we get for the kinetic energy of the particle : @xmath33 if the total energy of a particle is equal to kinetic energy and a potential @xmath34 : @xmath35 then the wave equation in the presence of an external potential is described by : @xmath36 the relation is equivalent to a time free schrdinger equation @xcite . it is important to consider , that this ( mechanical ) formulation of the wave equation explicitly relates physical potentials to kinetic energy and thus the wave length of a material wave ( as a result of the laplace operator acting on @xmath0 ) and that , therefore , in a region , where @xmath34 takes successively different values , every kinetic energy and thus wave length is generally possible . its interpretation in the context of material waves is not trivial , though . if the volume of a particle is finite , then wavelength and frequency become _ intrinsic _ variables of motion , which implies , due to energy conservation , that the wave function itself is a measure for the potential at an arbitrary point @xmath37 . if it is a measure for the physical conditions of the environment i.e. the intensity of @xmath34 , then it _ must _ have physical relevance . but if it has physical relevance , then the epr dilemma @xcite will be fully confirmed , and the result will favor einsteins s interpretation of quantum theory : quantum theory can not be complete under the condition that ( 1 ) particles do have finite dimensions , and ( 2 ) the schrdinger equation is a correct description of the variations of particle waves in the presence of external fields . the only alternative , which leaves quantum theory intact , is the assumption of particles with zero volume : an assumption which leads , in the context of electrodynamics , to the equally awkward result of infinite energy of particles . the first major result of this essay can therefore be formulated as follows : * the interpretation of particles as physical waves allows the conclusion , that quantum theory can not be , in a logical sense , a complete theoretical description of micro physical systems . there exists another problem usually hidden in the mathematical formalism of quantum theory . let an external potential be given by @xmath34 . transforming eq . [ ws019 ] into a moving coordinate system @xmath38 described by : @xmath39 we get the following equation : @xmath40 since the variable @xmath41 in this case is undefined , the relation contains an element of arbitrariness . it is not conceivable , from this deduction , how a time free schrdinger equation could be derived : if the time dependent part of @xmath0 is eliminated , the procedure is equivalent to a transformation into a moving coordinate system , but in this case the potential becomes a function of time . and if the wave equation is integrated in the system at rest , then the wave functions depend on time . from a physical point of view , the reduction of the schrdinger equation to its time free form actually eliminates all the time dependent interaction and adjustment processes bound to occur for _ every _ particle wave interacting with a changed environment . it may well be , that this disregard of development accounts for the commonly rather static concepts in quantum theory . within the framework of material waves , the arbitrariness of the schrdinger equation can be referred to the intrinsic potential of mass oscillations , not accounted for in quantum theory . and its consequence , as can equally be derived , is an uncertainty about the exact value of @xmath42 of the wave function @xmath43 . the uncertainty of @xmath42 together with the period of intrinsic potentials @xmath44 is then sufficient to derive heisenberg s uncertainty relations @xcite . to this aim we proceed from the schrdinger equation in the moving frame of reference eq . [ ws020 ] . setting the upper and lower limits of the potential @xmath45 to : @xmath46 and calculating the correlating @xmath47vectors of the wave function @xmath48 : @xmath49 we get the following relation : @xmath50 evaluating the relation in a one dimensional model and accounting for the yet unknown variation of the potential @xmath51 by : @xmath52 the uncertainty @xmath53 due to the undefined variable t will be : @xmath54 if the uncertainty originates from the neglected intrinsic potential @xmath55 ( see eq . [ ws011 ] ) , the uncertainty of the applied potentials is described by : @xmath56 and therefore the uncertainty of @xmath47 can be referred to the value of @xmath47 : @xmath57 the uncertainty does not result from wave features of particles , as heisenberg s initial interpretation suggested @xcite , nor is it an expression of a fundamental physical principle , as the copenhagen interpretation would have it @xcite . it is , on the contrary , an expression of the _ minimum error _ , inherent to _ any _ evaluation of the schrdinger equation , resulting from the fundamental assumption in quantum theory , i.e. the interpretation of particles as inertial mass aggregations . since the distance between two potential maxima is given by ( the value can deviate in two directions which yields total uncertainty of the @xmath58coordinate ) : @xmath59 the uncertainty may equally be written as : @xmath60 heisenberg actually derived @xcite the uncertainty relations @xmath61 . the fourier transform inherent to quantum theory suggests a correction of the constant eq . [ ws027 ] by a factor @xmath62 . in this case the deduction yields _ exactly _ the same result as the canonical formulation of quantum theory @xcite : @xmath63 the minimum error due to intrinsic potentials is therefore equal to the uncertainty of quantum theory , and the logical relation between schrdinger s equation and the uncertainty relations is therefore a relation of reason and consequence : _ because _ schrdinger s relation is not precise by the value of intrinsic potentials , the uncertainty relations _ must be _ generally valid . this is the second major result of the present essay : * _ because _ quantum theory neglects intrinsic characteristics of particle motion , the schrdinger equation is generally not precise by a value described by heisenberg s uncertainty relations . the result equally seems to settle the long standing controversy between the _ empirical _ and the _ axiomatic _ interpretation of this important relation : it is not empirical , since it does not depend on any measurement process ; but it is equally not a physical principle , because it is due to the fundamental assumptions of quantum theory . the solution to the problem of interpretation is therefore a somewhat wider frame of reference : although within the principles of quantum theory the relation is an _ axiom _ , it is nonetheless a _ result _ of fundamental theoretical shortcomings and not a physical principle . in the preceding section we could demonstrate , that quantum theory must not necessarily be a comprehensive representation of physical reality . it was shown , on the contrary , that the principal assumption of quantum theory , the assumption of inertial mass of particles , is a posteriori neither justified by the wave features of particles nor by the fundamental planck and de broglie relations . it is but _ one _ possible theoretical framework and , given the logical problems of a precise relation between measurements and reality arising from this concept , not even the best possible framework . on an analytical level the establishment of fundamental relations of quantum theory within a different framework and their significance that the uncertainty relations can also , and consistently , be interpreted as an expression of theoretical shortcomings due to a disregard of intrinsic properties served to establish the limitations of the current theoretical standard . they also served to define the logical _ location _ of quantum theory and to determine the regions of interest , where research is commendable if a theory _ outside _ the current framework is taken into consideration . from the viewpoint of theoretical synthesis we showed that the intrinsic properties of particles and their relations are a consequence of the experimental evidence of wave features of particles . the de broglie relations , in a non relativistic formulation , provided the essential laws by which these intrinsic variables are described . they allowed to establish ( 1 ) periodic mass oscillations within the particle , and ( 2 ) the existence of a periodic potential . these results were sufficient to deduce the fundamental relations of quantum theory _ and _ , at the same time , to show their inherent limitations . apart from showing the significance of the periodic potentials which will be done in the following sections the first two requirements mentioned in the introduction are therefore fulfilled . fundamental results of early quantum theory are reproduced ( schrdinger s relation as well as the uncertainty relations , while single experiments will be treated in the following sections ) , and the intrinsic wave features of particles were shown to be of _ physical _ significance , which strongly supports einstein s view in the epr paradox . classical electrodynamics is considered one of the most refined and powerful physical theories at hand . it served , during the last hundred years , ever again as a model for physical formalization , and its implications led to quite a few extensions of physical knowledge , not least albert einstein s theory of special relativity @xcite . a great deal of the fascination of classical electrodynamics results from its mathematical conciseness and brevity . the possibility to describe and account for most of the phenomena concerning interactions of charge by a set of basically four linear differential equations , the _ maxwell equations _ @xcite , must either seem a lucky coincidence or an expression of a deeply rooted analogy in physical reality . that classical electrodynamics can not be referred to mechanical conceptions , has been proved during the scientific controversy about the existence of the _ aether _ . this proposed medium of electromagnetic fields was never detected , its existence , furthermore , would contradict the experimental results of michelson and morley @xcite . from the theoretical side , the controversy was settled by einstein s theory of relativity . since the correctness of maxwell s equations is established beyond doubt , it remains an epistemological problem , why no attempt to refer them to fundamental physical axioms has succeeded so far @xcite . summarizing the current situation in two statements : ( 1 ) maxwell s equations are correct , and ( 2 ) they can not be referred to mechanical axioms , there exist equally two logical conclusions to the epistemological problem : ( 1 ) maxwell s equations are themselves of an axiomatic nature , and ( 2 ) the axioms , on which they are based , have not yet been detected . currently , the established opinion is ( 1 ) . basically , one could leave it at this state and employ the relations as correct algorithms . from the viewpoint of macro physics , the procedure is justified by the wealth of experimental verifications at hand . the same does not apply , though , to an application of electrodynamics in micro physical models . since the exact origin of the relations is obscure , it can not be guaranteed , that they do not , in some way , preclude characteristics of physical systems , which contradict other assumptions . obviously , there exist only two ways out of this logical dilemma : either it can be proved , that quantum theory is logically independent of electrodynamics a proof , which never has been given , or it can be established that electrodynamics itself is based on micro physical axioms : in this case the current procedure in qed @xcite is logically insufficient . the logical conclusion to our epistemological problem would then be : ( 2 ) the maxwell relations are an expression of intrinsic characteristics of particles . based on the intrinsic wave characteristics of free particles , developed in the previous section , the verification of ( 2 ) does not present an impossible problem . the maxwell equations , in material wave theory , are an expression of longitudinal variations of intrinsic variables of particles in motion . from the wave equation for longitudinal momentum @xmath64 : @xmath65 and the equation of continuity for momentum @xmath64 and density @xmath66 : @xmath67 it follows within a micro volume , where particle velocity @xmath24 is constant : @xmath68 the parameter @xmath69 is a dimensional constant employed to make mechanical units compatible with electromagnetic variables , its dimension is equal to density of charge . electric fields @xmath70 and magnetic fields @xmath71 are now defined in terms of momentum @xmath64 and the potential @xmath72 due to mass oscillations ( see the previous section ) : @xmath73 and substituting eq . [ ed003 ] into [ ed001 ] we get , with the help of [ ed004 ] and [ ed005 ] : @xmath74 and since the total intrinsic potential is a constant : @xmath75 the following relation holds in full generality within every micro volume : @xmath76 rotation of eq . [ ed004 ] yields , but for a constant , maxwell s first equation : @xmath77 to deduce the inhomogeneous equation we additionally have to employ the source equations of electrodynamic theory and the definition of @xmath78 fields : @xmath79 it is assumed that the material constants @xmath80 and @xmath81 also remain constant within the micro volume . substituting into eq . [ ed008 ] and computing the source of [ ed010 ] , the sum yields : @xmath82 and if the constant of integration is zero , the result will be maxwell s inhomogeneous equation : @xmath83 finally , by computing the time differentials of eqs . [ ed008 ] and [ ed009 ] , the wave equations for electromagnetic fields in a vacuum ( @xmath84 ) can be deduced : @xmath85 the results indicate , that electrodynamics and quantum theory do have a common root , the _ intrinsic _ properties of moving particles . from the viewpoint of material waves they are therefore different aspects of the same physical phenomenon , the wave features of mass , which is the third major result of this essay : * electrodynamics and wave mechanics are theoretical concepts based on the same physical phenomena , the intrinsic features of moving particles . logically speaking , the result means , that a method , which employs quantum theory to calculate physical variables by way of observables , and which identifies these observables measured variables as field variables in electrodynamics can not be correct . the energy principle of material waves can be identified as the classical lorentz standard by a combination of the continuity equation and the definition of electric fields including intrinsic potentials . since : @xmath86 the wave equation for linear motion @xmath87 leads to : @xmath88 and for uniform motion @xmath89 the potential can be related to momentum @xmath90 by : @xmath91 comparing with the expression in classical electrodynamics the vector potential @xmath92 , in classical electrodynamics an abstract conception , is proportional to momentum of particle motion : @xmath93 the wave features of photons , or the representation of electromagnetic fields by wave type field vectors , have long been the only aspect in consideration . the picture changed only after experimental research was directed towards the fundamental problems of interaction between matter and electromagnetic fields . due to planck s result on black body radiation and einstein s application of planck s formula to the problem of photo electricity @xcite , the energy of photons gained special attention , since it was ( 1 ) proportional to frequency and not , as field theory proposed , intensity , and ( 2 ) the energy quantum @xmath94 could no longer be distributed locally , because the energy changes in interactions required an exchange of a full number @xmath95 of energy quanta . in einstein s view the interaction suggested that energy of photons be distributed in single points within the electromagnetic fields of radiation . according to the copenhagen interpretation of quantum theory @xcite the two theoretical aspects wave features and energy quanta describe complementary facts of a complex physical reality not adaptable to any single theoretical framework . this interpretation , which focusses on the central notion of wave particle duality , states that all the qualities like polarization and interference , typical for a wave type theory , have to be treated by electrodynamics , while energy and interaction phenomena must be seen from the viewpoint of quantum theory . the logical justification for retaining the duality , which is in itself a logical contradiction if wave features and particle features were to prevail _ simultaneously _ , was seen by jordan @xcite and heisenberg @xcite in the impossibility of experiments i.e. physical processes where both aspects _ must _ simultaneously exist . this justification is not undisputed . renninger showed by a rather simple thought experiment @xcite , that an interference measurement involving two separate light paths does not allow for the probability interpretation of electromagnetic field vectors , since the result of the measurement can be altered in a deterministic manner by the insertion of optical devices into a single light path . the analysis of the experiment leads to the conclusion , that the wave features _ must have _ physical significance , which again seems to confirm the epr dilemma already quoted @xcite . based on this experiment renninger deduced the following statement about the features of photons ( translation by the author , emphasizing by renninger ) : _ every light quantum consists of an energy particle , which is guided by a wave without energy _ the important result , from our point of view , is the physical significance of wave features . the energy particle , which renninger refers to , is a direct result of einstein s interpretation of photo electricity , which equally determines the interpretation of the guiding wave as energy free . from the viewpoint of physical concepts , an energy free _ physical _ wave is not all too convincing . if a region of space is altered by a propagating wave , thus the obvious objection , then it can not be without a physical entity moving . but since all physical field variables are , in some way , related to energy , a physical entity propagating without any propagation of energy seems contradictory . on thorougher analysis the necessity to divide between two separate physical entities to account for energy as well as electromagnetic features of photons depends on the interpretation of the interaction process itself . since ( 1 ) the quantum of energy is fully transferred , and ( 2 ) the transfer interval is well below the level of measurements , the energy must be contained in an insignificantly small volume . the alternative being , as renninger pointed out , that the wave will be contracted with a velocity exceeding @xmath96 during every interaction process . it is this interpretation of the interaction process , which will be changed by material wave theory . to understand the decisive modification , we present here and without proof a result on the interactions of electrons and photons , which will be derived later on . assuming that the interaction interval equals , and setting @xmath97 for convenience equal to 1 , an intrinsic model of particles and fields will have to consider energy density at a specific point _ within _ the particle . if total energy density is constant and @xmath98 , then the time derivative of energy density equals : @xmath99 in this way the traditional features of photons are retained , while the meaning of the term _ energy quantum _ is modified : the energy quantum _ transferred _ in an interaction process equals @xmath100 , regardless of the actual volume of the photon , and equally regardless of the exact duration of the interaction process . as long as the total density of energy @xmath98 is constant , the relation holds for every single point of the underlying field and therefore for the whole region of the physical wave corresponding to a photon . now let us consider a case , where the initial photon is , by some optical device , split into two . if one of the partial volumes a half photon so to say then interacts with matter , the process is described by : @xmath101 the energy density transferred remains constant , but not total energy . the process therefore _ seems _ to be a confirmation of the principle , that all interactions occur in discrete energy quanta only , if discrete quantities of mass or charge are considered . only , in short , if the general outlook on processes is a mechanical one . it is no contradiction with einstein s calculation on photo electricity , to interpret it as an effect of _ energy densities _ and transfer rates . it only has to be considered , that the total volume affected by interactions will depend on the total volume of the fractional photon . from this result it can be concluded , that the term _ energy quantum _ is not appropriate for the physical process of interaction : since the total energy transferred depends on ( 1 ) the frequency of radiation , and ( 2 ) the volume of the photon , total energy transferred is the product of energy density proportional to frequency and volume which may be related to intensity . in an _ intrinsic _ formalism of interaction , furthermore , the total quantities transferred are only secondary , the more so , since electrodynamics already is a theory of intrinsic wave features , as deduced above . in material wave theory photons are distributions of mass and energy , and the guiding wave possesses the same dimensions as the energy distribution , while the guiding wave as well as mass movement contribute alike to total energy of the photon . to derive the essential features of photons we use the fundamental wave equation of intrinsic particle momentum . the exact dimensions as well as the problems of photon boundaries are shown to be of no effect for the results achieved . denoting an , initially , abstract momentum of a photon by @xmath102 , we proceed from the wave equation in a vacuum given by : @xmath103 with the identity for the laplace operator and the general relation between momentum @xmath102 and vector potential @xmath104 from eq . ( [ ed016 ] ) : @xmath105 the wave equation may be rewritten as : @xmath106 using the solution for the homogeneous maxwell relations ( the following deduction employs , for convenience , gaussian units ) : @xmath107 the wave equation yields the following result : @xmath108 we choose the specific solution by setting the term in brackets equal to zero and get therefore : @xmath109 the equation can be solved by any wave packet consisting of plain waves and complementary potentials given by : @xmath110 for a component @xmath111 of the wave packet the amplitudes must comply with : @xmath112 setting @xmath113 equal to @xmath114 and using the amplitudes of material waves as well as vacuum dispersion we recover einstein s energy relation @xcite : @xmath115 we define now a _ photon _ as any component of the wave packet . more specifically , we describe a photon by a plane material wave and its complementary electromagnetic potential . @xmath116 the _ energy density _ at a specific location of the photon is denoted by an electromagnetic potential and a potential due to the vector @xmath90 . if we identify @xmath90 as a longitudinal material wave , its energy density at any arbitrary point , the _ total potential _ , is double its value in classical mechanics . @xmath117 then the material aspect of a photon in motion is that of a material wave , while , due to additional electromagnetic characteristics , total energy density is constant and does not depend on location or time : @xmath118 the electromagnetic potential must correlate with electromagnetic field vectors . using electromagnetic potentials of a vacuum : @xmath119 and assuming for consistency with classical electrodynamics , that electromagnetic fields will be of transversal polarization , the transversal fields of a photon will be given by : @xmath120 @xmath121 transversal electromagnetic fields of a wave packet are therefore described by : @xmath122 the units of electromagnetic fields will be treated on a more fundamental basis in the following sections , presently they only signify a constant value of some suitable dimension . in the section on electrodynamics of his famous publication einstein @xcite referred the universal validity of maxwell s equations to their lorentz invariance . based on the theoretical framework of intrinsic features of variables , it is tempting to relate the lorentz invariance of maxwell s equations to the invariance of the wave equations under a lorentz transformation . if it can be established , that the lorentz transformation is also the transformation leaving intrinsic features of particles intact , then the kinematic section of the theory of relativity can also be seen as a necessary modification of mechanical concepts to account for electromagnetic phenomena . as the maxwell relations are not only valid for intrinsic features of photons but also electrons which derives from their independence of particle velocity u we proceed from the wave equation for the intrinsic momentum @xmath64 of a particle with arbitrary velocity @xmath123 in a moving frame of reference s with @xmath124 relative to a system s at rest . the wave equation in s then will be : @xmath125 with the standard lorentz transformation : @xmath126 the differentials in s are given by : @xmath127 density of mass and velocity in s are described by ( transformation of velocities according to the lorentz transformation , @xmath128 presently undefined ) : @xmath129 then the transformation of the wave equations yields two wave equations for momentum @xmath130 and density @xmath66 : @xmath131 energy measurements in the system s at rest must be measurements of the phase velocity @xmath132 of the particle waves . since , according to the transformation , the measurement in s yields : @xmath133 the total potential @xmath98 , measured in s will be : @xmath134 and if density @xmath135 transforms according to a relativistic mass effect with : @xmath136 then the total intrinsic potential measured in the system s will be higher than the potential measured in s : @xmath137 but the total intrinsic energy of a particle with volume @xmath18 will not be affected , since the x coordinate will be , again according to the lorentz transformation , contracted : @xmath138 the interpretation of this result exhibits some interesting features of the lorentz transformation applied to electrodynamic theory . it leaves the wave equations and their energy relations intact if , and only if energy _ quantities _ are evaluated , i.e. only in an integral evaluation of particle properties . in this case it seems therefore justified , to consider special relativity as a necessary adaptation of kinematic variables to the theory of electrodynamics . the same does not hold , though , for the intrinsic potentials of particle motion . any transformation into a moving reference frame in this case leads to an increase of the total potential @xmath98 , in the case of photons ( @xmath139 ) the intrinsic potential then will be infinite . the essential result of this analysis of special relativity may be expressed by : * the wave equations of intrinsic particle structures are lorentz invariant only for an integral evaluation of energy . the total intrinsic potentials increase for every transformation into a moving frame of reference . that maxwell s relations can be referred to intrinsic characteristics of free particles , proves two assumptions : ( 1 ) the relations are themselves not of an axiomatic nature , contrary to current opinions they can be referred to a sublayer of still more fundamental axioms . ( 2 ) quantum theory and electrodynamics are not logically independent , since both concepts are a formalization of wave features of particles . the features of photons in material wave theory equally seem to fit into the framework of classical electrodynamics : electromagnetic field vectors are transversal , while the longitudinal component is purely kinetic . both aspects together provide a model of photons , which is simultaneously an extension of electrodynamics ( since the variables of motion exist in addition to the electromagnetic field vectors ) , and an extension of quantum theory ( since the electromagnetic fields are not originally included in the kinetic model of quantum theory ) . relativistic electrodynamics only provides lorentz invariant formulations of intrinsic particle features , if energy quantities are evaluated . since the intrinsic potentials depend on the reference frame of evaluation and increase by a transformation into a moving coordinate system , the theory suggests that the infinity problems , inherent to relativistic quantum fields , may be related to the logical structure of qed . while the definition of magnetic fields eq . [ ed005 ] allows to deduce the maxwell equations and accounts for intrinsic properties of photons , it is incorrect , by a factor 2 , if magnetostatic interactions of electrons are considered . the definitions , describing the relations between the kinetic variables @xmath140 and the magnetic field @xmath141 , of photons and electrons correctly are written : @xmath142 the reason for this difference , which bears a superficial resemblance to the distinction between bosons and fermions ( s = @xmath143 ) , is that the two relations refer to different magnetic fields . on thorougher analysis , two different cases have to be considered : ( 1 ) the intrinsic magnetic fields , deriving from the kinetic properties of motion and oscillation , and ( 2 ) the external magnetic fields due to curvilinear motion of charge . while for photons the intrinsic electromagnetic fields have been known and dealt with in the framework of classical electrodynamics , for electrons these intrinsic fields can only be derived within material wave theory but not in quantum theory . and while the maxwell relations as an expression of intrinsic features are valid for both types of particles , the magnetostatic properties of charge apply only for electrons . the correction arises therefore from the treatment of different physical processes , and is easily understood as a reminiscence to the historical origins of the separate theories . it is tempting , at this stage , to refer the difference between fermions and bosons to the same historical origins , adopted for the concept of spin without further reflection . that the spin of particles can indeed be referred to magnetic fields , will be shown in section [ ep_spin ] . to account for magnetostatic interactions the definition of magnetic fields therefore has to be corrected . in _ homogeneous _ magnetic fields @xmath144 electron motion can be treated based on the changed definition of external magnetic fields eq . ( [ ep000 ] ) : @xmath145 the sources of @xmath146 vanish , since : @xmath147 the solution is equivalent to the traditional one : @xmath148 circular rotation in homogeneous magnetic fields is not subject to quantization , because the local component of de broglie s relation vanishes . particle density @xmath149 in any state of rotation is given by : @xmath150 then the relation between magnetic field @xmath71 and angular velocity @xmath26 will be , with eq . [ ep000 ] : @xmath151 and since @xmath152 it follows generally : @xmath153 considering the classical formulation for lorentz forces of charge in a magnetic field @xmath154 , we may write : @xmath155 which reduces in the average @xmath156 to : @xmath157 additionally , it has to be considered that rotation of mass in the rotating coordinate system gives rise to centrifugal forces described by : @xmath158 then the rotation of charge is , in the rotating coordinate system , free of forces , it is therefore its _ inertial _ state of motion in a magnetic field @xmath71 : @xmath159 but as centrifugal forces are inertial forces arising from the curvilinear path of motion , the same must apply to lorentz forces : a magnetic field determines rotation of charge , which experiences centrifugal and lorentz forces . which means , evidently , that lorentz forces must be inertial forces . the framework developed so far has some interesting consequences in electrostatic problems . defining a vector field @xmath160 by : @xmath161 and using the equation of continuity , which is equally valid for electron charge distributions : @xmath162 the vector field will be equivalent to : @xmath163 since momentum @xmath102 can be expressed in terms of @xmath164 , the wave equation also applies : @xmath165 if static problems are considered , @xmath166 will be independent of time , and using the definition of magnetic fields as well as ( [ ep002 ] ) , the statement can be rewritten in the following form : @xmath167 it is equivalent to _ ampere s law _ , if electrodynamic current density is defined by : @xmath168 in linear approximation , defining specific resistance @xmath169 by : @xmath170 and using the averages of density as well as neglecting boundary conditions , electromagnetic potentials inside a conductor will be due to sources of the electron wave vector @xmath42 : @xmath171 which means , otherwise , that the driving force of electrostatic current is the gradient of electron velocity . a possible way to show , how the system of physical units is affected by this , basically field theoretical approach to particle motion and interaction , is the following : by modeling a framework of particle motion consistent with classical electrodynamics , and comparing the energy expressions derived with energy of electron waves of hydrogen atoms , the unit of current and the electromagnetic coupling constants will be defined in terms of motion , they become _ mechanical _ units . if we consider the complementary fields of an electron or photon , we may describe their features by a vector field @xmath172 , which comprises , from an electrodynamic point of view , the dynamical situation : @xmath173 then electromagnetic fields are , similar to the formulations in classical electrodynamics , given by : @xmath174 the constants @xmath113 have to be added for reasons of consistency , since the traditional coupling of electromagnetic fields and dynamical variables can not be taken for granted . according to the definitions of @xmath175 and @xmath176 in the previous section , the @xmath113 will only be equal to unity , if @xmath177 $ ] is equal to @xmath178 $ ] . based on these definitions and the wave equation , maxwell s equations were accounted . if , on the other hand , @xmath177 $ ] is , like in classical electrodynamics , equal to energy density ( see eq . [ ed020 ] ) , then it equals @xmath179 , which has the dimensions of energy density but the direction of momentum . for reasons of consistency of the framework with classical electrodynamics the constants have to be set to unity and the vector field shall be equal to momentum per unit charge . @xmath180 then must be dimensionally equal to mechanical momentum . @xmath181 = [ p ] = \left[\frac{kg m}{s}\right]\ ] ] energy of the electromagnetic complement in terms of electromagnetic fields has to be modified accordingly . considering , that the time derivative of @xmath182 equals : @xmath183 and equally , that the energy of a particle is proportional to its frequency : @xmath184 energy density of electromagnetic complements , based on this definition of @xmath182 can be written : @xmath185 the correct relation between frequency and energy is then accounted for , while the constant @xmath180 may be estimated from hydrogen ground states . using the results of material wave theory for electron motion within a hydrogen atom @xcite and considering , that the amplitude of electromagnetic as well as kinetic energy must be double its kinetic value , @xmath180 is given by : @xmath186 \\ \hbar & = & \beta_{f } \sqrt{\frac{m_{e } } { 2 e^2}}\end{aligned}\ ] ] the unit ampere in this case is no longer free for definition , but must be expressed in terms of dynamical units m , kg , s. @xmath187 = \left[\frac{kg^{1/2}}{m s}\right]\ ] ] especially interesting is the meaning of coupling constants in this context . @xmath188 = \left[\frac{m^2}{s^2 } m^3\right ] = \left[u^2\ , v\right ] = \left[m u^2\ , \frac{v}{m}\right ] \nonumber\\ \left[\mu\right ] = \left[v\right ] \qquad \left[\alpha_{f}\right ] = \left[{\bf \epsilon}^{- 1}\right ] = \left[m u^2\ , \frac{v}{m}\right]\label{ep017}\end{aligned}\ ] ] where @xmath189 is sommerfeld s fine structure constant . the units signify , that the coupling constants actually describe the capacity of electron matter to store dynamical energy per unit mass : they give rise to the question , treated in the following sections , how quantum theory allows for energy or charge quantization . the field corresponding to electromagnetic fields @xmath190 and @xmath191 and in longitudinal direction will be defined by : @xmath192 it is not equal to intrinsic momentum , since it is harmonic in longitudinal direction . for the plane wave @xmath193 , the fields will be : @xmath194 on a more general basis we proceed from the lagrangian of electron photon interaction in the presence of an external field @xmath195 . the procedure is pretty standard : defining the lagrange density of a particle in motion , including external potentials and a presumed photon , a variation with fixed endpoints allows to calculate photon energy density in terms of the external potential and kinetic potential of the particle , as well as accelerations of the particle due to the potential applied . so far the result is the traditional one . but using a legendre transform for small velocity variations and calculating the hamiltonian density will show , that the interaction hamiltonian is equal to photon energy , which will allow , on a very general basis , to refer all electrostatic interactions to an exchange of photons . we define the lagrangian by the total energy of the moving electron , and include total energy of a presumed photon : @xmath196 @xmath197 using the hamiltonian principle for infinitesimal variations of the system with fixed endpoints : @xmath198 taking into account that @xmath199_{t_{1}}^{t_{2 } } } _ { = \,0 } - \int\limits_{t_{1}}^{t_{2 } } dt \frac{d}{d t } \left ( \frac{\partial \dot{x}_{i}^2}{\partial \dot{x}_{i } } \right ) \delta x_{i}\ ] ] the integral may be rewritten to : @xmath200 therefore the following statement is valid : @xmath201 the dynamic relation in the absence of a photon will be : @xmath202 the legendre transform of the electron photon system in an external field yields : @xmath203 for small variations of velocity the hamiltonian of the system is given by : @xmath204 the result seems paradoxical in view of kinetic energy of the moving electron , which does not enter into the hamiltonian . assuming , that an inertial particle is accelerated in an external field , its energy density after interaction with this field would only be altered according to its alteration of location . the contradiction with the energy principle is only superficial , though . since the particle will have been accelerated , its energy density _ must _ be changed . if this change does not affect its hamiltonian , the only possible conclusion is , that photon energy has equally been changed , and that the energy acquired by acceleration has simultaneously been emitted by photon emission . the initial system was therefore over determined , and the simultaneous existence of an external field _ and _ interaction photons is no physical solution to the interaction problem . the process of electron acceleration then has to be interpreted as a process of simultaneous photon emission : the acquired kinetic energy is balanced by photon radiation . a different way to describe the same result , would be saying that electrostatic interactions are accomplished by an exchange of photons : the potential of electrostatic fields then is not so much a function of location than a history of interactions . this can be shown by calculating the hamiltonian of electron photon interaction : @xmath205 since material wave theory is based on the intrinsic features of particles , it has to account for a problem , usually treated rather sloppily in quantum theory . we refer to the problem of particle scattering . if einstein s notion of energy as an intrinsic quality of motion is taken seriously - and the present paper seeks to accomplish the same for momentum : i.e. to establish momentum @xmath64 as an intrinsic quality of particle motion - then a thorough analysis of scattering processes raises problems neither quantum theory nor electrodynamics are qualified to solve . we assume two electrons of integral momenta @xmath206 and @xmath207 to interact ; during this process their path will be changed due to electrostatic interactions @xmath208 as well as their velocity . the decisive question , in this context , is the change of internal structures given by : @xmath209 where @xmath210 as well as @xmath211 are functions of time . electric and magnetic intensity @xmath212 and @xmath213 , and their relations were shown to be consequences of the internal structure of any single electron , electrodynamics in its traditional form therefore provides no answer to the question , by what means the alteration of internal structure is brought about . the argumentation of quantum theory in this respect is , strictly speaking , illegitimate , because it is not developed beyond the simple statement , that a specific momentum causes a specific wave packet . by what means this change is accomplished , remains unanswered . from the viewpoint of field theory , the medium of transmission can only be a field , whatever its qualities may be , and this field , by way of physical interactions , must cause the internal restructuring of the second particle from the definition of electric fields eq . [ ed004 ] and the equation of continuity for electron momentum and density , it may be concluded that ( in a one dimensional simplification ) : @xmath214 the time derivative , including the equation of continuity , then yields : @xmath215 identifying the external potential as the potential of a photon : @xmath216 the source term of the electron wave equation for any of the two particles can be determined : @xmath217 as quantization arises from the interaction of single points within the considered region ( see the following sections ) , conservation of electron momentum for the two particles and , consequently , for the two photons , must hold in a very general sense for an arbitrary location within the two photons . and using , again , the equation of continuity for intrinsic momenta and potentials of the photons , the sum of the two potentials must comply with : @xmath218 if the constant of integration equals zero , the total field of the two photons will vanish for every external observer : @xmath219 a specific feature of a component @xmath220 of the overall field will be , that it does not depend on the distance from its source @xmath221 : @xmath222 it is a consequence of ( [ ep031 ] ) , because if intensity would in some way depend on the distance @xmath223 , linear momentum could not remain constant for the system of two particles . if the external and static field @xmath224 is the field of an interacting particle , then emission and absorption will be balanced . the traditional hamiltonian , in this case , correctly describes dynamical variables of interaction , although only due to the - hidden - process of photon exchange . it is evident that neither emission , nor absorption of photons at this stage does show discrete energy levels : the alteration of velocity can be chosen arbitrarily small . the interesting question seems to be , how quantization fits into this picture of continuous processes . to understand this paradoxon , we consider the transfer of energy due to the infinitesimal interaction process . the differential of energy is given by : @xmath225 a denotes the cross section of interaction . setting @xmath226 equal to one period @xmath227 , and considering , that the energy transfer during this interval will be @xmath228 , we get for the transfer rate : @xmath229 since total energy density of the particle as well as the photon remains constant , the statement is generally valid . and therefore the transfer rate in interaction processes will be : @xmath230 the result was already quoted ( see section [ sec_ed ] ) . and it was equally emphasized , that the term energy quantum is , from the point of view taken in material wave theory , not quite appropriate . the total value of energy transferred depends , in this context , on the region subject to interaction processes , and equally , as will be seen further on , on the duration of the emission process . the view taken in quantum theory is therefore only a good approximation : a thorougher concept , of which material wave theory in its present form is only the outline , will have to account for _ every _ possible variation in the interaction processes . but even on this , limited , basis of understanding , planck s constant , commonly considered the fundamental value of energy quantization is the fundamental constant not of energy values , but of transfer rates in dynamic processes . constant transfer rates furthermore have the consequence , that volume and mass values of photons or electrons become irrelevant : the basic relations for energy and dispersion remain valid regardless of actual quantities . quantization then is , in short , a _ result of energy transfer _ and its characteristics . * if intrinsic features of particles and their interactions are considered , quantization has to be interpreted as a result of energy transfer processes and their specific properties . the photon interaction process leading to the acceleration of particles in an electrostatic field is accounted for by the field vector @xmath231 , of longitudinal orientation and including the effects of different field polarizations : @xmath232 elementary charge e shall be the flow through a surface of the field @xmath231 in unit time @xmath233 : @xmath234 integrating over a spherical surface with r = 1 and evaluating the time integral yields : @xmath235 the result is consistent with the source relation of electrostatic fields , since ( in si units ) : @xmath236 @xmath237 and the flow in unit time per unit volume is then : @xmath238 the calculation yields the numerically correct result . it is also dimensionally consistent , if it is considered , that it refers to unit volume and in unit time . the elementary quantum of charge is therefore not exactly a quantity of some physical variable , but , consistent with previous results , a measure for dynamical energy transfer rates per unit time . charge quantization is the consequence of _ photon interaction processes and their specific energy transfer rates_. * the elementary quantum of electron charge derives from energy transfer rates in electrostatic interactions , which are an expression of the interaction process accomplished by photons . using the result for @xmath180 in terms of e , planck s constant equally is not a fundamental constant but can be referred to dynamic interaction rates of particles and fields ( dimensions again corrected ) : @xmath239 comparing with the values for @xmath240 based on codata recommendation for a consistent system of fundamental constants @xcite , the result deviates by about 2.5 % : @xmath241 but considering , that the result was obtained by a completely different approach and , furthermore , that it is the first _ direct _ relation between the three fundamental constants @xmath242 without any additional parameter , the numerical value seems in acceptable accordance with current data . if the relation holds , then the consequences for the framework of quantum theory are significant . its essential conception , the interpretation of discrete frequency , mass and charge values as discrete quantities determining physical processes can , from this viewpoint , no longer be sustained . all processes analyzed so far , were found to be continuous in nature , and the appearance of discrete interaction quanta had to be referred to constant transfer rates applying to every single point of the underlying fields . * on the basis of electron photon interactions planck s constant in material wave theory only applies to interactions and transfer processes . a justification of discrete frequencies , mass or charge values from the viewpoint of allowed _ states _ can not be given in material wave theory . the result indicates , that the formalism of quantum theory lacks justification if it refers discrete energy levels to isolated particles . the result provides the logical reason for the deeply rooted epistemological problem of quantum theory expressed by heisenberg ( see ref . @xcite ) : if quantization is only appropriate for interactions , i.e. measurement processes , then the results of quantum theory can only hold for actual measurement processes . but since the formalism of quantum theory is based on single _ eigenstates _ , meaning states of isolated particles , this logical structure is not accounted for by the mathematical foundations of quantum theory . while , therefore , the mathematical formalism suggests a validity beyond any actual measurement , it can only be applied to specific measurements . what it amounts to , in short , is a logical inconsistency in the fundamental statements of quantum theory . the assumption of invariant charge , i.e. the conception , that charge remains invariant regardless of the interaction process or the surrounding , is not compatible with the results so far . if charge can only be determined by its photon interactions with other charge a result of the previous sections and if the term charge refers to the _ electromagnetic _ qualities of this interaction as described by the change of kinetic variables of charged particles then the _ charge _ of a particle must be referred to the dynamic qualities of interaction photons . as the dynamic qualities of a photon ( not accounting for transversal properties which will be treated in the next section ) are completely determined by ( 1 ) its frequency @xmath1 and ( 2 ) its momentum @xmath243 , the possible interactions of two particles of mass @xmath15 and charge @xmath244 e must be accomplished by a photon with linear momentum @xmath245 described by ( propagation assumed parallel to @xmath246 : @xmath247 the energy of the photon in every case equals @xmath23 , both forms comply with the wave equation for linear intrinsic momentum and yield identical transversal fields , the difference , though , is that the form @xmath248 can not be interpreted as a particle , because its momentum variable has the opposite direction of propagation . * the interpretation of electrostatic interactions as an exchange of photons , proper to material wave theory , has the consequence , that ( 1 ) the properties of single charges can only be referred to interactions with the environment , ( 2 ) photons can only be interpreted as particles if two identical charges interact , and ( 3 ) electrostatic potentials depend on the possibility of photon exchanges . while for two separate charges the third consequence is of no effect , it can not be assumed a priori , that aggregations of charge typical for nuclear environments will provide a mode of photon interactions . but if atomic nuclei do not provide a mode of photon interaction between individual nuclear particles , then electrostatic considerations do not apply to the problem of nuclear stability . so far the vector characteristics of electromagnetic fields have not been accounted for . the problem of electron photon interaction in this case requires a closer scrutiny of the energy transfer process itself . consistent with classical electrodynamics it may be postulated , that electromagnetic fields may be superimposed . in this case the vector field @xmath249 of an electron and interacting photon is given by : @xmath250 and consistent with the framework developed it may be stated , that total energy density at any location of the electron must be equal to : @xmath251 using ( [ ep014 ] ) and ( [ ep015 ] ) the energy relation yields : @xmath252 the energy is altered depending on the angle between electromagnetic field vectors of electron and interacting photon . the equation has the same structure as the energy relation in classical mechanics , it does not refer , though , to any longitudinal characteristic . even if longitudinal vectors are parallel , the difference in transversal fields must account for energy deviations . taking two identical @xmath47vectors and assuming , that every angle of deviation is equally probable , then the @xmath47vector after interaction will be in average : @xmath253 and if @xmath254 is considered as the undisturbed interaction energy , then the average deviations are given by : @xmath255 to estimate the importance of polarizations consider two electrons with arbitrary initial velocities and different orientation of intrinsic electromagnetic fields to interact . then the electromagnetic fields of the photons of interaction will show successively shifted angles and the kinetic characteristics of particle motion will be affected in a way not described by electrodynamics . it is this process , which must ultimately lead to a shift of scattering cross sections already predicted and confirmed by quantum theory . as a mathematical framework to calculate scattering processes on a dynamical basis complicated by the fact that the standard scattering approximations are bound to fail , since no potential can be defined for the angular dependency of photon interaction has not yet been developed , the exact changes compared to quantum theory can not currently be exhibited . david bohm devised a gedankenexperiment to differ between the theory of hidden variables and quantum theory @xcite , which was based on the assumption , that the direction of spin is a true but hidden variable . as bell formulated in his inequalities , the consequence of such an assumption leads to consequences , which should be , in principle , measurable @xcite . measurements of these consequences by aspect seemed to confirm , that a local theory of hidden variables is contradictory @xcite . any theory , thus the result , which assumes hidden variables of this type , is contradicted by measurements . this conclusion is not undisputed , as thompson recently showed by the chaotic ball model of this type of measurements @xcite . the usual interpretation , the author concluded , which assumes quantum theory to be superior to local hidden variable theories , contains a biased interpretation of experimental data . the theoretical framework of material waves , which can be seen as an extension of electrodynamics to intrinsic particle properties , is basically a local theory . particle spin is a local variable , and the theory is therefore a realistic and local theory of hidden variables , although , and this seems to be an important difference to bohm s concept , classical electrodynamics _ and _ quantum theory were shown to result from specific limitations . on these grounds the conflict of a measurement presumably excluding hidden variables in a local theory and a concept which explicitly goes beyond the limits of formalization of the quantum theory , on which evaluation of these measurements is based ( like the current framework ) , can only be clarified by exclusion . if these measurements are also in the current framework valid _ within _ the limits of quantum theory , then the current framework must be ruled out , since it contradicts experimental evidence of a highly consistent and successful theory . but if , on the other hand , these measurements do not yield valid results within the framework of quantum theory , then this experiment is not conclusive , because its interpretation presupposes theoretical formulations which have been violated . in this case an alternative interpretation of the measurements is possible , which is based on a different theory . although it can not currently be determined _ what _ was actually measured in aspect s experiments , it can be excluded that the measured result was obtained within the limits of quantum theory . and since bell s inequalities , which provided the theoretical background of interpretation , are based on the axioms of quantum theory , the interpretation of these results is questionable . we will return to these results in later publications . in the present framework spin is local and parallel to the polarization of the intrinsic magnetic field @xmath71 which can be shown as follows : the energy relations in classical electrodynamics yield , for a mass of magnetic moment @xmath256 in a magnetic field @xmath71 , the following energy : @xmath257 the following deduction is based on four assumptions : ( 1 ) the energy of an electron or photon is equal to the energy of the particle in quantum theory , ( 2 ) the magnetic field is the intrinsic ( photon ) or external ( electron ) magnetic field , ( 3 ) the frequency @xmath26 is also a frequency of rotation , and ( 4 ) the magnetic moment is described by the relations in quantum theory . with ( 1 ) the energy of photons ( total energy ) and electrons ( only kinetic energy ) will be : @xmath258 with ( 2 ) the magnetic fields of electrons and photons will be ( see eq . [ ep001 ] ) : @xmath259 and with ( 4 ) the magnetic moment will be ( @xmath260 denotes the gyromagnetic ratio ) : @xmath261 calculating the energy of interaction and considering , that the magnetic field in the current framework is @xmath96 times the magnetic field in classical electrodynamics ( see eq . [ ed009 ] ) , we get for the photon : @xmath262 with @xmath263 , which follows from the definition of magnetic fields ( see section [ ed_units ] ) , the relation can only hold for constant @xmath264 if : @xmath265 the energy relation of electrodynamics in this case is only consistent with the energy of photons of they possess intrinsic spin @xmath266 , a gyromagnetic ratio of 1 , and if the direction of spin polarization is equal to the direction of intrinsic magnetic fields . a consistent framework therefore requires ( 1 ) a localized spin , and ( 2 ) a spin @xmath266 of bosons . the same calculation for the properties of electron spin yields the following result : @xmath267 in principle the spin variable and the gyromagnetic ratio could be chosen like the values for photons . from the condition itself , the difference between bosons and fermions is not directly accessible . the difference can be accounted for , if the atomic model in quantum theory is considered : the theory of hydrogen fine structures by uhlenbeck and goudsmit @xcite requires two energy levels symmetric to the undisturbed one , thus a general multiplicity of 2 . the only solution , with this initial assumption , is a spin variable of @xmath268 and a gyromagnetic ratio of 2 . the direction of spin polarization is also for fermions equal to the direction of the magnetic field vector . generally the concept of spin seems questionable , because the frequency @xmath26 is not at all to be confused with rotation frequencies of a particle . consider , for example , a photon in uniform motion @xmath96 along a flat space time geometry . its intrinsic magnetic field does not determine any state of rotation , nor does the energy of a photon result from the interaction of a photon magnetic moment with a magnetic field . from the given deduction of spin properties of particles ( and the deduction seems highly appropriate , considering the effect of spin operators in the framework of quantum theory ) it seems , that the concept itself must be theoretically deficient . let us consider , on this basis , the result of any realistic measurement of photon spin in an epr like situation @xcite . the first problem , we are confronted with , is the exact definition of the spin value . from the given deduction it seems evident , that the local direction of the spin vector is either parallel or antiparallel to the intrinsic magnetic field . provided , we adopt the view of quantum theory for spin conservation of two photons , then spin up would mean , that the vector is antiparallel to @xmath71 , spin down means , it must be parallel . in this case we have two photons with an arbitrary angle @xmath269 of polarization in the state + 1 for photon one and -1 for photon two ( see fig . [ fig001 ] ) . = 1.0 the second problem originates from the periodic features of the intrinsic magnetic field . since the magnetic field vector is periodic with @xmath270 , the spin variable _ can not _ remain constant , but must oscillate from + 1 to - 1 . this oscillation requires , that we define the exact moment of measurement to obtain any valid result on the state of either particle . if the photon path remains constant in measurements , this requirement is fulfilled , but if , like in aspect s measurements @xcite , the measuring apparatus has to be rotated to account for different angles of observation , the effect on measurements seems not easy to estimate . this problem therefore leads to a problem of experimental methods , which is hidden in the usual argumentation of quantum theory . provided , the problem can be accounted for , our next problem is more substantial . since the spin variable oscillates , the measurement can only yield a definite result , if the variable is measured in the time interval @xmath271 , which requires , that the local resolution of the measuring apparatus is below @xmath272 . now as demonstrated in the deduction of the uncertainty relations , this interval is equal to the lowest limit of local resolution in the framework of quantum theory . thus we get the result , that a valid measurement of the spin variable must have a local precision which is explicitly beyond the fundamental uncertainty in quantum theory . if , therefore , the measurements yield any correlation related to physical processes , then the fundamental axioms of quantum theory must be violated . this result seems to turn the controversy about the epr paradox into a new direction . it can be established , that the mathematical framework of quantum theory , describing the state of either particle by a superposition of the two possible states @xmath244 1 is correct . but it can equally be established , that a significant correlation in spin measurements is not , as einstein suspected , a proof of hidden variables or a contradiction in terms of special relativity , but a proof of measurements below the fundamental level of uncertainty . if this sort of measurement is possible , then the uncertainty relations are violated . * if epr measurements are possible and yield a significant correlation for a pair of spin particles , then the uncertainty relations are violated . in section [ sec_ed ] it was established , that electromagnetic radiation consists of two distinct intrinsic variables : ( 1 ) a longitudinal momentum , and ( 2 ) a transversal electromagnetic field , both variables depending on the frequency of radiation . in this section we found , that the change of energy density due to the adsorption of a photon result in motion of the absorbing electron , the energy increase given by : @xmath273 if monochromatic x rays therefore interact with free electrons , i.e. electrons with a velocity small compared to @xmath274 , then the absorbing electrons should move in longitudinal direction with a velocity equal to : @xmath275 if this electron is decelerated , then it will emit electromagnetic radiation with a frequency described by : @xmath276 and , depending on the duration of the deceleration process , the radiation can possess arbitrarily low frequency . this _ primary _ emission therefore will not exhibit discrete frequency levels . but if _ secondary _ processes are considered , then the frequency change will depend on ( 1 ) the relative velocity between the source of radiation and the moving electron , and ( 2 ) the emission and absorption characteristics . assuming , that the emission and absorption processes are symmetric in terms of duration , then the frequency of secondary emission will depend on the original frequency as well as the angle of emission . the model presented is therefore a simplification , and it can not be excluded , that a precise and deterministic picture of the whole process will not yield a deviating result . but compared to the standard model in quantum theory @xcite , it is nevertheless an improvement , because it does not have to recur to mechanical concepts . if the electron moves in longitudinal direction and with a velocity equal to @xmath277 , the frequency @xmath278 of the source will be changed due to the electromagnetic doppler effect . the secondary absorption process therefore has to account for a frequency @xmath279 : @xmath280 an observer moving with the velocity @xmath277 of the electron will therefore measure changed frequencies and wavelengths of secondary emission . the wavelength @xmath281 will be : @xmath282 the second term , the shift of wavelength due to longitudinal motion of the electron relative to the source of radiation , is the _ compton wavelength _ of an electron @xcite . it is the general shift of wavelength for every electromagnetic field , and an expression of longitudinal motion . @xmath283 additionally it has to be considered , that the reference frame of observation is the same as the reference frame of the original radiation , thus a second frequency shift depends on the angle of observation @xmath269 . measured from the direction of field propagation this additional shift will be : @xmath284 and the overall shift of wavelength measured is therefore : @xmath285 the result is equivalent to compton s measurements . compton scattering , in our view , is therefore a result of ( 1 ) secondary absorption and emission processes , and of ( 2 ) successive doppler shifts due to longitudinal motion of the emitting electrons . that classical electrodynamics does not provide a theoretical model for compton scattering , can be understood as a consequence of neglecting longitudinal and kinetic features of photons . since electromagnetic radiation in electrodynamics is interpreted from transversal properties alone ( which are periodic ) , a uniform longitudinal motion due to absorption processes is not within its theoretical limits . the possibility to account for quantum effects with the modified and still continuous model of electrons is strongly supporting , in our view , the validity of the unification of electrodynamics and quantum theory by formalizing the intrinsic features of particles . as it turns out , retrospectively , the theory of material waves removes the two fundamental problems , which made a physical interpretation ( as opposed to the probability interpretation @xcite ) of intrinsic wave properties impossible . ( 1 ) the contradiction between phase velocity and mechanical velocity of a particle wave is removed by the discovery of intrinsic potentials . ( 2 ) the problem of intrinsic coulomb interactions is removed , because electrostatic interactions can be referred to an exchange of photons . a particle in uniform motion is therefore stable . to exhibit the full range of logical and physical implications of the theory developed , we repeat the essential results of this essay on material waves , displayed throughout the preceding sections of this paper : 1 . the interpretation of particles as physical waves allows the conclusion , that quantum theory can not be , in a logical sense , a complete theoretical description of micro physical systems . 2 . _ because _ quantum theory neglects intrinsic characteristics of particle motion , the schrdinger equation is generally not precise by a value described by heisenberg s uncertainty relations . electrodynamics and wave mechanics are theoretical concepts based on the same physical phenomena , the intrinsic features of moving particles . 4 . the wave equations of intrinsic particle structures are lorentz invariant only for an integral evaluation of energy . the total intrinsic potentials increase for every transformation into a moving frame of reference . if intrinsic features of particles and their interactions are considered , quantization has to be interpreted as a result of energy transfer processes and their specific properties . the elementary quantum of electron charge derives from energy transfer rates in electrostatic interactions , which are an expression of the interaction process accomplished by photons . 7 . on the basis of electron photon interactions planck s constant in material wave theory only applies to interactions and transfer processes . a justification of discrete frequencies , mass or charge values from the viewpoint of allowed _ states _ can not be given . the result indicates , that the formalism of quantum theory lacks justification if it refers discrete energy levels to isolated particles . the interpretation of electrostatic interactions as an exchange of photons , proper to material wave theory , has the consequence , that ( 1 ) the properties of single charges can only be referred to interactions with the environment , ( 2 ) photons can only be interpreted as particles if two identical charges interact , and ( 3 ) electrostatic potentials depend on the possibility of photon exchanges . 9 . if epr measurements are possible and yield a significant correlation for a pair of spin particles , then the uncertainty relations are violated . = 1.0 the theoretical framework thus provides definite answers to the epr paradox by rejecting the copenhagen interpretation of quantum theory @xcite , because ( 1 ) quantum theory is not a complete account of micro physical systems , since the wave function possesses physical significance , and ( 2 ) the theory is logically inconsistent , since it is applicable only to interaction processes , while its mathematical formulation suggests the results to derive from properties of individual particles . concerning renninger s objections on the quality of photons derived from thought experiments @xcite , it provides a model of photons where ( 1 ) the energy is homogeneously distributed over the whole region of electromagnetic fields , ( 2 ) the kinetic features of photons derive from a longitudinal component not accounted for in electrodynamics , and ( 3 ) the energy transfer in interactions applies for every single point of a given region , which yields the quantum effect of interaction . it is difficult to estimate either the impact or the subsequent changes due to a concept , which completely deviates from the mainstream of current physics . the aim of establishing and , to a limited extent , interpreting a physical reality by an alternative framework was to provide a means , by which experimental results can be seen from at least two different points of view . the fundamental idea , it seems , is not very different from the conceptions of einstein , schrdinger , bohm , or de broglie ( see the introduction ) , but the disregard of well established interpretations is considerably higher . this was not , initially , intended . it only turned out , that a theoretical concept limited to only minor changes will neither provide a scientifically satisfying solution since it still is based on some _ have to believe _ nor does it really offer any new insight into micro physical processes . it was not intended to recover all the results of current quantum physics , not least , because it would surely require many lifetimes to succeed . but it was intended , and , so i think , accomplished rather successfully , to extend the theoretical framework of micro physics beyond the uncertainty relations and the _ we can not know _ of quantum theory . whether the result presented it the best possible way out of the dilemma arising from a physics of quantities mechanics confronted with a world of intensities electrodynamics can not be currently estimated . to estimate the epistemological changes implied by the theoretical framework , fig . [ fig002 ] displays the logical structure of micro physics in material wave theory . to estimate , whether the new theory contradicts any experimental evidence within its range of formalization , it has to be considered , that it provides a meta theory combining the different frameworks of electrodynamics , quantum theory , and quantum electrodynamics in a single consistent framework . from the viewpoint of electrodynamics , it will reproduce any result the current framework provides , because the basic relations are not changed , merely interpreted in terms of electron and photon propagation . therefore an experiment , consistent with electrodynamics , will also be consistent with the new theory . in quantum theory , the only statements additional to the original framework are beyond the level of experimental validity , defined by the uncertainty relations . since quantum theory can not , in principle , contain results beyond that level , every experiment within the framework of quantum theory must necessarily be reproduced . that applies to the results of wave mechanics , which yields the eigenvalues of physical processes , as well as to the results of operator calculations , if von neumann s proof of the equivalence of schrdinger s equation and the commutator relations is valid . the only result , which could in this respect disprove the theory , is the spreading of a wave packet : but as this experimental result was the reason , that a _ physical _ interpretation of wave functions had to be discarded , the theory also in this respect seems to stand on firm ground . in quantum field theory , the same rule applies : the new theory only states , what is , in quantum theory , no part of a measurement result , therefore a contradiction can not exist . if , on the other hand , results in quantum electrodynamics exist , which are not yet verified by the new theory , then it is rather a problem of further development , but not of a contradiction with existing measurements . experimentally , the new theory can not be disproved by existing measurements , because its statements at once verify the existing formulations and extend the framework of theoretical calculations . the same does not hold , though , for the verification of existing theoretical schemes of calculation by the new framework . since the theory extends far beyond the level of current concepts , established theoretical results may well be subject to revision . the theory of material waves requires , that every valid solution of a microphysical problem can be referred to physical properties of particle waves . in cases , where the established theory provides a solution yielding , for example , correct eigenvalues but not wave functions which can be identified as _ physical _ waves , the result is , from this viewpoint , questionable . essentially , this is not limiting nor diminishing the validity of current results , since it extends the framework of micro physics only to regions , which so far have remained unconsidered . from the five tasks defined in the introduction as our personal criterium for a promising new approach , three have been accomplished by this publication . but as quantum theory , historically , was invented due to the puzzling results in atomic physics , the whole concept as such would be useless without a theory of atoms compatible with the framework developed . this theory has partly been published already ( see @xcite ) , partly it is still in the making although the essential results have been derived . it will be the aim of the next paper published in the internet , to give a thorough and up to date account of a model of hydrogen atoms in terms of material waves . schweber s. _ qed and the men who made it _ , princeton ( 1994 ) bohm d. _ causality and chance in modern physics _ , princeton ( 1957 ) de broglie l. in his foreword to @xcite dirac p.a.m . j. phys . _ * 5 * , 65 ( 1984 ) rseberg u. in gtschl j. ( ed . ) _ erwin schrdinger s world view _ , dordrecht ( 1992 ) p. 135 feynman r. _ qed _ , princeton ( 1988 ) p.9 schrdinger , e. _ ann . physik _ * 79 * , 361 ; 489 ( 1926 ) heisenberg w. `` remarks on the origin of uncertainty '' in price w. c. ( ed . ) _ the uncertainty principle and the foundations of quantum mechanics _ , wiley & sons , new york ( 1977 ) bohr n. `` discussion with einstein on epistemological problems in atomic physics '' , in _ a. einstein : philosopher scientist _ , schilpp p.a . ( ed . ) , harper and row , new york ( 1959 ) de broglie l. _ ann . phys . _ * 3 * , 22 ( 1925 ) born , m. _ z. physik _ * 37 * , 863 ( 1926 ) einstein a. , rosen n. , and podolsky b. _ phys . _ * 47 * , 180 ( 1935 ) kwiat p. , weinfurter h. , herzog t. , zeilinger a. , and kasevich m.a . lett . _ * 74 * , 4763 ( 1995 ) renninger m. _ zeitschr . f. phys . _ * 158 * , 417 ( 1960 ) bohm , d. _ phys . rev . _ * 85 * , 166 ; 180 ( 1952 ) bell j.s . _ physics _ * 1 * , 195 ( 1964 ) ; see also his remarks on hidden variable theories in _ rev . mod . phys . _ * 38 * , 447 ( 1966 ) davisson c. and germer l.h . rev . _ * 30 * , 705 ( 1927 ) hofer w. a. internal energy of free particles : de broglie waves in electrodynamics . tech . _ * 18 * , 157 ( 1995 ) hofer w. a. an inertial model of atoms : atomic spectra as a resonance problem . _ specs . _ , in print born m. , heisenberg w. , and jordan p. _ z. physik _ * 35 * , 557 ( 1926 ) planck m. _ ann . physik _ * 4 * , 553 ( 1901 ) heisenberg w. _ z. physik _ * 43 * , 172 ( 1927 ) cohen tannoudji c. , diu b. , and laloe f. _ quantum mechanics _ , wiley & sons , new york ( 1977 ) einstein a. _ ann . * 17 * , 891 ( 1905 ) sommerfeld a. _ elektrodynamik _ , akademische verlagsgesellschaft , reprint leipzig ( 1964 ) michelson , and morley . j. sci . _ * 34 * , 333 ( 1887 ) ; _ phil . mag . _ * 24 * , 449 ( 1887 ) jackson j. _ classical electrodynamics _ , wiley & sons , new york ( 1984 ) einstein a. _ ann . physik _ * 17 * , 132 ( 1905 ) jordan p. _ anschauliche quantentheorie _ , berlin ( 1936 ) p. 115 heisenberg w. _ die physikalischen grundlagen der quantentheorie _ , leipzig ( 1941 ) p. 7 and 107 renninger m. _ z. physik _ * 136 * , 251 ( 1953 ) a. einstein . * 20 * , 627 ( 1906 ) uipap document uip 20 . _ physica a _ * 93 * , 1 ( 1978 ) bohm d. _ quantum mechanics _ , prentice hall , eaglewood cliffs ( 1951 ) aspect a. _ trois tests experimenteaux des inegalites de bell par mesure de correlation de polarisation de photons _ , ph . d. thesis , universite de paris sud , centre dorsay ( 1983 ) thompson c. h. _ found . lett . _ * 9 * , 357 ( 1996 ) uhlenbeck g. e. and goudsmit s. a. _ naturwiss . _ * 13 * , 953 ( 1925 ) ; and _ nature _ * 117 * , 264 ( 1925 ) debye , p. _ phys . zeitschr . _ * 24 * , 161 ( 1923 ) compton , a. h. _ phys . * 22 * , 409 ( 1923 )

the motivation for the study reported in this paper comes from the celebrated and long - standing problem , originally posed by erds @xcite in 1946 , of obtaining a sharp lower bound for the number of distinct distances guaranteed to exist in any set @xmath0 of @xmath1 points in the plane . erds has shown that a section of the integer lattice determines only @xmath10 distinct distances , and conjectured this to be a lower bound for any planar point set . in spite of steady progress on this problem , reviewed next , erds s conjecture is still open . l. moser @xcite , chung @xcite , and chung @xcite proved that the number of distinct distances determined by @xmath1 points in the plane is @xmath11 , @xmath12 , and @xmath13 , respectively . szkely @xcite managed to get rid of the polylogarithmic factor , while solymosi and tth @xcite improved this bound to @xmath14 . this was a real breakthrough . their analysis was subsequently refined by tardos @xcite and then by katz and tardos @xcite , who obtained the current record of @xmath15 , for any @xmath16 , which is @xmath17 . in this paper we transform the problem of distinct distances in the plane to an incidence problem between points and a certain kind of curves ( helices or parabolas ) in three dimensions . as we show , sharp upper bounds on the number of such incidences translate back to sharp lower bounds on the number of distinct distances . incidence problems in three dimensions between points and curves have been studied in several recent works @xcite , and a major push in this direction has been made last year , with the breakthrough result of guth and katz @xcite , who have introduced methods from algebraic geometry for studying problems of this kind . this has been picked up by the authors @xcite , where worst - case tight bounds on the number of incidences between points and lines in three dimensions ( under certain restrictions ) have been obtained . the present paper serves two purposes . first , it studies in detail the connection between the distinct distances problem and the corresponding 3-dimensional incidence problem . as it turns out , there is a lot of interesting geometric structure behind this reduction , and the paper develops it in detail . we offer several conjectures on the number of incidences , and show how , if true , they yield the almost tight worst - case lower bound @xmath9 on the number of distinct distances . unfortunately , so far we have not succeeded in proving these conjectures . nevertheless , we have made considerable progress on the incidence problem itself , which is the second purpose of the study in this paper . we show how to adapt the algebraic machinery of @xcite to derive sharp bounds for the incidence problem . @xcite to derive sharp bounds for the incidence problem . these bounds are very similar to , and in fact even better than the bounds obtained in @xcite for point - line incidences , where they have been shown to be worst - case tight . however , they are not ( yet ) good enough to yield significant lower bounds for distinct distances . we believe that there is additional geometric structure in the particular problem studied here , which should enable one to further improve the bounds , but so far this remains elusive . the paper is organized as follows . we first describe the reduction from the planar distinct distances problem to the 3-dimensional incidence problem mentioned above . in doing so , we note and explore several additional geometric connections between the two problems ( as manifested , e.g. , in the analysis of _ special surfaces _ given below ) . we then present the tools from algebraic geometry that are needed to tackle the incidence problem ; they are variants of the tools used in @xcite , adapted to the specific curves that we need to handle . we then go on to bound the number of incidences . we first bound the number of rotations in terms of the number of parabolas , and then bound the number of incidences themselves . the latter task is achieved in two steps . we first use a `` purely algebraic '' analysis , akin to those in @xcite , to obtain a weaker bound , which we then refine in the second step , using more traditional space decomposition techniques . the final bound is still not as good as we would like it to be , but it shows that the case studied in this paper `` behaves better '' than its counterpart involving lines . [ [ distinct - distances - and - incidences - with - helices . ] ] distinct distances and incidences with helices . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we offer the following novel approach to the problem of distinct distances . [ [ h1-notation . ] ] ( h1 ) notation . + + + + + + + + + + + + + + let @xmath0 be a set of @xmath1 points in the plane with @xmath18 distinct distances . let @xmath19 denote the set of all quadruples @xmath20 , such that the pairs @xmath21 and @xmath22 are distinct ( although the points themselves need not be ) and @xmath23 . let @xmath24 denote the @xmath18 distinct distances in @xmath0 , and let @xmath25 . we have @xmath26 ^ 2 = \frac{\left[s(s-1)-x\right]^2}{x } .\ ] ] [ [ h2-rotations . ] ] ( h2 ) rotations . + + + + + + + + + + + + + + + we associate each @xmath27 with a ( unique ) _ rotation _ ( or , rather , a rigid , orientation - preserving transformation of the plane ) @xmath28 , which maps @xmath29 to @xmath30 and @xmath31 to @xmath32 . a rotation @xmath28 , in complex notation , can be written as the transformation @xmath33 , where @xmath34 and @xmath35 . putting @xmath36 , @xmath37 , we can represent @xmath28 by the point @xmath38 . in the planar context , @xmath39 is the counterclockwise angle of the rotation , and the center of rotation is @xmath40 , which is defined for @xmath41 ; for @xmath42 , @xmath28 is a pure translation . the _ multiplicity _ @xmath43 of a rotation @xmath28 ( with respect to @xmath0 ) is defined as @xmath44 the number of pairs @xmath45 such that @xmath46 . clearly , one always has @xmath47 , and we will mostly consider only rotations satisfying @xmath48 . as a matter of fact , the bulk of the paper will only consider rotations with multiplicity at least @xmath49 . rotations with multiplicity @xmath50 are harder to analyze . if @xmath51 then @xmath0 contains two congruent and equally oriented copies @xmath52 of some @xmath5-element set , such that @xmath53 . thus , studying multiplicities of rotations is closely related to analyzing repeated ( congruent and equally oriented ) patterns in a planar point set ; see @xcite for a review of many problems of this kind . [ [ anti - rotations . ] ] anti - rotations . + + + + + + + + + + + + + + + in this paper we will also consider _ anti - rotations _ , which are rigid , orientation - reversing transformations of the plane . any anti - rotation can be represented as a rotation , followed by a reflection about some fixed line , e.g. , the @xmath18-axis ( so , in complex notation , this can be written as @xmath54 ) . anti - rotations will be useful in certain steps of the analysis . [ [ h3-bounding - k . ] ] ( h3 ) bounding @xmath55 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + if @xmath51 then @xmath28 contributes @xmath56 quadruples to @xmath19 . let @xmath57 ( resp . , @xmath58 ) denote the number of rotations with multiplicity exactly @xmath5 ( resp . , at least @xmath5 ) , for @xmath8 . then @xmath59 [ [ h4-the - main - conjecture . ] ] ( h4 ) the main conjecture . + + + + + + + + + + + + + + + + + + + + + + + + + [ conj1 ] for any @xmath60 , we have @xmath61 suppose that the conjecture were true . then we would have @xmath62 ^ 2}{x } \le |k| = o(s^3 ) \cdot \left [ 1 + \sum_{k\ge 3 } \frac{1}{k } \right ] = o(s^3\log s ) , \ ] ] which would have implied that @xmath63 . this would have almost settled the problem of obtaining a tight bound for the minimum number of distinct distances guaranteed to exist in any set of @xmath1 points in the plane , since , as mentioned above , the upper bound for this quantity is @xmath10 @xcite . we note that conjecture [ conj1 ] is rather deep ; even the simple instance @xmath64 , asserting that there are only @xmath3 rotations which map ( at least ) two points of @xmath0 to two other points of @xmath0 ( at the same distance apart ) , seems quite difficult . in this paper we establish a variety of upper bounds on the number of rotations and on the sum of their multiplicities . in particular , these results provide a partial positive answer , showing that @xmath65 ; that is , the number of rotations which map a ( degenerate or non - degenerate ) triangle determined by @xmath0 to another congruent ( and equally oriented ) such triangle , is @xmath3 . bounding @xmath66 by @xmath3 is still an open problem . see section [ sec : impr ] for a simple proof of the weaker bound @xmath67 . [ [ lower - bound . ] ] lower bound . + + + + + + + + + + + + we next give a construction ( suggested by haim kaplan ) which shows : [ lem : lower ] there exist sets @xmath0 in the plane of arbitrarily large cardinality , which determine @xmath68 distinct rotations , each mapping a triple of points of @xmath0 to another triple of points of @xmath0 . * proof : * consider the set @xmath69 , where @xmath70 see figure [ lower ] . lower.pstex_t for each triple @xmath71 such that @xmath72 also belongs to @xmath73 , construct the rotation @xmath74 which maps @xmath75 to @xmath76 and @xmath77 to @xmath78 . since the distance between the two source points is equal to the distance between their images , @xmath74 is well ( and uniquely ) defined . moreover , @xmath74 maps the midpoint @xmath79 to the midpoint @xmath80 . we claim that the rotations @xmath74 are all distinct . indeed , suppose that two such rotations , @xmath74 and @xmath81 , for distinct triples @xmath82 , @xmath83 , coincide ; call the common rotation @xmath28 . we can represent @xmath28 as the rigid transformation which first translates the plane horizontally by distance @xmath84 , so that @xmath75 is mapped to @xmath76 , and then rotates it around @xmath76 by an appropriate angle @xmath85 , so that @xmath86 is mapped to @xmath78 . suppose first that @xmath87 . since @xmath88 , it maps @xmath89 to @xmath90 and then rotates this point by angle @xmath39 around @xmath76 , mapping it to a point outside the @xmath18-axis , contradicting the fact that @xmath81 maps @xmath89 to @xmath91 . if @xmath92 then we also must have @xmath93 , so @xmath94 . but then it is impossible to turn , around @xmath76 , the shifted point @xmath86 to @xmath78 and the shifted point @xmath95 to @xmath96 , by the same angle , a contradiction which shows that the two rotations are distinct . since there are @xmath97 triples @xmath82 with the above properties , the claim follows . @xmath98 * remarks . * * ( 1 ) * a `` weakness '' of this construction is that all the rotations @xmath74 map a _ collinear _ triple of points of @xmath0 to another collinear triple . ( in the terminology to follow , these will be called _ flat _ rotations . ) we do not know whether the number of rotations which map a _ non - collinear _ triple of points of @xmath0 to another non - collinear triple can be @xmath99 . we tend to conjecture that this is indeed the case . * ( 2 ) * we do not know whether conjecture [ conj1 ] is worst - case tight ( if true ) . that is , we do not know whether there exist sets @xmath0 , with @xmath100 arbitrarily large , so that there are @xmath101 distinct rotations , each mapping at least @xmath5 points of @xmath0 to @xmath5 other points of @xmath0 . [ [ h5-helices . ] ] ( h5 ) helices . + + + + + + + + + + + + + to estimate @xmath58 , we reduce the problem of analyzing rotations and their interaction with @xmath0 to an incidence problem in three dimensions , as follows . with each pair @xmath45 , we associate the curve @xmath102 , in a 3-dimensional space parametrized by @xmath103 , which is the locus of all rotations which map @xmath29 to @xmath31 . that is , the equation of @xmath102 is given by @xmath104 putting @xmath105 , @xmath106 , this becomes @xmath107 this is a _ helix _ in @xmath2 , having four degrees of freeedom , which we parametrize by @xmath108 . it extends from the plane @xmath42 to the plane @xmath109 ; its two endpoints lie vertically above each other , and it completes exactly one revolution between them . [ [ h6-helices - rotations - and - incidences . ] ] ( h6 ) helices , rotations , and incidences . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath110 be a set of rotations , represented by points in @xmath2 , as above , and let @xmath111 denote the set of all @xmath112 helices @xmath102 , for @xmath45 ( note that @xmath113 is permitted ) . let @xmath114 denote the number of incidences between @xmath110 and @xmath111 . then we have @xmath115 rotations @xmath28 with @xmath116 are not interesting , because each of them only contributes @xmath117 to the count @xmath114 , and we will mostly ignore them . for the same reason , rotations with @xmath118 are also not interesting for estimating @xmath114 , but they need to be included in the analysis of @xmath119 . unfortunately , as already noted , we do not yet have a good upper bound ( i.e. , cubic in @xmath1 ) on the number of such rotations . [ [ h7-incidences - and - the - second - conjecture . ] ] ( h7 ) incidences and the second conjecture . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + [ conj2 ] for any @xmath110 and @xmath111 as above , we have @xmath120 suppose that conjecture [ conj2 ] were true . let @xmath121 denote the set of all rotations with multiplicity at least @xmath5 ( with respect to @xmath0 ) . we then have @xmath122 from which we obtain @xmath123 thus establishing conjecture [ conj1 ] , and therefore also the lower bound for @xmath18 ( the number of distinct distances ) derived above from this conjecture . * remark . * conjecture [ conj2 ] can also be formulated for an _ arbitrary _ subset @xmath111 of all possible helices . note that two helices @xmath102 and @xmath124 intersect in at most one point this is the unique rotation which maps @xmath29 to @xmath31 and @xmath125 to @xmath126 ( if it exists at all , namely if @xmath127 ) . hence , combining this fact with a standard cutting - based decomposition technique , similar to what has been noted in @xcite , say , yields the weaker bound @xmath128 which , alas , only yields the much weaker bound @xmath129 , which is completely useless for deriving any lower bound on @xmath18 . ( we will use this bound , though , in section [ sec : conc ] . ) [ [ h8-from - helices - to - parabolas . ] ] ( h8 ) from helices to parabolas . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the helices @xmath102 are non - algebraic curves , because of the use of the angle @xmath39 as a parameter . this can be easily remedied , in the following standard manner . assume that @xmath39 ranges from @xmath130 to @xmath131 , and substitute , in the equations ( [ eq : helix ] ) , @xmath132 , @xmath133 , and @xmath134 , to obtain @xmath135 which are the equations of a _ planar parabola _ in the @xmath136-space . ( the parabola degenerates to a line if @xmath137 , a situation that we will rule out by choosing an appropriate generic coordinate frame in the original @xmath138-plane . ) we denote the parabola corresponding to the helix @xmath102 as @xmath139 , and refer to it as an _ @xmath140-parabola_. [ [ h9-joint - and - flat - rotations . ] ] ( h9 ) joint and flat rotations . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a rotation @xmath141 is called a _ joint _ of @xmath111 if @xmath28 is incident to at least three helices of @xmath111 whose tangent lines at @xmath28 are non - coplanar . otherwise , still assuming that @xmath28 is incident to at least three helices of @xmath111 , @xmath28 is called _ let @xmath142 be a rotation , incident to three distinct helices @xmath102 , @xmath124 , @xmath143 . from their equations , as given in ( [ eq : helix ] ) , the directions of the tangents to these helices at @xmath28 are @xmath144 put @xmath145 and @xmath146 . then the three tangents are coplanar if and only if @xmath147 simplifying the determinant , and recalling that @xmath148 , the condition is equivalent to @xmath149 in other words , the three helices @xmath102 , @xmath124 , @xmath143 form a joint at @xmath28 if and only if the three points @xmath150 ( and thus also @xmath151 ) are non - collinear . that is , we have shown : [ jointri ] a rotation @xmath28 is a joint of @xmath111 if and only if @xmath28 maps a non - degenerate triangle determined by @xmath0 to another ( congruent and equally oriented ) non - degenerate triangle determined by @xmath0 . a rotation @xmath28 is a flat rotation if and only if @xmath28 maps at least three collinear points of @xmath0 to another collinear triple of points of @xmath0 , but does not map any point of @xmath0 outside the line containing the triple to another point of @xmath0 . * remarks : * * ( 1 ) * note that if @xmath28 is a flat rotation , it maps the entire line containing the three source points to the line containing their images . specifically ( see also below ) , we can respectively parametrize points on these lines as @xmath152 , @xmath153 , for @xmath154 , such that @xmath28 maps @xmath152 to @xmath153 for every @xmath155 . * ( 2 ) * for flat rotations , we also need to ensure , for technical reasons , that the three ( or more ) helices incident to a flat rotation @xmath28 are such that their tangents at @xmath28 are all distinct . this fortunately is always the case . indeed , the preceding analysis is easily seen to imply that if @xmath102 and @xmath124 meet at @xmath28 then their tangents at @xmath28 coincide if and only if @xmath156 . but then @xmath102 and @xmath157 can not have a common point ( rotation ) unless @xmath158 too , i.e. , they are the same helix ; otherwise the common rotation would have to map @xmath29 to the two distinct points @xmath31 and @xmath126 , an impossibility . [ [ h10-special - surfaces . ] ] ( h10 ) special surfaces . + + + + + + + + + + + + + + + + + + + + + + + in preparation for the forthcoming algebraic analysis , we need the following property of our helices . let @xmath28 be a flat rotation , with multiplicity @xmath4 , and let @xmath159 and @xmath160 be the corresponding lines in the plane , such that there exist @xmath5 points @xmath161 and @xmath5 points @xmath162 , such that @xmath28 maps @xmath163 to @xmath164 for each @xmath165 ( and in particular maps @xmath159 to @xmath160 ) . by definition , @xmath28 is incident to the @xmath5 helices @xmath166 , for @xmath167 . let @xmath168 and @xmath169 denote unit vectors in the direction of @xmath159 and @xmath160 , respectively . clearly , there exist two reference points @xmath170 and @xmath171 , such that for each @xmath165 there is a real number @xmath172 such that @xmath173 and @xmath174 . as a matter of fact , for each real @xmath155 , @xmath28 maps @xmath175 to @xmath176 , so it is incident to @xmath177 . note that @xmath29 and @xmath31 are not uniquely defined : we can take @xmath29 to be any point on @xmath159 , and shift @xmath31 accordingly along @xmath160 . let @xmath178 denote the set of these helices . since a pair of helices can meet in at most one point , all the helices in @xmath178 pass through @xmath28 but are otherwise pairwise disjoint . using the re - parametrization @xmath179 , we denote by @xmath180 the surface which is the union of all the @xmath140-parabolas that are the images of the helices in @xmath178 . we refer to such a surface @xmath181 as a _ special surface_. an important comment is that most of the ongoing analysis also applies when only two helices are incident to @xmath28 ; they suffice to determine the four parameters @xmath182 that define the surface @xmath181 . we also remark that , although we started the definition of @xmath183 with a flat rotation @xmath28 , the definition only depends on the parameters @xmath184 , and @xmath169 ( and even there we have , as just noted , one degree of freedom in choosing @xmath29 and @xmath31 ) . if @xmath28 is not flat it may determine many special surfaces , one for each line that contains two or more points of @xmath0 which @xmath28 maps to other ( also collinear ) points of @xmath0 . also , as we will shortly see , the same surface can be obtained from a different set ( in fact , many such sets ) of parameters @xmath185 , and @xmath186 ( or , alternatively , from different flat rotations @xmath187 ) . an `` intrinsic '' definition of special surfaces will be given shortly . the surface @xmath181 is a cubic algebraic surface , whose equation can be worked out as follows . the equation of the parabola @xmath188 corresponding to @xmath177 is @xmath189 we can view this as a parametrization of @xmath181 using @xmath155 and @xmath190 as parameters . we can simplify these equations as @xmath191 where @xmath192 are quadratic polynomials in @xmath190 . eliminating @xmath155 from these equations gives us the first version of the equation of @xmath181 , which is @xmath193 this is a quartic equation , although it is only linear in @xmath194 and @xmath195 . note also that the cross - section of @xmath181 by any plane @xmath196 is a line , so @xmath181 is a ruled surface . we next reduce ( [ sigma1 ] ) to a cubic equation , as follows . let @xmath197 denote the coordinates of @xmath28 in the @xmath198-frame . we note that @xmath199 . this can be worked out explicitly , or concluded by noting that @xmath197 is a common point of all our parabolas , so @xmath197 can not determine @xmath155 , meaning that the coefficients @xmath200 and @xmath201 in ( [ parsig ] ) must both be zero . hence , each of the three polynomials @xmath202 , @xmath203 , and @xmath204 , appearing in the left - hand side of ( [ sigma1 ] ) , vanishes at @xmath205 , and is therefore divisible by @xmath206 . factoring @xmath206 out , we get a reduced equation for @xmath181 , of the form @xmath207 where @xmath208 and @xmath209 are linear in @xmath190 . recalling that @xmath210 an explicit calculation yields : @xmath211 an additional explicit calculation shows that @xmath212 ( to see , say , the first equality , we need to show that @xmath213 . writing @xmath214 , @xmath215 , where @xmath39 is the angle of rotation , and recalling that @xmath216 , the claim follows by straightforward trigonometric manipulations . ) this allows us to rewrite @xmath217 hence , the `` free '' term in ( [ sigma ] ) is the cubic polynomial @xmath218 @xmath219 @xmath220 we refer to the cubic polynomial in the left - hand side of ( [ sigma ] ) as a _ special polynomial_. thus a special surface is the zero set of a special polynomial . [ [ h11-the - geometry - of - special - surfaces . ] ] ( h11 ) the geometry of special surfaces . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + special surfaces pose a technical challenge to the analysis . specifically , each special surface @xmath181 captures a certain underlying pattern in the ground set @xmath0 , which may result in many incidences between rotations and @xmath140-parabolas , all contained in @xmath181 . the next step of the analysis studies this pattern in detail . uvuv1.pstex_t consider first a simple instance of this situation , in which two special surfaces @xmath181 , @xmath221 , generated by two distinct flat rotations @xmath28 , @xmath187 , coincide . more precisely , there exist four parameters @xmath182 such that @xmath28 maps the line @xmath222 to the line @xmath223 ( so that points with the same parameter @xmath155 are mapped to one another ) , and four other parameters @xmath224 such that @xmath187 maps ( in a similar manner ) the line @xmath225 to the line @xmath226 , and @xmath227 . denote this common surface by @xmath181 . since the surfaces coincide , the coefficients @xmath228 , @xmath229 for @xmath230 must be proportional to the coefficients @xmath231 , @xmath232 for @xmath233 . that is , we must have @xmath234 and @xmath235 , for some real @xmath236 . in other words , since @xmath238 are unit vectors , the angle bisector between @xmath168 and @xmath169 must coincide with that between @xmath239 and @xmath186 , as depicted in figure [ uvuv1 ] . moreover , as is easily checked , if we let @xmath240 be the intersection point of @xmath241 and @xmath242 , and let @xmath243 be the intersection point of @xmath244 and @xmath245 , then both @xmath28 and @xmath187 map @xmath240 to @xmath243 , and @xmath246 is contained in @xmath181 . ( see figure [ uvuv2 ] . ) indeed , @xmath187 lies on some parabola @xmath247 through @xmath28 which is contained in @xmath181 , and @xmath28 lies on some parabola @xmath248 through @xmath187 which is also contained in @xmath181 . since a pair of distinct @xmath140-parabolas meet in at most one point , the two parabolas must coincide , so @xmath249 and @xmath250 . however , by construction , @xmath251 lies on @xmath241 and @xmath252 lies on @xmath242 , so this common point must be @xmath240 , and , similarly , @xmath253 , as claimed . uvuv2.pstex_t since the preceding analysis applies to any pair of distinct rotations on a common special surface @xmath181 , it follows that we can associate with @xmath181 a common direction @xmath254 and a common shift @xmath255 , so that for each @xmath256 there exist two lines @xmath159 , @xmath160 , where @xmath28 maps @xmath159 to @xmath160 , so that the angle bisector between these lines is in direction @xmath254 , and @xmath28 is the unique rigid motion , obtained by rotating @xmath159 to @xmath160 around their intersection point @xmath257 , and then shifting @xmath160 along itself by a distance whose projection in direction @xmath254 is @xmath255 . the fact that the shifts of any pair of rotations on @xmath181 have the same @xmath254-component follows from the fact that they both map the intersection point @xmath240 of their source lines to the intersection point @xmath243 of their target lines ; consult figure [ uvuv2 ] . let @xmath181 be a special surface , generated by @xmath178 ; that is , @xmath181 is the union of all parabolas of the form @xmath188 , for @xmath154 . let @xmath258 be the common rotation to all these parabolas , so it maps the line @xmath259 to the line @xmath260 , so that every point @xmath175 is mapped to @xmath176 . let @xmath261 be a parabola contained in @xmath181 but not passing through @xmath258 . take any pair of distinct rotations @xmath262 on @xmath261 . then there exist two respective real numbers @xmath263 , such that @xmath264 , for @xmath265 . thus @xmath266 is the unique rotation which maps @xmath125 to @xmath126 and @xmath173 to @xmath174 . in particular , we have @xmath267 . this in turn implies that the triangles @xmath268 and @xmath269 are congruent ; see figure [ uvuv3 ] . uvuv3.pstex_t given @xmath125 , this determines @xmath126 , up to a reflection about @xmath270 . we claim that @xmath126 has to be on the `` other side '' of @xmath270 , namely , be such that the triangles @xmath268 and @xmath269 are oppositely oriented . indeed , if they were equally oriented , then @xmath258 would have mapped @xmath125 to @xmath126 , and then @xmath261 would have passed through @xmath258 , contrary to assumption . now form the two sets @xmath271 the preceding discussion implies that @xmath272 and @xmath273 are congruent and oppositely oriented . to recap , each rotation @xmath256 , incident to @xmath8 parabolas contained in @xmath181 , corresponds to a pair of lines @xmath274 with the above properties , so that @xmath28 maps @xmath5 points of @xmath275 ( rather , of @xmath276 ) to @xmath5 points of @xmath277 ( that is , of @xmath278 ) . if @xmath28 is flat , its entire multiplicity comes from points of @xmath0 on @xmath159 ( these are the points of @xmath276 ) which are mapped by @xmath28 to points of @xmath0 on @xmath160 ( these are points of @xmath278 ) , and all the corresponding parabolas are contained in @xmath181 . if @xmath28 is a joint then , for any other point @xmath279 outside @xmath159 which is mapped by @xmath28 to a point @xmath280 outside @xmath160 , the parabola @xmath247 is not contained in @xmath181 , and crosses it transversally at the unique rotation @xmath28 . note also that any pair of parabolas @xmath281 and @xmath282 which are contained in @xmath181 intersect , necessarily at the unique rotation which maps @xmath283 to @xmath284 and @xmath285 to @xmath286 . this holds because @xmath287 , as follows from the preceding discussion . [ [ special - surfaces - are - anti - rotations . ] ] special surfaces are anti - rotations . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath181 be a special surface , and let @xmath52 be the subsets of @xmath0 associated with @xmath181 , as in ( [ aandb ] ) . then there exists a single _ anti - rotation _ which maps @xmath272 to @xmath273 . conversely , any anti - rotation can be associated with a unique special surface in this manner . however , the number of incidences within a special surface may be larger than the incidence count of the anti - rotation with the appropriate variants of the @xmath140-parabolas : the former counts incidences between the points of @xmath272 ( or of @xmath273 ) and the lines that they determine , while the latter only counts the size of @xmath272 ( or of @xmath273 ) . [ [ an - alternative - analysis . ] ] an alternative analysis . + + + + + + + + + + + + + + + + + + + + + + + + recall the equation ( [ sigma ] ) of @xmath181 @xmath288 where , writing @xmath289 and @xmath290 , @xmath291 now let @xmath139 be a parabola contained in @xmath181 . substituting the equations ( [ parabola ] ) of @xmath139 into the above equation , we get @xmath292 - ( \lambda z + \mu ) \biggl [ ( a_2+b_2)z^2 - 2a_1z + ( b_2-a_2 ) \biggr ] + k(z ) \equiv 0 , \ ] ] where @xmath293 is the `` free '' cubic term in the equation of @xmath181 . a straightforward algebraic simplification of this equation yields @xmath294 + k(z ) \equiv 0 .\ ] ] in particular ( an interesting observation in itself , albeit obvious from the definition of @xmath295 ) , @xmath296 must be divisible by @xmath297 , with the remainder being a linear function of @xmath190 . eliminating this factor , we get @xmath298 for appropriate reals numbers @xmath283 , @xmath285 . now , writing @xmath299 and @xmath300 , where @xmath39 is the angle of rotation , and observing that @xmath301 the containment of @xmath139 in @xmath181 is equivalent to the two conditions @xmath302 for appropriate parameters @xmath303 . the geometric interpretation of the first condition is that the midpoint of @xmath304 has to lie on a fixed line @xmath305 ( whose direction , @xmath306 , is parallel to the angle bisector between the lines @xmath307 ( see figure [ uvuv2 ] ) . the second condition means that @xmath84 has a fixed component in the direction of @xmath305 . in other words , @xmath139 is contained in @xmath181 if and only if @xmath308 , where @xmath309 is the anti - rotation obtained as a reflection about @xmath305 followed by a shift parallel to @xmath305 . this constitutes an alternative derivation of the characterization of @xmath181 given above . [ [ h12-special - surfaces - and - parabolas . ] ] ( h12 ) special surfaces and parabolas . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + finally , we study intersection patterns involving special surfaces . let @xmath181 be a special surface as above , and let @xmath310 be another @xmath311-linear surface of the form @xmath312 . then either @xmath310 coincides with @xmath181 , or there is at most one parabola contained in both of them . indeed , the intersection of @xmath310 and @xmath181 is the curve satisfying @xmath313 this is a linear system in @xmath194 and @xmath195 . suppose first that its determinant , @xmath314 , does not vanish identically . then , with the exception of finitely many values of @xmath190 , we get a unique solution of the form @xmath315 , @xmath316 , which can describe at most one parabola . if the determinant vanishes identically , then the equation of @xmath310 can be written as @xmath317 , for an appropriate rational algebraic function @xmath318 . if @xmath310 and @xmath181 do intersect in a parabola , then we must have @xmath319 , so @xmath310 and @xmath181 coincide . @xmath98 as a corollary , we have : [ only1 ] let @xmath310 be an @xmath311-linear surface of the above form , and let @xmath28 be a flat rotation contained in @xmath310 . then either @xmath310 contains at least two of the parabolas incident to @xmath28 , and then it must coincide with the corresponding special surface @xmath181 , or @xmath310 contains at most one of these parabolas , so at least two other parabolas cross @xmath310 at @xmath28 . [ sspec ] a special surface can contain at most @xmath1 @xmath140-parabolas . * proof : * let @xmath310 be the given special surface . we claim that for each @xmath320 there can be at most one point @xmath321 such that @xmath322 . indeed , suppose that there exist two such points @xmath323 . since any pair of @xmath140-parabolas on @xmath310 intersect , @xmath324 and @xmath325 meet at a rotation @xmath28 , which maps @xmath29 to both @xmath326 and @xmath327 , an impossibility which completes the proof . @xmath98 [ surfpar ] the number of containments between @xmath328 @xmath140-parabolas and @xmath329 special surfaces is @xmath330 * proof : * as argued above , a special surface @xmath181 is characterized by an anti - rotation @xmath331 in the plane , specified by a line @xmath159 and a shift @xmath255 , such that @xmath332 is the point obtained by reflecting @xmath29 about @xmath159 and then by shifting the reflected point parallel to @xmath159 by distance @xmath255 . thus @xmath181 has three degrees of freedom , and can be parametrized by @xmath333 , where @xmath334 is the equation of @xmath159 and @xmath255 is the shift . we write @xmath335 to denote the special surface parametrized by @xmath336 . by construction , a parabola @xmath139 is contained in @xmath181 if and only if @xmath337 . we use the following parametric setup . we represent each special surface @xmath181 by the corresponding triple @xmath333 , and regard it as a point in parametric 3-space . each parabola @xmath139 is mapped to the locus @xmath338 of all ( points representing ) special surfaces containing @xmath139 . this is a curve in the @xmath333-space , given by the pair of scalar equations @xmath339 . this is a low - degree algebraic curve , whose concrete equations can be worked out explicitly , but we skip over this step . we thus have a system of @xmath329 points and @xmath328 such curves in 3-space , and we wish to bound the number of incidences between them . we have the additional property , noted in lemma [ only1 ] , that two curves meet in at most one point . by projecting these points and curves onto some generic 2-plane , one can easily show that that the number of incidences , and thus the number of original containments , is at most @xmath340 , as claimed . * remark . * if we represent each special surface by its corresponding anti - rotation , lemma [ surfpar ] simply bounds the number of incidences between @xmath329 anti - rotations and @xmath328 ( appropriately transformed copies of ) @xmath140-parabolas , and the bound noted in ( [ weak23 ] ) holds here as well . we begin by reviewing and extending the basic tools from algebraic geometry which have been used in @xcite and in @xcite . however , we develop them here in the context of incidences between points and our @xmath140-parabolas , rather than the context of points and lines considered in the previous papers . so let @xmath341 be a set of @xmath342 @xmath140-parabolas in @xmath2 . for each @xmath343 , we denote the plane containing @xmath344 by @xmath345 and its equation as @xmath346 , where @xmath347 is a linear polynomial . we represent @xmath344 as the intersection curve of @xmath346 and @xmath348 , where @xmath349 is one of the quadratic equations in ( [ parabola ] ) defining @xmath344 , say the first one . note that all the parabolas of @xmath341 cross every plane of the form @xmath196 , each at a single point . recalling the definitions in ( h9 ) , and similar to the case of lines , we say that a point @xmath29 is a _ joint _ of @xmath341 if it is incident to three parabolas of @xmath341 whose tangents at @xmath29 are non - coplanar . let @xmath350 denote the set of joints of @xmath341 . we will also consider points @xmath29 that are incident to three or more parabolas of @xmath341 , so that the tangents to all these parabolas are coplanar , and refer to such points as _ flat _ points of @xmath341 . we recall ( see ( h9 ) ) that any pair of distinct @xmath140-parabolas which meet at a point have there distinct tangents . first , we note that a trivariate polynomial @xmath251 of degree @xmath126 which vanishes at @xmath351 points that lie on a common parabola @xmath343 must vanish identically on @xmath344 . indeed , these points are common roots of @xmath251 and @xmath349 , restricted to the plane @xmath345 . by bzout s theorem @xcite , either these restricted polynomials have a common factor , or they have at most @xmath352 roots . since @xmath349 is irreducible , it must divide the restricted @xmath251 , so @xmath251 must vanish identically on @xmath344 , as claimed . [ [ critical - points - and - parabolas . ] ] critical points and parabolas . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a point @xmath29 is _ critical _ ( or _ singular _ ) for a trivariate polynomial @xmath251 if @xmath353 and @xmath354 ; any other point @xmath29 in the zero set of @xmath251 is called _ regular_. a parabola @xmath344 is _ critical _ if all its points are critical . the following proposition is adapted from @xcite . [ factor ] let @xmath355 and @xmath356 be two trivariate polynomials , of respective degrees @xmath5 and @xmath357 , so that there are @xmath358 parabolas of @xmath341 on which both @xmath359 and @xmath360 vanish identically . then @xmath359 and @xmath360 have a common factor . * proof . * assume that both @xmath355 and @xmath356 have a positive degree in @xmath18 ; this can always be enforced by an appropriate rotation of the coordinate frame . it is then an easy exercise to show that @xmath359 and @xmath360 have a common factor if and only if their resultant , when viewing them as polynomials in @xmath18 , is identically @xmath361 . recall that the resultant is a polynomial in @xmath362 and @xmath363 . ( the same holds when @xmath359 and @xmath360 have any number of variables , including @xmath18 , in which case the resultant is a polynomial in the remaining variables . ) for any fixed value @xmath364 of @xmath363 , @xmath365 and @xmath366 have at least @xmath358 common roots ( at the intersection points of the @xmath358 parabolas with @xmath367 ) , so , by bzout s theorem @xcite , they have a common factor . therefore , the resultant , with respect to @xmath18 , of @xmath365 and @xmath366 is identically @xmath361 ( as a polynomial in @xmath362 ) . since this is true for every value @xmath364 of @xmath363 , it follows that the resultant of @xmath355 and @xmath356 , with respect to @xmath18 , vanishes identically as a polynomial in @xmath362 and @xmath363 . therefore , @xmath355 and @xmath356 , as trivariate polynomials , have a common factor . @xmath98 [ prop1 ] let @xmath341 be as above . then any trivariate square - free polynomial @xmath251 of degree @xmath126 can have at most @xmath368 critical parabolas in @xmath341 . * proof : * we prove the claim by induction on the degree @xmath126 of @xmath251 . the claim holds trivially for @xmath369 , so assume that @xmath370 . assume first that @xmath251 is irreducible . apply proposition [ factor ] to @xmath251 and @xmath371 , say . both polynomials vanish identically on each critical parabola , and their respective degrees are @xmath126 and @xmath372 . if @xmath251 had more than @xmath368 critical parabolas then @xmath251 and @xmath371 would have a common factor , which is impossible since @xmath251 is irreducible . suppose next that @xmath251 is reducible ( but square - free ) , and write @xmath373 , so that @xmath359 and @xmath360 are non - constant polynomials which have no common factor ( since @xmath251 is square - free , this can always be done ) . denote the degrees of @xmath359 and @xmath360 by @xmath374 and @xmath375 , respectively ; we have @xmath376 . let @xmath344 be a critical parabola for @xmath251 . then either @xmath377 on @xmath344 or @xmath378 on @xmath344 ( or both ) . moreover , since @xmath379 on @xmath344 , it is easily checked that @xmath344 must satisfy ( at least ) one of the following properties : \(i ) @xmath380 on @xmath344 . \(ii ) @xmath344 is a critical parabola of @xmath359 . \(iii ) @xmath344 is a critical parabola of @xmath360 . indeed , if ( i ) does not hold , we have , without loss of generality , @xmath377 on @xmath344 , but @xmath360 vanishes only at finitely many points of @xmath344 . on any other point @xmath29 of @xmath344 we then must have @xmath381 , which implies that @xmath382 is identically zero on @xmath344 , so @xmath344 is critical for @xmath359 . this implies ( ii ) ; ( iii ) holds in the symmetric case where @xmath378 on @xmath344 but @xmath359 does not vanish identically on @xmath344 . by the induction hypothesis , the number of critical parabolas for @xmath359 is at most @xmath383 , and the number of critical parabolas for @xmath360 is at most @xmath384 . consider the parabolas that satisfy ( i ) and intersect all of them by any of the planes @xmath367 , as in the proof of proposition [ factor ] . all the intersection points are roots of @xmath385 and @xmath386 on this plane , and , as follows from the proof of proposition [ factor ] , these bivariate polynomials have no common factor ( or , more precisely , they can have a common factor only at finitely many values of @xmath363 ) . hence , by bzout s theorem , they have at most @xmath387 common roots . altogether , the number of critical parabolas for @xmath251 is at most @xmath388 @xmath98 [ prop2 ] let @xmath29 be a regular point of @xmath251 , so that @xmath389 on three parabolas of @xmath341 passing through @xmath29 . then these parabolas must have coplanar tangents at @xmath29 . * proof : * any such tangent line must be contained in the tangent plane to @xmath390 at @xmath29 . @xmath98 hence , a point @xmath29 incident to three parabolas of @xmath341 whose tangent lines at @xmath29 are non - coplanar , so that @xmath389 on each of these parabolas , must be a critical point of @xmath251 . [ prop4 ] given a set @xmath0 of @xmath357 points in 3-space , there exists a nontrivial trivariate polynomial @xmath391 which vanishes at all the points of @xmath0 , of degree @xmath126 , for any @xmath126 satisfying @xmath392 . * proof : * ( see @xcite . ) a trivariate polynomial of degree @xmath126 has @xmath393 monomials , and requiring it to vanish at @xmath357 points yields these many homogeneous equations in the coefficients of these monomials . such an underdetermined system always has a nontrivial solution . [ [ flat - points - and - parabolas . ] ] flat points and parabolas . + + + + + + + + + + + + + + + + + + + + + + + + + + call a regular point @xmath28 of a trivariate polynomial @xmath251 _ geometrically flat _ if it is incident to three distinct parabolas of @xmath341 ( with necessarily coplanar tangent lines at @xmath28 , no pair of which are collinear ) on which @xmath251 vanishes identically . let @xmath28 be a geometrically flat point of @xmath251 , and let @xmath394 be three incident parabolas on which @xmath251 vanishes . let @xmath395 denote the tangent line to @xmath396 at @xmath28 , and let @xmath397 denote a unit vector in the direction of @xmath395 , for @xmath398 . the second - order taylor expansion of @xmath251 at @xmath28 has the form @xmath399 for any vector @xmath254 , where @xmath400 is the _ hessian _ matrix of @xmath251 . @xmath401 is a quadratic polynomial ( in @xmath254 ) which approximates @xmath251 up to third order terms for sufficiently small values of @xmath402 . our goal is to construct , using this approximation and the fact that @xmath389 on three parabolas incident to @xmath28 , as above , a new polynomial , depending on @xmath251 , which vanishes at @xmath28 , and use this vanishing as a characterization of flat points . to do so , we need to make the analysis more specific , and taylor it to the special form of @xmath140-parabolas . let @xmath28 be a flat point , and let @xmath182 be the corresponding parameters in the @xmath138-plane ( so @xmath28 maps @xmath175 to @xmath176 for each @xmath154 ; cf . remark ( 1 ) at the end of ( h9 ) ) . let @xmath180 be the corresponding special surface spanned by the parabolas @xmath188 , for all @xmath155 ( here we vary @xmath155 continuously , but only finitely many corresponding parabolas belong to @xmath341 ) . since @xmath28 is flat , there exist at least three parabolas @xmath403 , @xmath398 ( all belonging to @xmath341 , contained in @xmath181 , and passing through @xmath28 ) , such that @xmath389 on each of them . let @xmath401 denote , as above , the quadratic polynomial which is the second - order taylor expansion of @xmath251 at @xmath28 . let @xmath404 be one of the above parabolas on which @xmath251 vanishes identically . for @xmath187 in the vicinity of @xmath28 , @xmath405 , so , for points @xmath187 near @xmath28 on @xmath344 , we have @xmath406 . let us continue to consider only points @xmath187 on @xmath344 . let @xmath197 ( resp . , @xmath136 ) be the coordinates of @xmath28 ( resp . , @xmath187 ) . the equations of @xmath344 ( see ( [ parabola ] ) ) are @xmath407 so we have @xmath408 which we can further rewrite as @xmath409 let us simplify these equations as @xmath410 where @xmath411 , @xmath412 , @xmath413 , and @xmath414 are all linear functions of @xmath155 . if we substitute these equations into the equation of @xmath401 , assume that @xmath190 is very close to @xmath205 , ignore terms which are at least cubic in @xmath206 , and use the fact that @xmath415 for any @xmath187 on @xmath344 sufficiently close to @xmath28 , we conclude that both the linear and the quadratic parts of @xmath416 ( in @xmath206 ) vanish identically . the linear part is @xmath417 and the quadratic part is @xmath418 hence we have @xmath419 note that both equations vanish for ( at least ) three distinct values of @xmath155 . since the first equation is linear in @xmath155 and the second is quadratic in @xmath155 , all the coefficients of both equations are identically zero . let us restrict ourselves to the coefficient of the linear term in the first equation and of the quadratic term in the second one . denote by @xmath420 ( resp . , @xmath421 ) the coefficient of @xmath155 in @xmath411 ( resp . , @xmath412 ) . then we have @xmath422 it is easily seen that @xmath420 and @xmath421 can not both be zero ( assuming a generic coordinate frame in the original @xmath138-plane ) , so , eliminating them gives @xmath423 which is the constraint we were after . in what follows , we refer to the left - hand side of ( [ sofsof ] ) as @xmath424 . that is , @xmath425 and this polynomial has to vanish at @xmath28 . we have thus shown : [ flat ] let @xmath251 be a trivariate polynomial . if @xmath28 is a regular geometrically flat point of @xmath251 ( with respect to three parabolas of @xmath341 ) then @xmath426 . * remark . * note that the left - hand side of ( [ sofsof ] ) is one of the three polynomials @xmath427 used in @xcite to analyze flat points in a 3-dimensional line arrangement . specifically , @xmath428 where @xmath429 is the unit vector in the @xmath363-direction ; the other two polynomials are defined analogously , using the other two coordinate vectors @xmath430 , @xmath431 . these polynomials form the _ second fundamental form _ of @xmath251 ; see @xcite for details . in particular , if the degree of @xmath251 is @xmath126 then the degree of @xmath424 is at most @xmath432 . in what follows , we call a point @xmath28 _ flat _ for @xmath251 if @xmath426 . we will need the following technical lemma . [ plinxy ] let @xmath251 be an irreducible trivariate polynomial , with the properties that ( i ) @xmath433 at each regular point @xmath28 of @xmath390 , and ( ii ) @xmath389 on at least two distinct intersecting @xmath140-parabolas of @xmath341 . then @xmath251 is a special polynomial . ( note that the converse of the lemma is trivial , because the second - order derivatives @xmath434 , @xmath435 , and @xmath436 are all identically zero for a special polynomial @xmath251 , and because of the way such polynomials are constructed . ) * proof : * fix @xmath437 and consider the restricted bivariate polynomial @xmath438 . clearly , @xmath439 on the plane @xmath440 . hence @xmath441 at each regular point @xmath442 of @xmath390 , and thus at each regular point of @xmath443 . ( note that a regular point of @xmath443 is also a regular point of @xmath251 , although the converse need not be true . ) note also that @xmath443 is an irreducible polynomial , except possibly for finitely many values of @xmath205 . as is well known @xcite , the curvature of the plane curve @xmath444 , at a regular point of @xmath443 , is given by @xmath445 hence this curve has zero curvature at every regular point of @xmath443 , and thus , being the zero set of an irreducible polynomial , it must be a single line . in other words , @xmath251 is linear in @xmath194 and @xmath195 for every fixed @xmath190 , except for finitely many values , implying that its equation is of the form @xmath446 , where @xmath447 , @xmath448 and @xmath449 are univariate polynomials . we now exploit assumption ( ii ) , denoting by @xmath181 the unique special surface determined by ( and containing ) the two given @xmath140-parabolas . the analysis in ( h12 ) then implies that @xmath181 coincides with the zero set of @xmath251 , so @xmath251 is indeed a special polynomial , as claimed . @xmath98 call an @xmath140-parabola @xmath343 _ flat _ for @xmath251 if all the points of @xmath344 are flat points of @xmath251 ( with the possible exception of a discrete subset ) . arguing as in the case of critical points , if @xmath344 contains more than @xmath450 flat points then @xmath344 is a flat parabola . as in @xcite , we next show that , in general , trivariate polynomials do not have too many flat parabolas . as before , we first establish this property for irreducible polynomials , and then extend the analysis to more general polynomials . [ lflat ] let @xmath251 be an irreducible trivariate polynomial of degree @xmath126 , which is not a special polynomial . then @xmath251 can have at most @xmath451 flat @xmath140-parabolas of @xmath341 . * proof : * suppose to the contrary that there are more than @xmath451 flat @xmath140-parabolas . as above , restrict @xmath251 and @xmath424 to a fixed plane @xmath452 of the form @xmath437 . the number of common roots of @xmath251 and @xmath424 on @xmath452 exceeds the product of their degrees . since this holds for every @xmath205 , proposition [ factor ] implies that they must have a common factor . since @xmath251 is irreducible , @xmath251 must be a factor of @xmath424 . this implies that all the ( regular ) points at which @xmath251 vanishes are flat . hence , by lemma [ plinxy ] , @xmath251 must be a special polynomial , a contradiction which completes the proof of the asserted bound . @xmath98 [ lflatx ] let @xmath251 be any trivariate square - free polynomial of degree @xmath126 with no special polynomial factors . then @xmath251 can have at most @xmath453 flat @xmath140-parabolas in @xmath341 . * proof : * if @xmath251 is irreducible , the claim holds by proposition [ lflat ] . otherwise , write @xmath373 where @xmath359 and @xmath360 are non - constant polynomials with no common factors ( and no special polynomial factors ) . let @xmath374 and @xmath375 denote their respective degrees , so @xmath376 . let @xmath28 be a regular flat point of @xmath251 . then either @xmath454 , or only exactly one of @xmath455 , @xmath456 vanishes . hence , if @xmath344 is a flat parabola for @xmath251 then either both @xmath359 and @xmath360 vanish identically on @xmath344 or exactly one of them vanishes identically on @xmath344 , while the other has only finitely many zeroes on @xmath344 . now , as already argued in the proof of proposition [ prop1 ] , there are at most @xmath387 parabolas of the former kind . to handle parabolas of the latter kind , consider a regular point @xmath28 of @xmath251 at which @xmath385 but @xmath360 is nonzero . a simple calculation yields : @xmath457 hence , at @xmath28 we have @xmath458 and therefore we have at @xmath28 , as is easily checked , @xmath459 that is , a regular flat point for @xmath251 , at which @xmath385 but @xmath360 is nonzero , is a regular flat point for @xmath359 , and a symmetric statement holds when @xmath386 but @xmath359 is nonzero . hence , any flat parabola of the latter kind is either a flat parabola for @xmath359 or a flat parabola for @xmath360 . arguing by induction on the degree , the number of flat parabolas for @xmath251 is thus at most @xmath460 and the lemma follows . in this section we extend the recent algebraic machinery of guth and katz @xcite , as further developed by elekes et al . @xcite , using the algebraic tools set forth in the preceding section , to establish the bound @xmath461 on the number of rotations with multiplicity at least @xmath49 in a collection of @xmath328 @xmath140-parabolas . [ thm : gk1 ] let @xmath341 be a set of at most @xmath328 @xmath140-parabolas in @xmath2 , and let @xmath110 be a set of @xmath357 rotations , each of which is incident to at least three parabolas of @xmath341 . suppose further that no special surface contains more than @xmath401 parabolas of @xmath341 . then @xmath462 . * remarks . * * ( 1 ) * the recent results of @xcite imply that the number of joints in a set of @xmath328 @xmath140-parabolas is @xmath463 . the proofs in @xcite are much simpler than the proof given below , but they do not apply to flat points ( rotations ) as does theorem [ thm : gk1 ] . since flat rotations are an integral part of the setup considered in this paper , we need to count them too , using the stronger theorem [ thm : gk1 ] . moreover , even if we were to consider only joint rotations , the analysis of their incidences with the @xmath140-parabolas will turn some of them into flat rotations ( by pruning some of the parabolas ) , so , as in @xcite , we will need to face flat rotations , no matter what . * ( 2 ) * by lemma [ sspec ] , we always have @xmath464 , and we also have @xmath465 , so the `` worst - case '' bound on @xmath357 is @xmath466 . * ( 3 ) * note that the parameter @xmath328 in the statement of the theorem is arbitrary , not necessarily the maximum number @xmath112 . when @xmath328 attains its maximum possible value @xmath112 , the bound becomes @xmath467 . the proof of theorem [ thm : gk1 ] uses the proof technique of @xcite , properly adapted to the present , somewhat more involved context of @xmath140-parabolas and rotations . * we first prove the theorem under the additional assumption that @xmath468 . the proof proceeds by induction on @xmath328 , and shows that @xmath469 , where @xmath272 is a sufficiently large constant whose choice will be dictated by the forthcoming analysis . the statement holds for all @xmath470 , for some constant @xmath471 , if we choose @xmath272 to be sufficiently large . fix @xmath472 , and suppose that the claim holds for all @xmath473 . let @xmath341 and @xmath110 be as in the statement of the theorem , with @xmath474 , and suppose to the contrary that @xmath475 . we first apply the following iterative pruning process to @xmath341 . as long as there exists a parabola @xmath343 incident to fewer than @xmath476 rotations of @xmath110 , for some constant @xmath477 that we will fix later , we remove @xmath344 from @xmath341 , remove its incident rotations from @xmath110 , and repeat this step with respect to the reduced set of rotations . in this process we delete at most @xmath478 rotations . we are thus left with a subset of at least @xmath479 of the original parabolas , each incident to at least @xmath476 surviving rotations , and each surviving rotation is incident to at least three surviving parabolas . for simplicity , continue to denote these sets as @xmath341 and @xmath110 . choose a random sample @xmath480 of parabolas from @xmath341 , by picking each parabola independently with probability @xmath155 , where @xmath155 is a small constant that we will fix later . the expected number of parabolas that we choose is @xmath481 , where @xmath482 is the number of parabolas remaining after the pruning . we have @xmath483 , because each surviving parabola is incident to at least @xmath476 surviving rotations , each incident to at least two other surviving parabolas ; since all these parabolas are distinct ( recall that a pair of parabolas can meet in at most one rotation point ) , we have @xmath484 . hence , using chernoff s bound , as in @xcite ( see , e.g. , @xcite ) , we obtain that , with positive probability , ( a ) @xmath485 . ( b ) each parabola @xmath343 contains at least @xmath486 rotations that lie on parabolas of @xmath480 . ( to see ( b ) , take a parabola @xmath343 and a rotation @xmath487 . note that @xmath28 will be incident to a parabola of @xmath480 with probability at least @xmath155 , so the expected number of rotations in @xmath488 which lie on parabolas of @xmath480 is at least @xmath489 . this , combined with chernoff s bound , implies ( b ) . ) we assume that @xmath480 does indeed satisfy ( a ) and ( b ) , and then ( recalling that @xmath490 ) choose @xmath491 arbitrary rotations on each parabola in @xmath480 , to obtain a set @xmath0 of at most @xmath492 rotations . applying proposition [ prop4 ] , we obtain a nontrivial trivariate polynomial @xmath493 which vanishes at all the rotations of @xmath0 , whose degree is at most the smallest integer @xmath126 satisfying @xmath494 , so @xmath495 for @xmath328 ( i.e. , @xmath471 ) sufficiently large . without loss of generality , we may assume that @xmath251 is square - free by removing repeated factors , we get a square - free polynomial which vanishes on the same set as the original @xmath251 , with the same upper bound on its degree . the polynomial @xmath251 vanishes on @xmath491 points on each parabola in @xmath480 . this number is larger than @xmath352 , if we choose @xmath155 sufficiently small so as to satisfy @xmath496 . hence @xmath251 vanishes identically on all these parabolas . any other parabola of @xmath341 meets at least @xmath486 parabolas of @xmath480 , at distinct points , and we can make this number also larger than @xmath352 , with an appropriate choice of @xmath155 and @xmath125 ( we need to ensure that @xmath497 ) . hence , @xmath251 vanishes identically on each parabola of @xmath341 . we will also later need the property that each parabola of @xmath341 contains at least @xmath498 points of @xmath110 ; that is , we require that @xmath499 , which will hold if @xmath500 . to recap , the preceding paragraphs impose several inequalities on @xmath125 and @xmath155 , and a couple of additional similar inequalities will be imposed later on . all these inequalities are easy to satisfy by choosing @xmath501 to be a sufficiently small positive constant , and @xmath125 a sufficiently large constant . ( these choices will also affect the choice of @xmath272see below . ) we note that @xmath251 can have at most @xmath502 special polynomial factors ( since each of them is a cubic polynomial ) ; i.e. , @xmath251 can vanish identically on at most @xmath502 respective special surfaces @xmath503 , for @xmath504 . we factor out all these special polynomial factors from @xmath251 , and let @xmath443 denote the resulting polynomial , which is a square - free polynomial without any special polynomial factors , of degree at most @xmath126 . consider one of the special surfaces @xmath505 , and let @xmath172 denote the number of parabolas contained in @xmath505 . then any rotation on @xmath505 is either an intersection point of ( at least ) two of these parabolas , or it lies on at most one of them . the number of rotations of the first kind is @xmath506 . any rotation @xmath28 of the second kind is incident to at least one parabola of @xmath341 which crosses @xmath505 transversally at @xmath28 . we note that each @xmath140-parabola @xmath344 can cross @xmath505 in at most three points . indeed , substituting the equations of @xmath344 into the equation @xmath507 of @xmath505 ( see ( [ sigma ] ) ) yields a cubic equation in @xmath190 , with at most three roots . hence , the number of rotations of the second kind is @xmath508 , and the overall number of rotations on @xmath505 is @xmath509 , since we have assumed in the present version of the proof that @xmath510 . summing the bounds over all surfaces @xmath505 , we conclude that altogether they contain @xmath511 rotations , which we bound by @xmath512 , for some absolute constant @xmath31 . we remove all these vanishing special surfaces , together with the rotations and the parabolas which are fully contained in them , and let @xmath513 and @xmath514 denote , respectively , the set of those parabolas of @xmath341 ( rotations of @xmath110 ) which are not contained in any of the vanishing surfaces @xmath505 . note that there are still at least three parabolas of @xmath515 incident to any remaining rotation in @xmath516 , since none of the rotations of @xmath516 lie in any surface @xmath505 , so all parabolas incident to such a rotation are still in @xmath515 . clearly , @xmath443 vanishes identically on every @xmath517 . furthermore , every @xmath518 contains at most @xmath126 points in the surfaces @xmath505 , because , as just argued , it crosses each surface @xmath505 in at most three points . note that this also holds for every parabola @xmath344 in @xmath519 , if we only count intersections of @xmath344 with surfaces @xmath505 which do not fully contain @xmath344 . hence , each @xmath517 contains at least @xmath520 rotations of @xmath516 . since each of these rotations is incident to at least three parabolas in @xmath515 , each of these rotations is either critical or geometrically flat for @xmath443 . consider a parabola @xmath518 . if @xmath344 contains more than @xmath352 critical rotations then @xmath344 is a critical parabola for @xmath443 . by proposition [ prop1 ] , the number of such parabolas is at most @xmath368 . any other parabola @xmath517 contains more than @xmath521 geometrically flat points and hence @xmath344 must be a flat parabola for @xmath443 . by proposition [ lflatx ] , the number of such parabolas is at most @xmath453 . summing up we obtain @xmath522 we require that @xmath523 ; that is , @xmath524 , which can be guaranteed by choosing @xmath155 sufficiently small . we next want to apply the induction hypothesis to @xmath515 , with the parameter @xmath525 ( which dominates the size of @xmath515 ) . for this , we first need to argue that each special surface contains at most @xmath526 parabolas of @xmath515 . indeed , let @xmath310 be a special surface . using ( [ sigma ] ) , eliminate , say , @xmath195 from the equation of @xmath310 and substitute the resulting expression into the equation of @xmath443 , to obtain a bivariate polynomial @xmath527 . let @xmath344 be a parabola of @xmath515 contained in @xmath310 . we represent @xmath344 by its @xmath194-equation of the form @xmath528 , and observe that @xmath527 vanishes on the zero set of @xmath529 . hence @xmath530 must be divisible by @xmath529 . note that , in a generic coordinate frame in the @xmath138-plane , two different parabolas can not have the same equation @xmath528 , because this equation uniquely determines @xmath531 , and @xmath532 , and then , in a generic frame , @xmath327 is also uniquely determined . note also that the degree of @xmath530 is at most @xmath533 , and that the degree of each factor @xmath529 is @xmath50 , implying that @xmath181 can contain at most @xmath534 parabolas of @xmath515 . an important observation , which we will use in the proof of general version of the theorem , is that the argument just given does not use the assumed bound on the number of @xmath140-parabolas contained in a special surface , but , rather , establishes this bound `` from scratch '' for the subproblem involving @xmath516 and @xmath515 . that is , even if the original problem does not satisfy the extra assumption in the restricted version , the subproblems that it generates always do satisfy it . hence , the maximum number of parabolas of @xmath515 contained in a special surface is at most @xmath535 , so , by the induction hypothesis , the number of points in @xmath516 is at most @xmath536 adding up the bounds on the number of points on parabolas removed during the pruning process and on the special surfaces @xmath505 ( which correspond to the special polynomial factors of @xmath251 ) , we obtain @xmath537 with an appropriate , final choice of @xmath155 , @xmath125 , and @xmath272 . this contradicts the assumption that @xmath538 , and thus establishes the induction step for @xmath328 , and , consequently , completes the proof of the restricted version of the theorem . [ [ proof - of - the - general - version ] ] proof of the general version : + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the proof proceeds almost exactly as the proof of the restricted version , except for the analysis of the number of rotations on the special surfaces @xmath505 . as noted above , we encounter this difference only once , in handling the original problem : when we apply the induction step , we always fall into the restricted setup . by assumption , each special surface @xmath505 contains at most @xmath401 @xmath140-parabolas . we modify the preceding analysis , so that each parabola is considered only once . that is , we iterate over the special surfaces in some order . when handling a surface @xmath505 , we consider only those @xmath140-parabolas that are not contained in any previously processed surface , and bound the number of rotations that they contain . then we remove these parabolas and rotations from further considerations and go to the next surface . as argued above , a special surface @xmath505 containing @xmath172 ( surviving ) parabolas contains at most @xmath539 rotations which lie on these parabolas ( and on no previously processed parabola ) . summing these bounds over all special surfaces , and using the fact that @xmath540 for each @xmath165 , we get an overall bound @xmath541 , as asserted . @xmath98 we summarize the remarks following theorem [ thm : gk1 ] , combined with lemma [ lem : lower ] , in the following corollary . [ roth ] let @xmath0 be a set of @xmath1 points in the plane . then there are at most @xmath3 rotations which map some ( degenerate or non - degenerate ) triangle spanned by @xmath0 to another ( congruent and equally oriented ) such triangle . this bound is tight in the worst case . in the following section we will continue to adapt the analysis of @xcite to obtain bounds on the number of incidences between helices ( @xmath140-parabolas ) and rotations with multiplicity @xmath542 , and , consequently , obtain bounds on @xmath543 , for any @xmath4 . in this section we further adapt the machinery of @xcite to derive an upper bound on the number of incidences between @xmath357 rotations and @xmath328 @xmath140-parabolas in @xmath2 , where each rotation is incident to at least three parabolas ( i.e. , has multiplicity @xmath542 ) . [ inc - gen ] for an underlying ground set @xmath0 of @xmath1 points in the plane , let @xmath341 be a set of at most @xmath342 @xmath140-parabolas defined on @xmath0 , and let @xmath110 be a set of @xmath357 rotations with multiplicity at least @xmath49 ( with respect to @xmath0 ) . then @xmath544 * remark . * as easily checked , the first term dominates the second term when @xmath545 , and the second term dominates when @xmath546 . in particular , the first term dominates when @xmath547 , because we have @xmath548 * proof : * the proof of theorem [ inc - gen ] proceeds in two steps . we first establish a bound which is independent of @xmath357 , and then apply it to obtain the @xmath357-dependent bound asserted in the theorem . for the first step , we have : [ inc-32 ] let @xmath341 be a set of at most @xmath342 @xmath140-parabolas defined on @xmath0 , and let @xmath110 be a set of rotations with multiplicity at least @xmath49 with respect to @xmath0 , such that no special surface contains more than @xmath491 parabolas of @xmath341 . then the number of incidences between @xmath110 and @xmath341 is @xmath463 . * proof . * write @xmath549 for short , and put @xmath550 . we will establish the upper bound @xmath551 , for some sufficiently large absolute constant @xmath273 , whose specific choice will be dictated by the various steps of the proof . suppose then to the contrary that @xmath552 for the given @xmath341 and @xmath110 . for @xmath343 , let @xmath553 denote the number of rotations incident to @xmath344 . we refer to @xmath553 as the _ multiplicity _ of @xmath344 . we have @xmath554 . the average multiplicity of a parabola @xmath344 is @xmath555 . we begin by applying the following pruning process . put @xmath556 . as long as there exists a parabola @xmath343 whose multiplicity is smaller than @xmath557 , we remove @xmath344 from @xmath341 , but do not remove any rotation incident to @xmath344 . we keep repeating this step ( without changing @xmath557 ) , until each of the surviving parabolas has multiplicity at least @xmath557 . moreover , if , during the pruning process , some rotation @xmath28 loses @xmath558 incident parabolas , we remove @xmath28 from @xmath110 . this decreases the multiplicity of some parabolas , and we use the new multiplicities in the test for pruning further parabolas , but we keep using the original threshold @xmath557 . when we delete a parabola @xmath344 , we lose at most @xmath557 incidences with surviving rotations . when a rotation @xmath28 is removed , the number of current incidences with @xmath28 is smaller than or equal to twice the number of incidences with @xmath28 that have already been removed . hence , the total number of incidences that were lost during the pruning process is a most @xmath559 . thus , we are left with a subset @xmath516 of the rotations and with a subset @xmath515 of the parabolas , so that each @xmath517 is incident to at least @xmath560 rotations of @xmath516 , and each rotation @xmath561 is incident to at least three parabolas of @xmath515 ( the latter is an immediate consequence of the rule for pruning a rotation ) . moreover , we have @xmath562 . it therefore suffices to bound @xmath563 . let @xmath564 . since at least three parabolas in @xmath515 are incident to each rotation in @xmath516 , it follows that each parabola in @xmath515 is incident to at most @xmath565 rotations of @xmath516 , and therefore @xmath566 . combining this with the fact that @xmath562 , we get that @xmath567 . we fix the following parameters @xmath568 for an appropriate absolute constant @xmath569 , whose value will be fixed shortly . clearly , @xmath570 , and we can also ensure that @xmath571 , i.e. , that @xmath572 , by choosing @xmath573 . furthermore , since @xmath567 we have @xmath574 . we construct a random sample @xmath575 of parabolas of @xmath515 by choosing each parabola independently at random with probability @xmath155 ; the expected size of @xmath575 is @xmath576 . now take @xmath18 ( arbitrary ) rotations of @xmath516 on each parabola of @xmath575 ( which can always be done since @xmath577 ) , to form a sample @xmath0 of rotations in @xmath516 , of expected size at most @xmath578 . for any parabola @xmath517 , the expected number of rotations of @xmath579 which lie on parabolas of @xmath575 is at least @xmath580 ( each of the at least @xmath557 rotations @xmath581 is incident to at least one other parabola of @xmath515 , and the probability of this parabola to be chosen in @xmath575 is @xmath155 ) . we assume that @xmath273 is large enough so that @xmath582 is larger than @xmath583 ( it suffices to choose @xmath584 ) . since @xmath585 , and the expected size of @xmath586 is @xmath587 , we can use chernoff s bound , to show that there exists a sample @xmath575 such that ( i ) @xmath588 , and ( ii ) each parabola @xmath517 contains at least @xmath589 rotations of @xmath516 which lie on parabolas of @xmath575 . in what follows , we assume that @xmath575 satisfies these properties . in this case , we have @xmath590 . now construct , using proposition [ prop4 ] , a nontrivial suqare - free trivariate polynomial @xmath251 which vanishes on @xmath0 , of smallest degree @xmath126 satisfying @xmath591 , so @xmath592 for @xmath328 sufficiently large ( for small values of @xmath328 we ensure the bound by choosing @xmath273 sufficiently large , as before ) . we will choose @xmath593 , so @xmath594 . as above , and without loss of generality , we may assume that @xmath251 is square - free : factoring out repeated factors only lowers the degree of @xmath251 and does not change its zero set . the following properties hold : ( a ) since @xmath595 , @xmath251 vanishes at more than @xmath352 rotations on each parabola of @xmath575 , and therefore , as already argued , it vanishes identically on each of these parabolas . ( b ) each parabola @xmath517 contains at least @xmath596 rotations which lie on parabolas of @xmath575 . since , as just argued , @xmath251 vanishes at these rotations , it must vanish identically on @xmath344 . thus , @xmath389 on every parabola of @xmath515 . before proceeding , we enforce the inequality @xmath597 which will hold if we choose @xmath255 so that @xmath598 . similarly , an appropriate choice of @xmath255 ( or @xmath273 ) also ensures that @xmath599 . we next consider all the special polynomial factors of @xmath251 , and factor them out , to obtain a square - free polynomial @xmath443 , of degree at most @xmath126 , with no special polynomial factors . as in the previous analysis , @xmath251 can have at most @xmath502 special polynomial factors , so it can vanish identically on at most @xmath502 special surfaces @xmath503 , for @xmath504 . let @xmath600 denote the set of those parabolas of @xmath515 which are not contained in any of the vanishing surfaces @xmath505 . for each parabola @xmath601 , @xmath443 vanishes identically on @xmath344 , and ( as argued above ) at most @xmath126 rotations in @xmath579 lie in the surfaces @xmath505 . hence , @xmath344 contains at least @xmath520 remaining rotations , each of which is either critical or flat for @xmath443 , because each such point is incident to at least three parabolas ( necessarily of @xmath602 ) on which @xmath603 . hence , either at least @xmath352 of these rotations are critical , and then @xmath344 is a critical parabola for @xmath443 , or at least @xmath521 of these rotations are flat , and then @xmath344 is a flat parabola for @xmath443 . applying propositions [ prop1 ] and [ lflatx ] , the overall number of parabolas in @xmath602 is therefore at most @xmath604 on the other hand , by assumption , each vanishing special surface @xmath505 contains at most @xmath491 parabolas of @xmath341 , so the number of parabolas contained in the special vanishing surfaces is at most @xmath605 , with our choice of @xmath255 . hence , the overall number of parabolas in @xmath515 is smaller than @xmath606 , a contradiction that completes the proof of theorem [ inc-32 ] . @xmath98 * proof of theorem [ inc - gen ] . * write @xmath549 for short . set @xmath607 and @xmath608 , for some sufficiently large constant @xmath125 whose value will be determined later , and apply the following pruning process . as long as there exists a parabola @xmath343 whose multiplicity is smaller than @xmath557 , we remove @xmath344 from @xmath341 , but do not remove any rotation incident to @xmath344 . similarly , as long as there exists a rotation @xmath141 whose multiplicity is smaller than @xmath609 , we remove @xmath28 from @xmath110 . of course , these removals may reduce the multiplicity of some surviving rotations or parabolas , making additional rotations and parabolas eligible for removal . we keep repeating this step ( without changing the initial thresholds @xmath557 and @xmath609 ) , until each of the surviving parabolas has multiplicity at least @xmath557 and each of the surviving rotations has multiplicity at least @xmath609 . we may assume that @xmath610 , by choosing @xmath125 suficiently large and using theorem [ thm : gk1](i ) . when we delete a parabola @xmath344 , we lose at most @xmath557 incidences with surviving rotations . when a rotation @xmath28 is removed , we lose at most @xmath609 incidences with surviving parabolas . all in all , we lose at most @xmath611 incidences , and are left with a subset @xmath516 of @xmath110 and with a subset @xmath515 of @xmath341 , so that each parabola of @xmath515 is incident to at least @xmath557 rotations of @xmath516 , and each rotation of @xmath516 is incident to at least @xmath609 parabolas of @xmath515 ( these subsets might be empty ) . put @xmath564 and @xmath612 . we have @xmath613 , so it remains to bound @xmath563 , which we do as follows . we fix some sufficiently small positive parameter @xmath501 , and construct a random sample @xmath614 by choosing each rotation of @xmath516 independently with probability @xmath155 . the expected size of @xmath615 is @xmath616 , and the expected number of points of @xmath615 on any parabola of @xmath515 is at least @xmath617 . chernoff s bound implies that , with positive probability , @xmath618 , and @xmath619 for every @xmath517 . we can therefore assume that @xmath615 satisfies all these inequalities . ( for the bound to apply , @xmath620 ( and @xmath357 ) must be at least some sufficiently large constant ; if this is not the case , we turn the trivial bound @xmath621 ( or @xmath622 ) on @xmath623 into the bound @xmath624 ( or @xmath625 ) by choosing the constant of proportionality sufficiently large . ) construct , using proposition [ prop4 ] , a nontrivial square - free trivariate polynomial @xmath251 which vanishes on @xmath615 , whose degree is at most the smallest integer @xmath126 satisfying @xmath626 , so @xmath627 assuming ( as above ) that @xmath620 is sufficiently large . choosing @xmath125 to be large enough , we may assume that @xmath628 . ( this will hold if we ensure that @xmath629 . ) this implies that @xmath251 vanishes at more than @xmath498 points on each parabola @xmath517 , and therefore it vanishes identically on each of these parabolas . as in the previous analysis , we factor out the special polynomial factors of @xmath251 , obtaining a square - free polynomial @xmath443 , of degree at most @xmath126 , with no special polynomial factors . let @xmath503 denote the special surfaces on which @xmath251 vanishes identically ( the zero sets of the special polynomial factors of @xmath251 ) , for some @xmath504 . let @xmath600 ( resp . , @xmath630 ) denote the set of those parabolas of @xmath515 ( resp . , rotations of @xmath516 ) which are not contained in any of the vanishing surfaces @xmath505 . put @xmath631 and @xmath632 . for each parabola @xmath601 , @xmath443 vanishes identically on @xmath344 , and , as argued in the proof of theorem [ thm : gk1 ] , at most @xmath126 rotations of @xmath579 lie in the surfaces @xmath505 . hence , @xmath344 contains more than @xmath520 rotations of @xmath633 , and , arguing as in the preceding proof , each of these rotations is either critical or flat for @xmath443 . hence , either more than @xmath352 of these rotations are critical , and then @xmath344 is a critical parabola for @xmath443 , or more than @xmath521 of these rotations are flat , and then @xmath344 is a flat parabola for @xmath443 . applying propositions [ prop1 ] and [ lflatx ] , the overall number of parabolas in @xmath602 is therefore at most @xmath634 we now apply theorem [ inc-32 ] to @xmath602 and @xmath633 , with the bound @xmath525 on the size of @xmath602 . the conditions of this theorem hold for these sets : clearly , each rotation in @xmath633 is incident to at least three parabolas of @xmath602 . for the other condition , we argue exactly as in the proof of theorem [ thm : gk1 ] , to conclude that any special surface can contain at most @xmath534 parabolas of @xmath515 , establishing the second condition of theorem [ inc-32 ] . this theorem then implies that the number of incidences between @xmath633 and @xmath602 , which is also equal to the number of incidences between @xmath633 and @xmath515 , is @xmath635 moreover , since each parabola of @xmath602 contains at least eight times more rotations of @xmath633 than of @xmath636 , this bound also applies to the number of incidences between @xmath636 and @xmath602 . it therefore remains to bound the number of incidences between @xmath636 and @xmath637 , namely , between the rotations and parabolas contained in the vanishing special surfaces @xmath505 . to do so , we iterate over the surfaces , say , in the order @xmath503 . for each surface @xmath505 in turn , we process the rotations and parabolas contained in @xmath505 and then remove them from further processing on subsequent surfaces . let us then consider a special surface @xmath505 . let @xmath638 and @xmath639 denote respectively the number of rotations and parabolas contained in @xmath505 , which were not yet removed when processing previous surfaces . the number of incidences between these rotations and parabolas can be bounded by the classical szemerdi - trotter incidence bound @xcite ( see also ( [ weak23 ] ) ) , which is @xmath640 . summing these bounds over all the special surfaces @xmath505 , and using hlder s inequality and the fact , established in lemma [ sspec ] , that @xmath641 , we get an overall bound of @xmath642 @xmath643 where we use the facts that @xmath644 and @xmath645 , which follow since in this analysis each parabola and rotation is processed at most once . the two linear terms satisfy @xmath646 ( the bound obtained in the pruning process ) , and @xmath647 since @xmath648 ; see remark ( 2 ) following theorem [ thm : gk1 ] . we are not done yet , because each rotation of @xmath636 is processed only once , within the first surface @xmath505 containing it . this , however , can be handled as in @xcite . that is , let @xmath28 be a rotation which was processed within the first surface @xmath505 containing it . suppose that @xmath28 also lies on some later surface @xmath649 , with @xmath650 , and let @xmath344 be a parabola contained in @xmath649 , which has not been removed yet ; in particular , @xmath344 is not contained in @xmath505 , and thus meets it transversally , so the incidence between @xmath344 and @xmath28 can be regarded as one of the transversal incidences in @xmath505 , which we have been ignoring so far . to count them , we simply recall that each parabola , whether of @xmath637 or of @xmath602 , has at most three transversal intersections with a surface @xmath505 ( see the proof of theorem [ thm : gk1 ] ) , for a total of at most @xmath126 crossings with all the vanishing surfaces . since each of these parabolas contains at least @xmath498 rotations of @xmath516 , those `` transversal incidences '' are only a fraction of the total number of incidences , and we simply ignore them altogether . to recap , we obtain the following bound on the number of incidences between @xmath516 and @xmath515 : @xmath651 adding the bound @xmath652 on the incidences lost during the pruning process , we get the asserted bound . @xmath98 it is interesting to note that the proof technique also yields the following result . [ mgek ] let @xmath341 be a set of @xmath328 @xmath140-parabolas and @xmath110 a set of points in 3-space which satisfy the conditions of theorem [ inc - gen](i ) . then , for any @xmath653 , the number @xmath654 of points of @xmath110 incident to at least @xmath5 parabolas of @xmath341 satisfies @xmath655 * proof : * write @xmath656 for short . we clearly have @xmath657 . theorem [ inc - gen ] then implies @xmath658 , from which the first two bounds follow . if @xmath659 we use the other bound ( in ( [ weak23 ] ) ) , to obtain @xmath660 , which implies that @xmath661 ( which is in fact an equivalent statement of the classical szemerdi - trotter bound ) . @xmath98 in this section we further improve the bound in theorem [ inc - gen ] ( and corollary [ mgek ] ) using more standard space decomposition techniques . we show : [ thm : impr ] the number of incidences between @xmath357 arbitrary rotations and @xmath328 @xmath140-parabolas , defined for a planar ground set with @xmath1 points , is @xmath662 where the @xmath663 notation hides polylogarithmic factors . in particular , when all @xmath547 @xmath140-parabolas are considered , the bound is @xmath664 * proof : * we dualize the problem as follows . we map each parabola @xmath139 to the point @xmath665 in @xmath666 . each rotation @xmath28 is mapped to a 2-plane @xmath667 , which is the locus of all points @xmath668 such that @xmath28 is incident to @xmath344 . this is indeed a 2-plane , because the equations of @xmath28 , either ( [ eq : helix ] ) in the @xmath103-frame , or ( [ parabola ] ) in the @xmath136-frame , are a pair of linear ( independent ) equations in @xmath108 . so in this new setup we have @xmath328 points and @xmath357 2-planes in 4-space , and we wish to bound the number of incidences between these points and 2-planes . we note that any pair of these 2-planes intersect in at most one point . ( the corresponding statement in the primal setup is that two rotations can be incident to at most one common @xmath140-parabola . ) to bound the number of incidences , we first project the points and 2-planes onto the 3-space @xmath669 . we claim that , with a generic choice of the coordinate frame in the original @xmath138-plane , the projected points remain distinct . indeed , a point @xmath108 , dual to an @xmath140-parabola @xmath139 , is projected to the point @xmath670 , so the projected point uniquely determines @xmath29 , and also @xmath31 , because we may assume that no two points of @xmath0 have the same @xmath18-coordinate @xmath326 . hence the projected points are all distinct . this is not necessarily the case for the 2-planes . indeed , consider a 2-plane @xmath667 . its projection onto the @xmath671-space is the plane satisfying the first equation of ( [ parabola ] ) , say , namely @xmath672 it is easily checked that this equation uniquely determines the @xmath194 and @xmath190 components of @xmath28 , leaving @xmath195 ( i.e. , the shift along the @xmath362-direction that @xmath28 makes after its initial pure rotation ) undetermined . thus it is possible that several distinct rotations , all with the same @xmath194 and @xmath190 components , are projected to the same 2-plane . this has the potential danger that the projection loses incidences , when several 2-planes , incident to a common point @xmath667 , get projected into the same plane , so that , instead of several incidences with @xmath667 in 4-space , we get only one incidence in the projection . nevertheless , this bad situation can not arise . this follows from the easy observation that two distinct rotations with the same @xmath194 and @xmath190 components can not both map a point @xmath673 into the same point @xmath674 . to recap , after the projection we get @xmath328 points and at most @xmath357 planes in @xmath2 , and our goal is to bound the number of incidences between them . more precisely , we want to bound only the number of original incidences . we note that each such incidence appears as an incidence in the projection , but not necessarily the other way around . we recall that , in general , the number of incidences between @xmath328 points and @xmath357 planes in 3-space can be @xmath622 in the worst case , because of the possibility that many points lie on a common line and many planes pass through that line . this situation can also arise in our setup , but we will apply a careful analysis to show that the number of original incidences that project to such a degenerate configuration is much smaller . we proceed as follows . we fix a parameter @xmath675 , to be determined shortly , and construct the following decomposition of 3-space . first , we note that the projected points @xmath670 have only @xmath1 distinct @xmath676-coordinates , which are the @xmath18-coordinates of the points of @xmath0 . similarly , they have only @xmath1 distinct @xmath326-coordinates . we partition the 3-space by a set @xmath677 of @xmath675 planes orthogonal to the @xmath676-axis , so that within each resulting slab the projected points have at most @xmath678 distinct @xmath676-coordinates . we construct a similar collection @xmath679 of @xmath675 planes orthogonal to the @xmath326-axis , so that within each resulting slab the projected points have at most @xmath678 distinct @xmath326-coordinates . we then choose a random sample @xmath680 of @xmath675 of the projected planes . we take the set @xmath681 of @xmath682 planes , construct their arrangement , and decompose each of its cells into simplices . we obtain @xmath683 simplices , and the construction and the standard @xmath684-net theory @xcite imply that , with high probability , the following properties hold for every simplex @xmath685 of the partition : ( i ) @xmath685 is crossed by at most @xmath686 projected 2-planes ; ( ii ) the projected points that fall into @xmath685 have at most @xmath678 distinct @xmath676-coordinates and at most @xmath678 distinct @xmath326-coordinates . further refining the simplices , if necessary , we can also assume that ( iii ) each simplex contains at most @xmath687 projected points . property ( ii ) is crucial . it asserts that the number of points of @xmath0 which induce the parabolas whose dual points project into a fixed simplex is at most @xmath688 ; more precisely , there are only @xmath678 `` source '' points of @xmath0 and only @xmath678 `` target '' points , so that each of these parabolas is of the form @xmath139 , where @xmath29 is one of the @xmath678 source points and @xmath31 is one of the @xmath678 target points . ( note , by the way , that the number of parabolas , @xmath687 , in volved in a subproblem is much smaller than the maximum possible value @xmath689 , when @xmath690 . ) we now apply theorem [ inc - gen ] to each simplex @xmath685 ; that is , to the set @xmath691 of those parabolas whose ( projected ) dual points lie in @xmath685 , and to the set @xmath692 of those rotations whose ( projected ) dual 2-planes cross @xmath685 . put @xmath693 and @xmath694 . we note that some rotations in @xmath692 may be incident to no more than two parabolas in @xmath691 ; these rotations contribute @xmath695 to the overall incidence bound . by theorem [ inc - gen ] we thus have is subsumed by the other terms , because of the possibility that some of the @xmath696 rotations are incident to only one or two parabolas in a subproblem . ] @xmath697 summing these bounds over all cells @xmath685 , we get an overall bound of @xmath698 @xmath699 where , as above , @xmath663 hides polylogarithmic factors . we also have to add to the bound incidences involving points , which are projections dual to parabolas , which lie on the boundaries of the cells of the cutting . let @xmath700 , the projection of a ( unique ) point @xmath701 be such a point . let @xmath359 denote the face whose relative interior contains @xmath401 . if @xmath359 is a 2-face of some simplex @xmath685 , we can associate @xmath401 with @xmath685 : except for the single plane containing @xmath359 , any other plane incident to @xmath401 must cross @xmath685 , and we can count the incidence within the subproblem of @xmath685 . the uncounted incidences , at most one per parabola , add up to at most @xmath328 . if @xmath359 is a vertex ( so @xmath702 ) then any plane through @xmath359 either bounds or crosses some adjacent simplex , so the total number of such incidences is @xmath703 . the harder situation is when @xmath359 is an edge . again , if a plane _ crosses _ @xmath359 at @xmath401 , we can count this incidence within any adjacent simplex , arguing as in the case where @xmath359 is a 2-face . the difficult case is when the plane _ contains _ @xmath359 , and we handle it as follows . it is simpler to consider @xmath359 as a full line of intersection of two sampled planes , rather than a single edge . ( the decomposition , though , has also other edges , obtained in the decomposition of arrangement cells into simplices ; these edges require a slightly different treatment , given below . ) let @xmath704 be the projected dual points that lie on @xmath359 , and let @xmath705 denote the parabola corresponding to @xmath706 , for @xmath707 . consider the rotations @xmath28 whose dual 2-planes project to planes containing @xmath359 . rotations @xmath28 of this kind which are incident to just one of the parabolas @xmath705 are easy to handle , because the number of incidences involving these rotations is at most @xmath357 ( for the fixed line @xmath359 ) , for a total of @xmath708 . consider then those rotations @xmath28 which are incident to at least two of the parabolas @xmath705 . since the points @xmath709 lie on a common line , it follows that the points @xmath163 are also collinear in the original @xmath138-plane , lying on a common line @xmath305 . the points @xmath164 are not necessarily collinear , but they have the property that , for any pair of indices @xmath710 , the ratio @xmath711 is fixed . see figure [ abab ] . abab.pstex_t now if @xmath28 is incident to two parabolas @xmath705 , @xmath712 , then @xmath28 maps @xmath163 to @xmath164 and @xmath713 to @xmath714 . in particular , @xmath715 . this , and the fact that @xmath711 is fixed , imply that @xmath28 maps @xmath305 to the line through @xmath164 and @xmath714 , and that the slope of this line has a _ fixed absolute value _ @xmath716 . hence , considering , with no loss of generality , only lines of the latter kind with positive slope , we can partition @xmath717 into equivalence classes , so that , for each class , all the corresponding points @xmath164 lie on a common line of slope @xmath716 . moreover , there is at most one rotation that is incident to at least two parabolas from the same class ( and no rotation can be incident to two parabolas from different classes ) . thus the total number of incidences of this kind , for the fixed @xmath359 , is at most @xmath155 . summing over all lines @xmath359 , we get a total of @xmath508 such incidences . in the preceding analysis we considered only intersection lines between sampled planes , but , as noted , the cutting has additional edges , interior to cells of the arrangement . we handle such edges in almost the same way as above . that is , we consider such an edge @xmath718 , and argue , exactly as above , that the number of original incidences involving points on @xmath718 and planes that contain @xmath718 is proportional to the number @xmath719 of points on @xmath718 plus the number @xmath720 of planes containing @xmath718 . ( incidences involving planes that cross @xmath718 are also handled exactly as above , wih the same resulting bound . ) the sum @xmath721 is still at most @xmath328 . for the other sum @xmath722 , we note that the number of edges @xmath718 is @xmath683 ( instead of @xmath723 in the preceding analysis ) , but each edge @xmath718 can be contained in at most @xmath686 planes , as follows easily from the @xmath684-net theory ( this holds with high probability , but we may assume that our sample does indeed have this property ) . hence , we have @xmath724 , the same bound as above . altogether , the number of incidences is thus @xmath725 we now choose @xmath726 this choice of @xmath675 makes the first and third terms in the incidence bound equal to each other , and they both dominate the second term , as is easily verified , using the fact that @xmath342 . note also that @xmath727 when @xmath728 assume first that @xmath357 lies in this range . then the incidence bound becomes @xmath729 when @xmath730 , we use @xmath731 and get the bound @xmath732 since @xmath733 , the second term dominates the two other terms , and the bound is thus @xmath734 . finally , when @xmath735 , we use the szemerdi - trotter bound in ( [ weak23 ] ) , which is easily seen to yield the bound @xmath508 . adding all these bounds , the theorem follows . @xmath98 using this bound , we can strengthen corollary [ mgek ] , as follows . [ mgek1 ] let @xmath341 be a set of @xmath328 @xmath140-parabolas and @xmath110 a set of rotations , with respect to a planar ground set @xmath0 of @xmath1 points . then , for any @xmath4 , the number @xmath654 of rotations of @xmath110 incident to at least @xmath5 parabolas of @xmath341 satisfies @xmath736 for @xmath547 , the bound becomes @xmath737 * proof : * the proof is similar to the proof of corollary [ mgek ] , and we omit its routine details . @xmath98 in this paper we have reduced the problem of obtaining a near - linear lower bound for the number of distinct distances in the plane to a problem involving incidences between points and a special class of parabolas ( or helices ) in three dimensions . we have made significant progress in obtaining upper bounds for the number of such incidences , but we are still short of tightening these bounds to meet the conjectures on these bounds made in the introduction . to see how far we still have to go , consider the bound in corollary [ mgek1 ] , for the case @xmath547 , which then becomes @xmath738 . ( here @xmath654 coincides with @xmath58 as defined in ( h3 ) . ) moreover , we also have the szemerdi - trotter bound @xmath739 , which is smaller than the previous bound for @xmath740 . substituting these bounds in the analysis of ( h3 ) and ( h4 ) , we get @xmath62 ^ 2}{x } \le |k| = n_{\ge 2 } + \sum_{k\ge 3 } ( k-1 ) n_{\ge k } = \ ] ] @xmath741 = n_{\ge 2 } + o(s^{29/9 } ) .\ ] ] it is fairly easy to show that @xmath119 is @xmath742 , by noting that @xmath119 can be upper bounded by @xmath743 , where @xmath744 is as defined in ( h1 ) . using the upper bound @xmath745 @xcite , we get @xmath746 thus , at the moment , @xmath119 is the bottleneck in the above bound , and we only get the ( weak ) lower bound @xmath11 on the number of distinct distances . showing that @xmath747 too ( hopefully , a rather modest goal ) would improve the lower bound to @xmath748 , still a rather weak lower bound . \(ii ) it gave us a new , and considerably more involved setup in which the new algebraic technique of guth and katz could be applied . as such , the analysis in this paper might prove useful for obtaining improved incidence bounds for points and other classes of curves in three dimensions . the case of points and circles is an immediate next challenge . another comment is in order . our work can be regarded as a special variant of the complex version of the szemerdi - trotter theorem on point - line incidences @xcite . in the complex plane , the equation of a line ( in complex notation ) is @xmath749 . interpreting this equation as a transformation of the real plane , we get a _ homothetic map _ , i.e. , a rigid motion followed by a scaling . we can therefore rephrase the complex version of the szemerdi - trotter theorem as follows . we are given a set @xmath110 of @xmath357 pairs of points in the ( real ) plane , and a set @xmath750 of @xmath328 homothetic maps , and we seek an upper bound on the number of times a map @xmath751 and a pair @xmath752 `` coincide '' , in the sense that @xmath46 . in our work we only consider `` complex lines '' whose `` slope '' @xmath251 has absolute value @xmath117 ( these are our rotations ) , and the set @xmath110 is simply @xmath753 . \(a ) obtain a cubic upper bound for the number of rotations which map only two points of the given ground planar set @xmath0 to another pair of points of @xmath0 . any upper bound smaller than @xmath754 would already be a significant step towards improving the current lower bound of @xmath17 on distinct distances @xcite . the bulk of the paper was written after the passing away of gyrgy elekes in september 2008 . however , the initial infrastructure , including the transformation of the distinct distances problem to an incidence problem in three dimensions , and many other steps , is due to him . as a matter of fact , it was already discovered by elekes about 10 years ago , and lay dormant since then , mainly because of the lack of effective tools for tackling the incidence problem . these tools became available with the breakthrough result of guth and katz @xcite in december 2008 , and have made this paper possible . thanks are due to mrton elekes , who was a driving force in restarting the research on this problem . n. h. katz and g. tardos , a new entropy inequality for the erds distance problem , in _ towards a theory of geometric graphs _ , j. pach , ed . , _ contemporary math . 342 , amer . soc . press , providence , ri , 2004 , pp . 119126 . a. pressley , _ elementary differential geometry _ , springer undergraduate mathematics series , springer verlag , london , 2001 . r. quilodrn , the joints problem in @xmath755 , arxiv:0906.0555 , posted june 2 , 2009 .

the standard model of elementary particles ( sm ) has been amazingly successful , and currently the large hadron collider ( lhc ) experiment is searching for higgs boson , which is the only particle that has not been observed yet in the framework of the sm . so far , some interesting hints of the higgs boson with a mass of about 125 gev have been reported at the lhc @xcite and its discovery may not be in a far future . the higgs boson with a mass of about 100 gev is totally consistent with the electroweak ( ew ) precision measurements . therefore , the sm well - describes the nature up to the ew scale . we , however , think that the sm is not the final theory of elementary particles . the hierarchy problem has been a strong driving force to think of the physics beyond the sm , and many ideas , such as supersymmetry , extra - dimension , and technicolor , have been proposed to solve the problem . the lhc experiment is an ideal place to probe the models for the hierarchy problem , and the searches are currently going on . the first round of the lhc searches , however , does not show any serious deviations from the sm prediction . it seems that this negative search result starts creating a tension between new physics models and the requirements for the hierarchy problem . therefore it will be good time to consider other approaches to think about new physics beyond the sm . there is a different approach to consider the physics beyond the sm , based on a consideration for unexplained experimental results . the muon anomalous magnetic moment ( muon g-2 ) is one of the most precisely measured observables @xcite . the theoretical prediction from the sm by several groups has suggested that there is a discrepancy between the experimental result and the sm prediction @xcite : @xmath5 where @xmath6 is a discrepancy between the experimental result @xmath7 and the sm prediction @xmath8 . there are many discussions @xcite on uncertainties of the sm prediction to understand whether this anomaly is real or not . if the discrepancy can not be explained by the sm , this would be an evidence of the physics beyond the sm . at present , there seems no satisfactory explanation for this anomaly within the sm . therefore , it is worth while considering the physics beyond the sm seriously in order to explain the anomaly . it is interesting to note that the anomaly of muon g-2 is the same size as 1-loop contributions induced by the ew gauge bosons in the sm . this suggests that in order to explain the anomaly , new particles with masses of ew scale are required if the interactions are of order of ew gauge couplings . although the anomaly of muon g-2 has been discussed in the context of new physics models ( for example , see ref . @xcite in mssm , ref . @xcite in extra dimension model , and ref . @xcite in little higgs model ) motivated mainly by the hierarchy problem , we take a `` bottom - up '' style approach . starting from introducing new interactions with muon to generate new contribution to muon g-2 , we try to capture important features of new physics models , although we can only discuss a part of complete models . in this paper , we consider models where muon has new yukawa type interactions with new particles to generate new contributions to muon g-2 . we find that in a model where the chirarity flip of muon occurs only in external line of muon g-2 diagrams , it is necessary to introduce the relatively large yukawa coupling and the ew scale new particles . on the other hand , in a model where the chirality flip can happen in the internal lines of muon g-2 diagrams , we can obtain favorable g-2 contributions without large yukawa couplings , and they are also consistent with the precision ew observables . we also expect these particles may be observed directly and/or indirectly at the lhc experiment . this paper is organized as follows . in the next section , we consider new physics models where muon has new yukawa couplings to explain the anomaly of muon g-2 . one model is that only right - handed muon has new yukawa coupling . another one is that both right- and left- handed muons have new yukawa interactions to enhance the contributions to the muon g-2 . we show the consistent region of parameter space . in sec . 3 , we discuss effects of the new physics on ew observables and whether the parameter space which is consistent with the muon g-2 is also consistent with the precision measurements . in sec . 4 , we discuss phenomenology of these new physics models at the lhc . especially we show the possible effect on the higgs boson decay of @xmath1 , and the direct production cross sections of new particles at the lhc . is devoted to summary . in order to explain the deviation from the sm prediction of muon g-2 by new physics , muon has to have new interactions with some charged particles . since the effective operator for muon g-2 couples to photon , the loop effects via the new interactions of muon with charged particles could induce the extra contribution to the muon g-2 . in this paper , as such new interactions , we consider new yukawa - type couplings with right - handed or / and left - handed muon . we consider two cases : ( 1 ) model where only right - handed muon has new yukawa interaction and ( 2 ) model where both right - handed and left - handed muons have new yukawa interactions . we consider the following yukawa interaction where the right - handed muon couples to new su@xmath11 singlet fermion @xmath10 and singlet scalar @xmath9 : @xmath12 here @xmath13 is the right - handed muon . qed charges of new fermion @xmath10 and scalar @xmath9 are @xmath14 and @xmath15 , respectively , in order to have the gauge invariant yukawa interaction . the masses of @xmath10 and @xmath9 are denoted by @xmath16 and @xmath17 , respectively . in order to simplify the model , we have assumed a @xmath18 parity , under which the sm particles are even and new particles @xmath9 and @xmath10 are odd . this type of the parity may be interesting for the dark matter . we also note that even if we impose the @xmath18 parity , the right - handed electron , for example , can have similar yukawa interaction with @xmath9 and @xmath10 . this will cause severe flavor mixing problem . we will not discuss the complete model here , however , we implicitly assume that @xmath10 or @xmath9 has approximate muon flavor number so that the flavor mixing is strongly suppressed . , width=566 ] since the right - handed muon couples to @xmath10 and @xmath9 , their radiative corrections induce the muon g-2 , @xmath19 as shown in fig . [ diagramg2 ] . the contribution @xmath19 is given by @xmath20 , \label{muong2}\end{aligned}\ ] ] where @xmath21 and @xmath22 are momenta of external muons and @xmath23 is a photon momentum and a limit @xmath24 is taken . here @xmath25 are so called passarino - veltman functions @xcite , and their definitions in this paper are shown in appendix . the explicit forms are @xmath26 where @xmath27 , @xmath28 and the higher order terms of @xmath29 are neglected . the first term in eq . ( [ muong2 ] ) comes from fig . [ diagramg2](a ) and the second is from fig . [ diagramg2](b ) . ) as a function of @xmath14 and @xmath17 . here we take @xmath30 and @xmath31 gev . we show contours for @xmath32 and @xmath33 , from right to left , corresponding to @xmath34 and @xmath35 deviation from the measured value , respectively.,width=453 ] in fig . [ fig_amu ] , we show the contribution to @xmath19 as a function of @xmath14 and @xmath17 . here we take @xmath30 and @xmath31 gev . we show contours for @xmath32 and @xmath33 , from right to left , corresponding to @xmath34 and @xmath35 deviation from the measured value , respectively . we note that @xmath36 dependence of @xmath19 is trivial , that is , @xmath19 is proportional to @xmath37 , as shown in eq . ( [ muong2 ] ) . for example , if one takes @xmath38 , the values of @xmath19 in fig . [ fig_amu ] reduce by a factor @xmath39 . one can see that the region with @xmath40 ( which corresponds to @xmath41 ) is disfavored by the data of muon g-2 . it is interesting to notice that the neutral fermion ( @xmath42 , corresponding to @xmath43 ) is difficult to explain the anomaly in any values of @xmath17 , and on the other hand , the neutral scalar ( @xmath44 , corresponding to @xmath45 ) can potentially accommodate the anomaly if the scalar @xmath9 is not so heavy . it is also interesting to note that the multi - charged fermion and scalar ( such as @xmath46 , corresponding to @xmath47 ) are also favored by the muon g-2 anomaly if @xmath17 is a right value for the anomaly . therefore , the anomaly of muon g-2 constrains the qed charges as well as the mass scale of new particles . 2 ( @xmath19 ) as a function of @xmath16 and @xmath17 for ( a ) @xmath45 and and @xmath30 ( b ) @xmath48 and @xmath49 , respectively . contours for @xmath32 and @xmath33 ( corresponding to @xmath34 and @xmath35 deviation from the measured value ) are shown , similar to fig . [ fig_amu ] . , title="fig:",width=321 ] 2 ( @xmath19 ) as a function of @xmath16 and @xmath17 for ( a ) @xmath45 and and @xmath30 ( b ) @xmath48 and @xmath49 , respectively . contours for @xmath32 and @xmath33 ( corresponding to @xmath34 and @xmath35 deviation from the measured value ) are shown , similar to fig . [ fig_amu ] . , title="fig:",width=321 ] in fig . [ fig_amu_2 ] , we show @xmath19 as a function of @xmath16 and @xmath17 for ( a ) @xmath45 @xmath50 and @xmath30 and ( b ) @xmath48 @xmath51 and @xmath49 . contours for @xmath32 and @xmath33 ( corresponding to @xmath34 and @xmath35 deviation from the measured value ) are shown , similar to fig . [ fig_amu ] . , one should simply multiply the value of @xmath19 in fig . [ fig_amu_2 ] ( a ) and ( b ) by a factor @xmath52 and @xmath53 , respectively . ] as one can see , new particles with their masses of about @xmath54 gev and the coupling @xmath55 are strongly favored by the anomaly of muon g-2 when @xmath45 . when the qed charge of @xmath10 decreases further negatively ( from @xmath56 to @xmath57 ) , the muon g-2 gets larger even if the yukawa coupling @xmath36 gets smaller , as shown in fig . [ fig_amu_2 ] ( b ) . since the anomaly of muon g-2 requires relatively light new particles ( @xmath58 ) and/or relatively large yukawa coupling ( @xmath59 ) in this scenario , the ew precision observables can be affected by these new particles and new interactions . therefore , we will check the effects of these particles on the ew observables in a next section . we would also like to point out that the anomaly of muon g-2 suggests the existence of relatively light new particles and/or multi - charged particles . therefore , it is very interesting to know whether these particles can be found directly or indirectly at the lhc . we will study the direct productions of these particles and the effects on higgs decay to @xmath60 at the lhc in a later section . unlike the previous case , here we consider that both right- and left - handed muons have new yukawa interactions : @xmath61 where @xmath62 , @xmath63 and @xmath9 are the second generation su(2 ) doublet lepton , su(2 ) doublet and singlet scalars , respectively , and @xmath10 is su(2 ) singlet fermion , whose mass is @xmath16 . here we have assumed the @xmath18 parity again in order to simplify the model . under the @xmath18 parity , the sm particles are even , and new particles @xmath9 , @xmath64 and @xmath10 are odd . we also implicitly assume the approximate muon flavor symmetry , under which @xmath10 or ( @xmath65 ) has muon flavor number , so that the flavor mixing of this type of new yukawa couplings are strongly suppressed . the model we discuss in this paper may be a part of the complete model . however , we think this part of the complete model is crucial for the muon g-2 . the qed charges of new particles are represented in term of qed charge of @xmath10 field ( @xmath14 ) as follows : @xmath66 since @xmath67 , @xmath9 and @xmath68 can mix each other . for example , the following gauge invariant term induces the @xmath69 mixing mass terms after the electroweak symmetry is spontaneously broken : @xmath70 here we parameterize the mass terms for @xmath9 and @xmath68 as follows : , mass term such as @xmath71 is also possible . here we neglect such a mass term for simplicity . even if we include such a mass term , our result does not change qualitatively . ] @xmath72 thus diagonalizing this mass matrix , we define the mass eigenstates @xmath73 as @xmath74 here @xmath75 is a mixing unitary matrix , which diagonalizes the mass matrix shown in eq . ( [ mass_matrix ] ) . the mass eigen values are taken to be @xmath76 . this type of mixing terms are important to induce large contribution to muon g-2 . the contributions to muon g-2 are summarized by @xmath77 where @xmath21 and @xmath22 are momenta of external muons , and @xmath23 is a photon momentum and a limit @xmath24 is taken . here we show the explicit expressions of above passarino - veltman functions : @xmath78 where @xmath79 and we neglected the higher order of @xmath80 . the effective operator which expresses the muon g-2 is written by @xmath81 where @xmath82 is a vacuum expectation value of higgs boson , and @xmath83 is a field strength of photon field , and @xmath84 is a typical scale related to new physics . as one see , the chirality of muon has to flip in this interaction . in the case where only right - handed muon has the new yukawa interaction , the chirality flipping of muon happens in the external muon line in the loop diagram . on the other hands , in the present case where both right- and left - handed muons have the new yukawa interactions , the chirality flipping can occur in the internal fermion line , which is proportional to the mass of the fermion @xmath10 . that is why there are terms which are proportional to @xmath85 in eq . ( [ muong2_model2 ] ) . this contribution enhances the effects of muon g-2 . therefore , it is very important to explain the anomaly of muon g-2 . ) as a function of @xmath16 and @xmath86 in case of ( a ) @xmath45 and ( b ) @xmath42 . contours for @xmath32 and @xmath33 ( corresponding to @xmath34 and @xmath35 deviation from the measured value ) are shown . here we assume that @xmath87 , @xmath88 . we also take ( a ) @xmath89 and ( b ) @xmath90 . , title="fig:",width=321 ] ) as a function of @xmath16 and @xmath86 in case of ( a ) @xmath45 and ( b ) @xmath42 . contours for @xmath32 and @xmath33 ( corresponding to @xmath34 and @xmath35 deviation from the measured value ) are shown . here we assume that @xmath87 , @xmath88 . we also take ( a ) @xmath89 and ( b ) @xmath90 . , title="fig:",width=321 ] in fig . [ muong2_model2 - 1 ] , we show @xmath19 as a function of @xmath16 and @xmath86 ( which is a mass of lighter scalar state ) in case of ( a ) @xmath91 and ( b ) @xmath42 . contours for @xmath92 , 10.1 , 18.1 , 26.1 , 34.1 , 42.1 and 50.1 ( corresponding to @xmath34 and @xmath35 deviation from the measured value ) are shown . here we took @xmath93 , @xmath94 , and ( a ) @xmath89 ( b ) @xmath90 . note that in order to generate the positive contribution to muon g-2 in fig . [ muong2_model2 - 1 ] ( a ) , the sign of new yukawa couplings had to be taken as @xmath95 . on the other hand , in fig . [ muong2_model2 - 1 ] ( b ) , the sign of @xmath96 should be positive . as shown in fig . [ muong2_model2 - 1 ] , compared to fig . [ fig_amu_2 ] , even smaller yukawa couplings and heavier new particles can accommodate the anomaly of muon g-2 because of the enhancement mentioned above . as can be seen from fig . [ muong2_model2 - 1 ] , when @xmath97 , the ew scale new particles are expected in order to explain the anomaly of muon g-2 . therefore , we will analyze the effects on the ew precision observables and show the consistent region of parameters in a next section . in the previous section , we showed that the relatively light new particles are required in order to explain the anomaly of muon g-2 . in addition , in the case where the only right - handed muon has new yukawa interaction , the new yukawa coupling should be relatively large for the anomaly of muon g-2 . therefore , we should check if the relatively light particles and the relatively large yukawa coupling are consistent with the ew precision measurements . although models which we consider in this paper may be only a part of complete models , we should know how these new particles affect ew observables . in this section , we adopt the formalism in refs.@xcite in order to include the oblique corrections as well as the vertex corrections in the ew observables . first we briefly summarize the formalism @xcite . in the presence of ew to tev scale physics , it is well known that the oblique corrections in gauge boson self - energy are important . they are parameterized by peskin - takeuchi s s , t , u parameters @xcite : @xmath98 here we use the notation @xmath99 and @xmath100 to refer to the cosine and sine of the weak mixing angle and @xmath101 . in addition , in refs.@xcite , @xmath102 and @xmath103 parameters are introduced in order to account for the smaller corrections : @xmath104 furthermore , new particles have effects on the running qed coupling constant @xmath105 @xcite : @xmath106 here @xmath107 , @xmath108 , @xmath109 , and @xmath110 are the leptonic , the five - flavor hadronic , the top quark and the new physics contributions to the running of the qed coupling constant respectively . the new physics contributions to the running of the qed coupling constant @xmath110 are defined as @xmath111 these oblique parameters contribute to ew observables , as shown in refs . @xcite . if muon has @xmath112 new yukawa couplings , extra yukawa couplings , for simplicity . ] it induces non - universal vertex corrections to the @xmath113 coupling . the standard model coupling of @xmath113 is given by @xmath114,\end{aligned}\ ] ] where @xmath115 are the standard model couplings , and their tree level contributions are @xmath116 and @xmath117 . when we take into account of corrections to @xmath113 vertex via the new yukawa coupling ( including the corresponding wave function renormalization ) , the @xmath113 coupling is modified by @xmath118.\end{aligned}\ ] ] here the vertex corrections generated by new particles are parameterized by @xmath119 . the corrections to @xmath120 vertex generates the corrections to the fermi constant @xmath121 and therefore , we parameterize it as @xmath122 . the @xmath122 for the new physics contributions from the vertex and box diagrams to the @xmath123-decay process is defined as @xmath124 here @xmath125 is the muon decay constant including effects of sm radiative corrections and new physics oblique corrections except the vertex and box corrections from new physics . .experimental data and theoretical predictions of electroweak observables . the predicted values of the sm and sample model discussed in sec . 3.2 are shown . [ cols="^,^,^,^,^,^",options="header " , ] of the ew observables listed in table . [ ewobs ] as a function of the higgs boson mass @xmath126 within the sm . ( b ) @xmath127 contours shown in dashed lines in s - t plane , assuming that other oblique corrections ( @xmath128 , @xmath103 , @xmath102 ) and vertex corrections are zero . reference higgs boson mass and top quark mass are taken to be @xmath129 gev and 173.2 gev , respectively . we also show the predicted s - t values ( shown in solid lines ) in the sm , varying the higgs boson mass and top quark mass from the reference values . , title="fig:",width=302 ] of the ew observables listed in table . [ ewobs ] as a function of the higgs boson mass @xmath126 within the sm . ( b ) @xmath127 contours shown in dashed lines in s - t plane , assuming that other oblique corrections ( @xmath128 , @xmath103 , @xmath102 ) and vertex corrections are zero . reference higgs boson mass and top quark mass are taken to be @xmath129 gev and 173.2 gev , respectively . we also show the predicted s - t values ( shown in solid lines ) in the sm , varying the higgs boson mass and top quark mass from the reference values . , title="fig:",width=302 ] using the formalism in ref.@xcite , one can calculate ew observables listed in table [ ewobs ] without assuming lepton universality . first , we show a result of the sm fit of the ew observables in fig . [ smfit ] ( a ) as a function of higgs boson mass @xmath126 . the best fit point is at @xmath130 gev . this is consistent with the result shown by the lep electroweak working group @xcite . in table [ ewobs ] , we show the sm fit when higgs boson mass is assumed to be 125 gev , which has been suggested by the latest lhc data . in fig . [ smfit](b ) , @xmath127 contours are shown in dashed lines in s - t plane . here we assume that other oblique corrections ( @xmath128 , @xmath103 , @xmath102 ) and vertex corrections are zero . reference higgs boson mass and top quark mass are taken to be 125 gev , 173.2 gev , respectively . the predicted values of s - t parameters are also shown by varying the higgs boson mass and top quark mass from the reference values . we note that the sm fit with lighter higgs boson is good except for the muon g-2 , and small positive s and t @xmath131 ) can decrease the @xmath127 further . therefore , we need new physics which largely contributes to muon g-2 , but whose effects on other ew observables are small . in this section , we analyze the ew observables in models discussed in the previous section . in this analysis , we assume @xmath132 gev . since the new scalar @xmath9 and new fermion @xmath10 are su(2 ) singlet , they do not couple to w boson . but if it has a qed charge , it can couple to photon and z boson . here we list the corrections to self - energy functions of gauge bosons in this model . the singlet fermion ( @xmath10 ) contributions are given by @xmath133 , \nonumber\\ \pi_{\gamma \gamma}^{(\chi)}(p^2)&= & -\frac{e^2}{4\pi^2}q_\chi^2\left [ m^2_{\chi}b_0(\chi,\chi)-p^2\left\{b_1(\chi,\chi ) + b_{21}(\chi,\chi)\right\ } -2(1-\epsilon)b_{22}(\chi,\chi ) \right ] , \nonumber \\ \pi_{\gamma z}^{(\chi)}(p^2 ) & = & \frac{geq_\chi^2s_w^2}{4\pi^2c_w } \left [ m_\chi^2b_0(\chi,\chi)-p^2\left\ { b_1(\chi,\chi)+b_{21}(\chi,\chi ) \right\ } -2(1-\epsilon)b_{22}(\chi,\chi ) \right ] . \nonumber\\ \label{singlet_chi}\end{aligned}\ ] ] here @xmath134 , which are passarino - veltman functions and are shown explicitly in appendix . we use the dimensional regularization in space - time dimension @xmath135 . the singlet scalar ( @xmath9 ) contributions are given by @xmath136 here passarino - veltman functions @xmath137 and @xmath138 are shown in appendix . we can easily show that peskin - takeuchi stu parameters in this case vanish ( @xmath139 ) , since the new particles do not have su(2)@xmath140 interactions . thus , the leading contributions to the oblique corrections are @xmath102 parameter and @xmath110 : @xmath141\right\}\right ] \nonumber \\ & & + \frac{s_w^4 q_\phi^2}{3\pi}\left[1-\frac{12m_\phi^2}{m_z^2 } \left\{1-\sqrt{\frac{4m_\phi^2}{m_z^2}-1 } \tan^{-1}\left[\frac{1}{\sqrt{\frac{4m_\phi^2}{m_z^2}-1}}\right]\right\}\right],\end{aligned}\ ] ] @xmath142\right ] \nonumber \\ & & -\frac{2\alpha q_\phi^2}{9\pi}\left [ 1-\frac{3m_\phi^2}{m_z^2 } + \frac{3}{4}\sqrt{\frac{4m_\phi^2}{m_z^2}-1 } \tan^{-1}\left[\frac{1}{\sqrt{\frac{4m_\phi^2}{m_z^2}-1}}\right]\right],\end{aligned}\ ] ] where we assume @xmath143 and @xmath144 . as discussed in the previous section , in order to explain the anomaly of muon g-2 , the new yukawa coupling in this model should be relatively large . in this case , the potentially large vertex corrections may be expected , as shown in fig.[vertex_fig ] . coupling.,width=566 ] the results are expressed by @xmath145,\end{aligned}\ ] ] where @xmath21 , @xmath146 and @xmath23 are muon , anti - muon and z - boson momenta , respectively . here passarino - veltman functions @xmath147 and @xmath148 are explicitly shown in appendix . . @xmath149 are shown as a function of @xmath16 and @xmath17 in case of @xmath45 and @xmath30 , where @xmath150 includes all corrections , and @xmath151 contains all corrections except for the vertex corrections . ( b ) @xmath127 as a function of @xmath16 and @xmath17 in case of @xmath45 and @xmath30.,title="fig:",width=302 ] . @xmath149 are shown as a function of @xmath16 and @xmath17 in case of @xmath45 and @xmath30 , where @xmath150 includes all corrections , and @xmath151 contains all corrections except for the vertex corrections . ( b ) @xmath127 as a function of @xmath16 and @xmath17 in case of @xmath45 and @xmath30.,title="fig:",width=302 ] . @xmath149 is shown as a function of @xmath16 and @xmath17 in case of @xmath48 and @xmath49 , similar to fig . [ vertex](a ) . ( b ) @xmath127 as a function of @xmath16 and @xmath17 in case of @xmath48 and @xmath49 . , title="fig:",width=302 ] . @xmath149 is shown as a function of @xmath16 and @xmath17 in case of @xmath48 and @xmath49 , similar to fig . [ vertex](a ) . ( b ) @xmath127 as a function of @xmath16 and @xmath17 in case of @xmath48 and @xmath49 . , title="fig:",width=302 ] the vertex corrections affect ew observables such as @xmath150 , @xmath152 , and @xmath153 . among these observables , we find that the effect on the @xmath150 is the most important in the fit because @xmath150 is most precisely measured . in fig.[vertex ] ( a ) , we show the effect of vertex corrections in @xmath150 , that is , @xmath154 as a function of @xmath16 and @xmath17 in case of @xmath45 and @xmath30 . here @xmath150 is a theoretical prediction in this model containing all corrections , on the other hand , @xmath151 contains all corrections except for the vertex corrections . the difference @xmath155 shows the effect of the vertex corrections . note that the size of experimental error of @xmath150 is @xmath156 for @xmath157 , as shown in table . [ ewobs ] . as one can see in fig . [ vertex](a ) , the vertex correction can change the prediction of @xmath150 about @xmath157 in the region where the muon g-2 can be explained by the new physics contributions . since the sm prediction of @xmath150 is smaller than that of the experimental result ( shown in table . [ ewobs ] ) , the vertex corrections do not help to improve the prediction of @xmath150 . therefore , the region of small @xmath10 mass is very constrained . in fig . [ vertex ] ( b ) , @xmath127 is shown as a function of @xmath16 and @xmath17 in case of @xmath45 and @xmath30 . as one can see from the figure , @xmath127 is large in the region of small @xmath16 because of the large vertex corrections , discussed above . in the region of larger @xmath16 and @xmath17 , it is difficult to explain the anomaly of muon g-2 because new particles are too heavy , and hence the @xmath127 gets larger . the minimum of @xmath127 is around @xmath158 gev and @xmath159 gev , and we see that the light new particles are favored by the ew observables including muon g-2 . in fig . [ vertex2 ] ( a ) , the effect of vertex corrections on @xmath150 , @xmath155 is shown in case of @xmath48 and @xmath49 , similar to fig . [ vertex ] ( a ) . the vertex corrections can be as large as @xmath157 of the experimental error of @xmath150 , and they increase the @xmath127 . in fig . [ vertex2 ] ( b ) , the @xmath127 is shown as a function of @xmath16 and @xmath17 . as one can see , the region of small @xmath16 is disfavored by the vertex corrections of @xmath150 , as shown in fig . [ vertex2 ] ( a ) . the minimum of the @xmath127 is around @xmath160 gev and @xmath159 gev . therefore , the ew precision measurements ( including muon g-2 ) prefer the relatively light new particles in this case . contours ( without including muon g-2 result ) in s - t plane ( @xmath175 to 34 in dashed lines ) and s - t values ( solid lines with points ) in this model . here we take @xmath89 , @xmath45 ( corresponding to @xmath176 and @xmath177 ) , and @xmath178 . two solid lines correspond to two different values of @xmath179 , @xmath88 and @xmath180 , varying @xmath181 from 200 gev to 400 gev at 50 gev step . , width=491 ] in fig . [ st_para1 ] , we show numerical result of @xmath127 contours in s - t plane ( without including muon g-2 result ) , shown in dashed lines . to draw the @xmath127 contours in s - t plane , we assume that other oblique corrections ( @xmath128 , @xmath102 , and @xmath103 ) and vertex corrections are zero . as one can see from the figure , slightly positive s and t ( @xmath182 and @xmath183 ) are favored by ew observables . in fig . [ st_para1 ] , we also show the predicted values of s - t parameters in this model , shown in solid lines . here we took @xmath89 , @xmath45 ( corresponding to @xmath176 and @xmath177 ) , and @xmath178 . two solid lines correspond to two different values of @xmath179 , @xmath88 and @xmath180 . the points on the solid lines represent the predicted values of s and t , varying @xmath181 from 200 gev to 400 gev at 50 gev step . when @xmath178 and @xmath88 , @xmath184 gev and @xmath185 gev . as roughly shown in eqs . ( [ st ] ) , when @xmath186 gev and @xmath177 , @xmath187 and @xmath188 , and hence @xmath189 and @xmath190 . as @xmath181 becomes larger and about 300 gev , both @xmath191 and @xmath192 get closer to zero because the doublet scalars are almost degenerate . then when @xmath181 is increased further , @xmath191 gets negative because @xmath193 , but @xmath190 . in fig . [ st_para1 ] , one can see this behavior . even when we increase @xmath179 , we can see that this behavior is almost same . except for @xmath90 and @xmath42 ( corresponding to @xmath194 and @xmath195 ) . , width=491 ] in fig . [ st_para2 ] , the same figure is shown except for @xmath90 and @xmath42 ( corresponding to @xmath194 and @xmath195 ) . when @xmath42 ( @xmath195 ) and @xmath186 gev , @xmath196 and @xmath187 , so that @xmath191 is negative . as @xmath181 gets larger , @xmath191 gets larger . for @xmath197 gev , @xmath198 because the doublet scalars are almost degenerate . for @xmath199 gev , @xmath191 becomes positive . this behavior is different from the previous case with @xmath177 since the sign of @xmath174 is different . note that @xmath192 is always positive unless su(2 ) scalar doublet is degenerate . as one can see in both cases , the small non - degeneracy of su(2 ) scalar doublet can make @xmath127 better . in our numerical analysis , we also include the vertex corrections in @xmath113 , @xmath200 and @xmath201 vertices even though the yukawa couplings @xmath202 and @xmath203 can be small in this model . the expression of the vertex corrections is summarized as follows : @xmath204,\\ \delta g_r^\mu & = & \frac{y_r^2}{16\pi^2}\left [ 2 \sum_{ij } \left\{\left(-\frac{1}{2}-q_\phi s_w^2\right)v_{2i}^ * v_{2j}-q_\phi s_w^2 v_{1i}^ * v_{1j}\right\ } v_{1i}v_{1j}^ * c_{24}(s_i,\chi , s_j ; p , q - p)\right.\nonumber \\ & & + \sum_i q_\chi s_w^2 \frac{1}{2}-2c_{24}-m_z^2 ( c_{12}+c_{23})+m_\chi^2 c_0\right\}(\chi , s_i,\chi;q - p , p ) \nonumber \\ & & \left . -\sum_i s_w^2|v_{1i}|^2(b_0+b_1)(s_i,\chi;p ) \right],\\ \delta g_l^{\nu_\mu } & = & \frac{y_n^2}{16\pi^2}\left [ 2\left(\frac{1}{2 } -s_w^2 q_{\phi_1 } \right ) c_{24}(\phi_1,\chi,\phi_1;q - p , p ) \right.\nonumber\\ & & + q_\chi s_w^2\left\ { \frac{1}{2}-2 c_{24}-m_z^2(c_{12}+c_{23})+m_\chi^2 c_0 \right\}(\chi,\phi_1,\chi;q - p , p ) \nonumber \\ & & \left . -\frac{1}{2}(b_0+b_1)(\phi_1,\chi;p ) \right],\end{aligned}\ ] ] where @xmath21 and @xmath23 are momenta for muon in @xmath205(muon neutrino in @xmath206 ) and z - boson , respectively . correction to @xmath207-vertex is written by @xmath208,\end{aligned}\ ] ] where @xmath21 , @xmath22 and @xmath23 are momenta for muon , muon - neutrino , and w - boson , respectively . here we use approximation @xmath28 . contours as a function of @xmath16 and @xmath86 in case of ( a ) @xmath45 and @xmath89 and ( b ) @xmath42 and @xmath90 . here we take @xmath87 and @xmath88.,title="fig:",width=302 ] contours as a function of @xmath16 and @xmath86 in case of ( a ) @xmath45 and @xmath89 and ( b ) @xmath42 and @xmath90 . here we take @xmath87 and @xmath88.,title="fig:",width=302 ] in fig . [ chi2_model2 ] , we show @xmath127 contours as a function of @xmath16 and @xmath86 in case of ( a ) @xmath45 and @xmath89 and ( b ) @xmath42 and @xmath90 . here we have taken @xmath88 , and @xmath87 . @xmath181 is chosen so that the @xmath127 becomes the smallest , provided that @xmath209 for @xmath45 and @xmath210 for @xmath42 respectively . in both cases shown in fig . [ chi2_model2 ] , the @xmath127 can be as small as 22 , which is smaller than those in the previous case where only right - handed muon has new yukawa coupling . as seen from fig . [ chi2_model2 ] , the relatively light @xmath10 and @xmath211 are favored by the ew observables . in table . 1 , we show the predicted values of the ew observables in case of @xmath42 , @xmath212 gev , @xmath213 gev , @xmath214 gev@xmath215 , and @xmath216 gev@xmath215 . as shown in table . 1 , the fit is quite good . therefore , it will be important to know what kind of effects we expect from these particles at the lhc . in a next section , we will discuss the possible effects of these new particles at the lhc . in the previous section , we show that in order to explain the anomaly of muon g-2 , it is strongly suggested that there should be the ew scale new particles . it is also suggested that multi - charged particles may be favored by the anomaly of muon g-2 . therefore , these particles may be reachable directly and/or indirectly at the lhc . here we discuss the effects of these new particles in the higgs decay @xmath1 and their direct productions at the lhc . the su(2 ) singlet and doublet scalars can couple to the higgs boson through the following scalar interactions : @xmath217 where @xmath218 is a standard model higgs boson doublet . in case where only su(2 ) singlet scalar @xmath9 exists , only coupling @xmath219 is non - zero , but in case where both su(2 ) singlet @xmath9 and doublet @xmath64 scalars are there , all couplings are allowed ( @xmath220 ) is zero , terms such as @xmath221 and @xmath222 are also possible . in our analysis , we assume that such terms are small , for simplicity . even if we include such terms , our result does not change qualitatively . ] . in the case where only right - handed muon has new yukawa interaction , there is su(2 ) singlet @xmath9 , and its interaction with higgs boson is @xmath223 where @xmath224 in this model and @xmath82 is a vacuum expectation value of higgs field , @xmath225 gev . when @xmath220 , which is the qed charge of @xmath9 , is non - zero , @xmath9 can contribute to the higgs decay @xmath1 . the decay width @xmath226 is given by @xmath227 where the amplitude @xmath228 at leading order is written by @xmath229 here the first , the second , and the third terms are top , w , and @xmath9 contributions , respectively , and @xmath230 . the function @xmath231 are given by @xmath232,\\ f_w(\tau)&=&2 + 3\tau^{-1}+3\tau^{-1}(2-\tau^{-1})f(\tau),\\ f_\phi(\tau_\phi ) & = & \tau^{-1}\left[-1+\tau^{-1}f(\tau)\right],\end{aligned}\ ] ] where @xmath233 for @xmath234 . branching ratio of @xmath1 is approximately given by @xmath235 because the total decay width does not change much , even if @xmath226 changes . here @xmath236 and @xmath237 are the sm predictions at the leading order . normalized by the sm prediction @xmath238 as a function of @xmath17 and @xmath239 . here we take @xmath132 gev . , width=377 ] notice that this ratio is a function of @xmath17 and @xmath240 when higgs boson mass @xmath126 is fixed . as we discussed earlier , in order to explain the anomaly of muon g-2 , larger @xmath220 may be favored . as can be seen in eq . ( [ hgg ] ) , larger @xmath220 induces larger effect on @xmath1 . in fig . [ higgs_lhc ] , the ratio @xmath241 is shown as a function of @xmath17 and @xmath239 . here we took @xmath132 gev . as one can see from the figure , the effect can be significant if @xmath239 is large and @xmath17 is @xmath242 gev . for example , as we discussed , when @xmath243 , @xmath244 and @xmath245 gev , @xmath246 . unfortunately , we can not predict the branching ratio @xmath247 in this model because @xmath248 is unknown . however , it is not surprising that @xmath247 is very different from the sm prediction . in the case where both right- and left - handed muons have new yukawa interactions , the su(2 ) singlet and doublet scalars have interactions with higgs boson , as shown in eq . ( [ higgs_int ] ) and they generates the following interactions : @xmath249 where @xmath250 they contribute to the decay width of @xmath1 , @xmath251 when @xmath45 , which corresponds to @xmath176 and @xmath177 ( as an example we discussed above ) , only @xmath252 can contribute to @xmath1 , because @xmath211 does not couple to photon . the result is same as one in fig . [ higgs_lhc ] for @xmath253 and @xmath254 . in this case , the effect can be significant if @xmath255 is large . for example , when @xmath256 gev and @xmath257 and @xmath56 , @xmath258 and 1.05 , respectively . when @xmath42 , which corresponds to @xmath194 and @xmath195 , scalars @xmath73 can contribute to @xmath1 . similarly to the previous case , if the higgs couplings @xmath259 and @xmath260 are large , the effect on @xmath1 can be significant . for example , when @xmath87 and @xmath261 , we obtain @xmath262 and @xmath263 for @xmath264 , on the other hand , @xmath265 and @xmath266 for @xmath267 . as a result , for @xmath178 , @xmath268 and @xmath269 for @xmath270 and @xmath56 ( 0.71 and 1.56 for @xmath271 and @xmath272 ) , respectively . although it is difficult to predict the decay branching ratio of @xmath1 in this model because it depends on many unknown parameters in scalar interactions , it is possible to have large effect on @xmath1 process . as we have discussed in the previous sections , the anomaly of muon g-2 strongly suggests the existence of ew scale new particles . models we discussed in this paper may not be complete , namely , there may be other new particles in the complete models . therefore , in this paper , we can not discuss the well - defined signatures at the lhc . however , we think that it is important to know how many events of new particles can be generated at the lhc . here we present the production cross sections of new particles at the lhc . in case of @xmath45 : @xmath273 , and ( b ) @xmath274 in case of @xmath195 : @xmath275 at the lhc ( tevatron ) , as a function of @xmath16 and @xmath276 , respectively . here we take @xmath277 tev and 14 tev for lhc and 1.96 tev for tevatron . , title="fig:",width=302 ] in case of @xmath45 : @xmath273 , and ( b ) @xmath274 in case of @xmath195 : @xmath275 at the lhc ( tevatron ) , as a function of @xmath16 and @xmath276 , respectively . here we take @xmath277 tev and 14 tev for lhc and 1.96 tev for tevatron . , title="fig:",width=302 ] in the case where the right - handed muon has new yukawa coupling , singlet scalar ( @xmath9 ) and singlet fermion ( @xmath10 ) exist in the model . when @xmath45 , only @xmath10 couples to the standard model gauge bosons ( @xmath278 and @xmath279 ) because @xmath44 . therefore , the fermion @xmath10 can be pair - produced via ew gauge interactions . we calculate the pair production cross section of @xmath10 at the lhc ( @xmath280 ) and at tevatron ( @xmath281 tev ) as a function of @xmath16 , as shown in fig . [ fig_lhc](a ) . to calculate the production cross sections , we have used the madgraph @xcite . in order to explain the anomaly of muon g-2 , the mass of the fermion @xmath10 should be of order of about 100 gev . from fig . [ fig_lhc](a ) , the production cross section of @xmath10 is in the range of 0.5 - 0.0005 pb ( 1 - 0.001 pb ) for @xmath282 gev and @xmath277 tev ( @xmath283 tev ) at the lhc and also 0.1 - 0.001 pb for @xmath284 gev and @xmath281 tev at the tevatron . if we consider a luminosity of 10 fb@xmath285 for @xmath277 tev and 14 tev , the number of events produced at the lhc is of order of 5 - 5000 and 10 - 10000 , respectively . therefore , number of signal events can be significant . in this case , if @xmath9 is lighter than @xmath10 , @xmath10 can decay to @xmath123 and @xmath9 . if @xmath9 is a stable neutral particle , the final signature may be a @xmath286 + missing energy . this will suffer from the standard model background such as @xmath287 production whose production rate is about 0.29 ( 0.51 ) pb for @xmath277 tev ( 14 tev ) at the lhc . therefore , the detail study will be very important to discuss the discovery potential of such a signal event . however , the detail signature will depend on the particle spectrum in the complete model . thus we will leave the detail study . if we consider multi - charged @xmath10 , the cross section increases by a factor @xmath288 , and the cross section further becomes larger , and hence the constraints from the lhc will be very important . pair production is also not negligible in this case . ] in the second model where one has a @xmath289 doublet scalar in addition to @xmath9 and @xmath10 , the scalars can be pair - produced via ew gauge interaction . for example , fig . [ fig_lhc ] ( b ) shows the production cross sections for the charged @xmath290 scalars as a function of @xmath276 in case of @xmath42 , which corresponds to @xmath194 and @xmath195 . we show the predicted values for @xmath277 tev and 14 tev at the lhc and @xmath281 tev at the tevatron . as can be seen from the figure , the cross sections are in the range of 0.1 - 0.0001 pb ( 0.2 - 0.0005 pb ) for @xmath291 gev at @xmath277 tev ( @xmath283 tev ) and also 0.02 - 0.0001 pb for for @xmath292 gev at the tevatron . for simplicity , we have assumed that there are no @xmath9-@xmath68 scalar mixing . in this case , @xmath293 are also produced via @xmath294-boson . if masses of @xmath289 doublet scalar are degenerated , we have checked that the cross section is similar to one in fig . [ fig_lhc ] ( b ) . in general , the signal events have the form @xmath295 where @xmath296 and @xmath297 . therefore , fig . [ fig_lhc ] shows the typical values of the cross section in this model . although we can not discuss the complete signatures because our sample models are not complete , we stress that the searches for the production processes of these new particles via ew interactions are very important in models where the anomaly of muon g-2 can be explained by these new particles . the current lhc experiment does not show any deviation from the sm predictions . this result puts severe constraints on new physics models . especially , colored particles are strongly constrained since the lhc as a hadron collider can easily produce the colored particles via qcd interactions . for example , in the constrained minimal supersymmetric standard model ( cmssm ) , gluino as well as the first and the second generation squarks are strongly constrained and their typical lower mass limits are about 1 tev @xcite . therefore , models motivated by the hierarchy problem are getting more constrained , and especially cmssm as a solution to the hierarchy problem is strongly disfavored . if we motivate new physics models from the reported anomaly of muon g-2 , our naive expectation is the existence of ew scale new particles because the size of anomaly of muon g-2 is as large as one induced by ew gauge bosons in the sm . in this paper , we analyzed two examples , where muon has new yukawa type interactions with new particles . one example was a model where right - handed muon has new yukawa interaction with su(2 ) singlet scalar and fermion . in this case , in order to explain the anomaly of muon g-2 as well as to satisfy the precision ew measurements , we found that the relatively large new yukawa coupling and ew scale new particles are strongly favored . we also noted that the multi - charged particles ( whose qed charges are large ) are favored . another example we analyzed was a model where both right- and left - handed muons have new yukawa interactions with su(2 ) doublet scalar and singlet scalar and fermion , so that the chirality flip of muon can occur in the internal fermion line of 1-loop muon g-2 diagrams . in this case , new contributions to muon g-2 are enhanced , and hence even smaller new yukawa couplings can explain the anomaly of muon g-2 . we also showed that the ew observables are also well - fitted by the new physics contributions . in this case , the ew scale new particles are favored by the data . although these two examples may be only a part of complete models , the muon g-2 as well as the ew observables can constrain the qed charge of new particles as well as the scale of new particles and the size of new yukawa couplings . since the ew scale new particles can explain the anomaly of muon g-2 and they can be consistent with the precision ew measurements , it is important to study the possible phenomenology at the lhc . since some of new particles have qed charges as well as couplings with higgs boson , the branching ratio of @xmath298 can be effected . especially , as we showed , multi - charged particles favored by the muon g-2 can significantly contribute to @xmath298 , and the effects can be sizable . therefore , higgs physics will have an impact on models we discussed here . we also computed direct production cross sections of new particles motivated by the muon g-2 . since models discussed here may not be complete , we could not discuss the complete signatures for the models . however , since the number of signal events are not negligible , there may be possibility to reveal new physics models for muon g-2 at the lhc . since we do not need the colored particles to explain the anomaly of muon g-2 , the search for new particles which only have ew interactions will be very important . finally , we comment on the possible completions of models discussed in this paper . the structure of the model where both right- and left - handed muons have new yukawa type interactions is quite similar to that of the minimal supersymmetric standard model . when @xmath195 , the scalar @xmath68 in su(2 ) doublet scalar @xmath64 and su(2 ) singlet scalar @xmath9 correspond to left- and right - handed smuons , respectively . the su(2 ) singlet fermion @xmath10 is a bino , for example . a part of a model discussed in ref . @xcite , which is motivated by neutrino mass and dark matter is also similar to the model we discussed here . therefore , our approach can capture crucial features of new physics models for the anomaly of muon g-2 and hopefully create an interesting way to find the physics beyond the standard model . _ note added _ after we completed this work , we learned that atlas and cms collaboration at lhc observed a new particle consistent with the sm higgs boson at a mass of about 125 gev @xcite . since we have taken the higgs boson mass to be 125 gev in our analysis , our conclusion does not change . this work is supported in part by grant - in - aid for scientific research from the ministry of education , science , sports , and culture ( mext ) , japan , no . 22540273 ( kt ) . this work is also supported in part by grant - in - aid for nagoya university global coe program , `` quest for fundamental principles in the universe : from particles to the solar system and the cosmos '' , from the mext , japan . in this paper , passarino - veltman functions @xcite are defined by @xmath299 \left[(k+p)^2-m_b^2+i \epsilon\right]},\nonumber \\ p^\mu b_1(a , b;p ) & = & 16\pi^2 \mu^{2\epsilon } \int \frac{d^n k}{i(2\pi)^n}\frac{k^\mu } { \left[k^2-m_a^2+i\epsilon\right]\left[(k+p)^2-m_b^2+i\epsilon\right ] } , \nonumber \\ p^\mu p^\nu b_{21}(a , b;p)&+&g^{\mu \nu}b_{22}(a , b;p ) \nonumber \\ & = & 16\pi^2 \mu^{2\epsilon } \int \frac{d^n k}{i(2\pi)^n } \frac{k^\mu k^\nu } { \left[k^2-m_a^2+i\epsilon\right]\left[(k+p)^2-m_b^2+i\epsilon\right ] } , \ ] ] @xmath300[(k+p_1)^2-m_b^2+i\epsilon][(k+p_1+p_2)^2-m_c^2+i\epsilon ] } , \nonumber \\ & & \left ( p_1^\mu c_{11}+p_2^\mu c_{12 } \right)(a , b , c;p_1,p_2 ) \nonumber \\ & = & 16\pi^2 \mu^{2\epsilon } \int \frac{d^n k}{i(2\pi)^n}\frac{k^\mu } { [ k^2-m_a^2+i\epsilon][(k+p_1)^2-m_b^2+i\epsilon][(k+p_1+p_2)^2-m_c^2+i\epsilon ] } , \nonumber \\ & & \left\{(p_1^\mu p_1^\nu c_{21 } + p_2^\mu p_2^\nu c_{22}+(p_1^\mu p_2^\nu+p_1^\nu p_2^\mu)c_{23 } + g^{\mu \nu}c_{24 } \right\ } ( a , b , c;p_1,p_2 ) \nonumber \\ & = & 16\pi^2 \mu^{2\epsilon } \int \frac{d^n k}{i(2\pi)^n}\frac{k^\mu k^\nu } { [ k^2-m_a^2+i\epsilon][(k+p_1)^2-m_b^2+i\epsilon][(k+p_1+p_2)^2-m_c^2+i\epsilon ] } , \nonumber \\\end{aligned}\ ] ] where we use dimensional regularization in @xmath301 dimensions . g. aad _ et al . _ [ atlas collaboration ] , phys . b * 710 * , 49 ( 2012 ) [ arxiv:1202.1408 ] ; s. chatrchyan _ et al . _ [ cms collaboration ] , arxiv:1202.1488 . see also recent tevatron result ; [ tevnph ( tevatron new phenomina and higgs working group ) and cdf and d0 collaborations ] , arxiv:1203.3774 . k. nakamura _ et al . _ [ particle data group collaboration ] , j. phys . g g * 37 * , 075021 ( 2010 ) ; b. l. roberts , chin . c * 34 * , 741 ( 2010 ) [ arxiv:1001.2898 ] . k. hagiwara _ et al . _ , j. phys . g g * 38 * , 085003 ( 2011 ) [ arxiv:1105.3149 ] . see also f. jegerlehner and a. nyffeler , phys . rept . * 477 * , 1 ( 2009 ) [ arxiv:0902.3360 ] ; m. davier _ et al . _ , eur . j. c * 71 * , 1515 ( 2011 ) [ erratum - ibid . c * 72 * , 1874 ( 2012 ) ] [ arxiv:1010.4180 ] ; f. jegerlehner and r. szafron , eur . j. c * 71 * , 1632 ( 2011 ) [ arxiv:1101.2872 ] ; t. teubner _ et al . _ , aip conf . proc . * 1343 * , 340 ( 2011 ) ; nucl . phys . proc . * 218 * , 225 ( 2011 ) . t. aoyama _ et al . _ , arxiv:1205.5370 ; arxiv:1205.5368 ; r. boughezal and k. melnikov , phys . b * 704 * , 193 ( 2011 ) [ arxiv:1104.4510 ] ; a. e. dorokhov , a. e. radzhabov and a. s. zhevlakov , eur . j. c * 71 * , 1702 ( 2011 ) [ arxiv:1103.2042 ] ; t. goecke , c. s. fischer and r. williams , phys . d * 83 * , 094006 ( 2011 ) [ arxiv:1012.3886 ] ; eur . j. a * 47 * , 28 ( 2011 ) [ arxiv:1009.5297 ] ; a. nyffeler , chin . phys . c * 34 * , 705 ( 2010 ) [ arxiv:1001.3970 ] ; a. nyffeler , phys . d * 79 * , 073012 ( 2009 ) [ arxiv:0901.1172 ] ; d. k. hong and d. kim , phys . b * 680 * , 480 ( 2009 ) [ arxiv:0904.4042 ] ; j. prades , e. de rafael and a. vainshtein , [ arxiv:0901.0306 ] ; a. e. dorokhov and w. broniowski , phys . d * 78 * , 073011 ( 2008 ) [ arxiv:0805.0760 ] ; j. bijnens and j. prades , mod . a * 22 * , 767 ( 2007 ) [ hep - ph/0702170 ] ; m. hayakawa _ et al . _ , pos lat * 2005 * , 353 ( 2006 ) [ hep - lat/0509016 ] ; k. melnikov and a. vainshtein , phys . rev . d * 70 * , 113006 ( 2004 ) [ hep - ph/0312226 ] ; j. h. kuhn , a. i. onishchenko , a. a. pivovarov and o. l. veretin , phys . d * 68 * , 033018 ( 2003 ) [ hep - ph/0301151 ] ; t. kinoshita and m. nio , phys . rev . lett . * 90 * , 021803 ( 2003 ) [ hep - ph/0210322 ] ; m. j. ramsey - musolf and m. b. wise , phys . lett . * 89 * , 041601 ( 2002 ) [ hep - ph/0201297 ] ; j. bijnens , e. pallante and j. prades , nucl . b * 626 * , 410 ( 2002 ) [ hep - ph/0112255 ] ; nucl . b * 474 * , 379 ( 1996 ) [ hep - ph/9511388 ] ; j. bijnens , e. pallante and j. prades , phys . * 75 * , 1447 ( 1995 ) [ erratum - ibid . * 75 * , 3781 ( 1995 ) ] [ hep - ph/9505251 ] ; i. r. blokland , a. czarnecki and k. melnikov , phys . lett . * 88 * , 071803 ( 2002 ) [ hep - ph/0112117 ] ; m. knecht , a. nyffeler , m. perrottet and e. de rafael , phys . rev . lett . * 88 * , 071802 ( 2002 ) [ hep - ph/0111059 ] ; m. knecht and a. nyffeler , phys . d * 65 * , 073034 ( 2002 ) [ hep - ph/0111058 ] ; z. bern , a. de freitas , l. j. dixon , a. ghinculov and h. l. wong , jhep * 0111 * , 031 ( 2001 ) [ hep - ph/0109079 ] ; m. hayakawa and t. kinoshita , phys . d * 57 * , 465 ( 1998 ) [ erratum - ibid . d * 66 * , 019902 ( 2002 ) ] [ hep - ph/9708227 ] ; m. hayakawa , t. kinoshita and a. i. sanda , phys . d * 54 * , 3137 ( 1996 ) [ hep - ph/9601310 ] ; m. hayakawa , t. kinoshita and a. i. sanda , phys . lett . * 75 * , 790 ( 1995 ) [ hep - ph/9503463 ] . t. moroi , phys . rev . d * 53 * , 6565 ( 1996 ) [ erratum - ibid . d * 56 * , 4424 ( 1997 ) ] [ hep - ph/9512396 ] . g. c. mclaughlin and j. n. ng , phys . b * 493 * , 88 ( 2000 ) [ hep - ph/0008209 ] . s. r. choudhury , a. s. cornell , a. deandrea , n. gaur and a. goyal , phys . d * 75 * , 055011 ( 2007 ) [ hep - ph/0612327 ] . k. kannike , m. raidal , d. m. straub and a. strumia , jhep * 1202 * , 106 ( 2012 ) [ arxiv:1111.2551 ] ; h. davoudiasl , h. -s . lee and w. j. marciano , arxiv:1205.2709 . g. t hooft and m. j. g. veltman , nucl . b * 153 * , 365 ( 1979 ) ; g. passarino and m. j. g. veltman , nucl . b * 160 * , 151 ( 1979 ) . g. -c . cho _ et al . _ , jhep * 1111 * , 068 ( 2011 ) [ arxiv:1104.1769 ] . g. -c . cho and k. hagiwara , nucl . b * 574 * , 623 ( 2000 ) [ hep - ph/9912260 ] . k. hagiwara _ et al . _ , z. phys . c * 64 * , 559 ( 1994 ) [ erratum - ibid . c * 68 * , 352 ( 1995 ) ] [ hep - ph/9409380 ] . m. e. peskin and t. takeuchi , phys . lett . * 65 * , 964 ( 1990 ) ; phys . d * 46 * , 381 ( 1992 ) . g. altarelli and r. barbieri , phys . b * 253 * , 161 ( 1991 ) ; g. altarelli , r. barbieri and s. jadach , nucl . b * 369 * , 3 ( 1992 ) [ erratum - ibid . b * 376 * , 444 ( 1992 ) ] ; g. altarelli , r. barbieri and f. caravaglios , phys . b * 349 * , 145 ( 1995 ) . i. maksymyk , c. p. burgess and d. london , phys . d * 50 * , 529 ( 1994 ) [ hep - ph/9306267 ] . tevatron electroweak working group , f. t. c. collaboration and d. collaboration , arxiv:1204.0042 . tevatron electroweak working group and for the cdf and d0 collaborations , arxiv:1107.5255 . see web page : http://lepewwg.web.cern.ch/lepewwg/ j. alwall _ et al . _ , jhep * 0709 * , 028 ( 2007 ) [ arxiv:0706.2334 ] . for example , see g. aad _ et al . _ [ atlas collaboration ] , arxiv:1206.1760 ; s. chatrchyan _ et al . _ [ cms collaboration ] , arxiv:1206.3949 . t. hambye , k. kannike , e. ma and m. raidal , phys . d * 75 * , 095003 ( 2007 ) [ hep - ph/0609228 ] . g. aad _ et al . _ [ atlas collaboration ] , [ arxiv:1207.7214 ] ; s. chatrchyan _ et al . _ [ cms collaboration ] , phys . b [ arxiv:1207.7235 ] .

tracking a lander on the surface of another planet with radio signals is an efficient way to observe its rotation ( dehant et al . , 2009 , 2011 ) . this can be done by measuring the doppler shift of the radio signal between the lander and a large antenna on earth , like the ones of the deep space network ( dsn ) . we consider here two - way x - band signal : the ground station transmits the radio signal to mars , which is coherently ( i.e. without any phase shift ) sent back to earth by the transponder onboard the lander . five martian landers have already been tracked from earth in order to get the martian rotation . the first landers were the two viking landers in 1976 - 1982 ( see for example borderies et al . , 1980 , yoder and standish , 1997 ) . 21 years later , the pathfinder lander stayed 3 months on mars surface in 1997 ( folkner et al . , 1997a and 1997b ) . more recently the two mars exploration rovers were tracked from earth when they were stationary ( le maistre , 2013 , kuchynka et al . , 2014 ) : spirit was stuck at the end of its life in 2009 , and opportunity stopped moving during the winter for energy saving during four months in 2012 . two martian missions including a geodesy experiment will be launched in the coming years . the first one is the nasa mission insight ( interior exploration using seismic investigations , geodesy and heat transport ) to be launched in 2018 . the insight spacecraft will have a small deep space transponder on board : the rise ( rotation and interior structure experiment , folkner et al . , 2012 ) transponder . the second mission is the esa mission exomars2020 that will include the lara ( lander radioscience , dehant et al . , 2009 ) transponder on the surface platform . this spacecraft is to be launched in 2020 . + the martian orientation parameters ( mop ) are the nutations , the precession , the length - of - day variations and the polar motion . the precession is the long - term drift of the rotation axis in space . nutations are seasonal periodic motions of this axis observed from space , while the polar motion is the seasonal motion of the rotation axis in a frame tied to the planet . the rotation rate variations , also called length - of - day variations , are the periodic variations of the martian diurnal rotation . the periods are also the harmonics of the annual period . since these rotational motions depend on the interior structure of mars , in particular the core dimension , density and state , and on the dynamics of the atmosphere and the ice caps and the co@xmath0 sublimation / condensation process , accurate measurements of these angles are very useful because additional constraints on the geophysical models could be provided . in this paper we derive expressions of the signature of the rotation parameters in the doppler and range observables . these expressions are useful in order to optimize the configuration parameters and to anticipate the observation times that maximize the signatures and the geophysical information . this is of course not only helpful for the insight and exomars mission planning but also for other geodesy experiments on other planets . + the paper is organized as follows : in section 2 , we present the different angles and variables that describe the geometry of the doppler observable . section 3 gives a summary of the models for the different mop . in section 4 , we give the analytical expressions of the signature of these mop in the doppler observable and their numerical values , while the section 5 focuses on the signatures in the range observable . the instantaneous doppler frequency can be interpreted as the space probe radial velocity with respect to the tracking station . as a first order approximation , it is the projection on the line - of - sight of the difference of the velocity vector between the emitter and the receiver . it is very sensitive to the configuration geometry , therefore we present in this section the different variables that are important for the signatures in the doppler observable . with respect to mars as a function of time . the sun - earth - probe ( sep ) angle is plotted on the upper graph in red with some arbitrary scaling , the black parts of the curve ( and the corresponding pink boxes ) being the points where the sep angle is smaller than @xmath1 , corresponding to radio data with a large noise due to the solar plasma . the gray boxes correspond to time intervals where the earth declination is smaller than 10@xmath2 . the insight and exomars nominal mission lifetime are also shown with the orange and red horizontal lines . , width=680 ] the exomars and insight missions have similar orbital configuration at launch and when arriving on mars . for both the distance between mars and the earth increases after the martian landing , being maximal a few months after the landing ( see fig . [ fig_dist ] , upper graph ) . the sun - earth - probe ( sep ) angle or elongation is small when the earth - mars distance is maximal , this is the conjunction time . the pink boxes on fig . [ fig_dist ] show where the sep angle is smaller than @xmath1 . at that time , the solar plasma perturbs the radio signal largely , and the noise on the radioscience measurements is much larger than the nominal noise . + the earth declination @xmath3 with respect to mars is the angle between the mars - earth direction and the equatorial plane of mars . because of the martian obliquity and the earth revolution in the ecliptic plane , the earth declination has variations due to the orbital motions of the two planets and due to the mars pole direction . it varies between -25@xmath2 and 25@xmath2 ( see fig . [ fig_dist ] , lower graph ) . when the earth - mars distance is minimal , corresponding to the opposition , the earth - mars direction reverses its rotation : for a few weeks , this direction goes from a prograde to a retrograde motion . this is due to the faster revolution velocity of the earth than for mars . at this time , in the earth s sky , mars is having an apparent retrograde motion with respect to the usual planetary rotation . at the same time , the earth declination shows some irregularities because of this retrograde motion . the maxima in the earth declination do not correspond to the distance minima because the direction of the martian spin axis has also to be taken into account . the maxima have a periodicity longer than the earth - mars synodic period ( 780 days ) . on the time scale of a spacecraft mission ( about 2 years ) , the mop can be safely neglected for the computation of the earth declination , the largest mop effect being the martian precession that changes the earth declination by less of @xmath4 . the gray boxes on fig . [ fig_dist ] for the earth declination correspond to time intervals where the earth declination is smaller than 10@xmath2 . in section 4 we will show that these intervals correspond to a lower sensitivity to mars spin axis motion in space . another useful equatorial coordinates that will appear in the mop signatures ( section 4 ) is the earth hour angle @xmath5 , the angle along the celestial equator from the mars local meridian to the hour circle passing through earth , in the retrograde direction . if @xmath6 is the earth right ascension , @xmath7 the sidereal mars rotation angle ( the angle from the vernal point to the prime meridian ) and @xmath8 the lander longitude ( positive east from the prime meridian ) , the earth hour angle is related to the other angles in the equatorial plane by the following equation @xmath9 the earth hour angle has a diurnal variation ( period of 24h 37min for mars ) , increasing every day non - linearly from 0 to 360@xmath2 , and depends on the lander position . every day the earth rises and crosses the lander sky . like the earth declination @xmath3 , the earth elevation @xmath10 has variations related to the orbital motion of the earth and mars . if @xmath11 is the lander latitude , the maximal elevation @xmath12 of the earth in the lander sky varies for each day between @xmath13 and the angle corresponding to the minimum between @xmath14 and @xmath15 . the rise instrument have two antennas designed to observe the earth within a @xmath1 to @xmath16 elevation range , with one antenna pointed nearly to the east and the other pointed nearly to the west . lara antennas nominal range shall be @xmath17 elevation , for all azimuthal directions ( @xmath18 ) , with a maximum gain direction at @xmath19 of elevation . the rotation matrix from the mars body - fixed reference frame to the inertial frame based on mars mean orbit of j2000 is ( folkner et al . , 1997a , konopliv et al . , 2006 , le maistre et al . , 2012 ) @xmath20 if @xmath21 is the lander position vector in the body - fixed reference frame , the position of the lander in the inertial frame @xmath22 is : @xmath23 @xmath24 , @xmath25 and @xmath26 are the classical elementary rotation matrices around the x , y , z axes respectively . as for the earth , five angles called here the martian orientation parameters ( mop ) are used to characterize the matrix @xmath27 : @xmath28 , @xmath29 , @xmath7 , @xmath30 and @xmath31 . @xmath30 and @xmath31 are the crust - fixed coordinates of mars spin axis related to the polar motion . @xmath7 is the martian sidereal rotation angle . its variations @xmath32 are called the length - of - day ( lod ) variations . @xmath33 and @xmath34 are the precession - nutation matrices depending on the obliquity @xmath28 between the instantaneous mars equator and the mean orbital plane at j2000 and on the longitude of the node of the mars equator @xmath29 . nutation in obliquity and longitude @xmath35 and @xmath36 are the periodic variations of @xmath28 and @xmath29 while the precession is @xmath37 the secular trend of @xmath29 . we can write @xmath38 @xmath39 is the diurnal rotation rate . the mop are small angles . + nutation is the periodic motion of the spin pole in space . the nutations for a rigid planet can be computed from celestial mechanics and ephemerides . a rigid nutation model is usually constructed as a sum of different frequencies with an amplitude and a phase for the nutation in obliquity ( @xmath35 ) and in longitude ( @xmath36 ) . in the following sections , we use the term `` nutation in longitude '' @xmath36 to mean the sum of the components at different frequencies of the nutation in longitude , and similarly the term `` nutation in obliquity '' @xmath35 as the sum of the main frequencies . for mars , the largest nutations are the annual and semiannual nutations . their maximal amplitude are about 550 and 1130 mas , corresponding to 9 and 19 m on mars surface . we use roosbeek ( 1999 ) rigid mars nutations model . however mars is not a rigid planet . therefore the nutation amplitudes depend on the interior structure and parameters , like the state ( liquid or solid ) of the core or its moment of inertia . we can model the non - rigid nutation amplitudes by applying a transfer function to the rigid nutations ( dehant et al . , 2000 , van hoolst et al . , 2000a , 2000b ) . this transfer function is based on the model developed for the earth by sasao et al . ( 1980 ) and includes the liquid core effects and other deformation effects . for example , if the core is liquid , then an additional free mode , the free core nutation ( fcn ) amplifies the nutation response if the fcn period is close to one of the nutation periods . current interior models estimate the fcn period to be between about -240 and -285 days . therefore the fcn can mostly enhance the small retrograde terannual ( 229 days ) or the large prograde semiannual nutation ( 343.5 days ) . considering the 3 sigma interval rather than the 1 sigma interval will make the fcn range larger than @xmath40 days and the fcn frequency can be even closer to the terannual frequency , increasing the amplification . in this study , we use a transfer function and amplifications computed from up - to - date interior models ( panning et al . , 2016 ) . the mantle mineralogy and the temperature profile are unique for all the models and the planet is assumed to be in hydrostatic equilibrium . the variable quantities in the interior models are the size of the core and the density and thickness of the crust . the models agree with the observed moment of inertia at one sigma . the frequencies of the nutations are not affected by the transfer function , therefore the modeling we use is a trigonometric expansion with known periods but unknown amplitudes . the amplifications are different for each nutation period and for the nutation in obliquity and in longitude ( see for example figure 10 of le maistre et al . , 2012 ) . depending on the size of the core , the maximal change of the nutations due to the liquid core is between @xmath41 mas and @xmath42 mas ( corresponding to about @xmath43 cm and @xmath44 cm on mars surface ) for the nutation in obliquity and between @xmath42 mas and @xmath44 mas or @xmath45 cm and @xmath46 cm for the nutation in longitude . the lower values correspond to a small core of @xmath47 km while the upper values correspond to a large core of about @xmath48 km . an important objective of the geodesy experiments is to measure the nutations and particularly the transfer function because it is good tool to constraint the interior structure of mars and the core dimension and density in particular . + the precession is the secular motion of the spin axis about the normal to the orbit plane . the spin axis completes one rotation in a period of about 170 300 years . the precession is due to the gravitational pull of the sun on the oblate figure and depends on the moment of inertia about the spin axis . the present uncertainty on the precession rate ( konopliv et al . , 2016 ) is 2.1 mas / y , corresponding to a period change of about 47 years . since the precession rate @xmath37 is proportional to @xmath49 where @xmath50 is mars moment of inertia , the estimation of the precession rate gives a constraint on the interior mass distribution . + the rotation rate variations or length - of - day ( lod ) variations @xmath32 are modeled by a trigonometric series with the annual , semiannual , terannual and quaterannual periodicities . the annual nutation has an amplitude of about @xmath51 mas or @xmath52 m on mars surface while the semiannual nutation is about @xmath53 mas or @xmath54 m ( konopliv et al . , 2016 ) . in the rotation matrix around the z - axis ( @xmath7 angle , see eq . [ eq_phi ] ) , above the lod variations , there is an additional term depending on the nutation @xmath55 , adding a nutation signature in the doppler observable . the lod variations are of interest for global climate models , to constrain the atmosphere and ice caps dynamics . + each polar motion ( pm ) coordinate @xmath30 and @xmath31 is modeled by a trigonometric series ( annual , semiannual , terannual ) and a chandler wobble term . the amplitudes of the seasonal terms of the polar motion are expected to be between 0 and 15 mas ( defraigne et al . , 2000 , van den acker et al . , 2002 ) . the period of the chandler wobble for mars is expected to be around 200 or 220 days . its amplitude has not yet been measured but is expected to be somewhere between 10 and 100 mas . the polar motion has both climatological and geodetic information ; particularly the chandler wobble period depends mainly on the dynamical flattening of the planet and it provides information on the planet s elasticity and on inelastic behavior . & max value & analytical expression & max value + rigid nutations in long . @xmath36 & 29.4 m & @xmath56 & 12.3 m + rigid nutations in obliq . @xmath35 & 10.3 m & @xmath57 & 10.3 m + & 0.4/0.7 m & & 0.4 m + & 0.2/0.4 m & & 0.2 m + precession : & 0.8 m & @xmath58 & 0.8 m + ( if @xmath59 2.1 mas / y ) & & @xmath60^{1/2}$ ] & + lod @xmath61 & 11.8 m & @xmath62 & 11.8 m + polar motion @xmath63 ( @xmath64 ) & 1.2 m & @xmath65 & 1.2 m + ( if chandler amp . @xmath66 m ) & & @xmath67 & + the maximal value of each mop on a long time interval is given in the second column of table [ tab1 ] , assuming the nominal model . the amplitude of the chandler wobble is assumed to be around @xmath68 meters . for the liquid core effect in the nutations , the first numbers ( 0.4 m for a large core and 0.2 m for a small core ) are the components of the nutation in obliquity , while the second ( 0.7 and 0.4 m ) are the components of the nutation in longitude . the precession rate is very large , we consider here only the effect of the present uncertainty on the precession rate at j2000 ( about 2.1 mas / y ) . the displacement due to this uncertainty after about 20 years after j2000 is about 0.8 m. each mop changes of the lander position . if the position change vector is the difference between the lander position with and without the mop and using the rotation matrix @xmath27 ( see equation [ eq_r ] ) , we compute the norm of this position change vector . the analytical expressions and the maximal value of the norm of this position change vector is given in the last columns of table [ tab1 ] . since we use first order approximation , all the displacements are proportional to the mop . some of the lander displacements are modulated by the diurnal rotation . we see that the lander displacement due to the lod variations is larger when the lander is closer to the equator . the maximal value of the displacement due to the rigid nutation is independent of the lander position ( because the square roots in table 1 are always equal or smaller than @xmath69 ) , but close to the pole , the displacement due to the nutations does not have any diurnal variations . there is a reduction of a factor 2.3 from the maximal value of the nutation in longitude @xmath36 to its change in the lander motion ( from @xmath70 m to @xmath71 m ) because of the obliquity of mars ( @xmath72 ) induces a projection in the lander displacement and because of the rotation angle component ( - @xmath73 ) that removes a part of the signal ( see equation [ eq_phi ] ) . for the other mop , the maximal displacement due to the mop is the same as the mop maximal value . depending on the size of the core , the maximal change of the lander position due to the liquid core is between 19 cm ( small core ) and 42 cm ( large core ) . the analytical expression is a combination of the transfer function and the change of lander position due to the rigid nutations in obliquity and in longitude , its expression is lengthly and therefore has not been included in table 1 . the nutations in longitude and in obliquity together changes the lander by about 12.6 m. in this section we give the analytical expressions of the mop signature in the doppler lander - earth observable (= direct - to - earth or dte observable ) . we consider an instantaneous doppler observable ( with an infinite speed of light and no integration time ) . the doppler measurements are highly sensitive to the position variations along the line - of - sight and not sensitive to position variations in the plane perpendicular to the line - of - sight . + the doppler measurements between the earth and a spacecraft on mars varies between about -17 and 17 km / s . this corresponds to the value of the relative velocity of the two planets around the sun . its main modulations come from the relative orbital motions of the earth and mars . other modulations come from the variation of the orientation and the rotation of mars and the earth . the diurnal rotations of mars and of earth impress a sinusoidal modulation upon the doppler signal , with a beat frequency of about 36 days . this period of 36 days corresponds to the mars - earth synodic period of rotation given that the rotation period of earth and mars are very close , 23h56min versus 24h37min ( le maistre 2013 ) . the doppler observable is positive when the lander is going away from the earth and is negative when getting closer . + the doppler observable @xmath74 can be modeled as the projection of the velocity difference @xmath75 on the line - of - sight @xmath76 , with @xmath77 being the angle between @xmath75 and the earth - lander line - of - sight . the signature of a parameter in the doppler observable is the observable estimated taking into account a parameter minus the observable without this parameter . to the first order in the mop , it can be expressed as ( yseboodt et al . , 2003 ) : @xmath78 where @xmath79 is the change in @xmath80 due to the mop and @xmath81 is the change in the line - of - sight direction due to the mop . the first term in eq . ( [ eq_signatu ] ) is the `` geometrical effect '' or change of the line - of - sight direction . this term expresses the change in direction of the emitter - receiver line induced by the rotation parameter multiplied by the amplitude of the velocity . the second term of eq . ( [ eq_signatu ] ) is the contribution of the velocity change induced by the rotation . both terms are proportional to mars radius and to the mop . the line - of - sight displacement because of the mop @xmath81 decreases if the emitter - receiver distance increases , therefore the geometrical effect is proportional to the relative velocity over the distance @xmath82 . the velocity effect , i.e. the change of the lander velocity @xmath79 , is independent of the emitter - receiver distance but is proportional to the diurnal rotation rate @xmath39 . if the relative velocity over the distance is smaller than the planet diurnal rotation rate , then the velocity effect dominates the signature in the doppler observable . this happens when the earth - planet distance is large . since this is the case for mars and because mars diurnal rotation rate @xmath39 is large , therefore this velocity effect is the one that is modeled in this section while the line - of - sight displacement is neglected . if the emitter - receiver distance is small , like for example if the receiver is on board an orbiter , then the dominant effect in the doppler signal is the change in the line - of - sight . + the signatures of the mop parameters in the observable have already been investigated in le maistre et al . ( 2012 ) for the doppler observable , konopliv et al . ( 2006 ) for the lod signature in the doppler observable and folkner et al . ( 1997a , eq . 1 ) for the range observable . previously , curkendall and mcreynolds ( 1969 ) gave first order analytical expressions for an orbiter - earth doppler and range observables and its link with the equatorial coordinates of the spacecraft while yseboodt et al . ( 2003 ) studied the mop signatures in the orbiter - earth doppler link in the frame of a geophysical mission including a network of landers on mars surface . kawano et al . ( 1999 ) studied the mop signatures in the same beam interferometry ( sbi ) observable for two landers ( or more ) on mars surface . + , the nutations in longitude @xmath36 , the liquid core effect in the nutations ( large core , free core nutation period of 242 days ) , a precession rate @xmath37 difference of 2 mas / y , the lod variations and the polar motion . on the left part of the graph , the insight lander is in western elysium planitia ( @xmath83n , @xmath84e ) and on the right part , the exomars lander is in oxia planum ( @xmath85n , @xmath86e ) . in the nutation plots , the dotted gray line shows the earth declination @xmath3 . the gray boxes correspond to time intervals where the earth declination is smaller than @xmath87 while the pink boxes show where the sep angle is smaller than @xmath1 . the red and purple curves are the surface displacements due to the nutation in longitude and nutation in obliquity , respectively . they are scaled . , width=755,height=604 ] the signatures of the different mop in the doppler observable are plotted on fig . [ fig_sensi ] . since we use first order approximation , the mop signatures are proportional to the mop . therefore the mop periodicities ( annual , semiannual , chandler period ... ) can also be seen in the signatures plot . the expected numerical values are summarized in section [ secamp ] . if @xmath35 is the nutation in obliquity and @xmath36 the nutation in longitude , their signature in the lander - earth doppler observable is @xmath88 where @xmath89 is mars radius , @xmath39 the diurnal rotation rate of mars , @xmath7 the diurnal rotation angle of mars , @xmath3 the earth declination with respect to mars , @xmath28 mars obliquity , @xmath5 the earth hour angle and @xmath6 the right ascension . the mathematical derivation of these expressions is given in appendix [ ap1 ] . + since the signature is proportional to the nutation , the annual , semiannual and terannual modulations in the @xmath35 and @xmath36 nutations also appear in the signature plots ( see also the red and purple lines in fig . [ fig_sensi ] ) . the time derivatives of the nutations @xmath90 and @xmath91 give second order effects in the signature since their period ( mostly annual and semiannual ) is much larger than the diurnal rotation rate @xmath39 . the nutation signature is proportional to the distance between the lander and the spin axis @xmath92 , therefore the nutation signature is maximal at the equator and null at the pole . the nutation signature in the doppler observable must be a function of the direction of its spin axis with respect to the earth direction , eqs . ( [ eq_signeps ] ) and ( [ eq_signpsi ] ) show that it is proportional to the sinus of the earth declination . a tilt of the rotation axis because of the nutation tilts the diurnal velocity vector and the resulting velocity difference is , to the first order , along the z ( polar ) axis in a body - fixed reference frame . the projection of this velocity vector on the line - of - sight is null if the earth in the equatorial plane of mars ( @xmath93 ) . therefore the doppler observable sensitivity to nutations is low when mars spin axis is nearly perpendicular to the mars - earth direction ( see the gray line and gray boxes in fig . [ fig_sensi ] ) . replacing @xmath94 by @xmath3 leads to a maximal error of 3 % because the earth declination is always smaller than mars obliquity ( @xmath95 ) . the factor @xmath94 in the nutation signature varies between @xmath96 and @xmath97 , therefore decreasing the signatures . in eqs . ( [ eq_signeps ] ) and ( [ eq_signpsi ] ) , there is a diurnal periodicity in the signature through mars rotation angle @xmath7 or through the local earth hour angle @xmath5 . the link between the angles in the equatorial plane is given in eq . [ eq_he ] . this diurnal variation is shown on the plot by the blue regions between the maximal and minimal curve . during one day , the signature of @xmath36 and @xmath35 can not be maximal at the same time , there is a time shift of 6 hours between the maxima of the two nutations . the earth elevation in the lander s sky does not appear in the nutation signature . however there is an indirect link between both quantities , because the earth elevation depends on the chosen timing of observation . for example if observations are chosen every early morning , the earth elevation will be always be small , and from day to day the value of the earth hour angle will almost be the same . the liquid core signature in the doppler observable is the difference between the doppler observable with and without the nutation transfer function , each part being the sum of the @xmath35 and @xmath36 contributions ( eqs . [ eq_signeps ] and [ eq_signpsi ] ) . the liquid core signature is much smaller than the rigid nutations signature because the liquid core amplification is also smaller . the signature of the non - rigid part of the nutations has more modulations than the other mop signatures because sometimes there is an amplification of the nutation in longitude due to the liquid core while later , it is the nutation in obliquity that is amplified . therefore , the liquid core signature is a combination of both signatures . if plotted on the same figure , the rigid and non - rigid nutation signatures will be very close and indistinguishable on fig . the dependence on the earth declination ( see the gray line and gray boxes ) and the lander latitude is the same as the rigid nutation signature . over an interval of a few years , the position of the maxima and of the roots moves if the free core nutation ( fcn ) period changes ( see fig . [ fig_liquid ] ) . ( scaled ) . , width=755,height=340 ] this is because if the fcn period is close to the terannual period ( @xmath98 days ) , the terannual nutation will be largely amplified while if the fcn period is close to the semiannual period ( @xmath99 days ) , the terannual nutation amplification will be small but the semiannual nutation will be larger . + again , the diurnal variation is shown on the plot by the blue area . therefore depending on the lander position , each day there is a different observation time that maximizes the liquid core signature . 12 hours after the time when the liquid core signature is maximum , the liquid core signature is minimum . 6 hours later , the liquid core signature is null . having a large negative signature gives a geophysical information as important as a large positive signature . the time where the liquid core signature is maximal during the day depends on the fcn value . the time difference is less than an hour for the fcn range investigated here . using a method similar to appendix [ ap1 ] , we find the expression of the precession rate @xmath100 signature in the doppler observable . if the product @xmath101 is assumed to be small , the precession signature is @xmath102 the signature increases with time but remains relatively small on an interval like the mission duration , which is short with respect to the precession period of about @xmath103 years . the precession signature has a diurnal variation via the earth hour angle @xmath5 . + the first term in eq . [ eq_signpsip ] is the dominant one close to the equator because @xmath104 . it depends on @xmath105 while the second term depends on @xmath106 . the maximal signature is almost constant with time ( see figure [ fig_sensi ] ) , there is a long period variation with a small amplitude due to @xmath107 in the first term of eq . ( [ eq_signpsip ] ) and due to @xmath108 in the second term . the second term ( but not the first one ) appears in the signature of the nutation in longitude @xmath36 . there is a large difference for the analytical expressions of @xmath36 and @xmath109 because the nutation in longitude appears twice in the rotation matrix @xmath27 ( in @xmath110 and in mars rotation angle rotation @xmath111 ) while the precession appears once in @xmath110 . the signature of martian rotation angle variations @xmath32 ( or lod variations ) in the doppler observable is @xmath112 there is a proportionality between the lod variations and its signature , therefore the main frequencies are annual , semiannual and quaterannual . similarly as for the nutation , the lod signature is proportional to the distance from the lander to the spin axis @xmath92 , therefore the signature is maximal at the equator and null at the pole . the lander longitude disappears from the signature . however there is an indirect effect because the earth hour angle @xmath5 depends on the longitude . the signature has again a diurnal periodicity via the earth hour angle . the lod signature is very little affected by the variations of the earth declination because the @xmath113 stays between @xmath114 and @xmath69 . + the diurnal velocity vector is in the equatorial plane of the body - fixed reference frame and is perpendicular to the lander position vector @xmath115 . the change of the velocity vector due to the lod variations is also in the equatorial plane and to the first order perpendicular to the diurnal velocity , it is in the direction of the lander position vector @xmath116 . the scalar product between the velocity difference and the line - of - sight ( los ) is maximized when the two vectors are in the same direction , therefore when the los is `` aligned '' with the mars - lander vector . each day , the lod signature is maximized when the mars - earth direction crosses the local meridian of the lander ( when the local earth hour angle is null ) or equivalently when the earth culminates in the lander sky . typically this happens near the local noon . when the lander is on the equator ( @xmath117 ) , the maximal earth elevation is @xmath118 . therefore the maximal lod signature in the doppler observable is proportional to @xmath119 and on a long time interval , the signature increases as the maximal earth elevation increases or equivalently as the earth declination decreases . when the lander is not on the equator , the maximal lod signature in the doppler observable is again when the local earth hour angle is null , corresponding to the maximal earth elevation during that day . however the lod signature is not anymore proportional to @xmath119 . @xmath120 the signature in the doppler observable of the equatorial plane coordinates of the lander @xmath121 and @xmath122 in the body - fixed reference frame is @xmath123 there is always a diurnal periodicity . on a daily basis , the @xmath121 signature is maximized when @xmath124 or @xmath125 while the @xmath122 signature is maximized when @xmath126 or @xmath127 . if we express this signature as a function of the lander latitude @xmath11 , the lander longitude disappears from the equation @xmath128 it is known that the doppler measurements are not very sensitive to the @xmath129-coordinate of the lander ( along the polar axis ) , and therefore solving for this coordinate is difficult ( le maistre , 2016 ) . a change in the @xmath129-coordinate only does not change the lander diurnal velocity . therefore its signature is very small because this is a second order contribution : the direction of the line - of - sight is changed a little bit because of the lander displacement in the polar direction . there is also a small second order velocity change . however if we want to know the third component of the lander position , we can impose that the lander has to stay on mars surface and there is a relationship between the three coordinates . this additional constraint related to the topography model helps to solve for the @xmath129-coordinate ( le maistre , 2016 ) . using the same equations , we can also evaluate the signature of a tidal displacement . a lander displacement along the x and y axis due to the tides has a signature in the doppler observable given by eq . [ eq_xysign ] . the signature of the polar motion @xmath30 and @xmath31 in the doppler observable is @xmath130 to the first order in the polar motion parameters , the first two coordinates of the position of the spin pole are @xmath131 , explaining the sign difference when comparing this equation to eq . [ eq_xysign ] . the diurnal periodicity in the signal comes from the earth hour angle @xmath5 . as for the lod variations , the signature is not very sensitive to the earth declination . during one day , the maxima and minima of the @xmath30 and @xmath31 signatures happen at different time , separated by 6 hours . the polar motion signature is a function of @xmath132 , therefore the signature increases with the lander latitude , being maximal at the poles . the reason is that the shift of the lander position due to the polar motion is @xmath133 . since the diurnal velocity vector is in the equatorial plane , because of the scalar product of the doppler observable , the part of the lander displacement that is important is also in the equatorial plane . this lander displacement in the equatorial plane is proportional to @xmath134 . equivalently to the sensitivity to the lander @xmath129-coordinate , the doppler observable is not sensitive to the @xmath129 displacement due to the polar motion . the maximal value of the mop signature in the doppler observable are given using the insight and exomars nominal missions configuration in table [ tablsi ] . the nominal mission time coverage is taken into account : 2 earth years for the insight lander starting end of 2019 , 1 earth year for the exomars lander starting beginning of 2021 . the insight landing site is expected to be close to the equator in western elysium planitia ( @xmath83n , @xmath84e ) while we pick the oxia planum location ( @xmath85n , @xmath86e ) for the exomars lander . the typical noise on doppler data is @xmath135 mm / s for a doppler counting interval of count interval of 60 seconds . @xmath136 the temporal evolution of the mop signature in the doppler observable is plotted on fig . [ fig_sensi ] . changing the landers position from one landing site to another position further from the equator does not affect a lot the value of the signatures ( this is barely noticeable on fig . 2 , see the left and the right parts ) , except for the polar motion signature . from tab . [ tablsi ] , we see that the maximal signatures are smaller for exomars than for insight , except for the polar motion signature , mostly because the exomars nominal mission does not cover one full martian cycle and because the chosen exomars location is farther from the equator than the insight landing site . the signature of the precession rate are very close because of a small decrease due to a higher latitude compensated by a small increase because the beginning of the exomars observation time is later with respect to j2000 . [ fig_sensilat ] shows that the maximal value of these signatures depends on the lander latitude . some rotation parameters like the nutation , the precession and the lod variations vary with @xmath137 with a larger signature at the equator while the polar motion varies with @xmath132 with a larger signature at the pole . and @xmath36 , the liquid core effect through nutations , the lod variations and the polar motion ( pm ) in the doppler observable as a function of the lander latitude . two different liquid core sizes have been considered . the two vertical lines show the landers latitude while the horizontal line shows the typical noise on doppler measurements ( @xmath135 mm / s).,width=502,height=302 ] + the maximal signature of the nutations in the doppler signal is up to @xmath138 mm / s , coming from the signature of the nutations in longitude @xmath36 . the nutation in obliquity signature is smaller ( up to @xmath139 mm / s ) , see fig . [ fig_sensilat ] . there is a difference of 6 hours between the maxima in the nutation in obliquity and in longitude , therefore they can not be maximal at the same time . the maximum of the nutation signature is smaller than the product of the maximal value of the component of equations ( [ eq_signeps ] ) and ( [ eq_signpsi ] ) ( i.e. @xmath140 and the maximum of @xmath141 or @xmath142 ) because the declination and the nutation do not have their maximum at the same time . + the liquid core signature in the doppler observable is much smaller than the nutation signature and largely depend on the characteristics of the core because of the resonant effects in the transfer function . the maximal signature of the liquid core in the doppler observable is plotted as a function of the fcn period on fig . [ fig_sensinut ] . the signature is larger ( up to 0.01 mm / s ) when the fcn is close to the terannual frequency . this can also be seen on fig . [ fig_liquid ] . a shift on the precession equal to the present uncertainty on the precession rate ( 2.1 mas / y , corresponding to a period shift of @xmath143 years , konopliv et al . , 2016 ) changes the doppler observable by about @xmath144 mm / s after @xmath145 years . + whatever the lander latitude below 80@xmath2 , the signature of the lod is well above the other signatures . it culminates at about @xmath146 mm / s for a lander located close to the equator . this explains why this rotation parameter has already been measured using the previous martian landers . + assuming a chandler wobble amplitude of @xmath68 meters , the maximal signature of the polar motion for a lander latitude of @xmath147 is 0.024 mm / s and @xmath148 mm / s for a lander latitude of @xmath149 . the signature increases as the latitude increase and an equatorial lander is not sensitive to the polar motion . if the lander is close to the pole , the maximal signature is close to 0.1 mm / s . + a tidal displacement of 1 cm on mars surface has a very small signature in the doppler observable of about @xmath150 mm / s for an equatorial lander . + the factor of proportionality between a mop ( expressed in surface displacement on the planet s surface ) and its signature in the doppler observable is close to @xmath39 for the lod variations , the x and y - coordinates , the tidal displacement and the polar motion . this factor is @xmath151 for the nutation in obliquity and @xmath152 for the nutation in longitude . for mars , the rotation rate is large ( @xmath153 rad / s ) . for example if a @xmath69 meter shift is applied to the x or y - coordinates , the doppler observable changes by maximum 0.07 mm / s . however if the z - coordinate of the lander changes by 10 meters , the signature in the doppler is negligible ( less than 0.001 mm / s ) . + the expressions given before are first order expressions linearized with respect to the mop . with respect to the full expression , it stays close up to 95 % of the signal . we also compared these analytical expressions to outputs from the gins software ( godsie par intgrations numriques simultanes ) , a numerical software able to process planetary geodesy data , and the numerical difference between the full expression and our analytical expression is small . we can apply the expressions of the rotational variations signature in the lander - earth direct radio link to other bodies of the solar system and estimate the order of magnitude of the signature , see tab . [ tab_ss ] . the lander has to be equipped with a transponder is on the planetary surface and the communication is direct - to - earth ( dte ) . the previous signature expressions can not be applied to doppler observable between a lander and an orbiter , or between the orbiter and the earth . the largest rotational variation of bodies in spin - orbit resonance is usually the annual forced libration . this libration amplitude depends on the interior structure and properties and if there is a liquid layer or not . additional forced librations with other frequencies also exist . the obliquity is another parameter important to observe , because if the body is in the cassini state , the equilibrium obliquity depends on the body moment of inertia . .amplitude of the libration at the orbital period and its maximal signature in the doppler observable between the earth and a lander on the planetary body . the lander is assumed to be on the equator which maximized the libration displacement . when the libration amplitude is not yet measured , we use estimation coming from a theoretical model . [ cols="^,^,<,^",options="header " , ] mercury is in a 3:2 spin orbit resonance and orbits around the sun with a period of 88 days . the main libration is the annual libration , with an amplitude of 38.5 as ( margot et al . , 2012 ) . using eq . ( [ eq_utsign ] ) , the maximal signature in the doppler observable of this libration is 0.56 mm / s . the earth declination seen from mercury varies between -11@xmath2 and 11@xmath2 . mercury has an obliquity of 2.04 arcmin ( margot et al . , 2012 ) . using eq . ( [ eq_signeps ] ) , the maximal signature of the obliquity in the doppler signal is about 0.33 mm / s , smaller than the libration signature . for mercury , like for mars , the change of the planet velocity due to the rotation parameter is also the dominant term in the doppler observable signature . however since the distance to earth is usually smaller for mercury than for mars , the geometrical effect for mercury is not as small as for mars . + since we evaluate only the change of the lander velocity because of the rotation parameter , the previous equations are valid if the lander is far from the receiving antenna on earth . if the lander - receiver distance is smaller , the change of the line - of - sight direction must be taken into account and increase the doppler observable signature . because this line - of - sight change is not modeled here and because the geometry of the earth - moon configuration is very different from the earth - planet configuration due to the synchronous rotation , we do not evaluate the signatures for a lunar lander . + if the lander in on phobos surface , the signature of the libration amplitude will be very large ( about 100 mm / s ) , because phobos rotation rate and its libration amplitude are both large . + for the icy moons around jupiter and saturn , the velocity effect is larger than the geometry effect . each libration signature in the doppler observable is proportional to the moon radius , to the diurnal frequency and to the moon libration amplitude . the variation of the earth declination does not largely affect the libration signature . for enceladus , the estimated signature is large ( about 28 mm / s ) because the measured libration amplitude is large ( @xmath154 , thomas et al . , 2016 ) . for the other moons , the libration amplitude has not yet been measured but we use order of magnitude from theoretical estimations . the signature in the doppler observable varies between 0.07 and 2.2 mm / s , see tab . [ tab_ss ] . these signatures are larger than the typical noise on doppler data ( @xmath135 mm / s ) . + a tidal displacement of 60 cm on europa surface has a signature in the lander - earth doppler observable of about @xmath155 mm / s for an equatorial lander . , the nutations in longitude @xmath36 , the liquid core effect in the nutations for a large core , a precession rate @xmath37 difference of 2 mas / y , the lod variations and the polar motion ( @xmath30 and @xmath31 ) . on the left part of the graph , the insight lander is in western elysium planitia ( @xmath83n , @xmath84e ) and on the right part , the exomars lander is in oxia planum ( @xmath85n , @xmath86e ) . in the nutation plots , the gray line shows the earth declination @xmath3 . the red and purple curves are the surface displacements due to the nutation in longitude and nutation in obliquity , respectively . , width=755,height=680 ] in this section , we give the expressions of the mop and lander position signature in the range observable . the instantaneous range observable is equivalent to the distance between the lander and the antenna on earth at a given time . the mop signature in the range observable @xmath156 is , to the first order in the rotation parameter , the vector change of position due to the mop @xmath157 projected on the earth - mars line - of - sight @xmath158 . @xmath159 this expression is independent from the earth - mars distance . the variables that appear in the following expressions are the position of the earth with respect to mars ( i.e. the earth declination @xmath3 , the right ascension @xmath6 and the hour angle @xmath5 ) and the lander position . because the doppler measurement can be seen as the time derivative of the range observable , the parameter signature in the doppler observable is the time derivative of the parameter signature in the range observable . therefore the global shapes of the signature of these two observables are very similar , compare figs . [ fig_sensi ] and [ fig_sensiran ] . the range signatures may have additional shift or long - term variations that disappear in the doppler observable signature . again the mop signature in the range observable has a diurnal variation and is proportional to the mop . the signature of the nutation in longitude @xmath36 and nutation in obliquity @xmath35 in the range observable is @xmath160 the first term of these expressions changes slowly and is null if the lander is close to the equator while the second term has a diurnal variation and is dominant . close to the equator , the dependence of the range signature to the declination is the same as the doppler signature : when the earth declination is null , the nutation signature in the range observable is null too . the liquid core signature in the range observable is the difference between the range observable with and without the nutation transfer function , each part being the sum of the contributions of the nutation in longitude and in obliquity ( eqs . [ eq_signraneps ] and [ eq_signranpsi ] ) . the signature of the precession in the range observable is @xmath161 the precession signature increases with time but on an interval like the mission duration which is short with respect to the precession period of about @xmath103 years , this increase is relatively small . the diurnal variations dominates the signature . the second and third terms are equivalent to the signature of the nutation in longitude @xmath36 . the additional term here with respect to the @xmath36 signature in the range observable ( the first one ) depends on @xmath105 while the others depend on @xmath106 . the largest term is the first one ( @xmath162 ) , then the third one . the second one ( the constant one ) is negligible . the signature of the lod variations in the range observable is @xmath163 when the earth culminates in the lander sky ( @xmath164 ) , the lod signature is null . the timing of the range maxima are therefore 6 hours later than the doppler maxima . the signature of the lander coordinates in the range observable is @xmath165 the signatures of the x and y - coordinates have a diurnal variation . the signature of the z - coordinate in the range observable , contrary to the doppler signature , is not null and has no diurnal variation . because of the @xmath108 , it has long term variations . therefore some range measurements when the earth declination is not null can be useful to constrain the z - coordinate of a lander . the signature of the polar motion in the range observable is @xmath166 the first part of these expressions ( @xmath167 or @xmath168 ) is large if the lander is close to the equator ( up to 0.3 m ) while the second term has a diurnal variation and is much smaller for an equatorial lander . the maximal values of the mop signatures in the range observable are given in tab . [ tablsiran ] . they can also be viewed on fig . [ fig_sensiran ] . @xmath169 as for the doppler signature , the parameter that has the largest signature in the range observable is the lod variations with a signature of about 11 meters . then the nutation has a maximal signature of about 3.7 to 5.5 meters . the nutation signature in the range observable is smaller than the nutation amplitude and the lander displacement given in table 1 because of the reduction proportional to the sinus of the earth declination . the liquid core signature is much smaller , about 0.17 meter . the polar motion signature is larger than the liquid core signature ( up to 0.7 meter ) . if the precession shift by 2.1 mas / y , the precession signature in the range observable is max 0.74 m around 2020 . using eq . [ eq_utransign ] and similarly as the doppler observable study , we can numerically evaluate the signature in the range observable of an libration at the orbital period for a lander on another planet or moon of the solar system . the typical noise on range data between the earth and a lander on another body is about a few meters . the maximal value of the rotation parameter signature is simply the maximal displacement on the body surface due to the libration . if we use the order of magnitude of the libration amplitude given in tab . [ tab_ss ] , the maximal signature in the range observable for an equatorial lander is given by the second column . it varies from a few meters up to 528 meters for enceladus . the signature for mercury is also large , 455 meters . the maximal signature of mercury obliquity in the range observable is about 270 m , smaller than the libration signature . all the signatures are modulated by the diurnal rotation and by variations at at long period due to the configuration geometry . the same beam interferometry ( sbi ) technique , also known as the inverse vlbi , is based on the simultaneous tracking of two martian landers with one large antenna on earth , for example a deep space network antenna , and measuring the delay between the two radio signals . the advantage of the sbi technique is that the two signals cross almost the same media at the same time , and therefore taking the difference cancels a large part of the common errors , like the earth atmospheric and ionospheric errors or the plasma delays . the sbi observable corresponds to the difference of the two ranges between the antenna and each lander . therefore the rotation parameter signature in the sbi observable is the difference between the rotation signature in the range observable of the two landers . this technique has never been used up to now it is expected to have an error budget up to 0.7 cm for martian landers ( gregnanin et al . , 2014 ) . the mop signature in the sbi observable depends on the latitude and longitude of each lander . we can replace the local angle @xmath5 by @xmath170 in the previous expressions and substitute the latitude and longitude of the lander by their rectangular coordinate @xmath171 . for example , the lod signature in the sbi observable is @xmath172 where @xmath173 and @xmath174 are the difference of x and y - coordinates of the two landers . the @xmath175 contribution does not appear in this signature . the nutation signature in the sbi observable is @xmath176 we see that the parts proportional to @xmath173 and @xmath174 are affected by a diurnal modulation and have a negligible contribution if the earth declination is small . the @xmath175 contribution has long term variations . these expressions are similar to the sbi sensitivity given in kawano et al . ( 1999 , eq . 28 ) . their variables describing the earth position with respect to mars are not equivalent to ours ( their model is simpler ) but it is possible to see the same modulations ( the diurnal and the long term ) by comparing the two methods . the numerical values of the mop signatures in the sbi observable depend of course of the separation vector between the two landers . but a general conclusion is that the signature is larger if the distance between the two landers is larger . using the analytical expressions of the mars orientation parameters ( mop ) signatures in the doppler observable , we can anticipate the moments where the mop signatures are maximal , on a daily basis and on longer interval and we can predict if a mop can be observed or not . the mop signature in the doppler and the range observable is proportional to the mop signal . therefore , the periodicities of the mop ( mostly annual , semiannual , terannual , the chandler period of about 200 days for the polar motion etc . ) can also be directly seen in the signatures . on a long time interval , for the mission operations planning , observation times should be privileged when the mop is large in order to get a larger signature . + additionally , some signatures like the nutations depend on the earth declination . the signature is larger if the earth is outside the equatorial plane of mars . the best time to observe the nutations is when the earth declination is large and when the sun - earth - probe angle is large because the noise on the doppler measurement due to the plasma will be smaller at that time . this corresponds to the beginning ( 2018.9 - 2019.25 ) and end ( 2020.3 - 2021 ) of the insight mission . for the exomars mission , this corresponds to approximately 2021 - 2021.1 and 2021.5 - 2021.8 . + in every signature , there is also a diurnal modulation because the main motion affecting the lander is the diurnal rotation of mars . therefore every day , each mop signature has a maximum and a minimum signature . these observation time should be favored . however , the daily maxima in the signatures of the nutations in obliquity and in longitude are separated by 6 hours , therefore it is impossible to maximize both at the same time . the daily maxima in the length - of - day variations signature in the doppler observable happen when the earth is at its highest point in the lander sky . + defining the best timing windows is useful for the mission planning of the geodesy experiments in the future insight and exomars missions . we will use these results when testing different strategies for adjustments of the rotational parameters . these strategies will be tested by performing numerical simulations for assessing the precision on the determination of the mop determination . we thank a. rivoldini for the computation of the transfer functions and rose - marie baland and sebastien le maistre for the helpful comments . this work was financially supported by the belgian prodex program managed by the european space agency in collaboration with the belgian federal science policy office . borderies , n. , g. balmino , l. castel , and b. moynot , study of mars dynamics from lander tracking data analysis , _ the moon and the planets , 22 _ , 191 - 200 , 1980 . curkendall , d. w. , and s. r. mcreynolds , a simplified approach for determining the information content of radio tracking data , _ j. spacecraft , 6_(5 ) , 520 - 525 , 1969 . defraigne , p. , o. de viron , v. dehant , t. van hoolst , and f. hourdin , mars rotation variations induced by atmosphere and ice caps , _ j. geophys . res . , 105_(e10 ) , 24563 - 24570 , 2000 . dehant , v. , t. van hoolst , and p. defraigne , comparison between the nutations of the planet mars and the nutations of the earth , _ survey geophys . , 21_(1 ) , 89 - 110 , 2000 . dehant , v. and 27 co - authors , lander radioscience for obtaining the rotation and orientation of mars , _ planetary and space science _ 57,10501067 , 2009 . dehant , v. , maistre , s.l . , rivoldini , a. , yseboodt , m. , rosenblatt , p. , hoolst , t.v . , mitrovic , m. , karatekin , o. , marty , j. , chicarro , a. , revealing mars deep interior : future geodesy missions using radiolinks between landers , orbiters , and the earth . _ planetary and space science _ 59 , 10691081 , 2011 . folkner , w. , r. kahn , r. preston , c. yoder , e. standish , j. williams , c. edwards , r. hellings , t. eubanks , and b. bills , mars dynamics from earth - based tracking of the mars pathfinder lander , _ j. geophys . res . , 102_(e2 ) , 4057 - 4064 , 1997a . folkner , w. , c. yoder , d. yuan , e. standish , and r. preston , interior structure and seasonal mass redistribution of mars from radio tracking of mars pathfinder , _ sci . , 278 _ , 1749 - 1751 , 1997b . folkner , w.m . , asmar , s.w . , dehant , v. , warwick , r.w . , the rotation and interior structure experiment ( rise ) for the insight mission to mars . in : lunar and planetary institute science conference abstracts , lunar and planetary institute technical report , vol . 43 , march , p. 1721 , 2012 . gregnanin , m. , m. yseboodt , v. dehant , l. iess and t. van hoolst , estimation of mars geophysical information through same beam interferometry , european planetary science congress ( epsc ) , vol . 9 , epsc2014 - 395 , 2014 . kawano , n. , m. hosokawa , h. hanada , and m. imae , inverse vlbi method for planetodesy , _ j. geod . _ , 45(3 ) , 181203 , 1999 . konopliv , a. , yoder , c. , standish , e.m . , yuan dah - ning , sjogren , w.l . , a global solution for the mars static and seasonal gravity , mars orientation , phobos and deimos masses , and mars ephemeris . _ icarus _ 182 ( 1),2350 , 2006 . konopliv , a.s . park , w.m . folkner , an improved jpl mars gravity field and orientation from mars orbiter and lander tracking data , _ icarus _ , 274 , 2016 . kuchynka , p. , folkner , w.m . , konopliv , a.s . , park , r.s . , le maistre s. , dehant , v. , new constraints on mars rotation determined from radiometric tracking of the opportunity mars exploration rover . _ icarus _ 229 , 340347 , 2014 . le maistre , s. , rosenblatt , p. , rivoldini , a. , dehant , v. , marty , j - c . , karatekin , . , lander radio science experiment with a direct link between mars and the earth . _ space sci . _ 68 , 2012 . le maistre , s. , the rotation of mars and phobos from earth - based radio - tracking observations of a lander , phd dissertation , universit catholique de louvain , belgium , available on dial.uclouvain.be , 2013 . le maistre , s. , insight coordinates determination from direct - to - earth radio - tracking and mars topography model , _ planet . space sci . _ 121 , 2016 . margot , j .- , s. j. peale , s. c. solomon , s. a. hauck ii , f. d. ghigo , r. f. jurgens , m. yseboodt , j. d. giorgini , s. padovan , and d. b. campbell , mercury s moment of inertia from spin and gravity data , _ j. geophys . _ , 117 , e00l09 , doi:10.1029/2012je004161 , 2012 . panning , m.p . , and 33 co - authors , planned products of the mars structure service for the insight mission to mars , submitted to space science reviews , 2016 . roosbeek , f. , analytical developments of rigid mars nutation and tide generating potential series , _ celestial mechanics , 75 _ , 287 - 300 , 1999 . sasao , t. , s. okubo , and m. saito , a simple theory on dynamical effects of a stratified fluid core upon nutational motion of the earth , proceedings of the agu symposium , no . 78 on _ nutation and the earth s rotation _ , fedorov ep et al . ( eds ) , kiew , d. reidel , norwell , mass . , 165 - 183 , 1980 . thomas , p.c . , r. tajeddine , m.s . tiscareno , j.a . burns , j. joseph , t.j . loredo , p. helfenstein , c. porco , enceladus s measured physical libration requires a global subsurface ocean , _ icarus , 264 _ , 2016 . van den acker , e. , t. van hoolst , o. de viron , p. defraigne , f. forget , f. hourdin , and v. dehant , influence of the winds and of the co@xmath0 mass exchange between the atmosphere and the polar ice caps on mars rotation , _ j. geophys . , 107_(7 ) , doi:10.1029/2000je001539 , 2002 . van hoolst , t. , v. dehant , and p. defraigne , chandler wobble and free core nutation for mars , _ planet . space sci . , 48_(12 - 14 ) , 1145 - 1151 , 2000a . van hoolst , t. , v. dehant , and p. defraigne , sensitivity of the free core nutation and the chandler wobble to changes in the interior structure of mars , _ phys . earth planet . inter . , 117 _ , 397 - 405 , 2000b . yoder , c. f. , and e. m. standish , martian precession and rotation from viking lander range data , _ j. geophys . res . , 102_(e2 ) , 4065 - 4080 , 1997 . yseboodt , m. , j .- barriot , and v. dehant , analytical modeling of the doppler tracking between a lander and a mars orbiter in terms of rotational dynamics , _ j. geophys . res . , 108 _ , 2003 . assuming there is no lod variations and no polar motion , we describe here the method to get the expression of the nutation signature in the doppler observable . however this development can be applied similarly to each mop . + if the vector @xmath177 is the fixed lander position in the body fixed reference frame , the lander position in the inertial frame is @xmath178 . its velocity in the inertial frame is : @xmath179 the terms proportional to @xmath100 and @xmath180 have been neglected here because they are much smaller than @xmath181 ( @xmath182 the diurnal rotation @xmath39 ) . to the first order in the nutations in obliquity @xmath183 , the lander velocity in a non - rotating frame based on the mars mean equator of j2000 ( mme2000 ) is @xmath184 the lander velocity @xmath185 is mostly a velocity in the xy plane . but the lander velocity induced by a nutation in obliquity , in the non - rotating frame based on the mars mean equator of 2000 , is mainly along the z axis . + similarly , the main component of the signature of the nutations in longitude @xmath36 in the lander velocity is also the z component , because the dependence of the velocity is only in the z coordinate . @xmath186 this z component ( to the first order in @xmath36 ) is not as straightforward as for @xmath35 because there are @xmath36 dependent terms both in the products of 4 rotation matrices and in the @xmath187 matrix , and there is cancellation between some of these terms . + in the same inertial frame , the tracking station - lander direction is aproximated by the earth - mars direction . @xmath6 is the earth right ascension . the mop signature in the doppler observable is the scalar product between the velocity difference @xmath189 = ( @xmath190 or @xmath191 ) and the los vector : @xmath192

the phenomenon of quantum interference can be used to control physical and chemical processes . one approach , the ` @xmath3 ' scheme , uses a two - color light field to interfere @xmath4- and @xmath5-photon transitions @xcite . interference between one- and two - photon transitions , for example , allows controllable polar asymmetry of photoelectrons in atomic ionization @xcite , controllable dissociation of hd@xmath6 @xcite , and controllable photocurrent injection in unbiased solids due to free carrier absorption @xcite , impurity - band transitions @xcite , and quantum well bound - continuum intersubband transitions @xcite . in biased asymmetric semiconductor double wells , ` @xmath7 ' interference allows control of carrier population and thz emission @xcite . interband ` @xmath7 ' interference in unbiased semiconductors , which is our interest here , allows independent control of electrical current injection @xcite and spin current injection @xcite . furthermore , in noncentrosymmetric semiconductors , it allows independent control of carrier populations ( i.e. , absorption ) @xcite and carrier spin polarization @xcite . in each scenario , the experimenter can control the interference by adjusting the phases of the two colors . the controllable optical phases are not the only source of phase between the transition amplitudes . in general , there is also a material - dependent intrinsic phase @xcite . phenomenologically , the intrinsic phase appears as a phase shift in the dependence of the process on a relative phase parameter of the optical fields . additionally , selectivity between two processes is possible when their intrinsic phases differ @xcite ; for example , ` 1 + 3 ' experiments on diatomic molecules have controlled the branching ratio of ionization and dissociation channels @xcite . the intrinsic phase can be strongly frequency dependent near resonances @xcite , and the hope that it might serve as a new spectroscopic observable @xcite has led to efforts to understand its physical origin . whereas a resonance is necessary for a phase shift to a ` 1 + 3 ' process @xcite , it is not necessary for a phase shift to a ` 1 + 2 ' process . for example , an intrinsic phase in the ` 1 + 2 ' photoionization of atoms is predicted from the simple model of a delta function potential @xcite . nevertheless , microscopic theories for the interband ` @xmath7 ' processes in bulk semiconductors have until now predicted trivial intrinsic phases @xcite . the photocurrent , for example , was predicted to be proportional to the sine of the relative phase parameter for all final energies @xcite . each of these theories use the independent particle approximation , in which the coulomb attraction between the injected electron and hole is neglected . that approximation is expected to be good for final energies well above the band gap , since in this case the electron and hole travel quickly away from each other . however , close to the band gap , one generally expects to see deviations from the independent particle approximation . in the one - photon absorption spectrum , for instance , it is well known that the electron - hole attraction is responsible for exciton lines below the band gap , and for an enhancement of the absorption above the gap known as _ sommerfeld _ or _ coulomb enhancement _ @xcite . the effect of the electron - hole attraction on one - photon absorption has been studied with various degrees of sophistication . on the one hand , modern _ ab initio _ calculations that include coulomb effects have recently given very good quantitative agreement with experimental spectra @xcite , although at the cost of significant computational overhead . on the other hand , simple models of wannier excitons can describe coulomb effects near the band edge of many direct gap semiconductors . these excitonic effects have long been understood qualitatively on the basis of a simple two - band model in the effective mass approximation @xcite , which is even quantitatively accurate for typical semiconductors @xcite . excitonic effects on nonlinear optical properties of bulk semiconductors have also been studied in the effective mass , wannier exciton approximation @xcite and only recently with _ ab initio _ methods @xcite . the two - photon absorption spectrum shows a different set of exciton lines and a coulomb enhancement that is weaker than its one - photon counterpart . one- and two - photon absorption spectra have been measured sufficiently often that excitonic effects on them are well established . in contrast , semiconductor ` 1 + 2 ' interference experiments have been done typically at only a single energy and typically many exciton binding energies above the band - gap . moreover , these initial experiments lacked an absolute calibration of the relative optical phase , and thus were insensitive to the intrinsic phase . such a calibration is possible @xcite , and could be used to verify the predictions we present here . a nontrivial intrinsic phase would have implications for the use of ` @xmath7 ' current injection as an absolute measurement of the carrier - envelope phase of an ultrashort optical pulse @xcite . in the present work , we extend the theory of ` 1 + 2 ' coherent control of bulk semiconductor interband transitions beyond the independent particle approximation , employing a set of approximations that are valid close to the @xmath8 point of a direct gap bulk semiconductor . our investigation is limited to a perturbative treatment in the fields . in this limit of low photoinjected carrier density , the only inter - particle interaction of importance is that between a single electron and hole . we show that , due to the electron - hole attraction , a nontrivial phase shift does in fact occur in the control of current and spin current , but not in the control of carrier population or spin polarization . the intrinsic phase can be understood in terms of partial wave phase shifts due to the coulomb potential between electron and hole . in addition , we find an enhancement of each process , and relate it to the coulomb enhancements of one- and two - photon absorption . in the next section , we establish notation necessary to describe the ` 1 + 2 ' processes in terms of one- and two - photon transition amplitudes . in section [ sec : model ] , we present the microscopic model . the transition amplitudes are worked out in section [ sec : transition_amplitudes ] . the final expressions for the ` 1 + 2 ' effects are given in section [ sec : results ] , and numerical results for gaas are presented . in section [ sec : discussion ] , further understanding of the enhancement and intrinsic phase is discussed , and we examine the ratios often used as figures of merit for ` 1 + 2 ' effects . intermediate state coulomb enhancement is examined in appendix [ appx_intstateenhancement ] . for reference , the current injection tensor is worked out for parabolic bands in appendix [ appx_etafree ] . the rate of photocurrent injection into an unbiased bulk semiconductor by a two color light field @xmath9 can be written @xmath10 where superscripts denote cartesian components and repeated indices are to be summed over @xcite . the fourth rank tensor @xmath11 , called the current injection tensor , describes the material response . it is purely imaginary in the independent particle approximation @xcite , but can be complex in general . we define the intrinsic phase @xmath12 of the component @xmath13 as @xmath14 so that it is zero or @xmath15 in the independent particle approximation . when the electron - hole interaction is included in the set of approximations used below , all the components of @xmath16 have the same phase . that is , @xmath17 the intrinsic phase @xmath18 appears as a phase shift in the dependence of the current injection on the phase of the optical fields . for co - linearly polarized fields ( @xmath19 and @xmath20 ) , for example , the current injection is @xmath21 we are ignoring scattering processes through which the current would relax to a steady state value under continuous illumination , or would decay to zero following pulsed excitation . the current injection discussed here can be used as a source term in hydrodynamic equations that treat the scattering phenomenologically @xcite or in microscopic transport equations @xcite . coulomb effects other than the excitonic effects we consider here play a role in scattering , especially at high densities of excited carriers . such coulomb effects are outside the scope of this paper . the current injection tensor can be written in terms of one- and two - photon transition amplitudes . we take as the initial state a clean , cold semiconductor . in a fermi s golden rule calculation for the ballistic current , light produces transitions to final states @xmath22 with velocity @xmath23 , probability amplitude @xmath24 , and energy @xmath25 above the ground state . the final states will be specified in detail in the next section ; here the label @xmath4 represents the set of quantum numbers for either an interacting or independent electron - hole pair . thus , @xmath26 where @xmath27 is the electron charge ( negative ) , @xmath28 is a normalization volume , and @xmath29 is the amplitude for an _ i_-photon transition . the @xmath30 take the form @xmath31 where the vector @xmath32 and second rank tensor @xmath33 depend only on properties of the material . we consider them in detail in section [ sec : transition_amplitudes ] . in the last equation of ( [ currentmicro ] ) , the first term is ( one - photon ) one - color current injection , the second term describes two - photon one - color current injection , while the interference term describes the ` 1 + 2 ' current . because of their different dependencies on the electric field amplitudes , these three terms can in principle be separated experimentally . but the first two terms vanish for centrosymmetric crystals , and the first vanishes even for zincblende crystals ; we neglect them here . excitonic effects on the first term were studied by shelest and ntin @xcite . the third term survives in all materials . comparing its expression with the phenomenological form ( [ current phenomenological ] ) , we have @xmath34 even if scattering from impurities and phonons is neglected , the injection current described by ( [ currentmicro ] ) does not capture the full current density ; there are also optical rectification and ` shift ' contributions to the current @xcite . the one - color varieties of these have been studied in some detail @xcite , but the ` 1 + 2 ' varieties have not . however , the different time dependencies of the three current contributions allows for their separate examination experimentally , at least in principle , and rough order - of - magnitude estimates indicate that typically the injection current will be the largest ; it is the only contribution to the current we treat here . as in ( [ currentmicro ] ) , the carrier injection rate can be written as @xmath35 where @xmath4 is the number density of injected electron - hole pairs . the first two terms in ( [ carrier injection rate ] ) are the usual one- and two - photon absorption rates , which we denote by @xmath36 and @xmath37 respectively . in terms of one- and two - photon coefficients @xmath38 and @xmath39 that depend only on the properties of the material , they can be written as @xmath40 and @xmath41 @xcite . from ( [ carrier injection rate ] ) , @xmath42 the third term in ( [ carrier injection rate ] ) , denoted @xmath43 , allows population control as discussed and observed by fraser _ et al _ @xcite . it can be written in terms of a third rank tensor @xmath44 as @xcite @xmath45 where @xmath46 it is purely real in the independent particle approximation @xcite . expressions such as ( [ currentmicro ] ) and ( [ carrier injection rate ] ) can also be written for carrier spin polarization and spin current @xcite . the interference terms of these describe ` 1 + 2 ' spin control @xcite and ` 1 + 2 ' spin current injection @xcite , which can be written in terms of material response pseudotensors @xmath47 @xcite and @xmath48 @xcite , respectively . in the independent particle approximation , @xmath47 is purely imaginary @xcite , while @xmath48 is purely real @xcite . the phases of the material response tensors @xmath16 , @xmath49 , @xmath50 , and @xmath51 are related to the phases of the one- and two - photon matrix elements , @xmath52 and @xmath33 . the one- and two - photon matrix elements also appear , respectively , in the one- and two - photon absorption coefficients , as can be seen from ( [ one_photon_schematic ] ) and ( [ two photon schematic ] ) . there have been many theoretical investigations of one- and two - photon absorption near the direct gap of bulk semiconductors that include excitonic effects @xcite . however , since one- and two - photon absorption are insensitive to the phases of the transition amplitudes , those calculations took no care to get the phases of the transition amplitudes correct . in the next two sections we find the transition amplitudes with the correct phases including excitonic effects we first review the two - band effective mass model of wannier excitons ; the two bands are nondegenerate conduction and valence bands that are parabolic and isotropic with a direct gap @xmath53 at @xmath54 ( the @xmath8 point ) @xcite . it has been used to study excitonic effects on one - photon absorption @xcite , two - photon absorption @xcite , and other nonlinear optical processes @xcite . we subsequently describe a generalization that accounts for degeneracy and multiple bands . it has been used for two - photon absorption @xcite , and has been implied whenever two - band results have been applied to actual semiconductors . the total hamiltonian of the system can be written in the form @xmath55 . here , @xmath56 is the field - free hamiltonian made up of the single - particle part @xmath57 and the part due to the coulomb interaction between carriers @xmath58 . using the minimal coupling hamiltonian , the optical perturbation takes the form @xmath59 , where @xmath60 is the vector potential associated with the maxwell electric field and @xmath61 is the velocity operator associated with @xmath62 . in the long wavelength limit , the position dependence of @xmath60 is neglected , and thus the second term in @xmath63 may be neglected , since it can simply be absorbed in an overall time - dependent phase of the full system ket and hence can not cause any transitions between states of @xmath62 . many approximate approaches to band structure calculation , including most pseudopotentials , and the truncation to a finite number of bands , implicitly assume an underlying field - free hamiltonian that is nonlocal ; there is then a correction to the interaction hamiltonian @xmath63 in the velocity gauge @xcite . however , we neglect such nonlocal corrections , as has been the practice in previous calculations of coherent control effects @xcite . the initial state is the `` vacuum '' @xmath64 ; it corresponds to completely filled valence bands and empty conduction bands . if the coulomb interaction were neglected in a two - band model consisting of valence ( @xmath65 ) and conduction ( @xmath66 ) bands , the final states would be of the form @xmath67 , where the operator @xmath68 creates an electron in an eigenstate of @xmath57 , a bloch state @xmath69 with band index @xmath4 and wavevector @xmath0 . the photon momentum has been neglected , consistent with the long wavelength approximation . the coulomb interaction couples states at different @xmath0 ; thus including @xmath58 the final states are of the form @xmath70 where @xmath71 labels the state ; its physical meaning is given below . in the effective mass wannier exciton approximation , the fourier transform @xmath72 which is the wavefunction for the relative coordinate between electron and hole , is a hydrogenic wavefunction satisfying @xmath73 where @xmath74 is the reduced mass in terms of the ( positive ) conduction and valence band effective masses , and @xmath75 is the coulomb potential , @xmath76 , screened by the static dielectric constant @xmath77 @xcite . the state has energy @xmath78 we choose the states to be normalized over the volume @xmath28 by @xmath79 and @xmath80 ; as a result @xmath81 is unitless , having the normalization @xmath82 . our focus in this paper is on the unbound solutions to ( [ exciton_schrodinger ] ) ; bound exciton states lack relative velocity between the electron and hole , and hence do not contribute to the ballistic current or spin current . for a fermi s golden rule calculation of the current or spin current , the unbound state must behave asymptotically like an outgoing plane wave in the relative coordinate between electron and hole ; @xmath83 is the wavevector of the outgoing plane wave . specifically , we must use `` ionization states '' rather than scattering states @xcite , as was done for atomic ` 1 + 2 ' ionization @xcite . they are related by @xmath84 ^{*}$ ] @xcite . calculations of one- or two - photon absorption are insensitive to an error in this choice of boundary condition , but the present calculation is not , since it is sensitive to the relative phase of the transition amplitudes . the ionization state wavefunctions that solve ( [ exciton_schrodinger ] ) can be expressed as a sum over partial waves , @xmath85 where @xmath86 is the exciton bohr radius , and @xmath87 and @xmath88 mean @xmath89 and @xmath90@xcite . the @xmath91 are legendre polynomials , @xmath92 is a confluent hypergeometric function , and @xmath93 is the gamma function . such a two - band model of wannier excitons is useful for the description of many optical properties . however , near the band gap at the @xmath8 point of a typical zincblende semiconductor there are , counting spin , eight bands : two each of conduction ( @xmath66 ) , heavy hole ( @xmath94 ) , light hole ( @xmath95 ) and split - off hole ( @xmath96 ) . other bands , especially the next higher conduction bands , can also be important for some processes , especially for population and spin control . the existence of multiple bands and band degeneracy modifies the exciton hamiltonian , i.e. , the operator acting on @xmath81 in the left side of ( [ exciton_schrodinger ] ) . in the effective mass approximation , using a basis of @xmath8 point states , the kinetic part of the wannier exciton hamiltonian has a matrix structure @xcite . even though this approximation neglects band warping , nonparabolicity , and inversion asymmetry , the hamiltonian lacks analytic eigenstates @xcite . this is essentially due to the degeneracy of the @xmath94 and @xmath95 bands at the @xmath93 point . as a result of the difference between @xmath97 and @xmath98 there is ` envelope - hole coupling , ' @xcite which is a spin - orbit - like coupling between the orbital angular momentum of the exciton envelope function and the total angular momentum of the valence band @xmath8 point bloch functions @xcite . baldereschi and lipari split the effective mass hamiltonian into a sum of terms based on symmetry , and showed that in a spherical approximation the envelope - hole coupling could be treated as a perturbation to the diagonal part , which has analytic , hydrogenic eigenstates @xcite . in order to extract the main physics , while preserving the simplicity of the two - band model , we neglect envelope - hole coupling entirely . in this approximation , ( [ exciton_schrodinger ] ) remains valid for each conduction - valence band pair , however one must use ` average ' effective masses for degenerate bands . specifically , the effective mass of the valence bands @xmath94 , @xmath95 , and @xmath96 is @xmath99 , where @xmath5 is the free electron mass , and @xmath100 is one of the luttinger parameters @xcite . the upper conduction bands have a different average effective mass . note that @xmath101 is independent of @xmath66 and @xmath65 within the set of bands @xmath102 . the effect of envelope - hole coupling has been studied for exciton bound states @xcite , but not for optical processes involving unbound excitons in the continuum . even within this model , the presence of multiple bands provides two types of terms in the sum over intermediate states in the two - photon amplitude : two - band terms , in which the intermediate and final states are in the same exciton series [ i.e. , two states of the form ( [ final state ] ) with the same @xmath66 and @xmath65 ] , and three - band terms , in which the intermediate and final states are in different series . three - band terms are important for some processes but not for others . for current control , three - band terms are important for cross - linearly polarized fields @xcite , and for spin current control , they are important for the spin current due to co - linearly polarized fields @xcite . three - band terms are essential for population and spin control @xcite . the velocity matrix elements involving the state @xmath104 are @xmath105 @xmath106 where @xmath107 is the velocity matrix element between bloch states . in the independent particle approximation , the transition amplitudes are @xmath108 and @xmath109 , where @xmath110 with excitonic effects included , using the perturbation @xmath111 to second order gives the transition amplitudes @xmath112 and @xmath113 where the sum over intermediate states is over both bound and free excitons . the two - photon transition amplitude is more difficult to deal with due to the sum over intermediate states ; however , in our set of approximations it has been done exactly @xcite . in order to proceed analytically , it is common to use ( [ vac_pairvelocity ] ) and ( [ twotwovelocity ] ) , and then make an expansion in @xmath0 of the velocity matrix elements @xmath114 about the @xmath8 point @xcite . however , due to the degeneracy at the @xmath8 point , the coefficients of such an expansion can depend on the direction of @xmath0 @xcite . to proceed , we note that wannier excitons have large spatial extent and hence only a small region of wavevectors is important for them , i.e. , @xmath115 is peaked in the region of @xmath0 near @xmath83 @xcite . this is especially true for final states with energies above the band gap . thus , we expand @xmath114 about the @xmath93 point , approached in the direction @xmath116 , @xmath117 where @xmath118 and @xmath119 . optical transitions due to the first term in ( [ bloch_v_expansion ] ) are called ` allowed ' , while those due to the second term are called ` forbidden ' . we restrict ourselves to materials for which the ` allowed ' valence to conduction band transition does not vanish . keeping only the ` allowed ' term in ( [ vac_pairvelocity]),@xcite @xmath120 for the intravalence and intraconduction band transitions , the first two terms of ( [ bloch_v_expansion ] ) in ( [ twotwovelocity ] ) give @xcite @xmath121 \int \frac{d^{3}r}{l^{3}}\left ( \psi _ { cv}^{\bm{\kappa } } \left ( \mathbf{r}\right ) \right ) ^{*}\psi _ { c^{\prime } v^{\prime } } ^{\bm{\kappa } ^{\prime } } ( \mathbf{r } ) \\ & -\left [ \delta _ { v , v^{\prime } } \bm{\nabla}_{\mathbf{k}}v_{cc^{\prime } } ^{i}\left ( \mathbf{\hat{\bm{\kappa } } } \right ) -\delta _ { c , c^{\prime } } \bm{\nabla } _ { \mathbf{k}}v_{v^{\prime } v}^{i}\left ( \mathbf{\hat{\bm{\kappa } } } \right ) \right ] i\int \frac{d^{3}r}{l^{3}}\left ( \psi _ { cv}^{\bm{\kappa } } \left ( \mathbf{r}\right ) \right ) ^{*}\bm{\nabla } \psi _ { c^{\prime } v^{\prime } } ^{\bm{\kappa } ^{\prime } } ( \mathbf{r } ) . \end{split}\ ] ] in particular @xcite , @xmath122 for ge and for simple models of zincblende semiconductors that neglect lack of inversion symmetry , the first term in ( [ v_intra ] ) always vanishes . this means that there are only ` allowed - forbidden ' two - photon transitions ( the interband transition is allowed , while the intraband transition is forbidden ) . when the first term is nonvanishing , there are ` allowed - allowed ' two - photon transitions . in principle , for materials that lack a center of inversion , there is also a small contribution to the ` allowed - forbidden ' two - photon transition from the first term in ( [ v_intra ] ) and the term in @xmath123 that comes from the second term in ( [ bloch_v_expansion ] ) ; we neglect it in what follows , but note that when compared to the dominant ` allowed - forbidden ' contribution that we consider here , it has a different coulomb enhancement but the same intrinsic phase ( see eq . 2.32 of rustagi _ et al_.@xcite ) . we write @xmath124 , and discuss the ` allowed - forbidden ' and ` allowed - allowed ' transitions separately . using ( [ ionization state ] ) , ( [ allowed_interband_v ] ) , and ( [ one photon rate ] ) , the one - photon transition amplitude is @xcite @xmath125 where only the ` allowed ' term is kept in @xmath126 . the transition is to an @xmath127-wave . the one - photon absorption coefficient is proportional to the norm of @xmath128 [ see ( [ carrier injection rate ] ) and ( [ one_photon_schematic])])@xcite . for ` allowed - forbidden ' two - photon transitions , substituting ( [ allowed_interband_v ] ) and the second term of ( [ v_intra ] ) into ( [ two photon rate ] ) , @xmath129 where @xmath130 and we have used the coulomb green function , @xmath131 which is known analytically @xcite . in particular , @xmath132 where we define @xmath133 @xmath134 is the exciton binding energy , and @xmath135 is a whittaker function with the integral representation @xmath136 since the green function depends only on the magnitude of @xmath137 , only the _ p_-wave of the final state survives the integral in eq . ( [ w_defined ] ) over the angles of @xmath137 , @xmath138 . the integral over @xmath88 can be done using @xcite @xmath139 the final result is @xmath140 where only the ` allowed - forbidden ' term is kept in @xmath141 , and @xmath142 for ` allowed - allowed ' two - photon transitions , substituting ( [ allowed_interband_v ] ) and the first term of ( [ v_intra ] ) into ( [ two photon rate ] ) , @xmath143 \mathbf{e}_{\omega } \cdot \mathbf{v}_{c^{\prime } v^{\prime } } \left ( \mathbf{\hat{\bm{\kappa } } } \right ) \\ & \times \hbar \int d^{3}r\left ( \psi _ { cv}^{\bm{\kappa } } \left ( \mathbf{r}\right ) \right ) ^{*}g_{c^{\prime } v^{\prime } } \left ( \mathbf{r},\mathbf{0};\hbar \omega -e_{c^{\prime } v^{\prime } } ^{g}\right ) . \end{split}\ ] ] since @xmath144 depends only on the magnitude of @xmath137 [ eq . ( [ coulombgreenfunction ] ) ] , only the @xmath127 part of the final state will survive the integration over angles of @xmath137 . again we use ( [ whittaker ] ) for the whittaker function . the integral over the magnitude of @xmath137 can be done using an identity obtained by taking a derivative with respect to @xmath145 of both sides of ( [ integral identity ] ) . finally , @xmath146 where only the ` allowed - allowed ' term is kept in @xmath141 , and @xmath147 this agrees with eq . 2.28 of rustagi @xcite , but note that we have defined @xmath148 , where @xmath149 is given , with a typographical error , in eq . 2.25 of that paper . the factors @xmath150 and @xmath151 , which appear in ( [ n_af ] ) and ( [ n_aa ] ) are the enhancements due to the coulomb interaction in the intermediate states ; they are discussed further in appendix [ appx_intstateenhancement ] . the one- and two - photon transition amplitudes were presented in the previous section on the basis of an expansion in @xmath0 of the bloch state velocity matrix elements . the ` allowed ' one - photon transition amplitude @xmath152 is in ( [ one photon allowed ] ) , the ` allowed - forbidden ' two - photon transition amplitude is in ( [ two photon a - f ] ) , and the ` allowed - allowed ' two - photon transition amplitude is in ( [ two photon a - a final ] ) . from them , @xmath153 and @xmath154 may be extracted by comparison with the definitions in ( [ d1definition ] ) and ( [ d2definition ] ) . the ` 1 + 2 ' current injection is dominated by interference of ` allowed ' one - photon transitions and ` allowed - forbidden ' two - photon transitions @xcite . substituting these into ( [ etaschematic ] ) , and using the gamma function identities @xmath155 and @xmath156 yields our final result for the current injection tensor @xmath157 where @xmath158 @xmath159 and @xmath160 with @xmath161 and @xmath162 where @xmath163 and @xmath164 note that only the ` allowed ' part of ( [ d1free ] ) and the ` allowed - forbidden ' part of ( [ d2free ] ) should be retained for a consistent solution . we have written ( [ eta_result ] ) to separate the parts due to the electron - hole interaction . in the independent particle approximation , the current injection tensor @xmath165 is @xcite @xmath166 it is evaluated for parabolic bands in appendix [ appx_etafree ] . ) ] and without [ eq.([eta_result_free ] ) ] excitonic effects . the grey dotted and dash - dotted lines are based on a parabolic band calculation of @xmath167 that only includes two - band terms ; the dotted line includes excitonic effects , while the dash - dotted line does not . the black solid and dashed lines are based on @xmath167 calculated with a nonperturbative solution of the @xmath168 @xmath169 hamiltonian ; the solid line includes excitonic effects , while the dashed line does not . [ fig_eta],width=307 ] for gaas , we present in fig . [ fig_eta ] the magnitude of @xmath170 , based on @xmath167 calculated by two methods . the first method , described in appendix [ appx_etafree ] , uses isotropic parabolic bands and includes only two - band terms ; it uses effective mass ratios for conduction , heavy hole , light hole , and split - off bands of @xmath171 , @xmath172 , @xmath173 , and @xmath174 respectively , @xmath175ev , the fundamental band gap @xmath176 is 1.519ev , and valence band spin - orbit splitting is 0.341ev @xcite . the second method solves the @xmath177 @xmath169 hamiltonian including remote band effects , but in a spherical approximation with warping and spin - splitting neglected by replacing @xmath178 and @xmath179 with @xmath180 @xcite ; the calculation is nonperturbative in @xmath0 ( hence it includes band nonparabolicity ) and it includes both two- and three - band terms in the two - photon amplitude @xcite . the solid and dotted lines in fig.[fig_eta ] are calculated with ( [ eta_result ] ) , and hence include excitonic effects ; the coulomb enhancement part of the calculation uses @xmath181mev @xcite and the band parameters listed above . note that the solid black line in fig.[fig_eta ] is inconsistent in the sense that the coulomb enhancement is based on an expansion in @xmath0 , whereas the free - particle result that it enhances is nonperturbative in @xmath0 ; nevertheless , such an approach has given good agreement with experiments for one- and two - photon absorption @xcite . the coulomb enhancement of @xmath16 can clearly be seen in fig.[fig_eta ] . there is a kink in each curve at excess photon energy 341 mev corresponding to the onset of transitions from the @xmath96 band . at higher energies , the coulomb enhancement of @xmath96 transitions is larger than the coulomb enhancements of @xmath94 and @xmath95 transitions , since the former transitions are to conduction band states with lower energy . hence , the kink in @xmath16 is enhanced by excitonic effects . ) in gaas due to excitonic effects . the solid line is calculated with eqs . ( [ delta explicitly defined ] ) and ( [ eta_result ] ) , and the dotted line is calculated with eq.([phase shift ] ) . the inset is eq . ( [ phase shift ] ) plotted in scaled units.,width=307 ] we extract the intrinsic phase of @xmath170 using ( [ delta explicitly defined ] ) . the solid line in fig.[fig_delta ] is the intrinsic phase of @xmath182 calculated for gaas with the nonperturbative @xmath168 @xmath169 hamiltonian band model ; the result for the parabolic band model is nearly identical . since we have used a spherical exciton model , the intrinsic phase is the same for all components of @xmath183 . the intrinsic phase has its maximum value of @xmath1 at the band edge , and goes to zero as the light frequency increases . the decrease is smooth except for a small kink at the onset of transitions from the @xmath96 band . in fact , for excess photon energies less than the split - off energy , the intrinsic phase has the simple analytic form @xmath184 equation ( [ phase shift ] ) is plotted as the dotted line in fig.[fig_delta ] ; compared to the solid line , it is identical below the onset of @xmath96 transitions , and it makes a good approximation above the the onset of @xmath96 transitions . since ( [ phase shift ] ) only depends on the excess photon energy scaled by the exciton binding energy , we plot it as a function of this scaled energy in the inset of fig . [ fig_delta ] ; it is useful for finding the intrinsic phase of materials other than gaas . in @xmath11 ( [ eta_result ] ) ] , the two- and three - band terms have different intermediate state coulomb enhancement @xmath185 . for many materials , however , @xmath185 is approximately equal for all the terms @xmath186 that contribute significantly to the total @xmath165 , as shown in appendix [ appx_intstateenhancement ] for gaas . thus , at photon energies for which transitions from the heavy- and light - hole bands dominate @xmath187 , the coulomb enhancement becomes approximately independent of the sum over bands and we can make the simplification @xmath188 where the intrinsic phase is given by ( [ phase shift ] ) , and @xmath189 [ eq . ( [ enhancement ] ) ] with @xmath190 , and the dotted line , applicable to carrier population and spin control is @xmath191 [ eq.([enhancement_aa ] ) ] with @xmath192,width=307 ] the coulomb enhancement factor @xmath193 is plotted in fig . [ fig_f ] with the approximation that @xmath190 ( see appendix [ appx_intstateenhancement ] ) . the ` 1 + 2 ' carrier population control is dominated by interference of ` allowed ' one - photon transitions and ` allowed - allowed ' two - photon transitions @xcite . substituting these into ( [ xischematic ] ) , and using the gamma function identity ( [ gamma identity ] ) , we find @xmath194 where @xmath195 and @xmath196 and @xmath197 are given by ( [ d2free ] ) and ( [ d1free ] ) . note that only the ` allowed ' part of ( [ d1free ] ) and the ` allowed - allowed ' part of ( [ d2free ] ) should be retained for a consistent solution . in the independent particle approximation , @xmath198 thus , population control has a coulomb enhancement due to excitonic effects , but no phase shift . note that ( [ population control result ] ) gives the population control tensor at final energies above the band edge . there can also be population control of bound excitons when both one- and two - photon transitions are to the same excitonic state . this can occur , for example , at @xmath127 excitons due to allowed - allowed two - photon transitions @xcite interfering with allowed one - photon transitions . if @xmath199 is approximately the same for all the terms that significantly contribute to @xmath200 , then , at photon energies for which transitions from the heavy and light hole bands dominate @xmath49 , the coulomb enhancement becomes approximately independent of the sum over bands , and we can make the simplification , @xmath201 where @xmath202 the coulomb enhancement factor @xmath203 is plotted in fig . [ fig_f ] with the approximation that @xmath192 ( see appendix [ appx_intstateenhancement ] ) . the ` 1 + 2 ' spin current is dominated by interference of ` allowed ' one - photon transitions and ` allowed - forbidden ' two - photon transitions , whereas ` 1 + 2 ' spin control is dominated by interference of ` allowed ' one - photon transitions and ` allowed - allowed ' two - photon transitions @xcite . under the approximations that led to ( [ eta_c_simple ] ) and ( [ population control simple ] ) , the spin current injection pseudotensor is @xmath204 where @xmath205 is given by ( [ enhancement ] ) , @xmath206 is given by ( [ phase shift ] ) , and @xmath207 is the spin current injection pseudotensor in the independent particle approximation . under similar approximations , the spin control pseudotensor is @xmath208 where @xmath191 is given by ( [ enhancement_aa ] ) , and @xmath209 is the spin control pseudotensor in the independent particle approximation . spin control , like carrier population control , has a coulomb enhancement but no phase shift . there can also be spin control of bound excitons , but it has not been included in ( [ zeta ] ) . we now examine the relationship between the coulomb enhancements of the ` 1 + 2 ' processes and of one- and two - photon absorption ; the latter are denoted by @xmath210 and @xmath211 so that for @xmath212 , @xmath213 . the relationship is particularly simple at photon energies for which transitions from the heavy- and light - hole bands are dominant and intermediate state coulomb enhancement is the same for each significant term in the sum over intermediate states . the coulomb enhancements for the ` 1 + 2 ' processes are then given by ( [ enhancement ] ) and ( [ enhancement_aa ] ) . for one - photon absorption , @xmath214 @xcite . in noncentrosymmetric semiconductors , two - photon absorption is dominated by allowed - allowed transitions just above the band gap , and by allowed - forbidden transitions at higher final energies ; the cross - over point in gaas is a few mev above the band gap @xcite . at photon energies for which allowed - allowed transitions dominate two - photon absorption , from ( [ two photon a - a final ] ) , @xmath215 and thus @xmath216 where @xmath217 , while at photon energies for which allowed - forbidden transitions dominate two - photon absorption , from ( [ two photon a - f ] ) , @xmath218 and thus @xmath219 note that , based on appendix [ appx_intstateenhancement ] , @xmath220 , which is the ratio of the two curves in fig . [ fig_f ] . in centrosymmetric semiconductors , there are no allowed - allowed transitions , and only ( [ f_geometric_mean_af ] ) applies . the ` 1 + 2 ' processes are often described by ratios . for example , a useful quantity to describe the current is the swarm velocity @xcite , defined as the average velocity per injected electron - hole pair @xmath221 the swarm velocity is a maximum when the relative intensities of the two colors are chosen such that @xmath222 ; returning to ( [ currentmicro ] ) , if one associates the one- and two - photon amplitudes with the arms of an effective interferometer , this condition corresponds to balancing that interferometer . for fields co - linearly polarized along @xmath223 , the maximum swarm speed is @xmath224 a useful quantity to describe pure spin currents is the maximum spin separation distance @xcite ; it is proportional to @xmath225 . as a consequence of ( [ f_geometric_mean_af ] ) , the maximum swarm speed , and the maximum spin separation distance , are _ unaffected _ by excitonic effects when allowed - forbidden transitions dominate two - photon absorption @xcite . however , close to the band edge , where allowed - allowed transitions dominate two - photon absorption , excitonic effects increase these ratios by a factor @xmath226 over their value in the independent particle approximation . in contrast , as a consequence of ( [ f_geometric_mean_af ] ) , excitonic effects do not affect the maximum control ratio for population and spin control ( @xmath227 and @xmath228 respectively @xcite ) close to the band edge and decrease them by a factor @xmath226 at higher photon energies for which allowed - forbidden transitions dominate two - photon absorption . in the terminology of seideman @xcite , the excitonic phase shift of the ` 1 + 2 ' current and spin current is a direct phase shift . it is due to the complex nature of the final state as it appears in the transition amplitudes . thus it can be understood in terms of the partial wave phase shifts of the final state caused by the coulomb potential between electron and hole . the coulomb interaction is rather unique due to its long range nature , so we first suppose the potential between the electron and hole falls off more rapidly than @xmath229 . in that simpler problem , the final state wave function is written as @xmath230 where the @xmath231 are real @xcite . if the potential between the particles is ignored , then the partial wave phase shifts , @xmath232 are zero . the allowed one - photon pathway reaches an @xmath127 wave , while the allowed - forbidden two - photon pathway reaches a @xmath145 wave . substituting this form for the wave function into the one- and two - photon transition amplitudes , yields @xmath233 for the one - photon rate , where @xmath234 is real and depends on @xmath235 , and @xmath236 for the two - photon rate , where @xmath237 is real and depends on @xmath235 and @xmath238 . here @xmath239 is the _ i_-photon transition amplitude when the potential between the particles is ignored . it is then straightforward to see from ( [ etaschematic ] ) that the relative shift of the partial waves is responsible for the phase shift of the current and spin current . that is , @xmath240 the use of ionization states as opposed to scattering states was important to get the correct sign of the intrinsic phase . with scattering states , one would find @xmath241 . due to the long - range nature of the coulomb potential , the partial wave phase shifts have a logarithmic @xmath88 dependent part , but it is the same for all partial waves and thus does not appear in the relative phase . the part of the coulomb partial wave phase shift @xmath242 that does not depend on @xmath88 is @xmath243 @xcite ; when inserted into ( [ related to partial waves ] ) , this reproduces ( [ phase shift ] ) . in contrast , the allowed - allowed two - photon pathway reaches an @xmath127 wave and thus there is no phase shift for population control or spin control . the expression ( [ related to partial waves ] ) for the intrinsic phase in terms of the scattering phases is particularly simple , since each pathway connects to only a single parity . this contrasts with ` 1 + 2 ' ionization from an atomic @xmath127 state , for which the one - photon transition is to a @xmath145 wave and the two - photon transition is to both @xmath127 and @xmath244 waves ; the intrinsic phase is thus a weighting of the @xmath145-@xmath127 and @xmath145-@xmath244 partial wave shifts @xcite . materials for which the first term in ( [ bloch_v_expansion ] ) is forbidden ( cu@xmath245o is an example ) have these same selection rules @xcite ; hence , they will have an intrinsic phase with a similar weighting . the absence of a phase shift in population control can be connected to a symmetry of the second order nonlinear optical susceptibility . from considerations of energy transfer and macroscopic electrodynamics , @xmath49 is related to the nonlinear susceptibility @xmath246 by @xmath247.\ ] ] in the independent particle approximation @xcite , @xmath248 ^{*},\ ] ] which is a generalization of overall permutation symmetry to resonant absorption . as a result of ( [ gopr ] ) , fraser _ et al_. showed that @xmath249 is proportional to @xmath250 , and is thus purely real @xcite . our result that @xmath49 remains real when excitonic effects are included suggests that ( [ gopr ] ) holds more generally . in fact , it can be shown that ( [ gopr ] ) holds for any hamiltonian symmetric under time - reversal so long as @xmath251 is not resonant . we have extended the theory of interband ` 1 + 2 ' processes in bulk semiconductors to include the electron - hole interaction . following previous theories @xcite , we have used a framework based on ( i ) a separation of the initial carrier photoinjection and the subsequent carrier scattering , and ( ii ) a perturbative expansion in the optical field amplitudes , with injection rates obtained in a fermi s golden rule limit for the bichromatic field . the injection rates for carrier population control , spin control , current injection , and spin current injection , have been described phenomenologically by tensors @xmath49 , @xmath51 , @xmath16 , and @xmath50 , respectively @xcite . like previous theories , we have used the long - wavelength limit , and neglect nonlocal corrections to the interaction hamiltonian . but whereas previous theories of ` 1 + 2 ' photoinjection used the independent particle approximation , we have included excitonic effects . we have shown that excitonic effects cause ( i ) an enhancement of each ` 1 + 2 ' process , and ( ii ) a phase shift for current injection and spin current injection . our main results , the modifications of the aforementioned tensors relative to the independent particle approximation are given in ( [ eta_c_simple ] ) , ( [ population control simple ] ) , ( [ mu ] ) , and ( [ zeta ] ) . these particularly simple results are valid at photon energies for which transitions from the heavy- and light - hole bands are dominant ; more general results are given for @xmath16 and @xmath49 in ( [ eta_result ] ) and ( [ population control result ] ) . our results are based on the effective mass model of wannier excitons ; degenerate bands are included , but we use a spherical approximation to the exciton hamiltonian , and we neglect envelope - hole coupling . this is a good approximation for many typical semiconductors , including gaas , since the electron - hole envelope function extends over many unit cells due to the screening of the coulomb interaction by the static dielectric constant @xcite . as a consequence of making the spherical approximation , the phase shifts and coulomb enhancements we find in this paper are independent of crystal orientation . also , our results are limited to low excess photon energy since ( i ) the wannier exciton hamiltonian assumes parabolic bloch bands , and ( ii ) we have truncated the expansion in @xmath0 of the bloch state velocity matrix elements , which is the basis of the transition amplitude expansion . by comparing the black dashed line and grey dash - dotted line in fig . [ fig_eta ] , one sees that higher order terms in @xmath0 ( for both bands and velocities ) are important in gaas for excess photon energies greater than about 200 mev . this can then be considered the limit of validity of our calculation . however , combining the coulomb enhancement calculated assuming parabolic bands with the nonperturbative independent particle approximation result ( as was done for the solid black line in fig.[fig_eta ] ) likely gives a good approximation for a few hundred more mev ; this was the case for one- and two - photon absorption @xcite . it is interesting to ask if there are other sources of intrinsic phases to the current ( or spin current ) besides the one that we have identified here , as these may produce spectral features in the intrinsic phase . one possibility is the coupling between bound @xmath96-@xmath66 excitons and the unbound @xmath94-@xmath66 or @xmath95-@xmath66 excitons , since it is known that the intrinsic phase can show spectral features near a resonance @xcite . another possibility is the envelope - hole coupling between the continua of unbound @xmath94-@xmath66 and @xmath95-@xmath66 excitons that was neglected in our treatment . the effect of coupled continua in the general theory of ` @xmath3 ' phase shifts has not been considered to date . finally , we note that the intrinsic phase and coulomb enhancement may be greater in reduced dimensional systems , which have greater exciton binding energies . the carrier - carrier coulomb interaction was included in the theory for ` @xmath7 ' control of electrons in biased asymmetric quantum wells , although the intrinsic phase was not studied @xcite . consider the functions @xmath252 and @xmath253 , which appear in ( [ n_af ] ) and ( [ n_aa ] ) ; we refer to them collectively as @xmath254 . first , note that due to the energy conserving delta function in ( [ etaschematic ] ) , @xmath255 will be equal to @xmath256 [ see eq . ( [ energy conservation ] ) ] , and thus @xmath254 is a function only of @xmath257 , @xmath53 , @xmath258 , @xmath259 , and @xmath260 . second , note that @xmath254 is defined so that if the electron - hole attraction is turned off , for example by letting @xmath261 , then @xmath262 @xcite . this allows @xmath254 to be identified as part of the coulomb enhancement . in particular , @xmath254 is the enhancement due to the coulomb interaction in the intermediate states ; if the coulomb interaction is neglected for the intermediate states , @xmath263 @xcite . [ note that lee and fan @xcite did not allow for @xmath264 in @xmath254 ( related to @xmath265 in their notation ) . ] since the integrand is smooth for the parameter range of interest , numerical integration of @xmath254 is straightforward ; however , it need not be undertaken . further simplification is possible since the parameter @xmath266 can be considered to be much less than one . since most materials have an exciton binding energy that is much smaller than the band gap , @xmath251 is detuned from the band edge by many exciton binding energies at photon energies consistent with the approximations made here . in gaas , for example , when @xmath267 is within 500mev of the gap , @xmath266 is at most @xmath268 . an expansion of @xmath252 for small @xmath266 , @xmath269 where @xmath270 , shows that @xmath252 is approximately @xmath271 and nearly constant as a function of @xmath257 . the same is true of @xmath253 , which has the expansion @xmath272 where , with @xmath273 , @xmath274 in fact , when @xmath275 , @xmath276 even to fourth order in @xmath277 . fig.[fig_n ] shows a numerical integration of @xmath252 using the parameters of gaas . [ intermediate state coulomb enhancement , see ( [ two photon a - f ] ) ] for gaas . the first ( second ) subscript is @xmath65 ( @xmath278 ) ; subscripts for @xmath279 are not shown ; @xmath280 denotes the heavy and light hole bands , and @xmath281 denotes the split off band.,width=307 ] the tensor @xmath282 , defined in ( [ eta_ccvv ] ) , can be used to calculate the current injection tensor with or without excitonic effects using ( [ eta_result ] ) or ( [ eta_result_free ] ) . it can be evaluated analytically in the approximation of parabolic bands . part of the result for the 8 band kane model has been given before but without the split - off band as an initial or intermediate state @xcite . we here give more detail , but only for the two - band terms . we denote the bands by a double index @xmath4 and @xmath127 , where @xmath4 is one of @xmath283 , and @xmath127 runs over the two spin states for each band . since the coulomb corrections to @xmath16 in ( [ eta_result ] ) do not depend on the spin index , we can include the sum over spin indices from ( [ eta_result ] ) in @xmath284 . and , since we are only calculating two - band terms , we set @xmath285 and @xmath286 . thus @xmath287 where @xmath288 for @xmath289 or @xmath94 , @xmath290 , while for @xmath291 , @xmath292 , where @xmath176 is the fundamental band gap , and @xmath293 is the spin - orbit splitting . the interband velocity matrix elements are approximated by their value at @xmath294 , but still depend on the direction of @xmath0 . in terms of the orthogonal triple of unit vectors @xmath295 , @xmath296 and @xmath297 , @xmath298 these matrix elements are : @xmath299 _ { s , s^{\prime } } \\ \mathbf{v}_{c , s;lh , s^{\prime } } & = & -\frac{1}{2}\sqrt{\frac{e_{p}}{3 m } } \left [ 2\mathbf{\hat{k}}\mathbb{i}+i\mathbf{\hat{l}}\sigma _ { y}-i\mathbf{\hat{m } } \sigma _ { x}\right ] _ { s , s^{\prime } } \\ \mathbf{v}_{c , s;so , s^{\prime } } & = & \sqrt{\frac{e_{p}}{6m}}\left [ \mathbf{\hat{k}}\mathbb{i}-i\mathbf{\hat{l } } \sigma_{y}+i\mathbf{\hat{m}}\sigma_{x}\right]_{s , s^{\prime } } , \end{aligned}\ ] ] where @xmath300 is the kane energy @xcite . here , @xmath301 is the @xmath302 identity matrix and @xmath303 are the pauli spin matrices . of course , for parabolic bands , the intraband matrix elements are @xmath304 , where @xmath305 is the effective mass of band @xmath4 . ( in the proper kane model , the effective masses are given in terms of the parameters @xmath176 , @xmath293 , and @xmath300 , but we treat them as additional parameters , which is equivalent to including remote band effects on the effective masses . ) the sums over spin then yield @xmath306 the remaining angular integrals can be done using @xmath307 the result for @xmath308 is @xmath309 , \label{etafree parabolic}\ ] ] where the tensor properties are @xmath310 and the term involving @xmath311 should not be included if @xmath312 . the free particle current injection tensor has also been investigated for parabolic bands , but with a simple three - band model @xcite . that model does not have the matrix elements @xmath313 and @xmath314 , and thus differs from our result for @xmath315 . this work was financially supported by the natural science and engineering research council , photonics research ontario , and the us defense advanced research projects agency . we gratefully acknowledge many stimulating discussions with daniel ct , james fraser , ali najmaie , fred nastos , eugene sherman , art smirl , marty stevens , and henry van driel .

low energy nucleon - antinucleon annihilation is a fertile area for studying hadron dynamics and particularly pionization . experiments give information on pion numbers , spectra , and correlations and seem to show little dependence on initial energy from annihilation at rest to kinetic energies of a few hundred mev . results do depend on the iso - spin and spin quantum numbers of the annihilating pair . understanding these results is a serious challenge to theory . in this paper we develop a description of nucleon - antinucleon annihilation into pions based on the creation of a coherent classical pion wave in the annihilation process and the quantization of this wave using the method of coherent states . about 40% of annihilations go through meson resonances , mostly rho and omega mesons , and then to pions @xcite . these channels do not lead to prompt pions and are not part of our treatment . our first studies show our picture accounts for the major phenomenological features of the pions seen in low energy annihilation . our picture divides the process into two steps . the first is the dynamics of the nucleon and antinucleon . we have little to say about this except to assume that once these two touch " and annihilation begins it proceeds very rapidly . this rapid annihilation leads to a pion pulse or pion wave that forms the basis of our coherent state . it is the description and consequences of that state and the way in which its features reflect the spatial and temporal evolution of the annihilation region as well as its quantum numbers that is our principal concern @xcite . our description of annihilation as occurring by the rapid radiation of pions , draws its inspiration from recent studies of annihilation in the skyrme model . et al _ @xcite have studied numerically the collision of two skyrmions , one of baryon number one and the other of baryon number minus one . they find that once there is significant overlap of the two , they disappear into pion radiation at nearly the causal limit . et al _ @xcite studied this rapid decay in a schematic model . they considered an initial well localized blob " of skyrmionic matter with baryon number zero and total energy about twice the single skyrmion energy , and studied its time evolution . they found that it decays into a pion wave nearly at the causal limit . that pion wave is a propagating , but well localized pulse . this picture of the annihilation process in terms of a rapid coherent radiation pulse is as far from the thermodynamic one as a broadcasting radio antenna is from a black body . the skyrmionic description is classical , and is meant as purely suggestive . does a classical starting point make sense for annihilation ? let us assume that the rapid conversion to pion radiation seen in the skyrme treatments is a feature of the real annihilation . then the pions are radiated from this bright flash " as a coherent pulse , or coherent pion quantum wave . since the mean number of pions radiated is not small , and the total energy to be radiated is large compared with the pion mass , @xmath0 , we can envisage this wave to be , in first approximation , a * classical coherent wave*. the remaining problem is then to quantize this wave . this we do with the well known method of coherent states . our picture , then , has two central assumptions . first we assume that rapid annihilation leads to a classical pion wave pulse . this is a consequence of the rapid time scale of the annihilation and the large energy ( compared to the pion mass ) released . second we assume that this classical wave can be quantized using coherent states . this is basically the first term in a description with expansion parameter @xmath1 , where @xmath2 is the mean number of pions . the error in our second assumption can therefore be estimated . it should be noted that our starting point arises in classical qcd , which is the non - perturbative domain of the theory . this seems to us the correct starting point for the study of annihilation . it may be that other strongly interacting systems are also best approached from this direction . it has recently been shown that that is true for the nuclear force @xcite . in such treatments it is the reintroduction of quantum aspects that is the problem . for radiation fields , coherent states are constructed to solve that problem . in the next section , we develop the coherent state formalism for the pions . we begin with a description of classical pion waves from a source , introduce coherent states , discuss the constraints of iso - spin and of energy and momentum conservation and give some statistical features of the coherent state . in section 3 we briefly investigate the consequences of our description for the annihilation data . we only touch on what could be done , but we do see that our approach generally agrees with experiment . finally section 4 discusses some directions for further work and presents a brief summary . we imagine nucleon antinucleon annihilation as occurring very rapidly from a small region of high energy density . in first approximation , we take this region to radiate a classical coherent pion wave . if we further neglect @xmath3 interactions and the back reaction of the radiated pions on the source , we can write the classical wave equation for the pion wave @xmath4 , as @xmath5 here @xmath6 is the source of the pion field , @xmath7 . this source will be assumed to be sharply localized in space and time . for the moment we neglect the isospin degree of freedom and will return to it later . we also leave out the factor of @xmath8 in front of the source that is familiar from electrodynamics , since it adds nothing here . eqn ( 1 ) gives the form of the pion wave as a classical wave radiated by the source @xmath9 . the details of the annihilation mechanism are contained in this source function . the equation for @xmath7 is easily solved in terms of @xmath9 by taking fourier transforms . we define the fourier transform of @xmath7 by @xmath10 and a corresponding function , @xmath11 as the fourier transform of @xmath6 . the pion field is then given by @xmath12 the pion radiation field comes entirely from the poles of the denominator in the integral . we must specify how those poles are to be treated in order to get causal radiation . doing so and using the pole contributions only to evaluate the @xmath13 integral in ( 3 ) , we obtain the approximate expression for the pion field , for @xmath14 , @xmath15 where @xmath16 . this form for the pion field is exact in the radiation region , but neglects non - radiating parts of the field coming from the @xmath13 dependence of @xmath17 . in the radiation zone there is no source and @xmath7 given in ( 4 ) satisfies the source free wave equation . it is this form for @xmath7 that we will take to represent the radiated pions of nucleon antinucleon annihilation . it is not the classical pion wave that is detected in annihilation experiments but rather the quanta of that field , the pions . we must therefore extract the quantum content of this classical wave . this is exactly what the coherent state method achieves . pioneered by glauber @xcite for optics in situations where the photon granularity is significant but not overwhelming , the coherent state formalism allows one to construct a quantum state that corresponds , in the large number of field quanta limit , to the classical wave field . coherent states for pions were studied by horn and silver @xcite , and many of the arguments given here were considered by them . they did not , however , study annihilation . because the pion field satisfies the wave equation without sources , it can be decomposed into a linear superposition of free waves with wave number @xmath18 and corresponding energy @xmath19 . that is the content of eqn ( 4 ) . we can bring in quantum mechanics for this wave by introducing the standard creation and annihilation operators for the modes @xmath18 , @xmath20 and @xmath21 . these satisfy the usual harmonic oscillator commutation relations , @xmath22 = \delta({\bf k}-{\bf k'}),\ ] ] with all other commutators zero . the quantum field operator corresponding to the pion field is then @xmath23 the coherent state associated with a given classical state , @xmath24 , is the quantum state that is an eigenstate of the positive frequency part of @xmath25 , @xmath26 , with the corresponding positive energy part of the classical wave as its eigenvalue . recall that the positive frequency part of the state has the @xmath27 dependence . in the quantum operator for the field , this is the part that goes with the annihilation operator . since the modes for each @xmath18 are independent , finding the eigenstate is equivalent to finding the eigenstate of one single annihilation operator . the full state can then be built up by superposing the independent modes . the normalized eigenstate , @xmath28 , with eigenvalue , @xmath29 , of a single mode annihilation operator , @xmath30 , ( @xmath31 ) , is given by @xmath32 where @xmath33 is the vacuum state . this equation for the eigenstate of @xmath30 is the key to the entire coherent state formalism . consider now the normalized quantum state @xmath34 defined by @xmath35 this is just a coherent state in which each mode @xmath18 carries weight @xmath36 . it is clear , for the positive frequency part of the field , that @xmath37 with the eigenfunction @xmath38 given by @xmath39 this @xmath38 satisfies the source free wave equation for any function @xmath40 . we then take @xmath34 to be the normalized quantum state corresponding to the classical wave @xmath38 . note that @xmath34 does not contain a fixed number of quanta , but is rather a coherent superposition of states of different numbers of quanta . the meaning of a coherent state arising from a classical source can be clarified by relating it to the standard s - matrix description of a field generated by a classical external source . the classical source , @xmath9 , coupled to the field , @xmath7 , creates from the vacuum the quantum state @xcite @xmath41 upon four dimensional fourier transformation and integration over @xmath42 , this reduces , up to a normalization , to eqn ( 8) . the exponent in ( 11 ) is just the s - matrix in the interaction representation , and makes clear both the connection of the s - matrix and coherent state approaches and why the resulting state has an indefinite number of quanta . in the electromagnetic case it seems natural to define states with indefinite numbers of photons , but even there if the wave has a fixed total energy only a fixed number of photons above some frequency can be detected . with pions the description in terms of a coherent mixture of states with different numbers of pions seems less comfortable , particularly in view of the finite pion mass . no one would suggest representing a pion wave with total energy near the pion rest mass as a classical object . in annihilation , however , the total energy released is about 13 times the pion rest mass , and the approximation of representing it by a classical coherent wave seems at least plausible . the purpose of this paper is to show that that description is not only plausible but useful and gives dynamical insight into the annihilation process and may have application in a number of other processes in which many pions are created . to make the pion coherent state correspond to the radiating solution @xmath24 of eqn ( 4 ) , that is to make @xmath38 in ( 9 ) or ( 10 ) correspond to the positive frequency part of @xmath7 , we must take in ( 8) @xmath43 with this choice , our quantum coherent state represents the classical radiating pion wave generated by the source @xmath6 . the pion is an iso - vector and comes in three charge states . in principle each charge state can have its independent source , corresponding classical wave , and coherent state . this is most easily formulated by making @xmath24 and @xmath6 of eqn ( 1 ) each iso - vectors . then each charge state could have a source density that varies arbitrarily in space and time . in fact nucleon antinucleon annihilation occurs in a state of fixed total iso - spin , 0 or 1 . the full pion wave must have the same total iso - spin . this condition of total iso - spin conservation imposes a * global * condition on the source and on the coherent state . since the different @xmath18 modes of the wave are independent , they must each combine to the correct total iso - spin , which means that the iso - spin source density can not depend on @xmath44 or @xmath45 . in the coherent state @xmath34 of ( 8) we expect in general that @xmath46 is an iso - vector dotted into the now iso - vector creation operator . but if we are to construct states of fixed iso - spin for any @xmath18 configuration , the iso - vector dependence of @xmath40 must be independent of @xmath18 . in that case @xmath40 is a constant vector in iso - spin space . the direction of that iso - vector depends on the source . for the moment let us consider a coherent state specified by a fixed unit vector in iso - spin space @xmath47 and an iso - scalar function @xmath46 . call that state @xmath48 . in place of ( 8) we write @xmath49 that is a coherent state constructed from a particular superposition of pions , that superposition made up of pions all pointing in the iso - direction of @xmath47 . this is not yet a state of fixed iso - spin . in fact it is easy to see that it contains states of all iso - spin . a state of fixed iso - spin must be projected from it . call the projected state of total iso - spin @xmath50 with z - component @xmath51 , @xmath52 . the projection is given by @xcite @xmath53 where @xmath54 is the usual spherical harmonic , and @xmath55 is a normalization constant . the state defined in ( 14 ) is orthogonal to states of different total iso - spin or z - component projection . to see this and to calculate the normalization let us study the inner product of two projected states . this is given by @xmath56 it is easy to show that @xmath57 for a strong field , large @xmath40 , this inner product is very sharply peaked around @xmath58 . we will show in the next section that @xmath59 where @xmath2 is the mean number of pions emitted in the annihilation . we will use this notation here both to save writing and to remind us that this integral , or @xmath2 , is large . to show the orthogonality in ( 16 ) write @xmath60 from this it is easy to show that @xmath61 from which the normalization can be read off . in particular for large @xmath2 , we can use the asymptotic form of the bessel function to yield for the normalization @xmath62 we should note that our iso - spin projection has the interesting property that states of even iso - spin contain only even numbers of pions and states of odd iso - spin only odd numbers . to escape this restriction , as the data does , we need to include the channels for annihilation into meson resonances . we would now like to examine the properties of the coherent pion state generated by the assumed initial source @xmath6 . we begin with a discussion of the case without iso - spin . this coherent state is given by ( 8) with @xmath46 given in terms of the fourier transform of @xmath9 as in ( 12 ) . although the state does not contain a fixed number of pions , we can ask what is the average number , @xmath2 . this is just the expectation value of the total pion number operator in the coherent state . we need @xmath63 ( note that we are working in the heisenberg representation so that the states are time independent while the operators in general carry a time dependence . however , the number operator is time independent . ) this expectation value is easily evaluated using the special properties of the coherent states to give ( 17 ) . thus we see that the integral that enters to normalize the coherent state in ( 8) , is just @xmath2 . it is large @xmath2 that corresponds to large field ( @xmath40 or @xmath9 large ) and thus to a good approximate correspondence between the classical and quantum descriptions . it is also clear from this that the average single pion momentum distribution is given by @xmath64 in terms of the square of the fourier transform of the source density . in the same way one obtains for the total energy released , the expectation of the hamiltonian , or @xmath65 in a quantum process , this energy is sharp and should be put equal to the total energy released in the annihilation . this fact serves as a convenient way to normalize @xmath40 . the sharpness of the state can be seen by calculating the dispersion in the number of pions emitted . one finds @xmath66 which shows that the fractional dispersion in @xmath67 is @xmath68 . this result also follows from the fact that the probability of finding a state of @xmath69 pions is given by a poisson distribution , as we shall show below . one can use the coherent state to calculate the amplitude for finding some particular quantum configuration . for example the amplitude for finding @xmath69 pions of momenta @xmath70 is given by @xmath71 the probability of finding @xmath69 pions of any momentum is then @xmath72 which is the poisson distribution . it is clear that the coherent state does not know " not to emit more pions than energy conservation permits . we will turn to these constraints of energy and momentum conservation below . our discussion here serves to emphasize that the coherent state approach is better at giving average information over the ensemble of annihilations than it is for particular quantum states . if we include iso - spin projection as in section 2.3 , we can also probe questions about iso - spin populations . for example let us study the average number of pions of a particular type in a projected coherent state with total iso - spin @xmath50 and z - component @xmath51 . this is given by @xmath73 where @xmath0 is the pion type ( @xmath74 ) and there is no sum over @xmath0 implied . using the forms of the projected states , this becomes @xmath75 using ( 18 ) and expressing the @xmath47 in terms of spherical harmonics , the integrals can be calculated in terms of standard identities . we continue to use ( 17 ) to express the integral over @xmath40 in terms of @xmath2 . we now use ( 17 ) to define @xmath2 but as we shall see it is still the average number of pions summed over all pion types . we then find @xmath76 in terms of a sum over clebsch - gordan coefficients . in obtaining ( 29 ) we have used the asymptotic forms of the bessel functions ( large @xmath2 ) so that there is no @xmath77 dependence in the normalization . for nucleon - antinucleon annihilation there are only two choices for @xmath50 , 0 or 1 . for the case of @xmath78 ( 29 ) reduces to @xmath79 for any @xmath0 . as we expect for a state of iso - spin zero the average number of pions is type independent , and each is one third of @xmath2 , making @xmath2 the true average number of pions summed over types , as promised . for @xmath80 there are three cases . for @xmath81 one finds @xmath82 we see that again the sum of averages over types is @xmath2 . note the dominance of @xmath83 , an important feature of the data . for @xmath84 one finds @xmath85 with the same result for @xmath86 . it is no surprise that for the @xmath78 case or for the @xmath80 , @xmath81 case the average number of @xmath87 and @xmath88 is the same . it is a bit more puzzling for @xmath80 , @xmath89 . what happened to charge conservation ? the initial @xmath89 state has charge @xmath90 . however , @xmath91 is a number of order @xmath2 . to that order all three @xmath51 values do correspond to states of average charge zero . charge conservation is a @xmath1 effect , and we are neglecting terms of that order . we will return to this approximation below . the same charge ratios can be obtained in ( 28 ) by making the overlap of ( 16 ) proportional to @xmath92 . we have already commented that this overlap is strongly peaked around @xmath58 for large @xmath2 . this delta function approximation is correct in the large @xmath2 limit and it makes calculations of the average pion number by type and of higher moments of the distribution very simple . we illustrate this delta function method in the calculation of isospin number correlations . for example the joint average of @xmath0 type and @xmath55 type pions in a state of total iso - spin @xmath50 with z - component @xmath51 is @xmath93 in the large @xmath2 limit where the overlap may be replaced by a delta function , we obtain , @xmath94 in this expression the components of @xmath47 are , in the usual spherical coordinates , @xmath95 higher order forms are obtained from the generalization of ( 33 ) and ( 34 ) . they come from higher powers of the number operator for the @xmath96th type in the expectation value with the coherent state replacement @xmath97 the integrals that appear in ( 34 ) are elementary , and hence these higher moments are easily calculated . greiner , gong and muller @xcite have recently introduced polynomial coherent states in a discussion of pion condensates in heavy ion collisions . they consider only @xmath78 states and calculate the iso - spin pair correlations defined as @xmath98 we obtain the same values as they do for @xmath99 for large @xmath2 , but our method is simpler . we can also use the method to calculate the pair correlations in other states . for example we find for @xmath80 , @xmath81 , @xmath100 and for @xmath80 , @xmath84 , @xmath101 . other moments and other amplitudes are also easily calculated and agree with those of greiner , gong and muller where calculated by them . thus we see that the coherent state approach makes definite predictions about average charge ratios in annihilations and can also be used to calculate higher moments of those distributions . it also predicts that the single pion momentum distribution should be independent of pion type . so far the coherent state we have defined does not have definite total energy or momentum . for example it contains contributions from states with only one pion or with arbitrary numbers of pions , neither of which is allowed by conservation of total energy and momentum . this fault is not serious for large @xmath2 , but is significant for the intermediate @xmath2 of annihilation . energy and momentum conservation can be imposed on the coherent state , as was shown by horn and silver @xcite , at the expense of some further complication of the formalism . let us begin by imposing energy and momentum conservation on the coherent state without iso - spin . the method for imposing energy and momentum on the coherent state is formally the same as the projection method for imposing fixed iso - spin . the projections commute , but taking them together complicates the physics . for iso - spin one first constructs a state with all the pions pointing in the same direction in iso - spin and then averages over that direction with the appropriate @xmath102 eigenfunction weight . to construct a state of fixed energy and momentum ( or four - momentum ) one first constructs a state of all the pions at a fixed place in space and time and averages over those places and times with the appropriate eigenfunctions of definite energy and momentum . to do this define the operator @xmath103 by @xmath104 then the ( un - normalized ) coherent state at @xmath44 and @xmath45 is @xmath105 it is more concise as well as more physical to use four dimensional notation . we call @xmath106 the four - vector of position with space part @xmath44 and time part @xmath45 . thus @xmath107 becomes @xmath108 . then the state of fixed total four - momentum , @xmath109 is given by @xmath110 to see how this works and to calculate the normalization , consider the overlap of two states of different four - momentum . one easily finds @xmath111 where @xmath112 and where the fourth component of @xmath113 under the integral is @xmath114 . a simple change of variable reduces the overlap to @xmath115 this clearly shows that states of different four - momentum are orthogonal , but the remaining integral is singular due to the undamped large @xmath106 behavior of the integrand . horn and silver @xcite also pointed out this problem . it arises from the first term in the expansion of the exponential . this term , a one " , corresponds to the no pion contribution to the annihilation process . since there is in fact no such contribution that conserves energy and momentum , we can subtract it out . similarly there is no one pion state that can contribute to annihilation and conserve four - momentum , thus we may also subtract out the one meson state and define a coherent state of total four - momentum @xmath109 that starts with two pions . this is given in terms of @xmath116 by @xmath117 we then find @xmath118 the integral now is convergent and we have lost nothing by the subtractions since we have only removed states that can not contribute physically . however the integral in ( 46 ) is still quite delicate and difficult to evaluate numerically because it contains delta functions coming from the @xmath119 integral . we can get around this problem and shed more physical light on the expression by expanding the exponential and interchanging the order of integration . let us call the integral in ( 46 ) @xmath120 . we can write @xmath121 using the definition of @xmath96 and interchanging integrations one finds @xmath122 to be @xmath123 . integrals of this form are easily done by monte carlo methods following a program given by barger and phillips . the individual terms in the @xmath69 sum in ( 47 ) represent the relative contributions to annihilation into @xmath69 pions . for fixed total energy the sum must terminate . for example , nucleon - antinucleon annihilation at rest can not go into more than 13 pions , and in that case @xmath124 must be zero for @xmath125 . in fact @xmath126 is just the fraction of annihilations that go into @xmath69 pions . thus the mean pion number is given by @xmath127 this is equal to the expression that one arrives at by taking the expectation value of the number operator in the state @xmath128 and normalizing . this expression can also be written @xmath129 note that @xmath130 in terms of the old expression for @xmath2 , ( 17 ) . for @xmath2 large , @xmath131 is large , and the integrands in ( 50 ) are dominated by the factors of @xmath132 and @xmath133 . the expression ( 50 ) can then be thought of in terms of a ratio of improper integrals @xmath134 these integrals can be approximated by the method of stationary phase . for annihilation at rest , the total energy released , @xmath135 , is large , and then the stationary phase point comes at @xmath136 . thus the @xmath131 in the numerator of ( 51 ) , but not in the exponent , can be evaluated at @xmath137 to give that the new @xmath2 is approximately equal to the old one . corrections are of order @xmath1 . we have verified this in a number of numerical examples . we see that four - momentum conservation can indeed be implemented for the coherent state , but for average quantities , and for @xmath2 large , that constraint is not important . as we will see below , it is important for the probability distribution and for higher moments of the distribution . it is clear that one can combine the method outlined above for projecting onto states of good four - momentum with the method of the previous section for projecting onto states of good iso - spin . we now turn to a discussion of how to do that . combining the methods of previous sections , we can construct a state of fixed total four - momentum @xmath109 and fixed iso - spin @xmath50 with z - component @xmath51 . we first generalize @xmath138 of ( 38 ) to give @xmath139 then the appropriate generalization of ( 44 ) to include iso - spin is @xmath140 giving the state ( not normalized ) @xmath141 the inner product of two such states belonging to different four - momentum and iso - spin gives a four dimensional delta function on four - momentum and kronecker deltas on iso - spin and its z - component times the following normalization integral @xmath142 this integral can be calculated by the same expansion methods used in ( 46 ) and ( 47 ) . in terms of the @xmath143 defined in ( 47 ) the normalization integral of ( [ eq : normal ] ) is given by @xmath144 where @xmath145 this integral can be done using the identity in ( 18 ) by differentiating ( 18 ) with respect to @xmath2 @xmath69 times and then setting @xmath2 to zero . one finds @xmath146 it should be recalled that sums over @xmath69 as in ( [ eq : fmsum ] ) have finite upper limits because of the constraint of energy conservation . the expected number of pions of type @xmath0 corresponding to ( 27 ) is now @xmath147 the joint average of @xmath0 type and @xmath55 type pions as in ( 33 ) is given by @xmath148 in obtaining ( [ eq : corr ] ) , we have neglected a term from the commutator that is down by @xmath1 . in these expressions , @xmath96 is still given as in ( 42 ) . these expectation values can be evaluated , as before , by expanding . once again they are expressed in terms of the @xmath124 . we have @xmath149 and @xmath150 where @xmath151 and @xmath152 let us examine these forms in the cases of physical interest . for @xmath78 , it is easy to show that @xmath153 then from ( 56 ) and ( 61 ) we get @xmath154 as we expect . for @xmath80 , @xmath81 , we find @xmath155 these do not seem to give the same charge ratios we obtained before in this state , but those were calculated in the large @xmath2 limit . in that limit the dominant terms in the sums of ( 56 ) and ( 61 ) come from terms with @xmath156 . thus the dominant terms in @xmath157 that matter have @xmath158 ( recall that @xmath77 is of order @xmath50 which is in turn of order @xmath159 ) . in this limit it is easy to show that @xmath160 with corrections of order @xmath161 . in that limit ( 67 ) yields the results of ( 31 ) . similarly for @xmath80 , @xmath84 , we find @xmath162 which yields ( 32 ) in the large @xmath2 limit , and shows how charge conservation enters for finite @xmath2 . one can calculate the two pion averages in the same way . the factors that enter become for @xmath80 , @xmath81 @xmath163 and for @xmath80 , @xmath84 @xmath164 which yield our previous results for the iso - spin pair correlations @xmath99 defined in sect . 2.5 in the large @xmath2 limit . the evaluation of the averages now depend on the functional form of @xmath46 . we turn to it in the next section . as a first orientation into the phenomenological content of the coherent pion wave description of annihilation , let us study the relation between the average pion number @xmath2 of ( 17 ) and the energy released , ( 23 ) , for a very simple model of @xmath40 . we do not take @xmath40 directly from the experimental pion spectrum since the pions from the meson resonance channels confuse the extraction . rather suppose we assume that the pion field source turns on at @xmath165 and then decays exponentially in time , and that it has a spherically symmetric yukawa shape . ( spherical symmetry is appropriate for annihilation at rest . ) that is we take for @xmath6 of ( 1 ) , @xmath166 where @xmath167 is the source strength and where we have put in a factor of @xmath45 to make the time dependence continuous at @xmath165 . using ( 12 ) this leads to @xmath168 where @xmath169 is a strength and where we have multiplied @xmath40 by @xmath170 to model the p - wave nature of pion emission . with this form , the integrals for the mean number of pions and for the mean energy ( 17 ) and ( 23 ) can not both be done analytically , but they can be easily evaluated numerically . this requires a choice of the parameters @xmath171 , @xmath172 and @xmath169 . the last can be fixed by requiring that the average energy be the energy released in annihilation . in units of the pion mass ( @xmath173 ) , and for annihilation at rest , this energy is 13.87 . for the range parameters , we take @xmath174 . this corresponds to an annihilation region with a time and distance scale of half a pion compton wave length - a reasonable size . it is not the purpose of this paper to make a careful study of the source function and to fit it to data . rather we simply want to demonstrate that a quite reasonable choice for functional form and size yields a correspondingly reasonable account of the data . thus we have not made an exhaustive parameter search . with our choice for parameters we find that ( 17 ) gives for the average pion number @xmath175 . with the same choice of parameters , the average pion number calculated with four - momentum conservation imposed as in ( 50 ) gives @xmath176 . this is certainly within @xmath1 of the unconstrained value , as promised . also @xmath2 about 6 agrees with the data , for reviews see , @xcite , particularly when it is realized that we are only modeling that part of annihilation that goes into uncorrelated pions , and not the part that goes into other mesons that subsequently decay into pions . since these typically give few pions , the average number without them will be somewhat higher than the total average number . although the average number from the poisson distribution and the distribution constrained by four - momentum conservation come out very close , the actual probability distributions in pion number are rather different . in fig . 1 we show the probability of finding @xmath69 pions as a function of @xmath69 for the poisson and for the m - sum " cases ( as defined in ( 47 ) and ( 48 ) ) . we see that the constrained case has a far narrower distribution than the poisson . for the poisson distribution the @xmath177 ( @xmath178 is defined in ( 24 ) ) is @xmath179 . for the distribution constrained by four - momentum conservation we find @xmath180 , which is quite close to the value deduced @xcite from experiment , @xmath181 . this value has a statistical error of at least five percent and an additional unknown error from modeling the distribution of neutrals . in fact the entire pion multiplicity distribution from the constrained calculation shown in fig . 1 is quite close to the experimental distribution @xcite , @xcite . our constrained distribution is also indistinguishable ( on the scale of the figure ) from a gaussian distribution with the same average number and same variance . that is what a statistical model would give . one can also use the combination of iso - spin projection and four momentum projection of sect . 2.7 to calculate the pion number distributions . they are shown in fig . 2a and 2b for @xmath78 and @xmath80 , @xmath81 respectively . recall that for @xmath78 only an even number of pions is possible while for @xmath80 there is only an odd number . the data is an appropriate average over these two , and again strongly resembles our figures @xcite , @xcite . annihilation into vector mesons must be added to evade this odd - even effect . let us now turn to a comparison of our picture with experiment for the average number of pions by iso - spin type . we compare with the ratios reported in sedlak and simak . to do this we need to estimate the relative population of @xmath78 and @xmath80 in proton antiproton annihilation at rest . we do so following the work of locher and zou @xcite . then using the charge ratios in states of good iso - spin we found in sect . 2.5 , and for vanishing total third component of iso - spin , we find @xmath182 with an error of about ten percent to be compared with the experimental number reported as @xmath183 @xcite , . this is remarkable agreement . note that these ratios do not depend at all on the details of our parameters , but do depend on the structure of the coherent state picture . in particular , it should be noted that the excess of @xmath83 s found experimentally comes out naturally in our treatment . given @xmath184 , we can calculate charge averages including four - momentum conservation as in sect . using the same parameters , we can find the average pion number by charge type and the variance in the various iso - spin states . these are shown in table 1 . the @xmath80 , @xmath84 case shows the effect of charge conservation , for @xmath84 , the average number of @xmath87 is one larger than the average number of @xmath88 . we also see that with four - momentum conservation the ratio of @xmath83 to @xmath87 in the @xmath80 , @xmath81 case is somewhat larger than in the unrestricted case . to make our formalism a real theory of the annihilation process we need to model annihilation in space and time as a classical source of pions . we have presented a very simple example as an orientation . more sophisticated models of the pion source could be made based on a dynamical theory of the annihilation process , for example the skyrme model calculation @xcite , or on nucleon - antinucleon potential model calculations . these could include the spatial asymmetry seen in these annihilation calculations @xcite . this asymmetry would lead to a different momentum distribution for the pions . alternately one could use ( 22 ) to extract the form of @xmath46 from the single pion momentum distribution data . we must also model the substantial part of annihilation that goes through meson resonances . we plan to return to these ideas in later work . given a form for @xmath46 , one can calculate not just single pion averages , but also correlations @xcite , @xcite and joint momentum distributions for multiple pions . the great advantage of the pion coherent state approach is that it makes detailed , specific and rather restrictive predictions for these observables and thus is easily verifiable , or perhaps falsifiable . clearly much work remains to be done to develop and test the coherent state approach developed here for pions radiated from nucleon antinucleon annihilation . predictions for higher moments of pion distributions from nucleon antinucleon states of sharp iso - spin have to be confronted with data . correlations have to be studied . theories of nucleon antinucleon interaction and of the development of the annihilation region have to be studied and interpreted in terms of a source of the pion radiation . it may be possible , for example , to model annihilation classically in the skyrme model including the non - linear interactions of the pions , and the effects of the annihilation back on the source . such a description would lead , far enough away from the annihilation region , to a free , non - interacting pion wave . this wave could then be the source of the coherent state . this approach is far simpler than a full quantum theory of the annihilation . beyond this study of the annihilation process , we must study the effect of other quantum numbers such as g - parity and angular momentum on annihilation in the coherent state description . finally a better understanding of the relation between annihilation into pion radiation and into other mesons is needed . we are exploring all these questions . should these detailed descriptions fail , it would be interesting to go beyond the coherent state picture by including squeezed states . one can also imagine other applications of the general methods advocated here . the basic idea is to look for any fast coherent processes in which the total energy radiated is large compared with the mass of the radiated quanta and in which the interaction among those quanta or by the quanta back on the source can be neglected . then the radiation process is described classically , and the subsequent wave quantized using coherent states . conserved quantum numbers have to be imposed on these states as we have imposed iso - spin . one can then explore averages , moments , momentum distributions and their relation to the radiating source , all as we did for annihilation . greiner , gong and muller @xcite have discussed one such possible application in heavy ion collisions , there are many more . there may be similar opportunities in the description of hadronization in qcd jets @xcite . other examples will occur to the reader . it may even be possible to lift the restriction of no self interaction of the field or of the field back on the source so long as one can treat these classically and then use coherent states to quantize only in the radiation zone . many qcd processes , like annihilation , may best be treated by starting in the classical non - perturbative domain of qcd . it may also be possible to extend these ideas to heavy ion collisions in general @xcite , where one could take into account the statistical nature of the process by introducing a density matrix and averaging over an appropriate ensemble of coherent states . such methods are well known in quantum optics @xcite , and may be an attractive alternative to the thermodynamic approach to heavy ion collisions . in summary we have seen that the large energy released and the possibility of very short reaction time suggest that a classical , coherent pion wave pulse is radiated in nucleon antinucleon annihilation . this wave can be quantized using the method of coherent states . from this description averages and higher moments of pion distributions in annihilation can be calculated . in particular by imposing iso - spin and four - momentum conservation , interesting correlations among pion charge types are easily calculated . momentum distributions contain information on the spatial and temporal distribution of the pion wave source . preliminary calculations show that all these features obtained from the coherent state picture agree with the data and therefore suggest that the coherent classical pion pulse picture of annihilation is valid for that part of annihilation that goes directly into pions . we also speculate on other applications of classical meson waves and their subsequent quantization . rda , fc , and jpd would like to thank the theory group of the division of nuclear and particle physics at the paul scherrer institute and for providing a very pleasant environment in which much of this work was done . rda and bs are supported in part by the united states national science foundation . the division de physique thorique is a research unit of the universities paris 11 and 6 associated to cnrs . rda and bs would like to thank dr . r. hollebeek for bringing reference to our attention . hg c. amsler and f. myhrer , ann . nucl . part . sci . * 41 * , 219 ( 1991 ) . + a brief account of this work has been published in r.d . amado , f. cannata , j - p . dedonder , m.p . locher and b. shao , lett . * 72 * , 970 ( 1994 ) sommermann , r. seki , s. larson and s.e . koonin , phys.rev . * d45 * , 4303 ( 1992 ) . + b.shao , n.r.walet , and r.d.amado , phys.lett . * b 303*,1 ( 1993 ) . walet and r.d . amado , phys.rev . * c47 * , 498 ( 1993 ) . + r.j . glauber , phys . rev . * 131 * , 2767 ( 1963 ) + d. horn and r. silver , annals of physics * 66 * , 509 ( 1971 ) . ginocchio and m.w . kirson , phys . letters * 44 * , 1744 ( 1980 ) + c. itzykson , j.b . zuber , quantum field theory , ( mcgraw - hill , singapore ( 1985 ) ) , chapter 4 . + the iso - spin projection presented here was independently and long previously developed for coherent states by j.c.botke , d.j.scalapino , and r.l.sugar , phys.rev . * d9 * , 813 ( 1974 ) . we apologize to these authors for having missed this work in our earlier publication . v. barger and r. phillips , collider physics , ( addison - wesley , redwood city , california ( 1987 ) ) , chapter 11 . + c. amsler and f. myhrer , ann . nucl . part . sci . * 41 * , 219 ( 1991 ) + j. sedlak and v. simak , sov . * 19 * , 191 ( 1988 ) . dover , t. gutsche , m. maruyama and a. faessler , prog . * 29 * , 87 ( 1992 ) . + c. ghesquiere in : proc.int.symp . on antinucleon - nucleon interaction , liblice - prague , 1974 , cern 74 - 18 , p. 436 boal , c.k . gelbke and b.k . jennings , rev.mod.phys . * 62 * , 553 ( 1990 ) + h.q . song , b.s . zou , m.p . locher , j. riedlberger and p. troul , z. phys . a - hadrons and nuclei * 342 * , 439 ( 1992 ) + m.p . locher and b.s . zou , z.phys a * 341 * , 191 ( 1992 ) . + c. greiner , c. gong and b. muller , phys.lett . b * 316 * , 226 ( 1993 ) . dokshitzer , v.a . khoze , s.i . troyan and a.h . mueller , rev . * 60 * , 373 ( 1988 ) . + j - p . blaizot and a. krzywicki , phys . rev . * d46 * , 246 ( 1992 ) . + figure captions + figure 1 . the probability distribution , @xmath185 , for finding @xmath69 pions from nucleon antinucleon annihilation at rest . the open circles are the poisson distribution and the solid squares are the distribution with the constraint of four - momentum conservation imposed . that distribution can not be distinguished ( on the scale of the figure ) from a gaussian of the same mean and variance . + figure 2 . the probability distribution , @xmath185 , for finding @xmath69 pions from nucleon antinucleon annihilation at rest with iso - spin projection . in all cases the constraint of four - momentum conservation has been imposed . figure 2a is for @xmath78 and figure 2b for @xmath80 , @xmath81 . the open circles are the probabilities with iso - spin projected and the solid squares without iso - spin projection . the solid squares are the same as the solid squares in figure 1 . note that in our model , for @xmath78 only an even number of pions is possible and for @xmath80 only an odd number . this makes the iso - spin projected distributions appear peaked . table 1 : the mean pion number and variance , and the mean pion number by charge type from nucleon antinucleon annihilation at rest in each of the three iso - spin states with the constraint of four - momentum conservation included .

d - branes have been studied for many years and used to reveal nonperturbative aspects of string theory . they are considered to be solitons in string theory . from the viewpoint of open string theory , ( for example , in the vacuum string field theory @xcite ) , d - branes emerge as soliton - like solutions of the equation of motion . the question we would like to address in this paper is what are d - branes in closed string field theory ? " . although several attempts have been made @xcite@xcite , d - branes have not been studied so much in the context of closed string field theories . actually , a fairly clear answer to the above question is given for noncritical strings . d - branes in noncritical string theories can be defined as in the critical ones @xcite . in ref.@xcite , fukuma and yahikozawa showed that the d - branes can be realized as solitonic operators which commute with the virasoro and @xmath3 constraints @xcite for the noncritical string theories . in ref.@xcite , it was shown that how such solitonic operators are realized in the string field theory of noncritical strings presented in ref.@xcite . states in which d - branes are excited can be made by acting these solitonic operators on the vacuum . what we would like to do in this paper is to construct such states in a critical bosonic string field theory . since the string field hamiltonian given in ref.@xcite consists of the joining - splitting interactions , it seems possible to generalize the construction of the operators given in ref.@xcite to the string field theories with the light - cone gauge type interactions . in this paper , we take the @xmath0 invariant string field theory @xcite as such a theory . the @xmath0 invariant string field theory is a covariantized version of the light - cone gauge string field theory @xcite and it was proved that the s - matrix elements coincide with those of the light - cone gauge one @xcite . what we will do is to construct solitonic operators made from the creation and annihilation operators of second - quantized strings . in order to deal with d - branes , we consider the closed strings whose wave functions are proportional to the boundary states . such states were shown @xcite to satisfy the idempotency equations . because of these relations , we expect that three string vertices for such strings look quite like the ones which appear in the noncritical string field theory in ref.@xcite . we will show that we can construct solitonic operators using such states . these operators can be considered as creation operators of d - branes ( or ghost d - branes recently proposed in ref.@xcite ) . acting them on the vacuum , we obtain brst invariant states , which can be regarded as states in which d - branes are excited . we calculate the vacuum amplitude and show that these operators create two d - branes . the organization of this paper is as follows . in section [ sec : noncritical ] , we will briefly explain the construction of the solitonic operators @xcite in the string field theory of noncritical strings @xcite . in section [ sec : ospsft ] , we will review the @xmath0 invariant string field theory @xcite@xcite . in section [ sec : boundary ] , we will define the boundary states and the creation and annihilation operators for the strings whose wave functions are proportional to such states , in the @xmath0 invariant string field theory . in section [ sec : solitonop ] , we will construct solitonic operators using the operators defined in section [ sec : boundary ] , following the construction in the noncritical case . regarding them as the creation operators for d - branes , we can get brst invariant states in which d - branes are excited . we compute the vacuum amplitude and find that the operators we construct create two d - branes or ghost d - branes . section [ sec : discussions ] will be devoted to discussions . in the appendices , we present the details of the calculations to derive the brst transformations for component fields of the string field which are proportional to the boundary states . noncritical strings can be described by the string field theory constructed in @xcite and its generalizations . for simplicity , let us consider the @xmath4 case . in this case , the only reparametrization invariant quantity which can specify the state of a closed string is its length @xmath5 . therefore we define the creation and annihilation operators @xmath6 of the string with length @xmath5 which satisfy @xmath7 = \delta ( l - l^\prime ) ~. \label{noncriticalcanonical}\ ] ] the correlation functions can be calculated using the stochastic hamiltonian @xmath8 where @xmath9 the processes which the first two terms in the hamiltonian represent are exactly the joining - splitting interactions . the third term corresponds to a tadpole term in which only strings with vanishing length are involved . here @xmath10 denotes the string coupling constant and @xmath11 denotes the cosmological constant on the worldsheet . in this formulation , the virasoro constraint for @xmath4 string theory can be written as @xmath12 where @xmath13 is a state which can be expressed by using the correlation functions . solitonic operators corresponding to d - branes can be constructed as operators which commute with @xmath14 @xcite . if such operators exist , by acting them on @xmath13 which is a solution of eq.([noncriticalvirasoro ] ) , one can generate other solutions . from the commutation relations @xmath15 & = & { } -\frac{1}{g_s}\int_0^\infty dl^\prime l^\prime\epsilon ( l^\prime ) \psi ( l+l^\prime ) ~ , \nonumber \\ { } \left[\ , \frac{1}{g_s^2}\int^{\infty}_{0 } dl^\prime l^\prime\epsilon ( l^\prime ) t(l^\prime ) \ , , \ , g_s\psi^\dagger ( l ) \ , \right ] & = & g_s\int_0^l dl^\prime l^\prime ( l - l^{\prime } ) \epsilon ( l^\prime ) \psi^\dagger ( l - l^\prime ) \nonumber \\ & & { } + \frac{2}{g_s}\int_0^\infty dl^\prime ( l+l^\prime ) \epsilon ( l+l^\prime ) \psi ( l^\prime ) ~ , \label{noncriticalcommutation}\end{aligned}\ ] ] it is straightforward to show that @xmath16 satisfies @xmath17 = \partial_\zeta \left(\partial_\zeta \tilde{\epsilon } ( \zeta ) \mathcal{v}(\zeta ) \right)~ , \label{noncritical1}\ ] ] where @xmath18 therefore @xmath19 commutes with @xmath14 if the limits of the integral are chosen appropriately . in perturbative calculations , the integration over @xmath20 is done by the saddle point method and we do not have to specify these limits . this operator can be identified with the creation operator of the zz - brane . @xmath21 in the exponent in @xmath22 has the effect of generating boundaries on the worldsheet with exactly the same weight as that for the boundary state of the zz - brane . moreover one can see that the solitonic operator increases the number of the eigenvalues of the matrix for the matrix model , by one . the calculations above are quite analogous to the ones in 2d free boson theory . @xmath23 , @xmath24 and @xmath14 can be compared to the oscillator modes of the boson and its energy - momentum tensor respectively . @xmath22 should correspond to the vertex operator with conformal weight @xmath25 . the condition that the right hand side of eq.([noncritical1 ] ) be a total derivative fixes the overall factor in the exponent of @xmath22 . actually , from this condition alone , there exists another choice for @xmath22 which is @xmath26 this operator should correspond to the ghost d - brane . what we would like to do in this paper is to generalize the above construction to critical strings . since the solitonic operator @xmath22 generates boundaries on the worldsheet , we should use the creation and annihilation operators of the critical strings whose wave functions are proportional to the boundary states , in place of @xmath24 and @xmath23 in the above construction . in ref.@xcite , it was shown that the boundary states @xmath27 satisfy the idempotency equation @xmath28 where @xmath29 denotes the product corresponding to a light - cone gauge type three string vertex . this equation implies that in the joining - splitting interaction for the strings whose wave function is proportional to the boundary states , what matter are only their lengths . therefore the three string vertex for such states is essentially the same as the one in the hamiltonian ( [ eq : stochastic ] ) for noncritical strings . hence it seems likely that the above construction works also for some string field theory of critical strings . the string field theory we consider in this paper is the @xmath0 invariant string field theory @xcite@xcite . in order to fix the notations used in this paper , let us recapitulate the formulation of this theory . siegel s procedure @xcite for covariantizing the light - cone gauge string field theory @xcite@xcite@xcite is to replace the @xmath30 transverse vector @xmath31 by the @xmath32 vector @xmath33 , where @xmath34 are grassmann even and the ghost fields @xmath35 and @xmath36 are grassmann odd . the metric of the @xmath32 vector space is @xmath37 the euclidean action is @xmath38 where @xmath39 denote the coordinates on the cylinder worldsheet . one has the mode expansion @xmath40 the nonvanishing canonical commutation relations are @xmath41 = i\eta^{\mu\nu}~ , \quad [ \alpha_n^\mu , \alpha_m^\nu ] = n\eta^{\mu\nu}\delta_{n+m,0}~ , \quad [ \tilde{\alpha}_n^\mu , \tilde{\alpha}_m^\nu ] = n\eta^{\mu\nu}\delta_{n+m,0}~ , \nonumber \\ & & \ { c_0,\bar{\pi}_0\ } = 1~ , \quad \ { \bar{c}_0,\pi_0\ } = 1~ , \quad \ { \gamma_n,\bar{\gamma}_m\ } = in\delta_{n+m,0}~ , \quad \ { \tilde{\gamma}_n,\tilde{\bar{\gamma}}_m\ } = in\delta_{n+m,0}\end{aligned}\ ] ] for @xmath42 . we also use @xmath43 the hilbert space for the string consists of the fock space of the oscillators and the wave function for the zero modes . we take the wave function to be a function of @xmath44 , i.e. we take the momentum representation for the zero modes . here @xmath45 denotes the string length , which is a variable characteristic of the string field theories with the joining - splitting interactions . in the momentum representation , the vacuum state @xmath46 in the first quantization is defined by @xmath47 the integration measure for the zero modes of the @xmath48-th string is written as @xmath49 the brst charge is defined @xcite@xcite as @xmath50 where @xmath51 and @xmath52 @xmath53 are the virasoro generators defined as @xmath54 here @xmath55 means the normal ordering of the oscillators in which the non - negative modes should be moved to the right of the negative modes . the brst charge ( [ eq : brst - charge ] ) is nilpotent : @xmath56 . the reflector is defined as @xmath57 where @xmath58 with @xmath59 we also introduce @xmath60 the reflector @xmath61 satisfies @xmath62 this yields @xmath63 @xmath64 satisfies similar identities . the bpz conjugate @xmath65 of @xmath66 is defined as @xmath67 from the definitions , we have @xmath68 and @xmath69 where @xmath70 denotes the grassmann parity of the string field @xmath71 . thus @xmath61 can be considered as the symplectic form for the string fields and @xmath72 is its inverse . the three string vertex is given by @xmath73 where @xmath74 denotes the interaction point and @xmath75 here @xmath76 denote the neumann coefficients associated with the joining - splitting type of three string interaction @xcite@xcite@xcite . invariance on the worldsheet . ] by using the three string vertex ( [ eq : v3 ] ) , the @xmath77-product @xmath78 of two arbitrary closed string fields @xmath71 and @xmath79 is defined by @xmath80 the @xmath77-product has following properties , @xmath81 the first equation is equivalent to @xmath82 and the second one is known as the jacobi identity . the action of the @xmath0 invariant string field theory is directly given by the @xmath0 extension from that of the light - cone gauge string field theory . this takes the form @xmath83~ , \label{eq : actionosp}\end{aligned}\ ] ] where @xmath84 denotes the proper time . the string field @xmath71 is taken to be grassmann even and subject to the level matching condition @xmath85 . note that in the interaction term the three string vertex @xmath86 is multiplied by the factor @xmath87 . this manipulation removes @xmath88 from the vertex @xmath86 , i.e. @xmath89 the action ( [ eq : actionosp ] ) is invariant under the brst transformation @xmath90 where the @xmath77-product is defined in eq.([eq : star - prod ] ) . the nilpotency of the brst transformation ( [ brs ] ) is assured by the nilpotency of @xmath91 and eqs.([eq : prop - ast ] ) . one can readily show that the action ( [ eq : actionosp ] ) is invariant under the brst transformation ( [ brs ] ) by using the nilpotency of the brst transformation ( [ brs ] ) and the fact that the action ( [ eq : actionosp ] ) can be expressed as @xmath92~. \label{eq : brstexact}\ ] ] in this string field theory , @xmath35 and @xmath36 play the role of the @xmath93 ghost in the usual theory . indeed with the following identifications @xmath94 with @xmath42 , @xmath95 becomes almost the same as the usual first - quantized brst operator . perturbative calculations can be done in a way similar to the one for the light - cone gauge string field theory . in the canonical quantization , we should think of the components of @xmath96 with positive @xmath45 as annihilation operators and those with negative @xmath45 as creation operators . the prescription for how to treat the physical on - shell states was given by ref.@xcite invariant string field theory as well . ] , and a proof was given to the fact that the s - matrix elements calculated using this theory coincide with those of the light - cone gauge string field theory . before concluding this section , one comment is in order . in refs.@xcite@xcite@xcite@xcite , gauge invariant actions were proposed and it was shown that the @xmath0 invariant theory can be obtained from them after gauge fixing . unfortunately the brst transformations which originate from these covariantized light - cone string field theories coincide with eq.([brs ] ) only for on - shell states . in this paper , we should deal with the boundary states which are off - shell , and consider eq.([brs ] ) as the brst transformation . the origin of this brst transformation eq.([brs ] ) may be understood by considering this system in terms of the bfv formalism . in principle , looking at the brst transformation itself , one can read off the constraints from which the brst transformation is constructed . in order to construct solitonic operators in the way mentioned at the end of section [ sec : noncritical ] , we should study the boundary states in the @xmath0 invariant theory and identify the creation and annihilation operators corresponding to such states . a problem is that the boundary states are not normalizable . we will introduce a brst invariant regularization and define normalizable states proportional to the boundary states . in what follows , we consider the toroidally compactified space - time characterized by @xmath97 @xmath98 to regularize the infrared divergence . in this situation , the zero modes of the matter sector are modified because of the momentum quantization and the windings . we briefly summarize the notations for the zero - mode part of the toroidally compactified matter sector . the zero - mode part of @xmath99 takes the form @xmath100 where @xmath101 and @xmath102 are complex coordinates on the cylinder worldsheet defined as @xmath103 and @xmath104 . the center - of - mass momentum @xmath105 is quantized and @xmath106 is related to the winding number @xmath107 as follows : @xmath108 @xmath109 and @xmath110 are defined as @xmath111 they obey the canonical commutation relations : @xmath112= \left [ x^{\mu}_{r}\ , , \ , p^{\nu}_{r } \right]=i \eta^{\mu\nu}$ ] , otherwise vanishing . thus the zero - mode sector consists of two canonical pairs . for later use , we introduce a new variable @xmath113 and we choose the basis of the zero - mode phase space to be @xmath114 . they satisfy @xmath115 = \left[y^{\mu}_{0}\ , , \ , q^{\nu}\right]=i\eta^{\mu\nu}$ ] . because of the quantization ( [ eq : wind ] ) of @xmath106 , the range in which @xmath116 ( conjugate to @xmath106 ) varies is finite as well as that of @xmath117 : @xmath118 let @xmath119 and @xmath120 be the eigenstates of the operators @xmath117 and @xmath116 with eigenvalues @xmath121 and @xmath122 . let @xmath123 and @xmath124 be the eigenstates of the operators @xmath105 and @xmath106 with eigenvalues @xmath125 and @xmath126 . we normalize these states as follows : @xmath127 it follows from eqs.([eq : range ] ) and ( [ eq : normes ] ) that @xmath128 in accordance with the modification above , the momentum dependent parts of the integration measure for the zero modes , the reflector @xmath129 and the three string vertex @xmath130 are respectively replaced as follows : @xmath131 in the last equation , the cocycle factor @xmath132 is necessary for the jacobi identity to be satisfied @xcite@xcite . ( see also the second paper in ref.@xcite . ) we consider the situation in which the d@xmath133-branes extend in the @xmath121 directions with @xmath134 . we refer to these directions as the neumann directions and denote them by @xmath121 ( @xmath135 . we refer to the directions transverse to the d - branes as the dirichlet directions and denote them by @xmath136 ( @xmath137 ) . in the first quantized approach to the closed string , the d - brane is described by the boundary state . the boundary state in the matter sector @xmath138 is expressed as the direct product of those for the neumann and the dirichlet sectors : @xmath139 where @xmath140 and @xmath141 are respectively the volumes of the neumann and the dirichlet directions , i.e. @xmath142 and @xmath143 . we notice that the zero - mode part of the state @xmath144 is @xmath145 and that of the state @xmath146 is @xmath147 , modulo normalization constants . in the following , we restrict ourselves to the case in which d@xmath133-branes are located at @xmath148 . let us turn to the ghost sector . we require that the dirichlet boundary condition should be satisfied by the ghost fields @xmath149 and @xmath150 at @xmath151 on the boundary state : @xmath152 in terms of the oscillation modes , these conditions read @xmath153 for @xmath154 . case of eq.([eq : bc2gh ] ) is not derived from eq.([eq : bc1gh ] ) , it holds automatically by definition of @xmath155 and @xmath156 . ] they coincide with the usual boundary conditions for the @xmath93 ghosts assuming eq.([ghostidentification ] ) . this implies that the boundary state @xmath157 is proportional to the state @xmath158 let us define @xmath159 as @xmath160 where @xmath161 is an arbitrary normalization constant . in what follows , we refer to this state as a boundary state . since the string field should have @xmath45 dependence , we need an @xmath45 dependent version of the boundary state . let us define @xmath162 as @xmath163 where the parameter @xmath5 is an eigenvalue of @xmath45 , i.e. @xmath164 . the boundary state ( [ eq : b0 ] ) is not normalizable . we need therefore regularize the divergence of the norm of this state , in order to treat it as a string field in string field theory . for this purpose we introduce a regularized boundary state @xmath165 by attaching a stub to the state @xmath166 as depicted in fig . [ fig : stab ] : @xmath167 a similar regularization is necessary even for on - shell physical states@xcite . this is a brst invariant regularization instead of @xmath168 because @xmath169 this regularization however does not work for our purpose . see the comment at the end of this subsection . ] because @xmath170 commutes with the brst charge @xmath91 , which can be seen from @xmath171 a subtlety seems to occur at @xmath172 , because @xmath45 appears in the form of the absolute value . as is usual in a light - cone formalism , we will exclude the modes with @xmath173 with some small @xmath174 from the spectrum . as we will see , this corresponds to an infrared cut - off on the worldsheet . we will study the theory perturbatively . therefore we will keep the most dominant contributions in the limit @xmath175 at each order in @xmath176 , in the following . . ( b ) the inner product @xmath177 . ] we regularize the state ( [ eq : bel ] ) accordingly : @xmath178 let us consider the inner product @xmath179 where @xmath180 and @xmath181 denote the integration measures for the zero modes defined in eq.([eq : zeromodemeasure ] ) . eq.([eq : bb-0 ] ) becomes the cylinder amplitude of a closed string with a fixed circumference @xmath182 propagating through a very short proper time @xmath183 . after the modular transformation for the non - zero mode part and the poisson resummation for the zero - mode part , we obtain @xmath184^{24 } } \nonumber\\ & & \times \prod_{\mu \in \mathrm{n } } \left [ \sum_{{\mathtt{w}}^{\mu}\in\mathbb{z } } e^{-\frac{|l|}{\epsilon } \frac{\pi^{2}{\alpha'}}{(r^{\mu})^{2 } } { \mathtt{w}}^{\mu 2 } } \right ] \times \prod_{i \in \mathrm{d } } \left [ \sum_{{\mathtt{n}}^{i}\in \mathbb{z } } e^{-\frac{|l|}{\epsilon}\frac{(\pi r^{i})^{2}}{{\alpha ' } } { \mathtt{n}}^{i2 } } \right]~. \label{eq : overlap2}\end{aligned}\ ] ] the dominant contribution @xmath185 in the limit @xmath186 originates from the propagation of the open string tachyon in the dual channel . we introduce the state @xmath187 defined as @xmath188 from eq.([eq : overlap2 ] ) , we find that @xmath189 thus the state @xmath187 is normalizable . a comment is in order . naively speaking @xmath190 because the wave function for @xmath162 lacks factors of @xmath191 and @xmath192 . however in eq.([eq : bb-0 ] ) , @xmath170 provides these and we get a nonvanishing answer for the inner product . now let us define the creation and annihilation operators of the closed strings whose wave functions are proportional to the boundary states . the states @xmath193 with @xmath194 are normalizable as stated above and the inner products ( [ eq : innerprod - n ] ) among them are non - degenerate . this enables us to choose a complete basis of the hilbert space which consists of these states and their orthogonal complement . taking also the states @xmath195 into account , we can expand @xmath196 as @xmath197~ , \label{eq : sf1}\ ] ] where ` @xmath198 ' denotes the contributions from the other states . the wave functions @xmath199 , @xmath200 , @xmath201 , @xmath202 etc . are the fields to be quantized in the second quantization . etc . depend on the proper time @xmath84 : @xmath203 . we suppress the proper time in the arguments for simplicity . ] @xmath199 and @xmath200 can be considered as the annihilation and creation operators for the closed strings corresponding to the boundary state . let us divide @xmath196 into the creation and annihilation parts as follows : @xmath204~,\nonumber\\ & & \left| \bar{\psi } \right\rangle = \int^{\infty}_{0 } dl \big [ \left| n(-l ) \right\rangle \bar{\phi}(l ) + \bar{\pi}_{0 } \left| n(-l ) \right\rangle \bar{\chi}(l ) + \cdots \big]~. \label{eq : sf2}\end{aligned}\ ] ] the string field @xmath196 satisfies the reality condition : @xmath205 where @xmath206 denotes the hermitian conjugate of @xmath207 , and @xmath208 denotes the bpz conjugate of @xmath207 defined in eq.([eq : bpz ] ) , respectively . since the bpz conjugation flips the sign of the string length @xmath45 , the reality condition ( [ eq : reality1 ] ) implies that @xmath209 combined with the relation @xmath210 eq.([eq : reality2 ] ) leads to @xmath211 the brst transformations @xmath212 and @xmath213 for the component fields @xmath199 and @xmath200 can be calculated from eq.([brs ] ) . considering the idempotency equations @xcite satisfied by the boundary states , we expect that the nonlinear terms in the transformation takes a very simple form . indeed we obtain @xmath214 \nonumber\\ & & { } + \cdots ~ , \nonumber\\ \frac{4c\epsilon^{3}}{g\mathcal{n } } \delta_{\mathrm{b } } \bar{\phi } ( l ) & = & { } -\frac{c}{g\mathcal{n}}\epsilon^{2 } \left ( \frac{\partial}{\partial l}+\frac{\pi^{2}}{2\epsilon } \right ) \left ( l\bar{\chi}(l ) \right ) + \int^{l}_{0 } dl_{1}\ , l_{1}(l - l_{1 } ) \bar{\chi}(l_{1 } ) \bar{\phi}(l - l_{1 } ) \nonumber\\ & & { } + \int^{\infty}_{0 } dl_{1}\ , l_{1}(l+l_{1 } ) \left [ \bar{\chi}(l+l_{1})\phi(l_{1 } ) + \chi(l_{1})\bar{\phi}(l+l_{1 } ) \right ] \nonumber\\ & & { } + \cdots ~ , \label{eq : brst}\end{aligned}\ ] ] where @xmath215 the derivation of eq.([eq : brst ] ) is presented in appendix [ sec : brst ] . as is intuitively clear , a boundary state split into two makes two boundary states and two boundary states joined together makes a boundary state . such contributions are written explicitly in eq.([eq : brst ] ) . however , a boundary state joined to a different state makes a state different from the boundary state . such contributions are denoted by ` @xmath198 ' . notice that each term in ` @xmath198 ' should be a product of one annihilation operator other than @xmath216 and one creation operator other than @xmath217 . in this section , we will construct solitonic operators made from the @xmath218 and @xmath219 and study their properties . let us canonically quantize the string fields defined in eq.([eq : sf1 ] ) first . the kinetic term of the action ( [ eq : actionosp ] ) can be written as @xmath220 in the same way as was performed in the light - cone string field theory @xcite@xcite , we obtain the canonical commutation relation @xmath221 = i(r , s ) \qquad \leftrightarrow \qquad \left [ \left|\psi\right\rangle_{r } , \left|\bar{\psi } \right\rangle_{s } \right ] = \left| r ( r , s ) \right\rangle~ , \label{eq : ccc1}\ ] ] where @xmath222 is defined as @xmath223 @xmath222 serves as the identity operator . in fact , the following relations hold for an arbitrary string field @xmath224 , @xmath225 multiplying the second equation in eq.([eq : ccc1 ] ) by @xmath226 from the left , we have @xmath227 = \frac{1}{\mathcal{n}^{2 } l_{r } } \delta ( l_{r}-l_{s})~. \label{eq : ccc2}\ ] ] one can also derive this commutation relation directly from the action ( [ eq : osp - kin ] ) expressed in terms of the component fields : @xmath228 the vacuum state @xmath229 in the second quantization is defined by @xmath230 this yields @xmath231 we take the normalization of the vacuum state @xmath229 as @xmath232 . the right hand sides of eq.([eq : brst ] ) look quite like those of eq.([noncriticalcommutation ] ) . indeed if we replace @xmath233 \ \rightarrow \ \frac{4c\epsilon^{3}}{g\mathcal{n}}\delta_{\mathrm{b } } \left ( \ \cdot \ \right)~,&\nonumber\\ & \displaystyle \frac{\sqrt{2}}{g_s\mathcal{n}l}\psi ( l ) \rightarrow \phi ( l)~ , \quad \frac{g_s}{\sqrt{2}\mathcal{n}}\psi^\dagger ( l ) \rightarrow \bar{\phi}(l)~ , \quad \epsilon ( l ) \rightarrow \bar{\chi}(l)~,&\end{aligned}\ ] ] in eq.([noncriticalcommutation ] ) , we get exactly the nonlinear terms involving @xmath202 on the right hand sides of eq.([eq : brst ] ) . moreover the commutation relation ( [ noncriticalcanonical ] ) becomes eq.([eq : ccc2 ] ) by such replacements . therefore we expect that operators in the following form can be used to construct brst invariant operators : @xmath234 \,\exp \left [ \mp \sqrt{2}\mathcal{n } \int^{\infty}_{0 } dl ' \ , e^{\zeta l ' } \phi ( l')\right].\ ] ] taking the linear terms on the right hand sides of eq.([eq : brst ] ) into account , we define @xmath235 \,\exp \left [ \mp \sqrt{2}\mathcal{n } \int^{\infty}_{0 } dl ' \ , e^{\zeta l ' } \phi ( l')\right ] e^{\pm \frac{c\epsilon^{2}}{\sqrt{2}g } \left(\zeta+\frac{\pi^{2}}{2\epsilon}\right)^{2}}~ , \label{eq : v - cal}\end{aligned}\ ] ] where @xmath236 is a constant . actually we can not make brst invariant operators from @xmath22 . rather we will show that we can construct brst invariant states by acting @xmath237 on the vacuum @xmath229 . as a warm - up , let us show that @xmath238 |0 \rangle \ ! \rangle \label{eq : d - branestate}\end{aligned}\ ] ] is brst invariant : @xmath239 the brst transformation can be calculated by using eq.([eq : brst ] ) as @xmath240 \,e^{\pm \sqrt{2 } \mathcal{n } \int^{\infty}_{0 } dl\ , e^{-\zeta l } \bar{\phi}(l ) } \ ; |0\rangle \!\rangle~. \label{eq : brstexp1}\end{aligned}\ ] ] here we have used @xmath241 and the fact that ` @xmath198 ' in eq.([eq : brst ] ) does not contribute because it includes annihilation operators other than @xmath218 . it is useful to introduce the laplace transforms @xmath242 and @xmath243 of @xmath200 and @xmath202 defined as @xmath244 the following identities hold , @xmath245 combining these relations with eq.([eq : brstexp1 ] ) , we obtain @xmath246 \right ) |0\rangle\!\rangle } \nonumber\\ & & = \frac{\partial}{\partial \zeta } \left ( \frac{\partial \tilde{\bar{\chi}}(\zeta)}{\partial \zeta } \exp\left[\pm \sqrt{2 } \mathcal{n } \int^{\infty}_{0}dl\ , e^{-\zeta l}\bar{\phi}(l ) \pm \frac{c\epsilon^{2}}{\sqrt{2}g } \left ( \zeta + \frac{\pi^{2}}{2\epsilon } \right)^{2}\right ] \right)\,|0\rangle\!\rangle~. \label{eq : brstexp3}\end{aligned}\ ] ] taking eq.([eq : d - branestate ] ) into account , we find that this equation implies the brst invariance ( [ eq : brsinv - dstate-0 ] ) of the state @xmath247 . @xmath247 can be considered as a state in which d - branes are excited . actually as we will see in the next subsection , two d - branes are there . in order to have more d - branes , we just have to operate @xmath237 successively on the vacuum @xmath229 , namely we construct states @xmath248 for @xmath249 . they are brst invariant , because @xmath237 is brst invariant modulo terms which annihilate the states @xmath250 , @xmath251 . this can be seen as follows . under the brst transformation ( [ eq : brst ] ) , @xmath252 transforms as @xmath253 + \left ( \pm \frac{2\sqrt{2}c\epsilon^{2}}{g } \zeta \frac{\partial \tilde{\chi}(\zeta)}{\partial \zeta } + \cdots \right ) \mathcal{v}(\zeta)~ , \label{eq : brst - v}\ ] ] where ` @xmath198 ' denotes the terms which includes annihilation operators other than @xmath254 , and @xmath255 and @xmath256 are the laplace transforms of @xmath254 and @xmath201 defined as @xmath257 we can prove eq.([eq : brst - v ] ) with the help of the relation @xmath258 the terms in the parenthesis @xmath259 in the second term on the right hand side of eq.([eq : brst - v ] ) commute with @xmath252 and annihilate @xmath229 . thus we find that the states of the form @xmath250 are brst invariant . in order to compare our description of d - branes constructed above with the usual one , let us compute the vacuum amplitude in the presence of d - branes in our formalism . since @xmath247 is considered as a state with some d - branes excited , the vacuum amplitude in the presence of d - branes can be given as @xmath260 where @xmath261 is the second - quantized hamiltonian . notice that the time variable in the @xmath0 invariant string field theory is not the physical time but with topological " nature because the hamiltonian @xmath261 is brst exact . this can be seen from the expression ( [ eq : brstexact ] ) . therefore @xmath262 can be transformed into a form which is independent of the value of @xmath263 . by comparing @xmath262 with the usual vacuum amplitude , we can see how many d - branes are there in the state @xmath247 . in order to do so , let us first perform the integration over @xmath20 in eq.([eq : d - branestate ] ) . perturbatively the factor @xmath264 \label{eq : gaussian}\ ] ] in the integrand in eq.([eq : d - branestate ] ) is the most dominant . therefore , through the saddle point approximation we obtain @xmath265 |0\rangle \ ! \rangle~ , \label{eq : saddle}\ ] ] where @xmath266 . one can find that @xmath267 plugging this relation into eq.([eq : saddle ] ) , we obtain the expression of the state @xmath268 in terms of the string state @xmath269 and @xmath270 as follows : @xmath271 |0\rangle \ ! \rangle~. \label{eq : d - branestate2}\ ] ] note that the divergent factor @xmath272 @xmath273 is miraculously canceled by the regularization factor @xmath274 in @xmath275 and we can express @xmath276 in terms of @xmath277 . now let us evaluate @xmath262 . perturbatively the lowest order contribution can be obtained by replacing the hamiltonian @xmath261 with its free part @xmath278 . substituting @xmath279~ , \nonumber\\ \hat{h}_{0 } & = & \int dr\ , { } _ { r}\!\langle \bar{\psi}| \ , \frac{l^{(r)}_{0}+\tilde{l}^{(r)}_{0}-2}{\alpha_{r } } \ , \left| \psi \right\rangle_{r}~,\end{aligned}\ ] ] into the expression and using the commutation relation ( [ eq : ccc1 ] ) , we find that @xmath280 |0\rangle\!\rangle \nonumber\\ & & = \left| \lambda ' \right|^{2 } \exp \left [ \frac{2}{\mathcal{n}^{2 } } \int^{\infty}_{0}\frac{dl}{l } \int^{\infty}_{0 } \frac{dl'}{l ' } \int dr \ ; { } ^{\epsilon}_{r}\ ! \langle b_{0}(-l')|\ , e^{-it\frac{l^{(r)}_{0}+\tilde{l}^{(r)}_{0}-2}{\alpha_{r } } } \ , |b_{0}(l ) \rangle^{\epsilon}_{r } \right]~. \label{eq : dd1}\end{aligned}\ ] ] here we have used the relation @xmath281 which follows from eq.([eq : ccc1 ] ) . the quantity in the exponent of eq.([eq : dd1 ] ) should be compared with the cylinder amplitude in the usual formulation . after the integrations over @xmath45 , @xmath191 , @xmath192 and @xmath282 , we obtain the integration measure for @xmath5 as @xmath283 since @xmath263 is fixed , this is exactly the integration over the parameter in front of @xmath284 . therefore the integration over @xmath5 is transformed into the one for the moduli parameter of the cylinder and the result is in a form which is independent of the value of @xmath263 . the overlap between the boundary states in the exponent on the right hand side of eq.([eq : dd1 ] ) can readily be obtained from eq.([eq : overlap2 ] ) by replacing @xmath285 in eq.([eq : overlap2 ] ) by @xmath286 . introducing @xmath287 , we obtain @xmath288~. \label{eq : dd2}\end{aligned}\ ] ] notice that the arbitrary normalization constant @xmath161 does not appear in this final answer . we can see that the exponent of eq.([eq : dd2 ] ) reproduces four times the annulus amplitude for one d - brane . this result implies that @xmath247 is the state in which two d - branes or ghost d - branes are excited . since @xmath219 generates boundaries on the worldsheet , depending on which sign in eq.([eq : v - cal ] ) is chosen , d - branes or ghost d - branes are excited . similar calculations yield that @xmath237 creates two d - branes or ghost d - branes . in this subsection , we have shown that the cylinder amplitudes for d - branes are reproduced in our formulation . what is remarkable is that the integration measure for the moduli of the cylinder appear from the integration over the length of the string . it would be an intriguing problem to check if the integration measures for the higher genus graphs appear in a similar way . in this paper , we have constructed solitonic operators which create d - branes . although we started from the nonnormalizable state ( [ eq : b0 ] ) , the divergent factors cancel with each other and the cylinder amplitude is reproduced . the cancellation occurred because of the factor @xmath289 $ ] in @xmath22 . it originates from the term @xmath290 in the brst charge ( [ eq : brst - charge ] ) . this term is peculiar to the @xmath0 invariant theory and it seems that our construction works only for this theory . the exponent of the exponential factor we are discussing may be interpreted as the potential for the open string tachyon . indeed @xmath20 appears in the form of @xmath291 in front of @xmath254 and can be considered as a constant tachyon background . it will be useful to interpret various quantities in our construction in terms of open string language using for example the methods in ref.@xcite . in our construction , we do not describe d - branes as solutions of equations of motion . rather we construct the solitonic operator @xmath19 , where the form of @xmath22 looks quite like the bosonization formula . another way to look at our results is as follows . we may write @xmath292 by introducing the state @xmath293 defined as @xmath294 this yields @xmath295\ ; |0\rangle\!\rangle~. \label{eq : d - b - bpsi}\ ] ] the state @xmath296 is regarded as a regularized version of the state @xmath297 . we find that the states @xmath298 and @xmath299 are annihilated by the brst charge @xmath91 : @xmath300 we can extend eq.([eq : d - b - bpsi ] ) by including the dependence on the annihilation modes @xmath301 into @xmath302 : |0\rangle\ ! \rangle~ , \label{eq : d - branestate3}\ ] ] where @xmath303 means the normal ordering in which the annihilation mode @xmath301 should be moved to the right of the creation mode @xmath304 . taking account of the relation ( [ eq : qbb ] ) , we notice that the exponent of eq.([eq : d - branestate3 ] ) takes a form quite similar to the interaction term of a closed string with a d - brane introduced by ref.@xcite into the action of the hikko closed string field theory @xcite . therefore we can say that d - branes are introduced as a source of closed strings . as we mentioned , we should introduce a cut - off @xmath174 so that @xmath305 . since @xmath5 works as a moduli , this cut - off will remove the singularity occurring in the limit where a part of the worldsheet becomes a very long cylinder . therefore every equation in this paper should be understood with such a cut - off being introduced . in the first relation in eq.([eq : fs ] ) , we have ignored the contribution from the boundary term @xmath306 . due to the cut - off , we should take care of the contribution of the boundary term near @xmath307 , which violates the brst invariance of the state @xmath247 . we should therefore employ the fischler - susskind mechanism @xcite , in order to remedy the violation of the brst invariance . this modifies the equations of motion for light states of closed string . this is also consistent with the picture of the d - brane as a source of closed strings . our treatment in section [ sec : solitonop ] is perturbative and ignores such higher order effects . there are several problems that remain to be studied . one immediate question is why our solitonic operators create two d - branes or ghost d - branes . we are not sure if there exist operators which create one d - brane or ghost d - brane . it will be an intriguing problem to look for such operators . in this paper , we have not fixed the sign in the exponents of eqs.([eq : v - cal ] ) and ( [ eq : d - branestate ] ) . from the calculation ( [ eq : dd2 ] ) , one can find that if the state @xmath247 with one sign is identified with ordinary d - branes , the state with the other sign should be identified with ghost d - branes @xcite . to determine which sign corresponds to which will be another interesting problem . of course , our results should be generalized to the superstring case . in order to do so , we should first construct the @xmath0 invariant theory for superstrings . we would like to thank k. araki , n. hatano , s. katagiri and t. saitou for discussions . this work is supported in part by grants - in - aid for scientific research 13135224 . in this appendix , we provide some details of the calculation to get eq.([eq : brst ] ) from the brst transformation ( [ brs ] ) for the string field @xmath308 given in eq.([eq : sf1 ] ) . for later convenience , we introduce the notation @xmath309 in order to get @xmath310 and @xmath311 , we consider the inner product of @xmath312 and eq.([brs ] ) . as for the left hand side of the brst transformation , one can readily find that @xmath313 let us turn to the right hand side of the brst transformation . first , we consider the linear term @xmath314 . using eq.([eq : reflection2 ] ) , we find that @xmath315 the state @xmath316 is expressed as @xmath317 where @xmath296 is the state introduced in eq.([eq : invbs ] ) . combined with eq.([eq : qbb ] ) , this leads to @xmath318 substituting this equation into eq.([eq : nqb-1 ] ) , we have @xmath319 on the rightest hand side in the above equation , only the @xmath320 component of @xmath196 provides a nonvanishing contribution because of the ghost zero - mode saturation , i.e. @xmath321 it is easy to show @xmath322 plugging eqs.([eq : n - pipi - n ] ) and ( [ eq : nbpiphi ] ) into eq.([eq : nqb-2 ] ) , we obtain @xmath323 \displaystyle -\frac{\mathcal{n}^{2}}{4\epsilon } l \left ( \frac{\partial}{\partial l } - \frac{\pi^{2}}{2\epsilon } \right ) \left ( l \bar{\chi}(-l ) \right ) & \mbox{for $ l<0 $ } \end{array } \right.~. \label{eq : nqb-3}\end{aligned}\ ] ] second , we consider the non - linear term @xmath324 . keeping the most dominant terms in the limit @xmath175 , we find that @xmath325\nonumber\\ & & \qquad \quad { } + \cdots~ , \label{eq : nphiphi-1}\end{aligned}\ ] ] where @xmath48 of @xmath326 can be any of @xmath327 . in going from the third line to the fourth line in the above equation , we expand @xmath328 in terms of the complete basis defined in subsection [ sec : expansion ] . in doing so , we have used the following idempotency equations @xmath329 for @xmath330 , and @xmath331 where @xmath332 and @xmath333 . these equations hold in the limit @xmath175 and can be proved by using the connection conditions satisfied by @xmath334 . ` @xmath198 ' on the rightest hand side stands for the contributions from the component fields other than @xmath335 and @xmath336 . one can easily see that these terms include one annihilation operator other than @xmath337 and one creation operator other than @xmath338 . we will ignore these contributions in the rest of this appendix . combining the ghost number conservation with the above arguments , we obtain the last equality in eq.([eq : nphiphi-1 ] ) . since we may choose an arbitrary @xmath48 for @xmath326 on the three string vertex @xmath339 as mentioned above , we set @xmath340 in the first term and set @xmath341 in the second term on the rightest hand side in eq.([eq : nphiphi-1 ] ) . from the definition ( [ eq : n0 - 1 ] ) of the state@xmath275 and the fact that the field @xmath88 satisfy the dirichlet boundary condition on the state @xmath342 , one can find that @xmath343 and thus @xmath344 this implies that in the leading order of @xmath285 , eq.([eq : nphiphi-1 ] ) becomes @xmath345 in the second equality in this equation , we have used the result obtained in appendix [ sec : liouville ] . we can further recast eq.([eq : eq : nphiphi-2 ] ) into @xmath346~,\label{eq : rhs - brs2}\end{aligned}\ ] ] where @xmath347 is the step function defined as @xmath348 for @xmath349 and @xmath350 for @xmath351 . combining eqs.([eq : brs - lhs ] ) , ( [ eq : nqb-3 ] ) and ( [ eq : rhs - brs2 ] ) , we obtain eq.([eq : brst ] ) . in this appendix , we will prove that the following relation holds in the leading order of @xmath285 : @xmath353 this equation was used in eq.([eq : eq : nphiphi-2 ] ) to derive eq.([eq : brst ] ) . therefore , by proving eq.([eq : vnnn ] ) , we can complete the derivation of eq.([eq : brst ] ) presented in the last appendix . from the definition ( [ eq : n0 - 1 ] ) , we find that we can prove the above equation by evaluating @xmath354 here we have introduced the integration measure @xmath355 for zero modes defined by removing the @xmath45 dependence from @xmath180 given in eq.([eq : zeromodemeasure ] ) : @xmath356 the amplitude ( [ eq : vbbb ] ) corresponds to the string diagram described in fig . [ fig : open](a ) . as was performed in ref.@xcite , we use conformal field theory ( cft ) technique to calculate eq.([eq : vbbb ] ) . in the @xmath0 invariant string theory , the ghost sector @xmath360 is described by the cft with central charge @xmath361 . the full theory , which consists of the matter sector @xmath362 in addition to the ghost sector , is therefore the cft with total central charge @xmath363 . since @xmath364 , the amplitudes of this system depend on the metric on the worldsheet . the cft we are dealing with consists of free bosons and fermions . therefore the metric dependence stems from the determinant of the laplacian on the worldsheet . it can be given by evaluating the liouville action on the worldsheet @xcite@xcite . the oscillator independent part of the three string vertex @xmath365 can be considered to be due to the contributions from the liouville action . @xmath365 corresponds to the diagram depicted in fig . [ fig : czplane](a ) . it is useful to pull this diagram ( @xmath366-plane ) back to the complex @xmath357-plane ( fig . [ fig : czplane](b ) ) through the mandelstam mapping @xcite , @xmath367 we fix the @xmath368 gauge symmetry on the worldsheet by setting @xmath369 , @xmath370 and @xmath371 . the interaction point is at @xmath372 , where we have @xmath373 . in order to avoid the singularity at the interaction point , we cut out a small circle of radius @xmath374 around the interaction point in the @xmath366-plane . we also cut the points corresponding to incoming and outgoing strings at @xmath375 by terminating each string at @xmath376 @xmath377 . these correspond to cutting circles out of the @xmath357-plane centered on @xmath378 with small radii @xmath379 @xmath380 , as represented in fig . [ fig : czplane](b ) . let us suppose that the @xmath366-plane is equipped with the flat metric : @xmath381 we will see that @xmath365 is constructed to reproduce the cft amplitudes on the @xmath366-plane with this metric . since the calculations are done on the @xmath357-plane , we should take care of the liouville action for the liouville field @xmath218 in eq.([rho - metric ] ) , because the determinant of the laplacian is expressed as @xmath382~,\ ] ] where @xmath383 denotes the variable parametrizing the boundary of the worldsheet @xmath384 ; @xmath385 denotes the geodesic curvature of the boundary defined as @xmath386 where @xmath387 is the unit vector tangential to the boundary while @xmath388 is normal , and @xmath389 denotes the covariant derivative associated with the metric @xmath390 . the dependence of @xmath391 on @xmath379 was calculated in appendix 11.a of ref.@xcite : ) is twice eq.(11.a.26 ) in ref.@xcite because we are dealing with the closed string case . ] @xmath392 where @xmath393 denotes the sum over all the values of @xmath394 and @xmath48 . by exploring the mandelstam mapping ( [ eq : mandelstam ] ) near the cuts , we find that @xmath379 depend on @xmath395 as follows : @xmath396 where @xmath397 using these results , we obtain @xmath398 thus we find that the determinant factor depends on @xmath395 in the following way , @xmath399 this reproduces the factor appearing in the three string vertex @xmath365 given in eq.([v30 ] ) , except for the fact that we should take the absolute value of the factor @xmath400 . thus we see that roughly speaking , @xmath365 is constructed to reproduce the cft amplitudes on the @xmath366-plane with the metric ( [ rho - metric ] ) . precisely speaking , @xmath401 corresponds to fig . [ fig : czplane](a ) . now we would like to calculate eq.([eq : vbbb ] ) in the limit @xmath402 . for this purpose , it is convenient to see the string diagram fig . [ fig : open](a ) from the point of view of the dual open string channel . in this channel , one can regard the worldsheet as being swept by four open strings interacting via mid - point type interaction fig . [ fig : open](b ) . in the limit @xmath402 , such open strings propagate through very long proper time and thus the most dominant contribution comes from the propagations of the open string tachyons . the propagator contributes the factor @xmath403 let us evaluate the determinant of the laplacian on the worldsheet depicted by fig . [ fig : open](b ) . the mandelstam mapping from the upper half @xmath357-plane into the open string 4-point @xmath404-plane is @xmath405 for the worldsheet described in fig . [ fig : open](b ) , we should choose @xmath406 let us choose @xmath407 @xmath408 . here we should take @xmath409 to treat the worldsheet that we are considering . the interaction point is @xmath410 which is a solution of @xmath411 . we introduce the parameter @xmath412 defined by @xmath413 in terms of @xmath412 , the interaction point is described by @xmath414 @xmath415 in our case , but let us treat @xmath412 as a free parameter in order to compare with the results of @xcite . in order to avoid the singularities , we excise small circles around the interaction point and external strings , as shown in fig . [ fig : open](c ) . we define @xmath416 accordingly . by using the metric ( [ rho - metric ] ) , we find that the moduli dependence of the determinant of the laplacian becomes @xmath417 + \frac{1}{6}\sum_{r=1}^4\ln\varepsilon_r^{\mathrm{open } } + \frac{1}{3}\ln\varepsilon_i^{\mathrm{open}}\nonumber\\ & \sim & \frac{1}{4}\ln\left(\alpha \cos\theta \sin\theta\right)~. \label{eq : lndetprime}\end{aligned}\ ] ] here we have used @xmath418 with @xmath419 combining eqs.([eq : leg ] ) and ( [ eq : lndetprime ] ) , we find that the amplitude corresponding to the pants diagram in the limit @xmath175 depends on @xmath395 , @xmath45 and @xmath412 as @xmath420 in ref.@xcite , the authors computed this quantity in the limit @xmath421 , by using the cremmer - gervais identity @xcite . by comparing our result with theirs , it is straightforward to determine the overall constant , and we obtain @xmath422 when we substitute eq.([eq : vbbb - result ] ) into eq.([eq : vnnn ] ) , the exponential factor on the right hand side in this equation is canceled by the regularization factors @xmath423 @xmath377 in @xmath424 . it is because @xmath425 . finally , carrying out the @xmath395 integration , we obtain eq.([eq : vnnn ] ) . a comment is in order . in eq.([eq : vnnn ] ) , there is a factor @xmath426 . the absolute value originates from the one in eq.([oscillatorindep ] ) . taking the absolute value is necessary to be consistent with @xmath427 which can be easily proved . l. rastelli , a. sen and b. zwiebach , `` string field theory around the tachyon vacuum , '' adv . theor . math . phys . * 5 * , 353 ( 2002 ) [ arxiv : hep - th/0012251 ] . + for reviews and references , see + w. taylor and b. zwiebach , `` d - branes , tachyons , and string field theory , '' arxiv : hep - th/0311017 . + a. sen , `` tachyon dynamics in open string theory , '' int . j. mod . a * 20 * , 5513 ( 2005 ) [ arxiv : hep - th/0410103 ] . k. hashimoto and h. hata , `` d - brane and gauge invariance in closed string field theory , '' phys . d * 56 * , 5179 ( 1997 ) [ arxiv : hep - th/9704125 ] . t. asakawa , s. kobayashi and s. matsuura , `` closed string field theory with dynamical d - brane , '' jhep * 0310 * , 023 ( 2003 ) [ arxiv : hep - th/0309074 ] . v. fateev , a. b. zamolodchikov and a. b. zamolodchikov , boundary liouville field theory . i : boundary state and boundary two - point arxiv : hep - th/0001012 . + a. b. zamolodchikov and a. b. zamolodchikov , `` liouville field theory on a pseudosphere , '' arxiv : hep - th/0101152 . + j. teschner , `` remarks on liouville theory with boundary , '' arxiv : hep - th/0009138 . + b. ponsot and j. teschner , `` boundary liouville field theory : boundary three point function , '' nucl . phys . b * 622 * , 309 ( 2002 ) [ arxiv : hep - th/0110244 ] . m. fukuma and s. yahikozawa , `` nonperturbative effects in noncritical strings with soliton backgrounds , '' phys . b * 396 * , 97 ( 1997 ) [ arxiv : hep - th/9609210 ] ; `` combinatorics of solitons in noncritical string theory , '' phys . b * 393 * , 316 ( 1997 ) [ arxiv : hep - th/9610199 ] . m. fukuma , h. kawai and r. nakayama , `` continuum schwinger - dyson equations and universal structures in two - dimensional quantum gravity , '' int . j. mod . a * 6 * , 1385 ( 1991 ) . + r. dijkgraaf , h. l. verlinde and e. p. verlinde , `` loop equations and virasoro constraints in nonperturbative 2-d quantum gravity , '' nucl . b * 348 * , 435 ( 1991 ) . m. hanada , m. hayakawa , n. ishibashi , h. kawai , t. kuroki , y. matsuo and t. tada , `` loops versus matrices : the nonperturbative aspects of noncritical string , '' prog . phys . * 112 * , 131 ( 2004 ) [ arxiv : hep - th/0405076 ] . n. ishibashi and h. kawai , `` string field theory of noncritical strings , '' phys . b * 314 * , 190 ( 1993 ) [ arxiv : hep - th/9307045 ] . + a. jevicki and j. p. rodrigues , `` loop space hamiltonians and field theory of noncritical strings , '' nucl . b * 421 * , 278 ( 1994 ) [ arxiv : hep - th/9312118 ] . m. kaku and k. kikkawa , `` the field theory of relativistic strings , pt . 1 : trees , '' phys . rev . d * 10 * , 1110 ( 1974 ) ; `` the field theory of relativistic strings . 2 : loops and pomerons , '' phys . d * 10 * , 1823 ( 1974 ) . i. kishimoto , y. matsuo and e. watanabe , `` boundary states as exact solutions of ( vacuum ) closed string field theory , '' phys . d * 68 * , 126006 ( 2003 ) [ arxiv : hep - th/0306189 ] ; `` a universal nonlinear relation among boundary states in closed string field theory , '' prog . phys . * 111 * , 433 ( 2004 ) [ arxiv : hep - th/0312122 ] . t. yoneya , `` extended fermion representation of multi - charge 1/2-bps operators in ads / cft : towards field theory of d - branes , '' jhep * 0512 * , 028 ( 2005 ) [ arxiv : hep - th/0510114 ] . t. okuda and t. takayanagi , `` ghost d - branes , '' arxiv : hep - th/0601024 . w. siegel and b. zwiebach , `` gauge string fields from the light cone , '' nucl . b * 282 * , 125 ( 1987 ) ; `` closed string fields via iosp(1,1/2 ) , '' phys . b * 184 * , 325 ( 1987 ) ; `` osp(1,1/2 ) string field theory with the world sheet metric , '' nucl . b * 288 * , 332 ( 1987 ) . + l. baulieu , w. siegel and b. zwiebach , `` new coordinates for string fields , '' nucl . b * 287 * , 93 ( 1987 ) . t. kawano , `` dilaton condensation in the covariantized light cone closed string field theory , '' prog . phys . * 88 * , 1181 ( 1992 ) . h. hata , k. itoh , t. kugo , h. kunitomo and k. ogawa , `` gauge string field theory for torus compactified closed string , '' prog . phys . * 77 * , 443 ( 1987 ) . m. maeno and h. takano , `` derivation of the cocycle factor of vertex in closed bosonic string field theory on torus , '' prog . phys . * 82 * , 829 ( 1989 ) . i. kishimoto and y. matsuo , `` cardy states as idempotents of fusion ring in string field theory , '' phys . b * 590 * , 303 ( 2004 ) [ arxiv : hep - th/0402107 ] ; `` cardy states , factorization and idempotency in closed string field theory , '' nucl . phys . b * 707 * , 3 ( 2005 ) [ arxiv : hep - th/0409069 ] . w. fischler and l. susskind , `` dilaton tadpoles , string condensates and scale invariance , '' phys . b * 171 * , 383 ( 1986 ) ; `` dilaton tadpoles , string condensates and scale invariance . 2 , '' b * 173 * , 262 ( 1986 ) .

quantum hall effects are characteristic of two - dimensional ( 2d ) electron gases in strong magnetic fields . the integral quantum hall effect ( iqhe ) @xcite is explained in terms of weakly - interacting electrons filling quantized landau levels ( ll ) produced by application of a strong magnetic field perpendicular to the 2d gas @xcite . in contrast , the fractional quantum hall effect ( fqhe ) @xcite can occur only as a result of strong electronic correlations in partially - filled landau levels @xcite . because of its massless chiral charge carriers and atomic - monolayer honeycomb lattice exhibiting sublattice , valley , and spin degeneracies , quantum hall effects in graphene could be much richer than in the semiconductor 2d electron gas , where there is no crystal structure and the only degeneracies are those of the ( non - relativistic ) landau levels and spin . for graphene in strong magnetic fields an integral quantum hall effect @xcite and a fractional quantum hall effect @xcite have been observed , with anomalous filling factors that reflect the unique degeneracies of the graphene electronic structure and the dirac - like nature of its electrons . the valley isospin and spin degrees of freedom imply that graphene in a strong magnetic field is described well by a low - energy hamiltonian that commutes approximately with the generators of an su(4 ) lie algebra . this su(4 ) symmetry allows classification of states in graphene , and can serve as the basis for computing explicit breaking of the su(4 ) symmetry by small non - symmetric terms in the effective hamiltonian . however , there is growing evidence that many states observed in modern experiments can not be described by explicit breaking of su(4 ) . for example , the ground state of graphene in a magnetic field at low temperature exhibits a rapid increase of the longitudinal resistance @xmath0 above a critical magnetic field @xmath1 @xcite . the value of @xmath1 decreases for increasingly cleaner samples , indicating that the resistance is not caused by impurity scattering but instead is intrinsic to the state itself @xcite . in quantum hall systems the currents are carried by edge states , so this insulating ground state must correspond to an emergent state that does not support edge currents produced by _ spontaneous _ ( not explicit ) breaking of the su(4 ) symmetry . thus the approximate su(4 ) symmetry of graphene can suggest possible low - energy collective modes exhibiting spontaneously broken symmetry that are important for the properties of graphene in a magnetic field , but the su(4 ) symmetry alone can not determine which of these modes is the ground state . until now , those broken - symmetry modes have been addressed quantitatively by numerical simulations employing limited numbers of states and particles , or by effective low - energy field - theory approximations . these calculations find various possible low - energy collective states resulting from spontaneous breaking of the su(4 ) symmetry with very similar energies . thus they have been unable to give a definitive answer to the nature of the insulating ground state . let us note that other approaches to the problems addressed here have been proposed ( for example , refs . @xcite ) . however , the present discussion will concentrate on methods based on approximate su(4 ) symmetry of the hamiltonian . an alternative and potentially more powerful application of symmetries has been employed extensively in both nuclear structure and condensed matter physics @xcite . fermion dynamical symmetry method _ truncates the hilbert space to a tractable collective subspace by positing a highest symmetry associated with the physical operators for the system , and constructing effective hamiltonians from polynomials in the casimir invariants of the highest symmetry s subgroup chains . in this approach it is possible not only to classify low - energy collective modes , but to solve analytically for the properties of these modes and to determine which lie lowest in energy , either exactly in particular symmetry limits , or approximately using generalized coherent state methods . in an earlier letter the first application of fermion dynamical symmetry methods to graphene was introduced and applied to determining the ground state in strong magnetic fields @xcite . this paper develops the full dynamical symmetry formalism upon which ref . @xcite rests . it will be shown that the highest symmetry is so(8 ) , with its generators identifiable with particle hole and pairing degrees of freedom that have been discussed previously in the physics of graphene . this symmetry will be shown to be isomorphic to an so(8 ) symmetry used extensively in nuclear structure physics , which permits already - developed mathematics to be appropriated for the graphene problem , and suggests instructive physical analogies between two very different physical systems . a generalized coherent state approximation will be introduced that corresponds to a hartree fock - bogoliubov ( hfb ) formalism subject to so(8 ) symmetry constraints . this permits quantitative evaluation of energy surfaces associated with spontaneously - broken symmetry . the so(8 ) highest symmetry will be shown to have an su(4 ) subgroup that recovers the known physics of su(4 ) quantum hall ferromagnetism as a special case , but implies in the more general case new low - energy physics that transcends su(4 ) quantum hall ferromagnetism . hence , a solvable and physically - illuminating approach to the rich low - energy structure of undoped graphene in strong magnetic fields will be proposed that reproduces known physics , but also suggests testable new physics in this complex system . comprehensive reviews of graphene physics may be found in refs . the presentation here will recall only a select set of features that will be relevant for subsequent discussion . undoped graphene is a 2-dimensional semiconductor with zero bandgap . it has a bipartite honeycomb lattice structure illustrated in fig . [ fig : a - b_sublattices ] that corresponds to two interlocking triangular sublattices , labeled a and b. the two - fold degree of freedom specifying whether an electron is on the a sublattice or b sublattice is a spin - like quantum number termed the _ sublattice pseudospin_. the sublattice pseudospin behaves mathematically like the actual spin of the electron ( which will be introduced later ) , but it is a separate degree of freedom . the dispersion of energy with momentum for undoped graphene in the absence of a magnetic field is illustrated in fig . [ fig : dispersiongraphene ] . 0pt 0.44 electronic dispersion of graphene calculated in a tight - binding model with no magnetic field . two inequivalent points in the brillouin zone are labeled @xmath2 and @xmath3 . near these k - points the dispersion becomes linear , leading to _ dirac cones , _ as shown in the expanded view . for undoped graphene the fermi surface lies at the apex of the cones , where the level density vanishes and the effective electronic mass tends to zero . thus low - energy electrons are governed approximately by a dirac equation for massless electrons . the two inequivalent points @xmath2 and @xmath3 are not connected by reciprocal lattice vectors . the corresponding two - fold k degree of freedom may be viewed mathematically as a 2-valued spin - like quantity , commonly termed the _ valley isospin _ because of the valley - like structure in the dispersion of fig . [ fig : dispersiongraphene ] around the k - points . for brevity the valley isospin will sometimes be termed simply _ isospin . _ obviously this two - fold `` isospin '' degree of freedom is a pseudospin that should be distinguished physically from the actual spin of the electron , and from standard usage of isospin quantum numbers in nuclear and particle physics . near the _ dirac cones _ ( inset to fig . [ fig : dispersiongraphene ] ) the dispersion is linear , the density of electronic states tends to zero , and the electrons are described by a massless dirac equation in which the fermi velocity plays the role that the speed of light would play in an actual relativistic system . thus the low - energy electrons for undoped graphene in zero magnetic field behave to good approximation as _ massless chiral fermions , _ with the chirality representing the projection of the sublattice pseudospin on the direction of motion , not the projection of the actual spin . the vanishing of the density of states at the dirac point ( fermi surface ) implies that the transport properties of graphene are different from either a metal or a semiconductor . our interest will be primarily low - energy states in a strong magnetic field . for non - interacting electrons , the quantized levels may be found by solving the dirac equation for massless fermions with a vector potential appropriate to the applied magnetic field . the dispersion of energy with magnetic field strength for massless dirac electrons is illustrated in fig . [ fig : llgraphene_n0](a ) . consider the @xmath4 state , which corresponds to half filling of the fourfold - degenerate @xmath5 ll in graphene , as illustrated in fig . [ fig : llgraphene_n0](b ) . the graphene honeycomb lattice is bipartite , as illustrated in fig . [ fig : a - b_sublattices ] . the @xmath5 ll is located exactly at the dirac point corresponding to @xmath6 . for low - energy excitations in each valley labeled by @xmath2 or @xmath3 , the inter - valley tunneling may be ignored and the electrons in the valley reside entirely on either the a or b sublattice . hence , for the @xmath5 ll labeling with the valley isospin ( indicating whether the electron is in a @xmath2 or @xmath3 valley ) is equivalent to labeling with the sublattice pseudospin ( indicating whether the electron is on the a or b sublattice ) . this is reminiscent of a nel state with spins on two different sublattices , with a nel order defined by the difference in spins on the a and b sublattices . a quantum hall effect is signaled by a plateau in the hall conductance @xmath7 having quantized values @xmath8 where the _ filling factor _ @xmath9 is defined by @xmath10 with @xmath11 the charge - carrier density , @xmath12 the magnetic field strength , and @xmath13 the magnetic flux density in units of the fundamental flux quantum @xmath14 . these plateaus indicate formation of an _ incompressible quantum liquid . _ this is a compact way to say that the ground state is separated from excited states by an _ energy gap , _ which inhibits compression because of the energy required for excitation across the gap . the integral quantum hall effect ( iqhe ) in graphene is similar to the integral quantum hall effect for non - relativistic electrons in that it corresponds to the formation of incompressible states resulting from the complete filling of landau levels by weakly - interacting electrons . however , there are two important differences between the iqhe states observed in graphene and those observed in conventional 2d semiconductor heterostructures : ( 1 ) in addition to the two - fold spin degeneracy ( in the absence of zeeman splitting ) , there is a two - fold valley degeneracy associated with the distinct @xmath2 and @xmath15 points in the first brillouin zone . thus the filling factor changes in steps of four between plateaus in the hall resistance for graphene . ( 2 ) for graphene the filling factor @xmath9 defined in eq . ( [ iqhegraph1.1 ] ) vanishes at the dirac point for particle hole symmetric half filling of the graphene lattice , since the electron density @xmath11 tends to zero there . hence , in the absence of a zeeman effect or strong electronic correlations , there is no integral quantum hall effect in graphene for @xmath16 . ( a ) relativistic dispersion for massless dirac electrons as a function of magnetic field strength @xmath12 . states are labeled by a principle quantum number @xmath17 and a quantum number @xmath18 indicating particle - like ( @xmath19 ) and hole - like ( @xmath20 ) states . each landau level labeled by @xmath17 has a high orbital degeneracy @xmath21 ( see eq . ( [ degen1.1 ] ) ) , and an additional 4-fold degeneracy associated with spin and valley isospin . ( b ) one configuration for occupation of the @xmath5 landau level in monolayer graphene . the splitting and occupation are schematic only . the ground state would be a superposition of such configurations and in the su(4 ) symmetry limit ( which obtains for coulomb - only interactions ) the 4 levels shown labeled by valley @xmath2 or @xmath3 and spin up or down would be degenerate . in realistic cases the splitting often is small , suggesting approximate su(4 ) symmetry . ] in graphene the analog of the integral quantum hall effect was first observed at filling factors @xmath22 by sweeping the field or the carrier density through the landau levels @xcite . this implies hall resistance quantization for filling factors in the sequence @xmath23 where @xmath11 is the charge carrier density and @xmath17 is the landau level index @xcite . this sequence is quite different from the integral quantum hall effect sequence observed in other 2d electron gases . however , it likely results from the same basic physics as the normal iqhe , modified by the 4-fold spin valley degeneracies of non - interacting , massless dirac electrons . the period @xmath24 in eq . ( [ iqhegraph1.2 ] ) is is a consequence of the approximate four - fold degeneracy of the graphene landau level , and the added @xmath25 ( which is not present in non - relativistic 2d systems ) is a berry phase effect that results from the special status of the @xmath5 state for massless dirac fermions : a quantum phase arises at the band degeneracy point associated with precession of the pseudospinor describing the 2-fold sublattice degree of freedom , which modifies the quantization condition for electronic orbits @xcite . the four - fold near - degeneracy of the landau level follows because the zeeman energy is small compared with the interaction energy , and the pseudospin degree of freedom representing the two inequivalent dirac cones at the corners of the brillouin zone ( k and k@xmath26 ) does not couple to external fields if the two sublattices are equivalent . because of the dirac - cone dispersion with the fermi level located at the apex of the cones ( see fig . [ fig : dispersiongraphene ] ) , low - energy excitations in graphene occur in regions of reduced electron density , which disfavors electron correlations . but by placing a strong perpendicular magnetic field on the system the resulting landau quantization ( corresponding semiclassically to requiring that an integral number of debroglie wavelengths wrap around a cyclotron orbit , and implying that an integral number of magnetic flux quanta pass through the area bounded by the orbit ) leads to a bunching of levels into regions of locally high degeneracy separated by gaps . these regions of locally high level density may exhibit conditions more favorable for the development of strong correlations between electrons . landau levels ( ll ) become strongly correlated when inter - ll excitations are of sufficiently high energy that they may be neglected and the low - energy excitations involve only intra - ll transitions . then the kinetic energy of the ll is a constant that can be omitted . this limit of strong electronic correlations has two important physical implications : ( 1 ) the approximate 4-fold spin valley degeneracy of the graphene landau levels leads to quantum hall ferromagnetic states that will be discussed further below . ( 2 ) the strong correlations can produce incompressible states at partial filling of the ll that are reminiscent of the fractional quantum hall effect ( fqhe ) in semiconductor devices . after the discovery of the integral quantum hall effect in graphene , experiments performed at higher magnetic field strengths observed fragile quantum hall states at filling factors @xmath27 . the @xmath28 states are thought to be the result of single - particle zeeman splitting of the landau levels but the states at filling factors @xmath29 are thought to be caused by electron electron interactions breaking degeneracies of the @xmath5 landau level @xcite . quantum hall incompressible states having actual fractional filling were later observed at filling factors such as @xmath30 @xcite . the fractional states follow the standard composite fermion model @xcite sequence for filling factors @xmath31 , but only even - numerator fractions are seen for @xmath32 . these sequences , and the energy gaps for the corresponding incompressible states , are thought to reflect the interplay of strongly - correlated chiral electrons and the characteristic internal symmetries of graphene that will be discussed further below @xcite . as has been seen , in the quantum hall effect for 2d electron gases produced in semiconductor devices , two fundamental classes of incompressible states are found : 1 . those where the requisite gaps are produced by complete filling of landau levels ( fermi energy lying between landau levels ) that are explained by weakly - interacting electrons subject to impurity - scattering localization . 2 . those where the gaps are produced by strong electron correlations within a partially - filled landau level . for the normal semiconductor quantum hall effect , the first case is commonly termed the integral quantum hall effect ( iqhe ) , because it leads uniquely to quantum hall plateaus at integral filling factors , and the second case is termed the fractional quantum hall effect ( fqhe ) , because it is characterized uniquely by quantum hall plateaus at fractional values of the filling factor . in graphene , the anomalous counting implied by the four - fold spin valley degeneracy for massless chiral fermions modifies this correspondence . one may again assume incompressible states divided into those that have gaps produced by weakly - interacting electrons filling landau levels completely and those that have gaps produced by correlations within a partially filled level . however ( 1 ) states of the first class occur at integral fillings but the integers of eq . ( [ iqhegraph1.2 ] ) are not those of the normal iqhe because of the anomalous counting implied by graphene s internal structure . ( 2 ) states of the second class may have either integral or fractional filling factors because of the anomalous counting . thus iqhe and fqhe may be used as shorthand labels in graphene , but it should be understood that what is meant by `` iqhe '' in graphene is the formation of incompressible states by weakly - interacting electrons completely filling landau levels , and by `` fqhe '' the formation of incompressible states by strongly - correlated electrons in a partially - filled level , irrespective of whether the observed filling factors are integers or rational fractions . our primary interest in this paper lies in those incompressible states formed by electron electron and electron lattice correlations in partially - filled landau levels , and thus in graphene `` fqhe '' states . in the normal two - dimensional electron gas produced in semiconductor devices the landau levels can contain @xmath33 states , where @xmath34 is the electronic charge and @xmath12 is the magnetic field . as has been seen , in graphene there is an additional 4-fold degeneracy associated with the spin and valley degrees of freedom . it is common to unite these four degrees of freedom in terms of an su(4 ) symmetry that is termed _ quantum hall ferromagnetism ( qhfm ) . _ su(4 ) symmetry for graphene in a strong magnetic field is expected when all four spin and valley levels are degenerate . two conditions must be satisfied for this condition to be fulfilled . 1 . landau - level mixing caused by inter - ll electronic transitions must be negligible . perturbations within a single ll that break the 4-fold spin isospin symmetry must be small . the resulting theory predicts quantum hall states that have no analog in semiconductor heterostructures . let us study these states by introducing an effective low - energy hamiltonian exhibiting approximate su(4 ) symmetry . the two - dimensional electronic spin degree of freedom and the two - dimensional electronic valley ( k ) degree of freedom are most elegantly expressed in terms of independent spin and valley isospin states . for the spin , introduce the pauli matrix vector @xmath35 , with the standard representation in terms of the @xmath36 matrices @xmath37 which obey the su(2 ) lie algebra @xmath38 } = i\epsilon_{ijk } \frac{\sigma_k}{2 } , \label{su2comm}\ ] ] where @xmath39 is the completely antisymmetric 3rd - rank tensor . the matrices ( [ algebra1.5 ] ) are assumed to operate on a spinor basis of spin - up and spin - down electrons @xmath40 a set of equations in the valley isospin space completely analogous to eqs . ( [ algebra1.5])([algebra1.6 ] ) for the spin space results if one defines the su(2 ) pauli - matrix representation for the valley isospin operators @xmath41 , @xmath42 which operate on the valley isospinor basis @xmath43 the operators @xmath44 and @xmath45 may now be used to define an effective low - energy hamiltonian having landau - level and internal spin and valley isospin degrees of freedom . the two largest energy scales for graphene in a strong magnetic field are the landau - level separation and the coulomb energy . at the charge neutral point ( fermi energy for undoped graphene ) the ll separation is approximately three times larger than the coulomb energy , which is in turn considerably larger than any additional terms in the interaction . therefore , a strategy is adopted here of ignoring excitations between landau levels and projecting onto the @xmath5 ll . at a quantitative level the landau level mixing can not be ignored ( see the discussion in ref . @xcite ) , but such an approximation gives the correct qualitative physics and the effect of excluded landau levels can be included to some degree by parameter renormalization , which will be sufficient for our purposes . within this single landau level the hamiltonian is assumed to be dominated by a long - range coulomb interaction that is su(4 ) symmetric , with shorter - range interactions in spin and valley degrees of freedom ( originating in both screened electron electron interactions and electron phonon interactions ) causing su(4 ) symmetry breaking . to implement this a graphene hamiltonian projected onto the @xmath5 landau level is adopted that was proposed in ref . @xcite and employed further in ref . @xcite,@xmath46 where the valley - independent coulomb interaction @xmath47 may be expressed as @xmath48 @xmath49 is the short - range , valley - dependent interaction , @xmath50 \delta(\bm r_i - \bm r_j ) , \label{so5_1.1c}\ ] ] where the pauli matrices @xmath51 operate on the valley isospin and @xmath52 and @xmath53 are coupling constants , and the zeeman energy @xmath54 is given by @xmath55 where @xmath56 is the bohr magneton , @xmath12 is the magnetic field strength , the pauli matrices @xmath57 operate on the electronic spin degrees of freedom , and the @xmath58 direction for the spin space is chosen to be aligned with @xmath12 . letting @xmath59 and @xmath60 , the set of 15 operators @xmath61 [ algebra1.4 ] is closed under commutation , defining an su(4 ) lie algebra that commutes with the coulomb interaction @xmath47 @xcite . thus , if @xmath49 and @xmath54 are small compared with @xmath47 in eq . ( [ so5_1.1a ] ) , the hamiltonian will have an approximate su(4 ) invariance . the operator @xmath62 represents the total spin and the operator @xmath63 represents the total valley pseudospin . in the @xmath5 landau level for graphene there is an equivalence between valley and sublattice degrees of freedom , so @xmath64 can be viewed as a nel vector in the @xmath5 landau level measuring the difference in spins on the a and b sublattices . the operators @xmath65 coupling spin and valley isospin will be discussed further below . the occurrence of su(4 ) symmetry and its explicit symmetry - breaking pattern depend on the values of the effective coupling parameters @xmath52 and @xmath53 . for the lattice spacings found in graphene , each can be estimated to be considerably smaller than the su(4)-symmetric coulomb term , so one may expect su(4 ) to be broken only weakly by explicit terms in realistic systems . four basic explicit symmetry - breaking patterns have been identified @xcite : 1 . for arbitrary non - zero values of @xmath52 and @xmath53 , the symmetry is broken to @xmath66 where @xmath67 is associated with global conservation of spin and @xmath68 with conservation of its @xmath58 component , and @xmath69 is associated with conservation of the @xmath70 component of the valley isospin . ( conservation of @xmath70 implies physically that the difference in electronic densities between the @xmath2 and @xmath3 sites is invariant , which might be expected to be true for low - energy states having minimal scattering between valleys . ) in the absence of zeeman splitting spin is conserved , but only the @xmath58 component of the valley isospin is conserved . the full hamiltonian ( [ so5_1.1a ] ) including the zeeman term conserves only the @xmath58 components of the spin and valley isospin . if @xmath71 but @xmath72 , the symmetry is broken to @xmath73 in the absence of zeeman splitting this corresponds to conserving independent spin rotations for the wavefunction in each valley labeled by @xmath2 and @xmath3 , but breaking of the valley isospin to u(1)@xmath74 . the full hamiltonian ( [ so5_1.1a ] ) including the zeeman term conserves only the @xmath58 components of the spin in each valley separately , and the @xmath58 component of the valley isospin . 3 . if @xmath75 , the symmetry is broken to @xmath76 corresponding to full spin and valley isospin rotational symmetry in the absence of zeeman splitting . the complete hamiltonian ( [ so5_1.1a ] ) including the zeeman term conserves the su(2 ) isospin symmetry but only the @xmath58 component of the spin . 4 . if @xmath77 , the hamiltonian commutes with @xmath65 , @xmath62 , and @xmath70 , and these 10 operators generate the lie group so(5 ) , so @xcite @xmath78 where the @xmath79 symmetry is generated by @xmath80 . thus , in the absence of zeeman splitting the system exhibits an so(5 ) symmetry involving both spin and valley isospin . the full hamiltonian ( [ so5_1.1a ] ) including the zeeman term conserves the @xmath58 component of spin and the @xmath79 symmetry . the so(5 ) subgroup plays the role of a transitional symmetry connecting the nel - like states associated with @xmath81 and the states associated with valley degrees of freedom @xmath82 and @xmath83 . the subgroup structure for these four patterns of explicit su(4 ) symmetry breaking is illustrated in fig . [ fig : dynamicalchains_qhf].dynamicalchains_qhf figures / dynamicalchains_qhf_vert_compact_nobkg.eps 0pt 0.40 explicit symmetry - breaking structure for su(4 ) quantum hall ferromagnetism described by the hamiltonian ( [ so5_1.1a ] ) . these symmetries and explicitly - broken symmetries have proven extremely useful in understanding the states of graphene in a strong magnetic field @xcite . let us now consider using symmetries in an even more powerful way than that discussed in the preceding section . specifically , let us attempt to describe the quantum hall ferromagnetic behavior of graphene using _ fermion dynamical symmetries _ of an effective hamiltonian operating in a highly - truncated collective subspace . this has the potential to prescribe dynamics as well as taxonomy and conservation laws , within a many - body model having analytical solutions that illuminate the physics of quantum hall states in graphene . for clarity of discussion a valence space will be assumed corresponding to a single landau level . to avoid cluttered notation the index @xmath17 labeling the landau level is suppressed and the fermion creation operator @xmath84 and the corresponding hermitian conjugate @xmath85 are introduced . the index @xmath86 takes one of two values ( @xmath87 ) labeling the valley isospin projections corresponding to valleys @xmath2 or @xmath3 , the electron spin polarization @xmath88 takes one of two values ( @xmath89 ) labeling spin - up or spin - down , and @xmath90 is a quantum number distinguishing degenerate states within a given landau level ( typically an angular momentum in symmetric gauge or a linear momentum in landau gauge ) . the fermionic operators @xmath91 and @xmath92 are assumed to obey @xmath93 where @xmath94 . by virtue of this anticommutator , the @xmath95 create a fermion in the state labeled by @xmath96 , the @xmath97 annihilate a fermion in the same state , and @xmath98 counts the number of fermions in the state labeled by @xmath96 . the four states representing possible combinations of @xmath86 and @xmath88 are displayed in table [ tb : quantumnumbermapping ] , ccccc ' '' '' valley & @xmath86 & @xmath88 & @xmath99 & @xmath100 + ' '' '' @xmath2 & @xmath19 & @xmath101 & @xmath102 & @xmath103 + [ ] @xmath2 & @xmath19 & @xmath104 & @xmath105 & @xmath106 + [ ] @xmath3 & @xmath20 & @xmath101 & @xmath107 & @xmath108 + [ ] @xmath3 & @xmath20 & @xmath104 & @xmath109 & @xmath110 + and correspond physically to the four possible combinations of the electron being in either the @xmath2 or @xmath3 valley with either spin up or spin down ( see fig . [ fig : llgraphene_n0]b ) . for brevity the label @xmath111 displayed in the last column of the table will often be used to distinguish these states . the four basis states labeled by @xmath100 are illustrated graphically in fig . [ fig : graphenebasis ] . figures / graphenebasis.eps 0pt 0.16 the four isospin spin basis vectors @xmath112 of table [ tb : quantumnumbermapping ] . table [ tb : quantumnumbermapping ] also displays a unique mapping of these four states to a label @xmath99 that takes the four possible projection quantum numbers @xmath113 of a fictitious angular momentum @xmath114 ; the motivation for this mapping will become apparent later . let us now introduce an operator @xmath115 that creates a pair of electrons , one in the @xmath116 level and one in the @xmath117 level , with the total @xmath90 of each pair coupled to zero term by term , @xmath118 and its hermitian conjugate @xmath119 , which annihilates a corresponding electron pair . each index @xmath100 or @xmath120 ranges over four values , implying 16 components in eq . ( [ algebra1.1 ] ) . however , the pair wavefunction must satisfy the pauli principle which , upon expanding the indices @xmath100 and @xmath120 using table [ tb : quantumnumbermapping ] , eliminates the four diagonal @xmath121 possibilities . furthermore , because of the antisymmetry requirement the pair creation operators are constrained by @xmath122 implying that only half of the remaining 12 operators @xmath123 are independent . thus eq . ( [ algebra1.1 ] ) defines six independent operators @xmath123 , with six independent hermitian conjugates @xmath124 . let us introduce in addition to these pairing operators the 16 particle hole operators @xmath125 through @xmath126 where @xmath127 is the kronecker delta and @xmath128 is the total degeneracy the single landau level ( see eq . ( [ degen1.2 ] ) ) . the commutators for the 28 operators @xmath124 , @xmath123 , and @xmath12 are found to be @xcite @xmath129 } & = -b_{db } \delta_{ac } - b_{ca}\delta_{bd } + b_{cb } \delta_{ad } + b_{da } \delta_{bc } \label{algebra1.3a } \\ { \left[b_{ab } , b_{cd}\right ] } & = \delta_{bc } b_{ad } - \delta_{ad}b_{cb } \label{algebra1.3b } \\ { \left[b_{ab } , a^\dagger_{cd}\right ] } & = \delta_{bc } a^\dagger_{ad } + \delta_{bd } a^\dagger_{ca } \label{algebra1.3c } \\ { \left[b_{ab } , a_{cd}\right ] } & = -\delta_{ac } a_{bd } - \delta_{ad } a _ { cb } , \label{algebra1.3d}\end{aligned}\ ] ] [ algebra1.3 ] which is isomorphic to an so(8 ) lie algebra . thus , the 28 members of the operator set @xmath130 exhibit so(8 ) symmetry under commutation . the ultimate justification for introducing this so(8 ) lie algebra will lie in the results to which it will lead . however , a general discussion of why the known physics of graphene suggests the efficacy of such an algebra in understanding the possible collective modes of the system may be found in appendix [ h : appendixa ] . the su(4 ) generators ( [ algebra1.4 ] ) may be expressed in terms of the operators ( [ algebra1.2 ] ) by employing eqs . ( [ algebra1.5])([algebra1.7e ] ) in eq . ( [ algebra1.4 ] ) and utilizing the equivalences in table [ tb : quantumnumbermapping ] to expand the indices for @xmath125 from eq . ( [ algebra1.2 ] ) . for example , consider the spin operator @xmath131 . from eq . ( [ algebra1.4a ] ) one may write @xmath132 & \left . \qquad\qquad\qquad -i c^\dagger_{-\uparrow m_k } c_{-\downarrow m_k}{^{\vphantom{\dagger}}}+ i c^\dagger_{-\downarrow m_k } c_{-\uparrow m_k}{^{\vphantom{\dagger}}}\right ) \\ & = -ib_{12 } + i b_{21 } -i b_{34 } + i b_{43}.\end{aligned}\ ] ] by such methods one finds that the spin operators ( [ algebra1.4a ] ) may be expressed in terms of the @xmath125 generators of so(8 ) as @xmath133 [ algebra1.8 ] the valley isospin operators ( [ algebra1.4b ] ) as @xmath134 [ algebra1.9 ] the nel vector of eq . ( [ algebra1.4c ] ) as @xmath135 [ algebra1.10 ] and the operators @xmath65 of eq . ( [ algebra1.4d ] ) as @xmath136 [ algebra1.11 ] the inverse transformations expressing the @xmath125 in terms of the @xmath137 are given in appendix [ h : appendixb ] . hence the su(4 ) algebra generated by the operators in eq . ( [ algebra1.4 ] ) is a subalgebra of the so(8 ) algebra , with its generators corresponding to particular linear combinations of the subset of so(8 ) generators defined by the particle hole operators @xmath125 in eq . ( [ algebra1.2 ] ) . in fact , the present formalism also could be constructed by starting with the su(4 ) algebra generated by the @xmath125 operators and expanding that to the so(8 ) algebra of eq . ( [ algebra1.3 ] ) by adding pairing operators , using the motivation discussed in appendix [ h : appendixa ] . one wishes to investigate possible collective modes for graphene electrons undergoing strong correlations within a single landau level . pairs of fermions often afford a convenient basis for discussing collective states , so let us consider some possible configurations involving pairs of electrons in graphene . let us first consider the degeneracies of undoped graphene placed in a strong magnetic field , confining the discussion to the case of a single landau level for simplicity . the single - particle states within the landau level will be assumed labeled by the quantum numbers @xmath138 , where @xmath17 is the principle quantum number labeling the landau level and @xmath139 is a quantum number distinguishing the degenerate states within the landau level . in the absence of spin and valley degrees of freedom , the states @xmath138 of the landau level are assumed to hold a maximum of @xmath140 electrons ( for consistency with earlier applications of the so(8 ) algebra in nuclear physics , @xmath141 will be defined to be the electron pair degeneracy , so that @xmath140 is the electron degeneracy ) . from the solution of the dirac equation in a magnetic field @xmath142 where @xmath12 is the strength of the magnetic field , @xmath143 is the area of the two - dimensional sample and @xmath144 defines the magnetic flux quantum . but graphene has in addition @xmath145 internal degrees of freedom associated with the @xmath146 space . thus there are four copies of each landau level in graphene and the total electron degeneracy @xmath147 is given by @xmath148 some pair degeneracies calculated from eq . ( [ degen1.2 ] ) as a function of domain size and magnetic field strength are displayed in table [ tb : degeneracytable ] , cccc ' '' '' domain size & @xmath149 & @xmath150 & @xmath151 + ' '' '' @xmath152 & @xmath153 & @xmath154 & @xmath155 + [ ] @xmath156 & @xmath157 & @xmath158 & @xmath159 + where it is assumed that the collective wavefunction is delocalized over the entire domain size . the _ fractional occupation _ @xmath160 of the single landau level [ not to be confused with the filling factor @xmath9 given in eqs . ( [ iqhegraph1.1 ] ) and ( [ fillingfactor ] ) ] may be defined as @xmath161 where @xmath17 is the number of electrons and @xmath162 is the number of electron pairs . for half - filling of the @xmath5 landau level located at the fermi surface ( corresponding to the ground state of undoped graphene ) the electron number @xmath163 is then @xmath164 these degeneracies and occupation numbers are just the standard results for relativistic landau levels in a 2d electron gas subject to a strong perpendicular magnetic field , but modified by the graphene spin and valley degeneracies . graphene exhibits both integral and fractional quantum hall effects but the filling factors are anomalous relative to those for standard quantum hall effects in semiconductor heterostructures . this is because of 1 . the 4-fold degeneracy associated with the spin and valley degrees of freedom , which introduces factors of four in the counting . 2 . the nature of the dirac solution , illustrated in fig . [ fig : llgraphene_n0 ] , for which the negative - energy solutions may be interpreted as electron holes , the positive - energy solutions as electrons , and the @xmath5 level is unique , being half - filled in the neutral ground state ( equivalently , it may be thought of as being shared equally by particles and holes ) . because of the particle - hole symmetry , the charge carriers change sign near the dirac points and the hall conductivity vanishes at charge neutrality ( the electron number density tends to zero at a dirac point ) . for this reason , the filling factor for graphene must be defined _ relative to the charge - neutral state . _ at charge neutrality the @xmath5 landau level is half filled and when the @xmath5 ll is completely full the filling factor is @xmath165 , from eq . ( [ iqhegraph1.2 ] ) . the quantum hall filling factor @xmath9 may be related to the landau level fractional occupation @xmath160 employed in the present formalism by @xmath166 therefore , half filling of the @xmath5 landau level corresponds to a fractional occupation @xmath167 but to a filling factor @xmath4 , @xmath168 corresponds to @xmath169 ( completely empty ) , @xmath170 corresponds to @xmath171 filling , @xmath172 corresponds to @xmath173 filling , and @xmath174 corresponds to @xmath175 ( completely full ) . consider the states created by repeated application of the pair creation operator @xmath115 defined by eq . ( [ algebra1.1 ] ) to the pair vacuum . it is useful to classify states according to a seniority - like quantum number @xmath176 defined to be the number of particles in the system not coupled to one of the pairs defined in eq . ( [ algebra1.1 ] ) . the @xmath177 subspace will be of particular interest since it will contain states of maximal collectivity with respect to the pairs ( [ algebra1.1 ] ) . an @xmath178-pair state in the @xmath177 subspace is given by @xmath179 where the total pair number @xmath178 is @xmath180 with @xmath17 giving the total number of electrons and @xmath181 giving the number of electron pairs created by @xmath182 operating on the vacuum state . for our discussion here it will always be assumed that one is dealing with @xmath177 states , corresponding physically to no broken pairs . from eq . ( [ algebra1.3 ] ) , the 16 operators @xmath125 are closed under commutation and form a @xmath184 subalgebra of the so(8 ) algebra ( [ algebra1.3 ] ) . let us now investigate the irreducible representations ( irreps ) that are associated with the @xmath185 subgroup chain in the @xmath177 representations . for @xmath177 at half filling , the number of pairs is @xmath186 and the highest - weight u(4 ) representation is given by @xmath187 . let us define a highest - weight ( hw ) state in the @xmath177 space , and choose it to correspond to the pair state with maximal value of @xmath99 from table [ tb : quantumnumbermapping ] , which results from placing one electron in the @xmath188 state and one electron in the @xmath189 state . thus , for @xmath190 pairs the highest weight state is given by @xmath191 where the sum runs over the @xmath192 states in the landau level labeled by the @xmath193 quantum number . the other states of the irreducible representation may then be created by the cartan dynkin algorithm , which consists of using raising and lowering operators in the weight space to construct successively all the other states beginning with the highest weight state @xcite . the state in eq . ( [ hwstate ] ) appears to have a quite complex form , involving a sum with number of terms equal to the pair degeneracy of the landau level raised to a power equal to the pair degeneracy ( with the pair degeneracy typically a large number ) . however , the actual structure of this state is considerably simpler than eq . ( [ hwstate ] ) would suggest because of the pauli principle . as an illustrative example of this assertion , let s construct explicitly the highest - weight state for the case @xmath194 , corresponding to @xmath195 pairs in a single landau level . writing the sum over @xmath196 in eq . ( [ hwstate ] ) out term by term gives @xmath197 where in raising the sum of operators inside the parentheses to the @xmath198 power , all products containing two or more creation operators with the same index vanish because of the pauli principle . similar considerations apply for arbitrary values of @xmath199 and in general the highest - weight state is given by @xmath200 where the simplification in going from the second to the third line is a consequence of the antisymmetry of the fermion creation operators ( the pauli principle ) implied by eq . ( [ anticommutator ] ) . thus the highest - weight state is a product state of pairs , one pair for each of the @xmath190 levels labeled by @xmath90 in the landau level . by the cartan dynkin algorithm , other states in the @xmath177 subspace can be constructed by applying successively to the highest - weight state appropriate lowering and raising operators . these will be functions of the generators @xmath125 , so for an arbitrary state @xmath201 in the weight space one has schematically @xmath202 where the function @xmath203 is specified by the cartan dynkin procedure . as an example , consider the action of the valley isospin lowering operator @xmath204 on the highest - weight state . from eqs . ( [ algebra1.9 ] ) and ( [ algebra1.2 ] ) , @xmath205 thus the state @xmath201 created by applying @xmath204 to @xmath206 is @xmath207 c_{{\scriptscriptstyle k\uparrow } m_k}^\dagger c_{{\scriptscriptstyle k\downarrow } m_k}^\dagger \ , { |0\rangle } \nonumber \\ & = \prod_{m_k } \left ( c_{3 m_k}^\dagger c_{2 m_k}^\dagger + c_{4 m_k}^\dagger c_{1 m_k}^\dagger \right ) { |0\rangle } , \label{u0pairs1.7}\end{aligned}\ ] ] where the simplifications are because the only terms that survive correspond to those where an annihilation operator in a factor inside the square brackets is exactly balanced by a creation operator from the factor outside the square brackets . likewise , the other states of the @xmath177 representation can be constructed by using successive applications of raising and lowering operators fashioned from the generators defined in in eqs . ( [ algebra1.8])([algebra1.11 ] ) . from the preceding discussion , for @xmath208 the states may be written as @xmath209 { |0\rangle } , \label{upairs1.8}\end{aligned}\ ] ] where @xmath210 denote valley isospin projection quantum numbers and @xmath211 denote spin projection quantum numbers . this is the same form as the most general collective pair state used by kharitonov @xcite in his classification of possible broken symmetry states for the @xmath5 landau level in graphene ( see eq . ( [ basis1.0 ] ) in the appendix ) . thus , for undoped graphene the general pairing wavefunction ( [ u0pairs1.0 ] ) characteristic of the so(8 ) @xmath183 su(4 ) dynamical symmetry is in fact _ equivalent to the product form ( [ upairs1.8 ] ) _ employed in standard discussion of quantum hall ferromagnetism , for which the summations are over the internal @xmath212 rather than landau @xmath213 degrees of freedom . the equivalence of eqs . ( [ upairs1.8 ] ) and ( [ u0pairs1.0 ] ) , despite their superficially very different forms , is a fundamental consequence of the pauli principle acting in the collective fermionic pair subspace , which greatly restricts allowed pair configurations . the equivalence established above implies that the present so(8 ) formalism can be used to derive the framework used in ref . @xcite to describe possible collective states in graphene , which establishes an essential connection between the standard discussion of su(4 ) quantum hall magnetism in graphene and the present more general so(8 ) dynamical symmetry formalism . as will be demonstrated further below , this permits the so(8 ) formalism to ( 1 ) encompass the established physics of su(4 ) quantum hall ferromagnetism , ( 2 ) extend quantum hall ferromagnetism to suggest possible additional collective graphene physics beyond the su(4 ) limit , and ( 3 ) provide _ analytical solutions _ for the states suggested by spontaneous breaking of su(4 ) symmetry that can not be obtained using su(4 ) symmetry and must be addressed numerically , or with effective field theory approximations , within standard su(4 ) quantum hall ferromagnetism . furthermore , it will be seen below that the equivalence of product and paired forms for the wavefunction implies a deep formal connection between the collective states resulting from strong electron correlations in graphene landau levels and the collective states produced by strong nucleon correlations in nuclear structure physics , and a suggestive formal connection to the properties of the strongly - correlated electrons responsible for high - temperature superconductivity . the preceding discussion has established that the fermion dynamical symmetry method applied to undoped graphene in a strong magnetic field has one dynamical symmetry chain @xmath185 that recovers exactly su(4)-symmetric quantum hall ferromagnetism . since the @xmath125 operators introduced in eq . ( [ algebra1.2 ] ) form an su(4 ) subgroup of so(8 ) that is in one - to - one correspondence with the operators used to formulate the effective low - energy hamiltonian ( [ so5_1.1a ] ) , [ h : fdsgraphene ] implies that all of the physics associated with this effective hamiltonian that has been discussed in the prior literature ( see @xcite and references cited therein ) is implicit in the present formalism . furthermore , the discussion of [ h : pairrepresentations ] shows that the pair basis ( [ u0pairs1.0 ] ) of the truncated collective subspace for the so(8 ) fermion dynamical symmetry is in fact identical to the most general wavefunction ( [ upairs1.8 ] ) that has been proposed @xcite for collective states breaking the su(4 ) symmetry , despite its superficially very different form . thus the hilbert - space truncation implied by the collectively - paired so(8 ) subspace ( [ u0pairs1.0 ] ) recovers the understanding in the existing literature of the classes of states to be expected from spontaneous breaking of the su(4 ) symmetry by valley - dependent correlations . however , the existing discussions of these collective states in terms of broken su(4 ) symmetry have been largely qualitative , and have turned to numerical simulations to discuss the actual structure and energy of the states . it will now be demonstrated that the present formalism is capable not only of classifying , but also of addressing the quantitative nature of those collective states _ in analytical fashion . _ furthermore , it will be shown that the so(8 ) highest symmetry implies subgroup chains in addition to @xmath214 that are associated with spontaneous breaking of the symmetry by correlations and have not been discussed in the previous literature and that may play a role in graphene . let us begin that discussion by first transforming to a more convenient representation of the so(8 ) generators . this new representation will be physically equivalent to the original representation , but will offer some advantages in interpretation , and will expose an unexpected relationship between graphene physics and that of a very different field , nuclear structure physics . for the pair creation operators defined in eq . ( [ algebra1.1 ] ) , each electron creation operator @xmath91 carries both spin and valley isospin ; hence the products @xmath215 correspond to a clebsch gordan series representing sums of terms having different values of total spin and total isospin . likewise , in the particle hole operators of eq . ( [ algebra1.2 ] ) each creation operator @xmath91 and annihilation operator @xmath92 carries spin and isospin , so the product @xmath216 in eq . ( [ algebra1.2 ] ) represents a superposition of states carrying different total spin and total valley isospin . these representations with indefinite spin and isospin will be termed _ uncoupled representations . _ on physical grounds , the spin is expected to be conserved ( if the zeeman term in the hamiltonian is neglected ) and the valley isospin is expected to be approximately conserved for low - energy excitations . thus , it is desirable to use the uncoupled representation of the pairing and particle hole operators to construct new _ coupled representations _ that have good total spin and good total valley isospin quantum numbers for bilinear operators . using standard angular momentum coupling theory @xcite , an electron pair creation operator coupled to good spin and valley isospin may be defined by @xmath217 where @xmath143 is the total spin of the pair with @xmath218 its projection , @xmath219 is the total valley isospin of the pair with @xmath220 its projection , and @xmath221 are clebsch gordan coefficients for the angular momentum sum @xmath222 that couple the pair to good total spin or total valley isospin . antisymmetry implies that the the pair wavefunction can have only @xmath223 , @xmath224 ; or @xmath225 , @xmath226 ( spin - triplet , isospin - singlet ; or spin - singlet , isospin - triplet pairs ) . explicitly the possibilities are @xmath227 { a_{\scriptscriptstyle 00}^{\raisebox{0.2em}{$\scriptstyle\dagger$ } { \scriptscriptstyle 0 1 } } } = a_{14}^\dagger + a_{23}^\dagger \quad { a_{\scriptscriptstyle 01}^{\raisebox{0.2em}{$\scriptstyle\dagger$ } { \scriptscriptstyle 0 1 } } } = \sqrt 2 a_{13}^\dagger \qquad { a_{\scriptscriptstyle 0 - 1}^{\raisebox{0.2em}{$\scriptstyle\dagger$ } { \scriptscriptstyle 0 1 } } } = \sqrt2 a_{24}^\dagger , \end{array } \label{coupled1.3}\ ] ] with the hermitian conjugates of eq . ( [ coupled1.3 ] ) giving the six corresponding pair annihilation operators in coupled representation . it is useful to define an alternative set of six coupled pairing operators @xmath228 and @xmath229 according to @xmath230 d_0^\dagger = \tfrac{1}{\sqrt2 } \,{a_{\scriptscriptstyle 00}^{\raisebox{0.2em}{$\scriptstyle\dagger$ } { \scriptscriptstyle 0 1 } } } = \tfrac{1}{\sqrt2 } \left ( a_{14}^\dagger + a_{23}^\dagger \right ) \\[7pt ] d_1^\dagger = \tfrac{1}{\sqrt2 } \ , { a_{\scriptscriptstyle 01}^{\raisebox{0.2em}{$\scriptstyle\dagger$ } { \scriptscriptstyle 0 1 } } } = a_{13}^\dagger \quad d_{-1}^\dagger = \tfrac{1}{\sqrt2 } \,{a_{\scriptscriptstyle 0 - 1}^{\raisebox{0.2em}{$\scriptstyle\dagger$ } { \scriptscriptstyle 0 1 } } } = a_{24}^\dagger \\[7pt ] d_2^\dagger = \tfrac{1}{\sqrt2 } \ , { a_{\scriptscriptstyle 10}^{\raisebox{0.2em}{$\scriptstyle\dagger$ } { \scriptscriptstyle 1 0 } } } = a^\dagger_{12 } \quad d_{-2}^\dagger = \tfrac{1}{\sqrt2 } \,{a_{\scriptscriptstyle -10}^{\raisebox{0.2em}{$\scriptstyle\dagger$ } { \scriptscriptstyle 1 0 } } } = a_{34}^\dagger \end{array } \label{coupled1.4}\ ] ] and the six corresponding hermitian conjugates @xmath143 and @xmath231 . the physical meaning of these pairs may be deduced by constructing the corresponding electronic configurations . consider @xmath232 . from eqs . ( [ coupled1.4 ] ) and ( [ algebra1.1 ] ) @xmath233 where in the last step the correspondence between the index @xmath234 and the valley ( @xmath2 or @xmath3 ) and spin ( @xmath235 ) labels in table [ tb : quantumnumbermapping ] has been invoked . this implies that @xmath236 creates a state with one spin - up and one spin - down electron on each equivalent site @xmath2 in the brillouin zone , as illustrated schematically in fig . [ fig : actionddagger2 ] . figures / actionddagger2.eps 0pt 0.35 the action of the pair creation operator @xmath237 on the vacuum state is to create a charge density wave with a spin - singlet pair on each site @xmath2 and no electrons on the @xmath3 sites . this is a component of a lattice - scale charge density wave , since the charge differs by two electronic units between adjacent sites . likewise , one finds that @xmath238 creates a charge density wave as in fig . [ fig : actionddagger2 ] , but with the spin - singlet pairs on the @xmath15 sites . the pair configurations produced by all generators of eq . ( [ coupled1.4 ] ) operating on the pair vacuum @xmath239 are summarized in fig . [ fig : s_d_pairs_brillouin ] . figures / s_d_pairs_brillouin_withq.eps 0pt 0.30 configurations created by the @xmath228 and @xmath240 pair creation operators of eq . ( [ coupled1.4 ] ) , and for the linear combinations @xmath241 of eq . ( [ qdef ] ) , operating on the pair vacuum @xmath239 . the upper left box illustrates the valley isospin labeling . a total of six valence electrons as assumed to be distributed on sites in the brillouin zone . in each pair configuration , location of the dots ( @xmath2 or @xmath3 site ) indicates the valley isospin , spin - up electrons are indicated by up arrows , and spin - down electrons are indicated by down arrows . also shown are the configurations generated by the linear combinations @xmath242 which will be useful in later discussion . kharitonov has given a general classification of low - lying collective modes for the @xmath5 landau level of graphene in terms of collective pairs @xcite . the collective pairs created by the so(8 ) pair generators in fig . [ fig : s_d_pairs_brillouin ] are similar physically to the pairs identified by kharitonov , as will now be described . ( 1 ) the configuration generated by @xmath243 is to the difference of two terms , each with alternating spin - up and spin - down on adjacent sites , implying that all spins on the a sublattice ( identified with valley @xmath2 ) point in one direction and all spins on the b sublattice ( identified with valley @xmath15 ) point in the opposite direction . each term corresponds to a spin density wave ( af order ) , with a nel vector defined by the difference in total spins on the two sublattices serving as an order parameter , but because of the difference of the two terms the net af order for this configuration is zero ( see [ sh : orderparameters ] and table [ tb : orderparms ] ) . ( 2 ) the configurations generated by @xmath244 are spin - singlet charge density waves , with alternating charges of two and zero units on adjacent sites . an appropriate order parameter is the difference in charge between the a and b sublattices . ( 3 ) the configurations generated by @xmath245 have one spin on each site , all pointing in the same direction ; this is a ferromagnetic state , with the net spin as an order parameter . ( 4 ) the configuration generated by @xmath246 is the same as that generated by @xmath243 , except for a positive sign for the second term . this also implies alternating spins on adjacent sites and af order for each term , but the total af order vanishes because of the contribution of the two terms . ( 5 ) the configurations corresponding to @xmath247 are states with af order characterized by the difference in spins on the two sublattices labeled by @xmath2 and @xmath3 . thus the coupled - representation pairs carrying good spin and valley isospin quantum numbers in eq . ( [ coupled1.4 ] ) represent physical degrees of freedom already discussed in the literature as candidate collective modes representing spontaneous breaking of the su(4 ) graphene symmetry by interactions in a single partially - filled landau level . it is desirable to express the particle hole generators of eq . ( [ algebra1.2 ] ) in coupled representation . let us begin by introducing a set of operators @xmath248 with the definition @xmath249 where @xmath250 and @xmath251 take the values of the fictitious angular momentum projection @xmath99 in table [ tb : quantumnumbermapping ] , providing a labeling equivalent to that of @xmath100 and @xmath120 in @xmath125 , with @xmath250 or @xmath251 values @xmath252 mapping to @xmath100 or @xmath120 values @xmath253 , respectively . for example , from table [ tb : quantumnumbermapping ] , @xmath254 and @xmath255 label the same quantity , which is defined in eq . ( [ algebra1.2 ] ) . from the selection rules for angular momentum coupling in eq . ( [ multipole1.1 ] ) , the index @xmath256 can take the values @xmath257 , with @xmath258 projections @xmath259 for each possibility , which gives a total of 16 operators @xmath260 . by inserting the explicit values of the clebsch gordan coefficients the 16 independent @xmath261 may be evaluated in terms of the @xmath125 . @xmath262 p^1_0 & = \sqrt{\tfrac{9}{20}}(b_{11 } - b_{44 } ) + \sqrt{\tfrac{1}{20}}(b_{22 } -b_{33 } ) \nonumber \\ & = \sqrt{\tfrac{9}{20}}(n_1 - n_4 ) + \sqrt{\tfrac{1}{20}}(n_2 -n_3 ) \nonumber \\[2pt ] p^1_1 & = -\sqrt{\tfrac{3}{10 } } \,b_{12 } - \sqrt{\tfrac{4}{10 } } \,b_{23 } -\sqrt{\tfrac{3}{10}}\ , b_{34 } \nonumber \\[2pt ] p^1_{-1 } & = \sqrt{\tfrac{3}{10 } } \,b_{21 } + \sqrt{\tfrac{4}{10 } } \,b_{32 } + \sqrt{\tfrac{3}{10}}\ , b_{43 } \nonumber \\[2pt ] p^2_0 & = \tfrac{1}{2 } ( b_{11 } - b_{22 } + b_{44 } - b_{33 } ) \nonumber \\ & = \tfrac12(n_1 - n_2 + n_4 - n_3 ) \nonumber \\[2pt ] p^2_1 & = \tfrac{1}{\sqrt{2 } } ( b_{34 } - b_{12 } ) \qquad p^2_{-1 } = \tfrac{1}{\sqrt2 } ( b_{21 } - b_{43 } ) \label{pofb } \\[2pt ] p^2_2 & = -\tfrac{1}{\sqrt{2 } } ( b_{13 } + b_{24 } ) \qquad p^2_{-2 } = -\tfrac{1}{\sqrt2 } ( b_{31 } + b_{42 } ) \nonumber \\[2pt ] p^3_0 & = \sqrt{\tfrac{1}{20 } } ( b_{11 } - b_{44 } ) + \sqrt{\tfrac{9}{20 } } ( b_{33}-b_{22 } ) \nonumber \\ & = \sqrt{\tfrac{1}{20 } } ( n_1 - n_4 ) - \sqrt{\tfrac{9}{20 } } ( n_2 - n_3 ) \nonumber \\[2pt ] p^3_1 & = -\sqrt{\tfrac15 } b_{12 } + \sqrt{\tfrac35 } b_{23 } - \sqrt{\tfrac15 } b_{34 } \nonumber \\[2pt ] p^3_{-1 } & = \sqrt{\tfrac15 } b_{21 } - \sqrt{\tfrac35 } b_{32 } + \sqrt{\tfrac15 } b_{43 } \nonumber \\[2pt ] p^3_2 & = \sqrt{\tfrac12 } ( b_{24 } - b_{13 } ) \qquad p^3_{-2}= \sqrt{\tfrac12 } ( b_{42 } - b_{31 } ) \nonumber \\[2pt ] p^3_3 & = -b_{14 } \qquad p^3_{-3 } = b_{41 } , \nonumber\end{aligned}\ ] ] where the @xmath263 are number operators for each of the four states and the total particle number @xmath17 is the sum over the four states labeled by @xmath100 in table [ tb : quantumnumbermapping ] , @xmath264 it will be convenient notationally to sometimes replace the operator @xmath265 with the operator @xmath266 , according to the relationship @xmath267 where @xmath147 is the degeneracy of the space for the particles that participate in the so(8 ) symmetry . physically @xmath268 is one half the particle number measured from half filling ( which corresponds to @xmath269 ) . because the six operators defined by eq . ( [ coupled1.4 ] ) , their six hermitian conjugates , and the 16 operators defined by eq . ( [ multipole1.1 ] ) are independent linear combinations of the so(8 ) generators defined in eqs . ( [ algebra1.1 ] ) and ( [ algebra1.2 ] ) , the 28 operators @xmath270 also close an so(8 ) algebra under commutation . the so(8 ) commutation relations for the coupled representation @xmath271 are given explicitly by @xcite @xmath272 } & = -2s_0 \label{commcoupled1.1 } \\ { \left[d_{\mu ' } , d^\dagger_{\mu}\right ] } & = -2\delta_{\mu\mu ' } s_0 + \sum_{t { \rm\ odd } } ( -1)^{\mu ' } \nonumber \\ & \qquad \times \ { c_{2 , \scriptscriptstyle-\mu ' 2 \mu}^{\,t,\scriptscriptstyle\mu-\mu ' } } { \left\ { \begin{matrix } 2&2&t\\#4&\tfrac32&\tfrac32 \end{matrix}\right\ } } p_{\scriptscriptstyle\mu,-\mu'}^t \label{commcoupled1.2 } \\ { \left[d_\mu^\dagger , s\right ] } & = p_\mu^2 \label{commcoupled1.3 } \\ { \left[p^r_\mu , s^\dagger\right ] } & = 2 \delta_{r2}d^\dagger_\mu + 2 \delta_{r0 } \delta_{\mu0 } s^\dagger \label{commcoupled1.4 } \\ { \left[p_{\mu'}^r , d_\mu^\dagger\right ] } & = 2(-1)^{\mu ' } \delta_{r2 } \delta_{-\mu\mu ' } -4\sqrt{5(2r+1 ) } \nonumber \\ & \qquad\times\ { c_{r \scriptscriptstyle\mu ' 2 \mu}^{2,\scriptscriptstyle\mu+\mu ' } } { \left\ { \begin{matrix } 2&2&r\\#4&\tfrac32&\tfrac32 \end{matrix}\right\ } } d_{\scriptscriptstyle\mu+\mu'}^\dagger \label{commcoupled1.5 } \\ { \left[p^r_{\mu ' } , p^s_\mu\right ] } & = 2 ( -1)^{r+s } \sqrt{(2r+1)(2s+1 ) } \,\,\sum_t { c_{r { \scriptscriptstyle\mu ' } s { \scriptscriptstyle\mu}}^{\,t , { \scriptscriptstyle \mu+\mu ' } } } \nonumber \\ & \times \left[1-(-1)^{r+s+t}\right ] { \left\ { \begin{matrix } r&s&t\\#4&\tfrac32&\tfrac32 \end{matrix}\right\ } } p^{\,t}_{\scriptscriptstyle\mu+\mu ' } \label{commcoupled1.6}\end{aligned}\ ] ] [ commcoupled ] where @xmath266 is defined in eq . ( [ p0n ] ) and @xmath273 denotes the wigner 6-@xmath274 symbol @xcite for the recoupling of three angular momenta to good total angular momentum . the action of the so(8 ) pair creation operators on the pair vacuum @xmath178 times creates a @xmath275-particle state @xcite , @xmath276 where the total number of pairs is @xmath277 . the portion of the full hilbert space that is spanned by the states ( [ collective1.1 ] ) will be termed the _ collective subspace . _ it will play an important role in subsequent discussion where it will be shown that the so(8 ) symmetry may be used to construct effective hamiltonians that are diagonal in this space , and that the generators of so(8 ) do not couple the subspace to the remainder of the space . the reason for our alternative labeling of the states in table [ tb : quantumnumbermapping ] in terms of the index @xmath99 , and our particular choices of phases and normalizations in equations , can now be made clear . with these labelings and choices the six coupled particle particle operators @xmath228 and @xmath278 defined in eq . ( [ coupled1.4 ] ) , their six hermitian conjugates @xmath143 and @xmath231 , and the 16 coupled particle hole operators @xmath279 defined in eq . ( [ multipole1.1 ] ) , are mathematically in one - to - one correspondence with the 28 generators for the ginocchio so(8 ) model @xcite and the so(8 ) fermion dynamical symmetry model @xcite . these have found broad application in nuclear structure physics @xcite and may be viewed as a microscopic justification for the interacting boson model ( ibm ) @xcite , which is one of the most commonly used phenomenological models in nuclear structure physics . this correspondence has three important implications : ( 1 ) mathematically , the group - theoretical methodology obtained for so(8 ) already in nuclear physics applications may be appropriated for use in the graphene problem . ( 2 ) physically the nature of the generators for the nuclear physics and graphene so(8 ) symmetries are fundamentally different , but analogs of physical interpretations applied already for nuclear physics so(8 ) symmetries may shed light on the graphene problem . ( 3 ) philosophically , the so(8 ) correspondence between graphene and nuclear structure physics implies a satisfying convergence of mathematical reasoning and physical abstraction in two completely different scientific subfields . this convergence will be elaborated further in [ h : analogy_su4 ] . the so(8 ) group has various subgroups ( subsets of generators closed under commutation ) and these in turn may have other subgroups . these sequences of subgroups define subgroup chains . these chains will be discussed first in terms of the nuclear physics basis @xmath280 , and then in terms of a new basis that is mathematically equivalent but is physically better suited to describing the physics of graphene . in the nuclear physics basis @xmath281 is the particle number and generates a group @xmath282 , while @xmath283 is proportional to the total angular momentum and generates a group @xmath284 . in the nuclear physics context the total angular momentum and the particle number are expected to be conserved exactly for all physical states . thus one seeks subgroup chains of so(8 ) that end in the subgroup @xmath285 corresponding to charge and angular momentum conservation . three so(8 ) subgroup chains satisfy these conditions . from eq . ( [ commcoupled1.1 ] ) the quasispin generators @xmath288 close an @xmath289 algebra that is a subalgebra of so(8 ) , and from eq . ( [ commcoupled1.4 ] ) the operators @xmath260 with @xmath290 close an so(5 ) algebra and commute with these su(2 ) quasispin generators . thus , one subgroup of so(8 ) is @xmath291 where so(5 ) is generated by the 10 operators @xmath292 , where @xmath259 takes the @xmath258 values @xmath293 and the _ quasispin group _ @xmath289 is generated by @xmath294 , where @xmath295 . furthermore , the three generators @xmath296 are components of the total angular momentum @xmath297 and generate an @xmath284 subgroup of so(5 ) , and @xmath266 generates a @xmath282 subgroup of @xmath289 corresponding to conservation of charge . thus one subgroup chain is @xmath298 & \supset \underset{\scriptscriptstyle\ { p^1\ } } { { { \rm so(3)}}_l}\times\underset{\scriptscriptstyle \ { s_0 \ } } { { { { \rm u(1)}}_{{\scriptstyle\rm c } } } } , \label{chains1.2n}\end{aligned}\ ] ] where the generators of each subgroup are indicated in brackets below the subgroup and the product group on the last line corresponds to conservation of total angular momentum and particle number . the groups su(4 ) and so(6 ) share the same lie algebra . in nuclear physics it is more common to refer to this group as so(6 ) , but to maintain a parallel with the ensuing treatment of graphene it will be labeled su(4 ) in the present discussion . from eq . ( [ commcoupled1.6 ] ) , the 16 operators @xmath300 are closed under commutation , corresponding to @xmath301 where the generator of @xmath282 is @xmath302 and the 15 operators @xmath303 are the generators of su(4 ) . furthermore , the subset of @xmath304 with odd @xmath256 are generators of the so(5 ) symmetry discussed above and so generate an so(5 ) subgroup of this su(4 ) group . hence a second subgroup chain is @xmath305 & \supset \underset{\scriptscriptstyle \ { p^1 , \ , p^3\}}{{{\rm so(5 ) } } } \times \underset{\scriptscriptstyle \ { s_0\}}{{{{\rm u(1)}}_{{\scriptstyle\rm c } } } } \supset \underset{\scriptscriptstyle\ { p^1\ } } { { { \rm so(3)}}_l}\times\underset{\scriptscriptstyle \ { s_0 \ } } { { { { \rm u(1)}}_{{\scriptstyle\rm c } } } } , \label{chains1.3n}\end{aligned}\ ] ] where eq . ( [ p0n ] ) has been used to replace @xmath306 with @xmath266 . from eqs . ( [ commcoupled1.2 ] ) , ( [ commcoupled1.5 ] ) , and ( [ commcoupled1.6 ] ) , the 21 operators @xmath307 close an @xmath308 subalgebra of so(8 ) and the subset @xmath309 closes an @xmath310 subalgebra of so(7 ) . thus a third subgroup chain is given by @xmath311 the relationships of the nuclear analog so(8 ) subgroup chains described above are summarized in fig . [ fig : dynamicalchains ] . nuclear analog so(8 ) subgroup chains with generators in the coupled representation @xmath312 given by eqs . ( [ multipole1.1 ] ) and ( [ coupled1.4 ] ) and obeying the lie algebra of eq . ( [ commcoupled ] ) . generators are indicated below each group factor and @xmath266 and @xmath313 may be interchanged using eq . ( [ p0n ] ) . the subgroup structure expressed in this basis is in one - to - one correspondence with the so(8 ) fermion dynamical symmetry model @xcite of nuclear structure physics . however , for the description of graphene it is more useful to transform this basis to the new one employed in fig . [ fig : dynamicalchains_graphene ] using eqs . ( [ ngx ] ) , which gives a more direct physical interpretation of quantities important in graphene physics . ] the subgroup chains in [ h : subgroupchainsnuclear ] were expressed in terms of the nuclear physics basis @xmath314 . this basis demonstrates the deep connection between graphene quantum hall physics and nuclear structure physics , and is suitable mathematically to describe graphene quantum hall effects . however , it is not well suited physically to interpreting the graphene quantum hall effects for three reasons . ( 1 ) the relationship between the generators of the @xmath185 subgroup in the nuclear physics basis and those of the su(4 ) quantum hall ferromagnetism basis defined in eq . ( [ algebra1.4 ] ) is not clear , which hinders interpretation of the present results in terms of preceding results found in the graphene literature . ( 2 ) charge and electronic spin are expected to be conserved in graphene ( if the zeeman term is neglected ) , but none of the generators @xmath315 can be interpreted as spin in the application to graphene ( instead , spin is a linear combination of these generators ) . for physical applications it is desirable to employ a basis in which the relevant conservation laws are manifest . ( 3 ) in addition to the exact conservation laws expected for charge and spin in graphene , it is expected on physical grounds that for low - energy excitations the scattering between valleys is strongly suppressed and the difference in electron densities between neighboring valleys should be very nearly conserved . this difference is expressed by the @xmath58 component of the valley isospin @xmath70 , and the corresponding approximate invariance is reflected in a @xmath316 symmetry generated by @xmath70 . but @xmath70 is not proportional to any of the @xmath304 generators ( it is a linear combination of these generators ) , so this approximate invariance is not manifest in the nuclear su(4 ) basis . thus a new basis will be employed for the so(8 ) generators in application to graphene for which the particle number ( charge ) operator @xmath17 or @xmath266 and the 12 pairing operators @xmath317 are retained , but the 15 su(4 ) generators @xmath318 in the nuclear representation are replaced with the 15 su(4 ) generators @xmath319 with @xmath320 defined in the graphene representation given in eq . ( [ algebra1.4 ] ) , @xmath321 \ { { \mathscr{s}}_\alpha,\ , t_\alpha,\ , n_\alpha,\ , { \pi_{\alphax } } , \,{\pi_{\alphay } } , \ , s_0 , \ , s , \,s^\dagger , \,d{^{\vphantom{\dagger}}}_{\mu},\ , d^\dagger_\mu \ } { _ { \mbox{\scriptsizegraphene so(8 ) } } } . \end{array}\ ] ] the transformation from the @xmath322 generators to the @xmath323 generators is given in appendix [ h : appendixb ] . in the new basis it will be convenient to take as order parameters @xmath324 [ orderp1.1 ] where @xmath325 is the number operator counting particles in basis state @xmath326 and the expectation value is taken with respect to the collective wavefunction . physically ( 1 ) the net spin is measured by @xmath327 , which characterizes _ ferromagnetic _ order . ( 2 ) the difference in charge between the a and b sublattices is measured by @xmath328 , which characterizes _ charge density wave _ order . ( 3 ) the difference in spins between the a and b sublattices is measured by @xmath329 , which characterizes _ antiferromagnetic _ ( nel or spin density wave ) order . the order parameters evaluated for the states of fig . [ fig : s_d_pairs_brillouin ] are displayed in table [ tb : orderparms ] . lccc ' '' '' state & @xmath327 & @xmath328 & @xmath329 + ' '' '' @xmath330 & @xmath331 & @xmath331 & @xmath331 + [ ] @xmath332 & @xmath331 & @xmath333 & @xmath331 + [ ] @xmath334 & @xmath331 & @xmath103 & @xmath331 + [ ] @xmath335 & @xmath333 & @xmath331 & @xmath331 + [ ] @xmath336 & @xmath103 & @xmath331 & @xmath331 + [ ] @xmath337 & @xmath331 & @xmath331 & @xmath331 + [ ] @xmath338 & @xmath331 & @xmath331 & @xmath103 + [ ] @xmath339 & @xmath331 & @xmath331 & @xmath333 + in the new basis it will be assumed that both charge and spin are exactly conserved for the physical states of the model in the absence of the zeeman term @xmath54 , that the charge and the @xmath58-component of spin are exactly conserved if the zeeman term is included in the hamiltonian , and that @xmath70 is conserved , where appropriate . neglecting the zeeman term , the spin charge symmetry corresponds to a group structure @xmath340 , where @xmath341 is generated by the spin operators and @xmath282 is generated by the particle number operator . thus one seeks subgroup chains of so(8 ) that end in the subgroup @xmath340 corresponding to charge and spin conservation , and in some of these chains a u(1)@xmath74 subgroup implying conservation of @xmath70 will also be required . the group and subgroup structure in the new basis is illustrated in fig . [ fig : dynamicalchains_graphene ] , where seven nontrivial subgroup chains may be identified that begin with so(8 ) and end with the symmetry @xmath342 corresponding to conservation of charge and @xmath58-component of the spin in the magnetic field . @xmath343 [ groupchains ] where for brevity all u(1 ) factors are dropped in the notation and su(2 ) means su(2)@xmath344 corresponding to conservation of spin . each of these corresponds to a different dynamical symmetry that is realized for particular choices of parameters for the so(8 ) hamiltonian , and that yields _ exact manybody solutions _ using the dynamical symmetry methodology . let us now discuss in more detail three subgroup chains of so(8 ) . the quasispin generators @xmath345 close an @xmath289 algebra that is a subalgebra of so(8 ) , and the operators @xmath346 close an so(5 ) algebra and commute with these su(2 ) quasispin generators . putting this together , one subgroup of so(8 ) is @xmath347 furthermore , the three generators @xmath348 are components of the total spin and generate an @xmath341 subgroup of so(5 ) , and @xmath266 generates a @xmath282 subgroup of @xmath289 corresponding to conservation of charge . thus one subgroup chain is @xmath349 & \supset { { { \rm su(2)}}_{\sigma}}\times{{{\rm u(1)}}_{{\scriptstyle\rm c } } } , \label{chains1.2}\end{aligned}\ ] ] where the product group on the last line corresponds to conservation of spin and charge . this subgroup chain with its corresponding generators is illustrated in fig . [ fig : dynamicalchains_graphene ] . alternatively , so(5 ) may be broken according to the pattern @xmath350 & \supset { { { \rm su(2)}}_{\sigma}}\times { { { \rm u(1)}}_{{\scriptstyle\rm c } } } , \label{chains1.2b}\end{aligned}\ ] ] which also conserves spin and charge , and is illustrated in fig . [ fig : dynamicalchains_graphene ] . a @xmath351 subgroup of so(8 ) may be obtained by removing the 12 pairing operators from the so(8 ) generator set . the @xmath352 subgroup is generated by the particle number ( charge ) and the su(4 ) subgroup is generated by the 15 remaining operators , which are defined in eq . ( [ algebra1.4 ] ) in the current basis . there are several options for chains corresponding to further subgroups . ( 1 ) the subset @xmath353 defines generators of the so(5 ) symmetry discussed above and so forms an so(5 ) subgroup of this su(4 ) group . hence one su(4 ) subgroup chain is @xmath354 which is displayed in fig . [ fig : dynamicalchains_graphene ] . ( 2 ) physically , the total spin is conserved . if there is little inter - valley scattering one may also assume the spin within each @xmath2 valley and each @xmath3 valley to be separately conserved , corresponding to a @xmath355 symmetry . thus a second su(4 ) subgroup chain corresponds to @xmath356 & \supset { { { \rm su(2)}}_{\sigma}}\times { { { \rm u(1)}}_{{\scriptstyle\rm c } } } , \label{chainy1.1}\end{aligned}\ ] ] where @xmath316 is generated by @xmath70 . this chain also is displayed in fig . [ fig : dynamicalchains_graphene ] . ( 3 ) finally , one can imagine that su(4 ) is broken into subgroups corresponding to simultaneous conservation of both spin and valley isospin , giving a third su(4 ) subgroup chain @xmath357 as illustrated in fig . [ fig : dynamicalchains_graphene ] . comparing fig . [ fig : dynamicalchains_graphene ] with fig . [ fig : dynamicalchains_qhf ] , it is apparent that the three @xmath358 subgroup chains defined in this section correspond to the three symmetry - breaking patterns described in [ h : symmetriesfqhegraphene ] and discussed in refs . @xcite for the su(4 ) quantum hall ferromagnetism model . the portion of the so(8 ) subgroup structure leading to su(4 ) quantum hall ferromagnetism is indicated by the shaded boxes with dashed outer boundaries in fig . [ fig : dynamicalchains_graphene ] . thus it is seen explicitly that the special case corresponding to the @xmath358 subgroup chains of the present model imply the results of refs . @xcite . the 21 operators @xmath359 close an @xmath360 subalgebra of so(8 ) , and the subset of generators @xmath361 close an @xmath362 subalgebra of so(7 ) . thus a third subgroup chain is given by @xmath363 as illustrated in fig . [ fig : dynamicalchains_graphene ] . this subgroup chain is of particular interest because it will define a _ critical dynamical symmetry _ that represents an entire phase exhibiting critical behavior and interpolating between two other phases . let us use the subgroup structure of the preceding section to obtain exact solutions of the correlated many - body problem in these dynamical symmetry limits . the basic idea is to use the casimir invariants of the subgroup chains like those described in [ h : subgroupchainsgraphene ] and illustrated in fig . [ fig : dynamicalchains_graphene ] to label states . then model hamiltonians constructed only from the casimir invariants of a single subgroup chain permit analytical solution of the effective schrdinger equation in that symmetry limit @xcite . specifically , if a hamiltonian @xmath364 can be expressed as a function of the casimir invariants of some subgroup chain @xmath365 , where the @xmath366 represent casimir operators for the groups @xmath367 , then the system is said to possess a dynamical symmetry associated with the subgroup chain . the discussion will be simplified by restricting to the lowest - order casimir invariant for each group , which corresponds physically to omitting @xmath17-body interactions with @xmath368 . elementary properties of lie groups then permit the eigenvalues @xmath369 and eigenfunctions @xmath370 of this hamiltonian to be expressed in closed form as @xmath371 where the @xmath372 stand for the quantum numbers required to specify the irreducible representations ( irreps ) of the groups @xmath367 . the physical properties of the corresponding states can then be elucidated by using the methods of lie groups and lie algebras to evaluate matrix elements for observables . in this way , one generally finds that the dynamical symmetries associated with subgroup chains of some highest symmetry define collective ( emergent ) states that correspond to particular patterns of spontaneous symmetry breaking . in terms of the generators ( [ coupled1.4 ] ) , ( [ multipole1.1 ] ) , and ( [ algebra1.4 ] ) , the quadratic casimir operator @xmath373 for the so(5 ) subgroup is @xmath374 where @xmath375 . the corresponding casimir operator for the @xmath376 subgroup is @xmath377 the quadratic casimir operator for the so(7 ) subgroup is @xmath378 the casimirs for the su(2 ) subgroups are @xmath379 and for the u(1 ) charge subgroup the invariant operator is trivially the single generator , @xmath380 or some power of it . finally , the quadratic casimir operator for the full so(8 ) group may be expressed as @xmath381 the casimir operators that appear in each of these subgroups chains and the relevant quantum numbers labeling the states for each dynamical symmetry are summarized in table [ tb : so8subgroups ] . llllll ' '' '' group & dim & generators & quantum numbers@xmath382 & casimir operators @xmath383 & casimir eigenvalues@xmath384 + ' '' '' @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath390 + [ ] @xmath391 & @xmath392 & @xmath393 or & @xmath394 & @xmath395 & @xmath396 + [ ] & & @xmath397 & & & + [ ] @xmath398 & @xmath399 & @xmath400 or & @xmath401 & @xmath402 & @xmath403 + [ ] & & @xmath404 & & @xmath405 & + [ ] u(4 ) & @xmath406 & @xmath407 & @xmath408 & @xmath409 & @xmath410 + [ ] @xmath411 & @xmath412 & @xmath413 or & @xmath414 & @xmath415 & @xmath416 + [ ] & & @xmath417 & & @xmath418 & + [ ] @xmath289 & @xmath108 & @xmath419 & @xmath9 & @xmath420 & @xmath421 + [ ] @xmath341 & @xmath108 & @xmath422 & @xmath423 & @xmath424 & @xmath425 + [ ] @xmath426 & @xmath108 & @xmath63 & @xmath86 & @xmath427 & @xmath428 + [ ] @xmath282 & @xmath103 & @xmath266 & @xmath17 & @xmath266 & @xmath429 + + + + + + + + in the next section these results for the casimir operators will be used to construct the most general hamiltonian permitted in the truncated collective space for specific dynamical symmetries . as has been seen , in a particular dynamical symmetry limit the most general hamiltonian can be constructed from a sum of casimir invariants for the groups contained in the corresponding subgroup chain . for so(8 ) dynamical symmetries the most general hamiltonian in the absence of the zeeman term is represented by the linear combination @xmath430 where @xmath431 is assumed constant in the symmetry limit , the casimir operators @xmath373 have been summarized in the preceding section , and @xmath432 does not appear explicitly because it has been eliminated by the constraint @xcite : @xmath433 hamiltonians representing specific dynamical symmetry limits then correspond to particular choices of the coefficients @xmath434 in this general expression . it may be shown that the most general hamiltonian can also be expressed in the compact form ( see eqs . ( 4.1 ) of ref . @xcite ) @xmath435 where @xmath436 is assumed constant in a symmetry limit and where the different dynamical symmetry limits correspond to specific choices for the values of the parameters @xmath437 , @xmath438 , and @xmath439 . the last term is expressed in terms of the @xmath304 from the nuclear basis . it can be converted to the graphene basis by inverting eqs . ( [ ngx ] ) of appendix [ h : appendixb ] to solve for the @xmath304 . let us now discuss each of the so(8 ) dynamical symmetries and their physical interpretations . for brevity , let us refer to 1 . the dynamical symmetry structure associated with eqs . ( [ chains1.2 ] ) and ( [ chains1.2b ] ) as the _ @xmath440 dynamical symmetry , _ 2 . the dynamical symmetry structure associated with eqs . ( [ chains1.3])([chainy1.2 ] ) as the _ su(4 ) dynamical symmetry , _ and 3 . the dynamical symmetry structure associated with eq . ( [ chains1.4 ] ) as the _ so(7 ) dynamical symmetry . _ initially the role of the zeeman term ( which would break the full spin symmetry down to conservation of its @xmath58 component ) will be ignored and it will be assumed that the chains end in the subgroup @xmath441 corresponding to the physical requirement that spin and charge be conserved exactly . the dynamical symmetry chains given in eqs . ( [ chains1.2])([chains1.2b ] ) and illustrated in fig . [ fig : dynamicalchains_graphene ] correspond to two alternative ways of choosing subgroups of @xmath442 : @xmath443{45}{$\supset$}}}$ } \ { { { \rm su(2)}}_{\sigma}}\times { { \rm su(2)}}{_{\mbox{\scriptsizep}}}\ \raisebox{-0.8ex}{${\mathbin{\rotatebox[origin = c]{-45}{$\supset$}}}$ } \\ { { \rm so(5)}}\times{{\rm su(2)}}{_{\mbox{\scriptsizep } } } \hphantom{{\mathbin{\rotatebox[origin = c]{45}{$\supset$}}}\ { { { \rm su(2)}}_{\sigma}}\times{{{\rm su(2)}}_{\tau}}{\mathbin{\rotatebox[origin = c]{-45}{$\supset$ } } } } \ \ { { { \rm su(2)}}_{\sigma}}\times { { { \rm u(1)}}_{{\scriptstyle\rm c}}}\\ \hphantom{{{\rm so(5)}}\times{{{\rm su(2)}}_{\tau}}\ } \raisebox{1.0ex}{${\mathbin{\rotatebox[origin = c]{-45}{$\supset$}}}$ } \makebox[\widthof{$\ { { { \rm su(2)}}_{\sigma}}\times{{{\rm su(2)}}_{\tau}}\ $ } ] [ c]{${{\rm so(5)}}\ \times \ { { { \rm u(1)}}_{{\scriptstyle\rm c}}}$ } \raisebox{1.0ex}{${\mathbin{\rotatebox[origin = c]{45}{$\supset$}}}$ } . \end{array } \label{so5dyn1.1}\ ] ] in the upper branch of the middle step the so(5 ) symmetry is broken to its @xmath444 spin subgroup , with @xmath289 unbroken . physically this corresponds to conservation of the spin associated with the @xmath445 angular momentum algebra and the pseudospin associated with the @xmath446 pair algebra , but not the full so(5 ) symmetry . in the lower branch of the middle step , the so(5 ) symmetry remains intact and @xmath143-pair pseudospin @xmath444 is broken to @xmath447 charge . in the final subgroup of both chains , only the spin and charge remain as conserved quantities . the hamiltonian in the @xmath448 dynamical symmetry limit corresponds to eq . ( [ mostgeneralham ] ) with the restriction that @xmath449 , @xmath450 where terms that are constant within a given representation have been omitted . the most general so(8 ) state in the @xmath177 collective subspace is given by eq . ( [ collective1.1 ] ) and corresponds to a superposition of @xmath143 and @xmath231 pairs . schematically , @xmath451 with @xmath228 and @xmath452 defined in eq . ( [ coupled1.4 ] ) and @xmath453 , where @xmath178 is the total pair number . on the other hand , the most general states corresponding to the various subgroup chains illustrated in fig . [ fig : dynamicalchains_graphene ] correspond to pair superpositions having specific constraints on the relative contribution of @xmath143 and @xmath454 pairs . the collective subspace for the @xmath455 subgroup of so(8 ) is of the form ( see eq . ( 9.13 ) of ref . @xcite ) @xmath456 implying that it is a superposition of @xmath143 pairs . conceptually , the wavefunction of the @xmath455 subgroup chain for @xmath177 is obtained from the most general state in the collective subspace by converting all of its @xmath454 pairs to @xmath143 pairs . as was seen in [ sh : coupledpairrep ] and fig . [ fig : s_d_pairs_brillouin ] , the @xmath143 and @xmath454 pairs correspond to coherent superpositions of particular electronic distributions in spin and valley pseudospin . thus , specific @xmath143 and @xmath454 pair content for so(8 ) dynamical symmetry subgroup chains implies specific collective modes associated with coherent distribution of the electrons in spin and valley space . it has been noted above that in the @xmath457 dynamical symmetry limit the ground states correspond to a superposition of @xmath143 pairs . the nature of this collective state may be inferred from fig . [ fig : s_d_pairs_brillouin ] and is illustrated in fig . [ fig : s_d_pairs_brillouin_so5 ] . the collective state corresponding to the @xmath458 ground state . the state is an antisymmetric combination of a spin density wave with spin - up on the a lattice sites and spin - down on the b sites , and a spin density wave with spin - up on the b sites and spin - down on the a sites . ] the su(4 ) dynamical symmetry corresponds to the three so(8 ) subgroup chains @xmath459 that were introduced in eqs . ( [ chains1.3])([chainy1.2 ] ) and fig . [ fig : dynamicalchains_graphene ] . as already noted , these three dynamical symmetry chains are in one to one correspondence with the explicit symmetry breaking patterns that have been identified for su(4 ) quantum hall ferromagnetism . the most general su(4 ) wavefunction for @xmath178 pairs in the @xmath177 collective so(8 ) subspace is given by @xcite @xmath460^p { |0\rangle } , \label{su4pairstate}\ ] ] which corresponds physically to a restriction of the general so(8 ) wavefunction ( [ so8pairstate ] ) to a specific superposition of @xmath143 and @xmath454 pairs . the wavefunction for the @xmath362 subgroup of su(4 ) is given by eq . ( [ so5pairstate ] ) . the wavefunction of the parent su(4 ) group is a superposition of @xmath143 and @xmath454 pairs but the @xmath448 subgroup has a wavefunction containing only @xmath143 pairs . the so(7 ) dynamical symmetry corresponds to the so(8 ) subgroup chain @xmath461 that was introduced in eq . ( [ chains1.4 ] ) and displayed in fig . [ fig : dynamicalchains_graphene ] . the hamiltonian in the so(7 ) dynamical symmetry limit corresponds to eq . ( [ mostgeneralham ] ) with the restriction that @xmath462 , @xmath463 where terms have been dropped that are constant within a given representation . from the nuclear physics analog so(8 ) symmetry @xcite , one may surmise that so(7 ) will play the role of a _ critical dynamical symmetry _ interpolating smoothly between the collective states corresponding to the @xmath411 dynamical symmetry and the collective states corresponding to @xmath398 dynamical symmetry . such critical dynamical symmetries have been discussed previously in both nuclear physics @xcite and for the strongly - correlated electrons leading to cuprate and iron - based high - temperature superconductivity @xcite . they may be viewed as the generalization of a quantum critical point to an entire _ quantum critical phase , _ and may represent a fundamental organizing principle for quantum critical behavior . the physical implications of this so(7 ) critical dynamical symmetry for graphene quantum hall physics will be discussed further below . the dynamical symmetry limits discussed above represent special solutions resulting from particular choices of the coupling parameters appearing in the hamiltonian . for arbitrary choices of the coupling parameters the solutions will correspond generally to superpositions of the different symmetry - limit solutions and will not have exact analytical forms . in this more general case it is quite feasible to obtain solutions numerically , since the collective subspace is highly truncated relative to the full hilbert space . however , there is a powerful alternative approach : the _ generalized coherent state approximation , _ which permits _ analytical solutions _ for arbitrary choices of the coupling parameters in the hamiltonian . for the so(8 ) lie algebra introduced in this paper for graphene the gilmore perelomov algorithm @xcite may be implemented to obtain solutions in terms of a set of generalized coherent states . these solutions represent the most general hartree fock bogoliubov theory that can be formulated in the space , subject to a dynamical symmetry constraint @xcite . the solutions of this symmetry - constrained hartree fock bogoliubov ( schfb ) theory correspond to determining the stable points of energy surfaces , which represent the coherent - state expectation values of the effective hamiltonian on the coset space . thus the coherent state solutions also represent a microscopically - derived implementation of ginzburg landau theory . these coherent state solutions are uniquely well suited to study the interplay of competing spontaneous symmetry breaking in determining the ground state of the system and its properties . the coherent states associated with the full set of subgroup chains in fig . [ fig : dynamicalchains_graphene ] will be discussed in future work . in this paper the power of the method will be illustrated succinctly by restricting to the coherent states associated with the subgroup chains of so(8 ) that contain the so(5 ) subgroup , as illustrated in fig . [ fig : dynamicalchains_coherentso5 ] . 0pt 0.20 subgroup chains included in the present coherent state analysis . thus the corresponding coherent state solutions will represent a superposition of the symmetry - limit solutions for the @xmath464 { { \rm so(8)}}\supset{{\rm su(4)}}\supset{{\rm so(5)}}\qquad { { \rm so(8)}}\supset{{\rm so(7)}}\supset{{\rm so(5)}}\end{array}\ ] ] dynamical symmetries . these coherent state solutions will be seen to have the following properties : 1 . the @xmath391 dynamical symmetry will play the role of a critical dynamical symmetry interpolating between @xmath411 and @xmath398 symmetry - limit solutions . because of fundamental symmetries obeyed by the wavefunction , the coherent state solutions may be parameterized in terms of a single collective parameter @xmath465 that governs the mixture of the @xmath143 and @xmath454 pairs defined in eq . ( [ coupled1.4 ] ) contributing in the ground state . the collective parameter @xmath465 may also be interpreted physically in terms of the pair configurations displayed in fig . [ fig : s_d_pairs_brillouin ] . the so(8 ) coherent states corresponding to the symmetry structure in fig . [ fig : dynamicalchains_coherentso5 ] have been developed previously in ref . @xcite for nuclear physics applications and will be adapted extensively to development of the present formalism . let us now consider the energy surfaces that may be computed from the coherent states , which link the so(8 ) solutions to ginzburg landau theory . within the coherent state formalism the ground state energy may be determined through the variational requirement @xmath466 where @xmath467 is the coherent state , @xmath468 is the so(8 ) hamiltonian , @xmath469 and the coefficients @xmath437 and @xmath470 are functions of the effective interaction . for the dynamical symmetry chains having so(5 ) as a subgroup the energies take the general form @xcite @xmath471 , \label{generalenergyformula}\ ] ] where the group - dependent parameters @xmath472 , @xmath473 , @xmath474 , @xmath475 , and @xmath476 are given in table [ tb : stiffnessparameters ] . cccccc ' '' '' g & @xmath472 & @xmath124 & @xmath477 & @xmath478 & @xmath479 + ' '' '' su(2 ) & @xmath437 & @xmath480 & @xmath481 & @xmath482 & @xmath483 + [ ] so(5 ) & @xmath484 & @xmath485 & @xmath486 & @xmath487 & @xmath488 + [ ] su(4 ) & @xmath489 & @xmath490 & @xmath491 & @xmath492 & @xmath331 + [ ] so(7 ) & @xmath438 & @xmath493 & @xmath494 & @xmath495 & @xmath496 + our primary interest in this discussion is in the ground state properties of graphene in a strong magnetic field . the ground states in the coherent state approximation at fixed @xmath497 will be given by those values of @xmath498 that correspond to minima of the energy surface @xmath499 . these are determined by the values of @xmath465 satisfying @xmath500 evaluating these constraints for the energy surfaces ( [ generalenergyformula ] ) , one finds that the minima @xmath501 are given by @xcite @xmath502 the coherent state energy surfaces for the @xmath448 , so(7 ) , and su(4 ) symmetry limits computed from eq . ( [ generalenergyformula ] ) using the entries in table [ tb : stiffnessparameters ] are shown as functions of @xmath465 for several values of the fractional occupation @xmath160 in fig . [ fig : coherentenergysurfaces_graphene ] . coherent state energy surfaces as a function of the order parameter @xmath465 for three of the so(8 ) dynamical symmetry limits . the curves are labeled by the fractional occupation @xmath503 defined in eq . ( [ fillingfraction ] ) , where @xmath17 is the particle number and @xmath147 is the maximum number of particles that can be accommodated in the landau level . the formalism is particle hole symmetric so curves for fractional occupations with @xmath504 are equivalent to those shown above , but with the fractional occupation counted in terms of number of holes . for example , the curves for @xmath505 and @xmath506 are equivalent . for the direct product group @xmath507 the energy surface is the sum of contributions to eq . ( [ generalenergyformula ] ) from su(2 ) and so(5 ) . as will be demonstrated later , the su(2 ) contribution typically dominates that of so(5 ) . ] there one sees that indeed the minima for the @xmath448 and so(7 ) limits are at @xmath508 , and the minima for the su(4 ) limit are at @xmath509 . although the minimum energies for both the @xmath448 and so(7 ) limits are consistent with @xmath510 , fig . [ fig : coherentenergysurfaces_graphene ] shows that these symmetries differ fundamentally in the localization of the minimum . for @xmath448 the energy surface has a deep minimum at @xmath510 but for so(7 ) the energy surface is very flat around @xmath508 , with a broad range of @xmath465 giving essentially the same ground state energy . this highly - degenerate so(7 ) ground state has significant physical implications that will be discussed further below . the so(8 ) generalized coherent state is equivalent to the hartree fock bogoliubov ( hfb ) approximation subject to a symmetry constraint . since hfb is a bcs - type approximation married to a hartree fock mean field , its solutions correspond to symmetry - breaking intrinsic states . in particular , the bcs - like state conserves the physical particle number only on average , and the hartree fock mean field may break both translational and rotational invariance . let us address these issues for the so(8 ) coherent state . the fractional uncertainty in electron number @xmath511 for the so(8 ) coherent state is given by @xcite @xmath512 where @xmath513 is the value of @xmath465 at the minimum energy , given by eq . ( [ betaminima ] ) in the symmetry limits . expressing eq . ( [ particlenumberfluctuation ] ) in terms of the fractional occupation @xmath514 , in the @xmath448 and su(4 ) limits , respectively , one obtains @xmath515_{{\rm so}_5\times{{\rm su(2 ) } } } \hspace{-5pt } = \sqrt{\frac{1-{f}}{{f}\omega } } \qquad \left[\frac{\delta n}{n}\right]_{{\rm su}_4 } \hspace{-5pt } = \sqrt{\frac{1 - 2{f}}{{f}\omega } } , \label{particlenumberfluctuation3}\ ] ] from these results one may notice two important things . 1 . the fluctuation in particle number is large at low degeneracy @xmath128 but decreases with increasing @xmath128 . if su(4 ) symmetry is realized @xmath516 decreases with increasing @xmath160 and _ vanishes identically _ at @xmath167 for any @xmath128 , which corresponds to the fractional occupation for the ground state of undoped graphene . thus it is expected that the current theory applied to graphene has negligible particle number fluctuation @xmath516 in the su(4 ) limit . in the @xmath448 limit the particle number fluctuation @xmath516 remains finite for all @xmath160 but it becomes very small as @xmath128 becomes large , particularly near @xmath517 . thus it too may be neglected in the large-@xmath128 limit . comparison with table [ tb : degeneracytable ] suggests that graphene quantum hall experiments involve sufficient degeneracy that particle number fluctuation in the coherent state solution is not significant . the coherent state approximation represents a mean field localized in spatial position and orientation , so it violates translational and rotational invariance . however , because the crystal is generally macroscopic , the net violation of these symmetries may be expected to be negligible . one concludes that for applications of coherent state methods to graphene , violations of particle number conservation , rotational invariance , and translational invariance are negligible in realistic systems . the generalized coherent state method has been used above to calculate total energy surfaces for quantum hall states in graphene , but one also may use the coherent state wavefunctions and appropriate operators to calculate matrix elements of other relevant observables . this section addresses the nature of the wavefunction and the matrix elements that can serve as order parameters . a significant consequence of the so(8 ) dynamical symmetry structure displayed in fig . [ fig : dynamicalchains_coherentso5 ] is that the phases may be distinguished in terms of a single order parameter and its fluctuations , which may be taken to be @xmath465 . let us now characterize in more depth the physical meaning of this order parameter . in [ sh : orderparameters ] an antiferromagnetic order parameter @xmath329 was defined . in the coherent state approximation the onset of af order is signaled by an energy - surface minimum at a finite value of @xmath465 . because @xmath518 [ see eq . ( [ ngx9 ] ) ] , the antiferromagnetic order parameter @xmath329 is related to the coherent state af order parameter @xmath465 by the intrinsic state matrix element of @xmath519 @xcite @xmath520 where @xmath489 is the coupling strength for the @xmath521 term in the hamiltonian . each value of @xmath465 corresponds to a unique value of @xmath329 , so @xmath465 is a measure of antiferromagnetic order . the location of the maxima may be obtained by setting the derivative with respect to @xmath465 of eq . ( [ intrinsic1.1 ] ) equal to zero , which yields that @xmath522 for a given @xmath17 occurs at a @xmath465 of @xmath523 where eq . ( [ betaminima ] ) was used to make the identification in the last step . thus , if @xmath524 the maximum value of @xmath329 maps to a value of @xmath465 that corresponds to a minimum of the energy surface ( a ground state ) in the su(4 ) limit . substituting eq . ( [ betaminima ] ) for @xmath465 in eq . ( [ intrinsic1.1 ] ) , for @xmath524 ground states the af order parameter @xmath329 depends on the electron number @xmath17 as @xmath525 the so(8 ) model is particle hole symmetric so @xmath17 or @xmath526 count electrons up to half filling and holes for greater than half filling . hence the maximum af collectivity occurs for half filling of the single valence landau level . as was discussed in [ h : conservationlaws ] , the coherent state wavefunction corresponding physically to @xmath527 pairs conserves particle number only on average and so is a superposition of terms having different numbers of pairs . in eq . ( 5.27 ) of ref . @xcite the so(8 ) coherent state is decomposed into terms of definite pair number @xmath528 according to @xmath529 where @xmath530 denotes an intrinsic state with order parameter @xmath465 and closed forms for @xmath531 and @xmath532 are given in ref . @xcite . according to eq . ( 5.28 ) of ref . @xcite , the values of @xmath533 that correspond to the minima of the potential energy surface at @xmath534 for the @xmath448 limit and @xmath535 for the @xmath411 limit [ see eq . ( [ betaminima ] ) ] , respectively , are @xmath536 so the so(8 ) coherent state wavefunction ( [ coherentfn1.1 ] ) in the @xmath448 and su(4 ) limits , respectively , becomes @xmath537 \displaystyle { |{\rm su}_4\rangle } = \sum _ { p } { c_p}\left ( s^\dagger \pm d_0^\dagger \right)^p { |0\rangle } = 2\sum _ { p } { c_p}\left ( q_\pm^\dagger \right)^p { |0\rangle } , \end{array } \label{coherentfn1.4}\ ] ] where eq . ( [ qdef ] ) was used . as discussed in [ h : conservationlaws ] , fluctuations in particle number are negligible in the large-@xmath128 limit for so(8 ) coherent states , implying that the summations in eqs . ( [ coherentfn1.4 ] ) become dominated by terms with @xmath538 . thus for large @xmath128 the coherent state wavefunctions are well approximated up to a normalization by @xmath539 [ largeomegwf ] as seen from table [ tb : orderparms ] , the su(4 ) state of eq . ( [ largeomegwf.b ] ) is a coherent superposition of @xmath540 or @xmath541 pairs , each contributing vanishing ferromagnetic order @xmath327 and charge density wave order @xmath328 , but non - zero af order @xmath329 . conversely , the @xmath448 state of eq . ( [ largeomegwf.a ] ) is a coherent superposition of @xmath143 pairs , each with vanishing @xmath327 , @xmath328 , and @xmath329 . thus the @xmath448 and su(4 ) limits of the so(8 ) symmetry are distinguished by the order parameter @xmath329 , which is zero in the @xmath448 state and is non - zero in the su(4 ) state . equivalently , the coherent state order parameter @xmath465 vanishes in the pure @xmath448 limit and is equal to @xmath542 in the pure su(4 ) limit [ see eq . ( [ betaminima ] ) and fig . [ fig : coherentenergysurfaces_graphene ] ] . equation ( [ generalenergyformula ] ) depends only on even powers of @xmath465 so the sign for the two possible spontaneously broken symmetry solutions does not affect the energy . for undoped graphene the fermi surface corresponds to the @xmath543 curves of fig . [ fig : coherentenergysurfaces_graphene ] . these are shown in fig . [ fig : coherentenergysurfaces_graphenegs ] for the @xmath448 , so(7 ) , and su(4 ) limits . ground - state energy surfaces in coherent state approximation for three of the so(8 ) dynamical symmetry limits . ( a ) the @xmath448 limit . ( b ) the so(7 ) limit . ( c ) the su(4 ) limit . the diagrams indicate schematically the corresponding wavefunctions , as suggested by eqs . ( [ coherentfn1.1])([largeomegwf ] ) and fig . [ fig : s_d_pairs_brillouin ] . ] * the @xmath448 limit in fig . [ fig : coherentenergysurfaces_graphenegs](a ) has its minimum @xmath513 at @xmath544 . it corresponds to states of the form ( [ largeomegwf.a ] ) , with vanishing expectation values of @xmath545 , @xmath70 , and @xmath546 . * the su(4 ) limit of fig . [ fig : coherentenergysurfaces_graphenegs](c ) corresponds to states of the form ( [ largeomegwf.b ] ) , with minima located at @xmath547 when the symmetry is broken spontaneously by choosing one of these possibilities ( they are equivalent since the energy depends only on even powers of @xmath465 ) , the resulting state has @xmath548 ( no spin or isospin order ) , but @xmath549 ( spin density wave or af order ) . * the so(7 ) limit of fig . [ fig : coherentenergysurfaces_graphenegs](b ) corresponds to a critical dynamical symmetry that interpolates between the @xmath448 and su(4 ) states through critical fluctuations in the antiferromagnetic order . thus the so(8 ) dynamical symmetry limits illustrated in fig . [ fig : coherentenergysurfaces_graphenegs ] represent a rich set of collective states that can be distinguished by the expectation value and fluctuations associated with the order parameter @xmath465 . the so(8 ) coherent state solution can be used to study transitions among the phases defined in fig . [ fig : coherentenergysurfaces_graphenegs ] . for the @xmath177 space ( no broken pairs ) assumed here , @xmath550 yields a constant that is neglected and eq . ( [ coherent1.12 ] ) may be expressed as @xmath551 the last two terms yield constants when evaluated in a given representation , and @xmath552 is found to contribute negligibly to the total energy compared with @xmath553 . therefore , it will be instructive to set @xmath554 and study the approximate so(8 ) hamiltonian @xmath555 from eqs . ( [ casimir_so5])([casimir_so8 ] ) , one finds that @xmath556 { \langles^\dagger s\rangle } + { \langle p^2{\hspace{-0.1em}\cdot\hspace{-0.1em}}p^2\rangle } \sim { \langle{c_{\scriptscriptstyle{{\rm so(7)}}}}\rangle } \end{array}\ ] ] if constants are neglected . thus the model hamiltonian ( [ modelhamiltonian ] ) may be tuned to favor the @xmath448 , so(7 ) , or su(4 ) phases by varying the ratio of the coupling parameters @xmath437 and @xmath489 . to study the quantum phase transitions of the so(8 ) model with the approximate hamiltonian ( [ modelhamiltonian ] ) , it is convenient to define a parameter @xmath557 and to rewrite eq . ( [ modelhamiltonian ] ) as @xmath558 thus the value of @xmath559 tunes the hamiltonian ( [ modelhamiltonian2 ] ) between @xmath507 and su(4 ) phases via an intermediate so(7 ) phase exhibiting quantum critical behavior . 1 . if @xmath560 the ground - state energy surface is approximated by fig . [ fig : coherentenergysurfaces_graphenegs](a ) , with a minimum at @xmath544 , no antiferromagnetic order , and @xmath507 symmetry . 2 . if @xmath561 the ground - state energy surface is approximated by fig . [ fig : coherentenergysurfaces_graphenegs](c ) , with an energy minimum at @xmath562 implying @xmath411 symmetry and antiferromagnetic order . 3 . if @xmath563 , the ground - state energy surface is approximated by fig . [ fig : coherentenergysurfaces_graphenegs](b ) and the system exhibits so(7 ) critical dynamical symmetry , with large fluctuations in the af order parameter @xmath465 . let us now use the hamiltonian ( [ modelhamiltonian2 ] ) to study the quantum phase transitions and spontaneously broken symmetry of the so(8 ) model . if terms involving @xmath564 and @xmath565 are ignored ( as justified above ) , eq . ( [ generalenergyformula ] ) with the parameters in table [ tb : stiffnessparameters ] imply that @xmath566 hence the energy surfaces corresponding to the hamiltonian ( [ modelhamiltonian2 ] ) may be expressed as @xmath567 the variation of the energy surface computed from eq . ( [ coherent1.13d ] ) with the control parameter @xmath568 for half filling ( ground state for undoped graphene ) is shown in fig . [ fig : coherentenergysurfaces_control2d ] ; quantum phase transitions with coupling strength as control parameter : coherent state energy surface as a function of the control parameter @xmath569 for a degeneracy parameter @xmath570 and a fractional occupation @xmath571 . the solid green curves from @xmath572 correspond to approximate @xmath448 symmetry , the solid blue curves for @xmath573 correspond to approximate su(4 ) symmetry . the dashed red curves near @xmath563 correspond to an approximate so(7 ) symmetry mediating the quantum phase transition from @xmath448 to su(4 ) symmetry . ] by tuning the control parameter from @xmath574 to @xmath561 , one sees that the system undergoes a quantum phase transition near @xmath575 from an approximate @xmath457 state with energy minimum at @xmath534 [ see fig . [ fig : coherentenergysurfaces_graphenegs](a ) ] to an approximate su(4 ) state having energy minima at @xmath576 [ see fig . [ fig : coherentenergysurfaces_graphenegs](c ) and eq . ( [ betaminima ] ) ] . for @xmath563 the system has an approximate so(7 ) dynamical symmetry [ see fig . [ fig : coherentenergysurfaces_graphenegs](b ) ] , with no well - defined minimum for the energy as a function of @xmath465 , but with large fluctuations in @xmath465 implied by a highly - degenerate ground state . for fixed values of the coupling parameters @xmath437 and @xmath489 , phase transitions may be mediated by changing the particle occupancy . [ fig : coherentenergysurfaces_vs_n_2d ] illustrates for different values of @xmath497 at fixed @xmath577 . one sees that as the particle number is increased the system makes a transition from approximate @xmath448 symmetry with @xmath534 to su(4 ) symmetry with @xmath524 through a critical so(7 ) symmetry for which the energy is highly degenerate in @xmath465 . in the coherent state approximation the dynamical symmetry structure of fig . [ fig : dynamicalchains_coherentso5 ] has been examined and not the full group structure given in fig . [ fig : dynamicalchains_graphene ] . for the group chains that contain the so(5 ) subgroup , the only physical implication is to omit the physical effect of zeeman splitting from dynamical symmetry hamiltonian [ which would break the so(5 ) subgroup down into a u(1 ) subgroup generated by the @xmath58 component of the physical spin ] . our primary concern in this discussion is the structure associated with the @xmath5 landau level , for which the effect of the zeeman term is expected to be small ( see [ h : applicationsgraphene ] ) . thus , one may view the effect of the zeeman term as a perturbation on the results obtained thus far that will act only on the spin part of the wavefunction . as kharitonov @xcite has already discussed , the competition of the zeeman term with the valley interactions will convert the antiferromagnetic solution into a canted antiferromagnetic solution . quantum phase transitions with particle number as control parameter : coherent state energy surfaces for different filling fraction @xmath578 at fixed values for @xmath437 and @xmath489 with @xmath579 . the solid green curves for @xmath580 correspond to approximate @xmath448 symmetry , the solid blue curves for @xmath581 correspond to approximate su(4 ) symmetry . curves near @xmath582 ( dashed red ) correspond to an approximate so(7 ) symmetry mediating the quantum phase transition from @xmath448 to su(4 ) symmetry . ] placing a strong magnetic field on graphene leads to high level degeneracy at energies corresponding to the quantized landau levels . in general , interacting electronic systems with high level degeneracy near the fermi surface can produce ( even for relatively weak interactions ) a rich variety of collective states that differ qualitatively from the non - interacting ground state . these states correspond to a spontaneous breaking of the symmetry and generally can not be obtained through small perturbations of the weakly - interacting ground state since they are emergent in nature . the su(4 ) symmetry of quantum hall ferromagnetism gives rise to a ground state symmetry reflected in the su(4)-symmetric hamiltonian ( [ so5_1.1a ] ) and a possible symmetry breaking structure that has been outlined in [ h : symmetriesfqhegraphene ] . however , the symmetry - breaking patterns illustrated in fig . [ fig : dynamicalchains_qhf ] represents _ perturbations around the symmetric ground state ( explicit symmetry breaking ) . _ they can not capture the nature of these possible spontaneously - broken symmetry states , since the broken - symmetry states may differ fundamentally from the possible states in fig . [ fig : dynamicalchains_qhf ] : the states corresponding to the most general linear combination ( [ upairs1.8 ] ) represent a complex superposition of many su(4)-symmetric components and generally can not be classified by pure or any simple linear combination of su(4 ) irreducible representations . since the nature of the broken symmetry states can not be studied directly within the su(4 ) framework because they are unlikely to be anywhere near eigenstates of an su(4)-symmetric hamiltonian , the broken - symmetry states have typically been studied numerically , or by effective field theory methods . however , as has been shown , the kinds of collective configurations that have been proposed as candidates for low - lying broken - symmetry states in graphene ( see fig . [ fig : s_d_pairs_brillouin ] and @xcite ) bear strong resemblance to _ eigenstates _ of so(8 ) dynamical symmetry chains . thus , the present so(8 ) symmetry holds the promise of providing _ analytical _ solutions for possible broken - symmetry states in graphene . this is the most important result of the present paper . at specific filling factors the ground state of graphene will be determined by the competition among the su(4 ) symmetry breaking terms . the most obvious su(4)-anisotropic effect is the zeeman term , which favors a spin - polarized state , but the graphene sublattice structure introduces additional interactions that favor ground states without spin polarization that are characterized by spin density wave or charge density wave order at the lattice scale . the competition between zeeman - term spin polarization and the lattice - scale polarizations can be studied by changing the in - plane component of the magnetic field relative to the perpendicular component , since this changes the zeeman energy but not the orbital energies @xcite . such studies indicate that for the higher - energy landau levels the zeeman term is dominant , producing spin ferromagnets that have skyrmionic excitations at half filling , but in the @xmath5 landau level the lattice - scale interactions dominate the zeeman interaction and drive the system into a spin - unpolarized state @xcite . the remainder of this discussion will concentrate on these spin - unpolarized collective states that are candidates for the ground state in the @xmath5 landau level for charge - neutral graphene , with the zeeman interaction viewed as a perturbation on a collective structure that is dominated by lattice - scale interactions . consideration in this paper will be restricted further to those states that can arise from the dynamical symmetries of fig . [ fig : dynamicalchains_graphene ] that contain the so(5 ) subgroup ( those displayed in fig . [ fig : dynamicalchains_coherentso5 ] ) . it has been shown that so(8 ) describes analytically a number of spontaneously - broken - su(4 ) candidates for the states observed in modern experiments such as those described in refs . these solutions provide a spectrum of excited states as well as ground states . the excited states will not be discussed here , except to note that all ground state solutions have a gap to electronic and collective excitations and so correspond to incompressible states . the general theory to be discussed in forthcoming papers can accommodate fm , cdw , and af states , but for the dynamical symmetries containing so(5 ) that were the focus here , all solutions may be classified by a _ single parameter _ @xmath465 measuring af order : su(4 ) states have finite af order but no cdw or fm order , @xmath448 states have no af , cdw , or fm order , and so(7 ) states correspond to a critical dynamical symmetry interpolating between su(4 ) and @xmath448 with large af fluctuations but no static af order , and with no cdw or fm order . in a strong magnetic field the zero - energy state of graphene has four - fold spin and valley degeneracy per landau level , and ( neglecting the lattice - scale interactions ) near the sample boundary one might expect the zero landau level to split into one positively - dispersing ( electron - like ) and one negatively dispersing ( hole - like ) mode for each spin projection . this would suggest a ground state having a bulk energy gap at charge neutrality but with electron - like and hole - like states of opposite spin polarization crossing at the edge of the sample ( with the edge - state structure being analogous to the quantum spin hall effect ) @xcite . however , experiments indicate that the ground state of charge - neutral graphene becomes _ strongly insulating _ at high magnetic fields @xcite . the detailed nature of this state remains uncertain , but it is generally expected to correspond to a spontaneously broken symmetry caused by the strong coulomb interactions among the electrons in the zero landau level . transport properties are not manifest in the algebraic solutions presented here but the coherent state approximation is equivalent to symmetry - constrained hartree fock bogoliubov ( hfb ) theory @xcite , suggesting that so(8 ) theory can be mapped onto hartree fock ( hf ) transport calculations . hf calculations for armchair nanoribbons found that af and cdw states similar to ours have no edge currents @xcite . it may be speculated that our af states also are insulating and thus strong candidates for the high - field ground state , but confirmation requires more work . solutions depend on @xmath437 and @xmath489 in eq . ( [ modelhamiltonian ] ) , which define effective interactions in the truncated space [ highly renormalized relative to parameters in eq . ( [ so5_1.1a ] ) ] . they may be fixed by systematic comparison with data , enabling a robust prediction for the nature of the ground and other low - energy states . one expects modest impurity levels to modify the effective interaction parameters but to leave dynamical symmetries intact . a unified model of conventional superconductivity and high - temperature superconductors having an @xmath385 highest symmetry has been proposed @xcite . for sufficiently strong on - site coulomb repulsion ( true for the cuprates and approximately true for the fe - based compounds ) , the most general so(8 ) symmetry is reduced to its su(4 ) subgroup @xcite , which forms the basis of an su(4 ) dynamical symmetry model of high - temperature superconductivity . this su(4 ) model has three dynamical symmetry subchains ending in an @xmath583 subgroup representing conservation of spin and charge . physically , these su(4 ) subgroup chains represent 1 . an su(2 ) pseudospin subgroup chain that describes a @xmath584-wave singlet superconductor ( sc ) . an so(4 ) subgroup chain that describes an antiferromagnetic mott insulator ( af ) . an so(5 ) subgroup chain representing a critical dynamical symmetry that is soft with respect to af and sc fluctuations and interpolates between the af and sc collective modes . thus , in the su(4 ) model of high-@xmath585 superconductivity the so(5 ) subgroup chain plays a similar physical role as the so(7 ) subgroup chain of the graphene so(8 ) dynamical symmetry model and the so(7 ) subgroup chain of the nuclear so(8 ) model . in all three cases the subgroup chains represent the generalization of a quantum critical point to an entire quantum critical phase that exhibits large fluctuations ( in order - parameter space ) connecting collective modes defined by other dynamical symmetries of the problem . in the high-@xmath585 case the modes connected by the @xmath586 critical dynamical symmetry are antiferromagnetism and superconductivity , in graphene the critical @xmath587 dynamical symmetry connects states with and without to nel ( af ) order , and in the nuclear structure case the critical @xmath587 dynamical symmetry connects collective states that differ in the relative contributions of nucleon pairs carrying total angular momentum 0 and total angular momentum 2 . thus physically these three applications of dynamical symmetry have little in common , but one sees that mathematically they have a deep similarity . although the microscopic physics differs fundamentally , at the level of the observed emergent collective modes in the system one sees that graphene quantum hall states , high - temperature superconductors , and broad classes of nuclear collective states have a unified description in terms of dynamical symmetries associated with very similar compact lie algebras . this remarkable similarity is illustrated in fig . [ fig : nuc_sc_graphene_chains ] the well - known quantum hall ferromagnetic su(4 ) symmetry of graphene in strong magnetic fields has been extended by adding to the particle hole operators that generate su(4 ) a set of six creation and six annihilation operators that create or destroy fermion pairs in either a total valley isospin triplet , total spin singlet state , or a total valley isospin singlet , total spin triplet state ( the only possibilities allowed by the pauli principle ) . this extended set of operators is shown to close an so(8 ) algebra under commutation , which is formally analogous to the so(8 ) algebra of the fermion dynamical symmetry model of nuclear structure physics . this permits immediate adaptation of mathematical tools developed in nuclear physics to the graphene problem . the previously - known su(4 ) quantum hall ferromagnetism symmetry is recovered as one subgroup , but one finds a richer set of low - energy collective modes associated with the full subgroup structure of so(8 ) . by exploiting the established methodology of fermion dynamical symmetries , it was possible to decouple a collective - pair subspace from the full hilbert space of the problem , permitting exact , analytical , many - body solutions to be obtained in several physically - interesting limits . in addition to exact solutions in specific dynamical symmetry limits , a generalized so(8 ) coherent state approximation has been introduced that permits a broad range of solutions to be obtained even when not in the dynamical symmetry limits . the pairs spanning the collective subspace are shown to be analogous to pairs that have already been discussed at a qualitative level in the graphene literature as defining the possible broken symmetry ground states in the presence of strong electron electron and electron phonon correlations in the @xmath5 landau level . the development here places these pairs on a firm , unified mathematical footing and permits analytical solutions to be developed that explore the possible collective states that previously required numerical simulation for their quantitative description . finally , it has been shown that there are uncanny dynamical symmetry analogies among broken symmetry states for graphene in a strong magnetic field , high temperature superconductors , and strongly collective states in atomic nuclei . on the one hand this has the practical utility of allowing technology already developed in one field to be adapted easily to another . on the other hand , it implies a deep and intriguing mathematical affinity among physical problems that are not usually viewed as having more than a superficial connection . _ acknowledgments._* * this work is supported by the basque government ( grant no . it472 - 10 ) , the spanish micinn ( project no . fis2012 - 36673-c03 - 03 ) , and the basque country university ufi ( project 11/55 - 01 - 2013 ) . partial support was provided by lightcone interactive llc . the formalism discussed in this paper was introduced by postulating a set of physical operators that were shown explicitly to generate an so(8 ) lie algebra under commutation . it is illuminating to consider a somewhat different perspective on the motivation for introducing an so(8 ) symmetry for the graphene problem . it is well known that in a fermionic space having @xmath275 degrees of freedom the most general set of bilinear products @xmath588 of creation annihilation operators and their hermitian conjugates generates an su(@xmath275 ) lie algebra under commutation @xcite . ( physically , the restriction to bilinear products corresponds to limiting consideration to two - body interactions . ) likewise , it is well known that adding to this particle hole operator set the most general pair creation and annihilation operators @xmath589 and @xmath590 extends the su(@xmath275 ) algebra to so(@xmath591 ) . this extension is useful because sometimes more is less . the advantage of expanding the space from su(@xmath275 ) , with @xmath592 generators , to so(@xmath591 ) , with @xmath593 generators , is that the added pair operators permit the definition of a ( collective ) subspace of the full hilbert space spanned by the products of pair creation operators acting on the pair vacuum . if an effective hamiltonian is then constructed by writing the most general polynomial in the casimir invariants of all groups in the subgroup chains of the highest symmetry so(@xmath591 ) ( restricted to quadratic casimir invariants if one considers only two - body interactions ) , it will correspond to the most general hamiltonian that can be written in the collective subspace , and will be diagonal in the collective subspace basis for specific dynamical symmetry subgroup chains . thus , the schrdinger equation can be solved analytically in the symmetry limits defined by each dynamical symmetry subgroup chain , and even away from the symmetry limits it can be solved analytically in coherent - state approximation . the su(@xmath275 ) particle hole algebra alone can be used to construct a hamiltonian that commutes with its generators , and the corresponding schrdinger equation can be solved analytically . but when one considers the realistic case of adding symmetry - breaking terms to the hamiltonian that do not commute with the generators of su(@xmath275 ) , the best that can be done analytically is to assume that these terms are small and that the physical solutions can be treated as small perturbations around the symmetric solution . for the non - perturbative case where the added terms lead to spontaneously broken symmetry and new possible ground states that are not connected perturbatively to the symmetric ground state , one has no systematic way to construct the new ground state from the symmetric one except to guess it . but since the su(@xmath275)-symmetric solution is not connected analytically to non - perturbative broken symmetry solutions , the symmetries of the unperturbed ground state could be a poor guide to guessing the nature of the broken - symmetry states and one must rely on numerical solutions or other approximations and not symmetry to determine their properties . let us now apply this general discussion to quantum hall magnetism in graphene . as discussed in [ h : symmetriesfqhegraphene ] , it is widely accepted that an approximate su(@xmath275 ) particle hole symmetry with @xmath594 is relevant in the @xmath5 landau level because the dominant long - range coulomb interaction is su(4)-symmetric . however , the short - range terms that break this symmetry in the effective hamiltonian prevent the su(4 ) symmetry from providing a solution for the broken symmetry ground state , unless it is assumed that those terms only perturb the su(4)-symmetric solution ( small explicit symmetry breaking ) . but experimental evidence suggests that the true ground state of graphene in a strong magnetic field at low temperature breaks su(4 ) symmetry _ spontaneously , not explicitly _ ( see the introduction ) , and therefore is produced by a non - perturbative effect that can not be explained in terms of small fluctuations around the su(4)-symmetric solution . the possible ( spontaneously ) broken - symmetry states for the @xmath5 landau level have been described in terms of the most general sets of electron pairs occupying the two valley isospin and two spin degrees of freedom @xcite , @xmath595 { |0\rangle } , \label{basis1.0}\ ] ] where the vacuum state @xmath239 corresponds to completely filled landau levels for @xmath596 and completely empty landau levels for @xmath597 . each factor in the product @xmath598 creates a pair of electrons in the state @xmath599 at orbital @xmath600 of the @xmath5 ll , with @xmath601 equal to sublattice a or b , @xmath211 equal to spin up or down , and with the valley isospin and sublattice pseudospin identified : @xmath602 and @xmath603 . but the pair creation operators @xmath215 in eq . ( [ basis1.0 ] ) and their hermitian conjugates are not generators of su(4 ) and do not commute with the su(4)-symmetric hamiltonian , and the collective states of the form ( [ basis1.0 ] ) that are of interest in the present context are unlikely to represent small fluctuations around the su(4)-symmetric solution . thus the nature of these collective states is not determined by the su(4 ) symmetry and had to be investigated by numerical calculations using a small basis in prior work @xcite . on the other hand , the so(8 ) pair generators introduced in eq . ( [ coupled1.4 ] ) are included in the most general collective pairs generated by the @xmath604 operators in eq . ( [ basis1.0 ] ) , and include the collective degrees of freedom discussed in ref . @xcite ( see [ sh : otherstates ] and fig . [ fig : s_d_pairs_brillouin ] ) . thus , the so(8 ) fermion dynamical symmetry permits the nature of possible spontaneously broken symmetries to be investigated in terms of _ symmetry properties _ that permit analytical solutions for the broken - symmetry states . hence , the formalism described in this paper may also be introduced by the following logic . the su(4 ) particle hole symmetry generated by the operators @xmath125 defined in eq . ( [ algebra1.2 ] ) is known to provide a good starting point for graphene quantum hall states dominated by the long - range coulomb interaction , but does not describe quantitatively the broken - symmetry modes discussed by kharitonov @xcite in terms of collective pairs resulting from short - range correlations . motivated by the preceding discussion in this appendix , the su(4 ) generator set may be extended to include the possible pair creation and pair annihilation operators operating in the space corresponding to the indices in the su(4 ) basis . by the general su(@xmath275 ) @xmath605 so(@xmath591 ) extension discussed above , this gives the so(8 ) lie algebra of eq . ( [ algebra1.3 ] ) . hence , expanding the algebra from su(4 ) to so(8 ) introduces the capability to explore _ analytically _ the possible collective states following from perturbation of the su(4 ) quantum hall ferromagnet by short - range interactions that break su(4 ) both explicitly and spontaneously . this appendix collects some useful transformations among the several bases that have been employed in this paper . for brevity , in the following @xmath606 will be termed the nuclear so(8 ) basis and @xmath607 will be termed the graphene so(8 ) basis . in transforming from the nuclear so(8 ) basis to the graphene so(8 ) basis the particle number ( charge ) operator @xmath17 or @xmath266 and the 12 pairing operators @xmath608 are retained , but the 15 su(4 ) generators @xmath609 in the nuclear representation are replaced with the 15 su(4 ) generators @xmath610 defined in the graphene representation of eq . ( [ algebra1.4 ] ) . the explicit transformation from the @xmath611 generators to the @xmath612 generators is given by @xmath613 \label{ngx10 } \\ { \pi_{yx } } & = \tfrac{i}{2}\left [ \sqrt{\tfrac{2}{5 } } p^3_{-3 } + p^3_3 \right . \nonumber \\ & \left . + \sqrt{\tfrac{2}{5}}\left(p^1_{-1 } + p^1_1\right ) - \sqrt{\tfrac{3}{5}}\left(p^3_1 + p^3_{-1}\right)\right ] \label{ngx11 } \\ { \pi_{zx } } & = -\tfrac{1}{\sqrt2 } \left ( p^3_2 + p^3_{-2}\right ) \label{ngx12 } \\ { \pi_{xy } } & = \tfrac{i}{2}\left[\sqrt{\tfrac{2}{5 } } p^3_{-3 } + p^3_3 \right . \nonumber \\ & \left.- \sqrt{\tfrac{2}{5}}\left(p^1_{-1 } + p^1_1\right ) + \sqrt{\tfrac{3}{5}}\left(p^3_1 + p^3_{-1}\right)\right ] \label{ngx13 } \\ { \pi_{yy } } & = \tfrac{1}{2}\left [ \vphantom{\sqrt{\tfrac{2}{5 } } } -p^3_{-3 } + p^3_3 \right . \nonumber \\ & \left.- \sqrt{\tfrac{2}{5}}\left(p^1_1 - p^1_{-1}\right ) + \sqrt{\tfrac{3}{5}}\left(p^3_1 - p^3_{-1}\right)\right ] \label{ngx14 } \\ { \pi_{zy } } & = -\tfrac{i}{\sqrt2 } \left ( p^3_2 - p^3_{-2}\right ) \label{ngx15}\end{aligned}\ ] ] in eqs . ( [ algebra1.8])([algebra1.11 ] ) the graphene basis @xmath137 has been expressed in terms of the generators @xmath125 defined in eq . ( [ algebra1.2 ] ) . the inverse transformations giving the @xmath125 in terms of the @xmath614 are where the unlisted operators may be obtained from @xmath616 and the diagonal operators have been assumed to obey the u(4 ) constraint @xmath617 with @xmath618 the total particle number and @xmath128 the total pair degeneracy given by eq . ( [ degen1.2 ] ) . since from eq . ( [ orderp1.1 ] ) the order parameters for the quantum hall ground states are functions of the expectation values for the number operators @xmath619 specifying the population of the four basis states in table [ tb : quantumnumbermapping ] and fig . [ fig : graphenebasis ] labeled by the index @xmath100 , it is useful to have explicit expressions for them in terms of the @xmath260 operators . these are @xmath620 n_2 & = \tfrac14 n + \tfrac{\sqrt5}{10}p^1_0 - \tfrac12 p^2_0 - \tfrac{3\sqrt5}{10}p^3_0 \label{ptoi2 } \\[2pt ] n_3 & = \tfrac14 n - \tfrac{\sqrt5}{10}p^1_0 -\tfrac12 p^2_0 + \tfrac{3\sqrt5}{10}p^3_0 \label{ptoi3 } \\[2pt ] n_4 & = \tfrac14 n - \tfrac{3\sqrt5}{10}p^1_0 + \tfrac12 p^2_0 -\tfrac{\sqrt5}{10}p^3_0\end{aligned}\ ] ] f. t. arecchi , e courtens , r. gilmore , and h. thomas , phys . * a6 * , 2211 ( 1972 ) . r. gilmore , ann . phys . * 74 * , 391 ( 1972 ) . a. m. perelomov , commun . * 26 * , 222 ( 1972 ) . r. gilmore , rev . de fisica * 23 * , 143 ( 1974 ) . zhang , d. h. feng , and r. gilmore , _ rev . mod.phys._ * 62 * , 867 ( 1990 ) . a. f. young , c. r. dean , l. wang , h. ren , p. cadden - zimansky , k. watanabe , t. taniguchi , j. hone , k. l. shepard and p. kim nature physics * 8 * , 550556 ( 2012 ) .

the cosmic evolution survey ( cosmos ) is a treasury program on the _ hubble space telescope _ ( _ hst _ ) , awarded a total of 590 _ hst _ orbits , carried out in cycles 12 and 13 ( scoville et al . 2007a , 2007b ; koekemoer et al . 2007 ) . in total , a sky area of 1.64 square degree is covered with advanced camera for surveys ( acs ) f814w filter around the central position @xmath6 and @xmath7 . note that we originally proposed to map a @xmath8 square degree field . however , due to the observational constraints , the sky area of 1.64 square degree was mapped ( koekemoer et al . 2007 ) . on the other hand , the subaru cosmos 20 project has covered the whole 2 square degree field . the comparison between the _ hst _ acs field and the subaru cosmos 20 field is shown in figure [ fig : field ] . a point source limiting magnitude is down to @xmath9 ( @xmath10 , 024 diameter aperture ) . these acs observations provide us a large sample of galaxies with a spatial resolution of 0.1 arcsec covering a redshift range between @xmath11 to @xmath12 ( e.g. , taniguchi et al . 2009 ; murata et al . 2014 ; kobayashi et al . 2015 ) . ( red line ) and the acs field of @xmath13 ( blue line ) overlaid on the ia427 band image.[fig : field],width=302 ] the main purpose of the cosmos project is to understand the evolution of galaxies , active galactic nuclei ( or super massive black holes ) , and dark matter halos together with the evolution of large scale structures in the universe . in order to carry out this project , we also need multi - wavelength data from x - ray , ultraviolet though optical to infrared and radio . indeed , such multi - wavelength campaign has been made intensively : x - ray ( hasinger et al . 2007 ; elvis et al . 2009 ) , ultraviolet ( zamojski et al . 2007 ) , optical ( taniguchi et al . 2007 ; capak et al . 2007 ) , infrared ( sanders et al . 2007 ) , and radio ( schinnerer et al . 2007 ; smoli et al . 2012 ) . among them , optical imaging observations made by use of suprime - cam ( miyazaki et al . 2002 ) on the subaru telescope ( kaifu et al . 2000 ; iye et al . 2004 ) are highly useful to investigate both photometric properties and photometric redshifts of the galaxies found in the cosmos field ( mobasher et al . 2007 ; ilbert et al . 2009 ; salvato et al . 2009 ) . in our subaru cosmos 20 project , we used 20 filters in the optical covering from 400 nm to 900 nm : six broad - band ( @xmath0 , @xmath1 , @xmath2 , @xmath3 , @xmath4 , and @xmath5 ) , twelve intermediate - band ( ia427 , ia464 , ia484 , ia505 , ia527 , ia574 , ia624 , ia679 , ia709 , ia738 , ia767 , and ia827 ) , and two narrow - band filters ( nb711 and nb816 ) . this is the origin of the project s name , `` subaru cosmos 20 '' . since we have already given a detailed description on the broad - band and nb816 imaging of the cosmos field ( taniguchi et al . 2007 ; hereafter paper i ) , we present details of observations , data reductions , calibration , and quality assessment for the twelve intermediate - band filters and nb711 in this paper ( see also capak et al . 2007 ) . as for the nb711 imaging of the cosmos field , see also shioya et al . ( 2009 ) and kajisawa et al . ( 2013 ) . in appendix , we present a summary of the optical properties of the suprime - cam ia filter system . throughout this paper , we use the ab magnitude system . the twelve intermediate - band filters are selected from the suprime - cam ia filter set ( hayashino et al . 2000 ; taniguchi et al . the spectral resolution of all the ia filters is @xmath14 , being just intermediate between typical broad - band filters ( @xmath15 ) and narrow - band filters ( @xmath16100 ) . therefore , imaging with multi - ia filters is equivalent to low - resolution spectroscopy with an @xmath17 ( see , for example , yamada et al . it is also mentioned that the use of ia filters makes it possible to detect very strong emission - line objects ( galaxies or active galactic nuclei ) . although such objects tend to be rare , some examples have been discovered to date : ultra strong emission line galaxies ( usels ) defined as @xmath18 ( kakazu et al . 2007 ) , and green pea objects found in the galaxy zoo project ( cardamone et al . 2009 ) . since such very strong emission lines in galaxies affect broad - band colors of the galaxies ( e.g. , nagao et al . 2007 ) , careful analyses are recommended to make any sample selection of a particular class of galaxies . for example , in the case of color selection of very high redshift galaxies at @xmath198 , strong emission line galaxies at @xmath20 with little stellar continuum can act as interlopers ( see taniguchi et al . 2010 ; atek et al . 2011 ) . on the other hand , such very strong emission line galaxies themselves are important populations , because most of them are very metal poor galaxies ( kakazu et al . 2007 ; amorn et al . 2014 , 2015 ) . therefore , surveys of such objects contribute to the understanding chemical evolution of galaxies . in our subaru cosmos 20 project , we also use the two narrow - band filters , nb711 and nb816 . imaging with such narrow - band filters provides us samples of targeted emission line galaxies . for example , nb816 has been used to find ly@xmath21 emitters at @xmath22 ( e.g. , hu et al . 2004 , 2010 ; shimasaku et al . 2005 ; murayama et al . 2007 ) . however , the use of nb816 also provides us samples of h@xmath21 emitters at @xmath23 ( shioya et al . 2008 ) and [ o ii ] emitters at @xmath24 ( takahashi et al . 2007 ; ideue et al . 2009 , 2012 ) . in the case of cosmos project , nb711 has also been used to sample both ly@xmath21 emitters at @xmath25 ( shioya et al . 2009 ) and [ o ii ] emitters at @xmath26 ( kajisawa et al . 2013 ) . therefore , if we combine imaging surveys with multiple nb filters , we can trace the cosmic star formation history from high to low redshifts ( e.g. , hopkins 2004 ; shioya et al . 2008 ) . moreover , multi - band optical imaging such as our subaru cosmos 20 improves the accuracy of photometric redshifts of galaxies ( ilbert et al . 2009 ) and active galactic nuclei ( salvato et al . 2009 , 2011 ) . the accurate photometric redshifts for the large sample of galaxies allow us to map the large - scale structure at various redshifts and to study the environmental effects of the galaxy evolution ( feruglio et al . 2010 ; scoville et al . one can also combine such photometric redshifts with a smaller spectroscopic sample to estimate the overdensity of galaxies with high accuracy ( e.g. , kova et al . 2014 ) and to measure the clustering strength of agns ( georgakakis et al . the cosmos field covers an area of @xmath27 , centered at @xmath28 and @xmath29 . the suprime - cam consists of ten @xmath30 ccd chips and provides a very wide field of view , @xmath31 in @xmath32 pixels ( 0202 pixel@xmath33 ) ( miyazaki et al . although the field of view of the suprime - cam had been the widest one among available imagers on the 810 m class telescopes before hyper suprime cam on the subaru telescope ( miyazaki et al . 2012 ) , we needed multiple pointings to cover the entire cosmos field . in our previous suprime - cam observations , we used the two dithering patterns , pattern a and pattern c ( paper i ) . pattern a consists of @xmath34 sets ( 48 shots in total ) to cover the whole cosmos field ( see figure 1 in paper i ) . this dithering pattern is a half - array shifted mapping method to obtain accurate astrometry and a self - consistent photometric solution across the entire field . another dithering method is pattern c that consists of @xmath35 sets ( 36 shots in total ) to cover the whole cosmos field efficiently ( see figure 2 in paper i ) . it is noted that both patterns are designed to take care of spatial gaps ( @xmath36@xmath37 or @xmath38@xmath39 ) between the ccd chips of the suprime - cam . after our previous observations , we confirmed that observations with pattern c only are enough to obtain accurate photometry and astrometry . the flat frames generated using only pattern c in the broad bands are consistent with those created with both pattern a and pattern c within 1% root mean square . therefore , our new observations with the intermediate and narrow - band filters were made by using pattern c. this made our observations more efficient than our previous observations . in this paper , we used twelve intermediate - band filters ( ia427 , ia464 , ia484 , ia505 , ia527 , ia574 , ia624 , ia679 , ia709 , ia738 , ia767 , and ia827 ) and one narrow - band filter ( nb711 ) . note that we intended to use another nb filter , nb921 , whose effective wavelength and the full width at half maximum ( fwhm ) are @xmath40 and @xmath41 , respectively ( kodaira et al . 2003 ; kashikawa et al . 2004 ; taniguchi et al . 2005 ) . however , we did not have enough time to take any nb921 data . the filter response curves including the ccd sensitivity and the atmospheric transmission are shown in figure [ filter_response ] for the twelve ia filters and nb711 ( see also section 3.2 for details ) . in figure [ filter_iabbnb3 ] , we also show those for the 20 filters used in subaru cosmos 20 . in order to see the wavelength coverage fairly , all the response curves are normalized ; i.e. , all the peak values are set to be unity . note that the current ccd chips installed on suprime - cam are different from those used in our subaru cosmos 20 project . @xmath42the figure `` i '' given in the last of i d number means an intensive program . @xmath43@xmath44 , and @xmath45 . @xmath46suprime - cam imaging of the _ hst _ cosmos 2-degree acs survey deep field ( intensive program ) . @xmath47wide - field search for ly@xmath21 emitters at @xmath48 in the _ hst_/cosmos field . @xmath49cosmos-21 : deep intermediate & narrow - band survey of the cosmos field ( intensive program ) . @xmath50cosmos-21 : deep intermediate - band survey of the cosmos field . @xmath42the figure `` i '' given in the last of i d number means an intensive program . @xmath43@xmath44 , and @xmath51 . @xmath46first half night was used in every night . @xmath47compensation nights because of the poor weather in s04b-142 jan and feb runs . @xmath49observing time exchange with dave jewitt ( ifa , uh ) . @xmath50second half night was used in every night . @xmath52three hours @xmath53 3 . our suprime - cam observations of the cosmos field have been made during a period between 2006 january and 2007 march , consisting of two open - use observing programs : s05b-013i ( cosmos-21 : deep intermediate & narrow - band survey of the cosmos field ) and s06b-026 ( cosmos-21 : deep intermediate - band survey of the cosmos field ) . the first one is an intensive program on the subaru telescope . fourteen and half nights were allocated for these two proposals ( tables [ observational_programs ] and [ observational_runs ] ) . with 28.5 nights ( including 4 compensation nights ) allocated for our cosmos observations in paper i , 43 nights were allocated in total . our observations in the available nights shown in table [ observational_runs ] were made under the photometric conditions except for the ia505- and ia679-bands observations on feb . 24 , 2006 . we observed the spectrophotometric standard stars immediately before and after the target observations for the photometric calibration . these standard stars have been observed at various airmass , defocussing the telescope to avoid saturation . the observed standard stars are summarized in table [ standard_star ] . although we also observed gd 71 with several ia filters , all the data were saturated and we did not use gd 71 for the photometric calibration . [ standard_star ] all the individual ccd data were reduced using imcat with the same manner as the suprime - cam broad - band data of the cosmos survey ( capak et al . 2007 ) . at first , we performed the bias subtraction and the masking of bad or saturated pixels . then the flat fielding was carried out with the median dome flats . we subtracted the median sky frames from the flat - fielded object frames to remove the night sky illumination and fringe pattern . the residual background was measured in a grid of @xmath54 pixel squares after masking objects and subtracted . after the sky subtraction , we calculated an astrometric solution for each ccd chip in all frames using the cosmos astrometric reference catalog . this catalog was build in 2004 using the megacam @xmath55-band data ( capak et al . 2007 ) , a dataset with bright enough saturation magnitude on individual exposures ( @xmath56 ) to be registered on classical astrometric references , and deep enough ( reaching @xmath57 ) to allow for the registration of all other subaru and acs cosmos data . the cosmos astrometric reference catalog was build iteratively via the following four steps : 1 ) the astrometric solution for the megacam images was computed using the astrometrix software ( radovich et al . 2001 ) , that was at the time part of the terapix pipeline ( scamp was only introduced in 2005 ) , using the usno - b1.0 catalog ( monet et al . 2003 ) as reference . 2 ) the @xmath55-band catalog obtained from the megacam stack was cross - matched to the cosmos vla pre survey ( schinnerer et al . 2004 ) , and we measured offsets of @xmath58 mas and @xmath59 mas , without any indication of variation across the field . 3 ) the measured offsets were applied to the input usno - b1.0 catalog , and the astrometric solution of the megacam @xmath55-band images was recomputed in the same manner as step 1 . 4 ) the final cosmos astrometric reference catalog was obtained from the @xmath55-band megacam stack produced at step 3 . the internal accuracy of the cosmos astrometric reference , i.e. , the maximum value of the residual of the astrometric solution derived by astrometrix when fitting the shifted usno - b1.0 catalog is @xmath60 mas and @xmath61 mas , and the absolute offset to the full survey vla astrometry is @xmath62 mas and @xmath63 mas . all our cosmos 20 images were forced to the cosmos astrometric reference using a 35th order polynomial . the polynomial order was increased until the astrometric errors were consistent with the seeing size in the data . the internal scatter of the resulting astrometry is always less than 0.2 arcsec . we then performed the scattered light correction for the flat as described in capak et al . the dome and sky flat can be affected by the scattered light at 35% level . the correction factor for this effect was calculated in each @xmath54 pixel grid so that the background subtracted fluxes of an object at different positions of the detector in the different frames have the same values . objects in the all frames were simultaneously used in the fitting procedure for each band . in this process , we also added the additional correction factor for each frame to take account of the effects of the airmass and non - photometric condition . these correction factors for the scattered light in each region and for the atmospheric condition in each frame were simultaneously determined in the fitting for each band ( equations ( 1 ) and ( 2 ) in capak et al . thus we made the corrected flat frames and applied them to the object frames . then the frame to frame offsets of the background - subtracted fluxes of objects were examined as a function of airmass and modified julian date . if the data followed the airmass trend estimated from the standard star observations within the expected 12% error due to point spread function ( psf ) variation , the data were deemed photometric . if data stopped following the expected trend , or did not follow it for a night , the data were deemed non photometric . the non - photometric frames were scaled to the mean of the airmass corrected photometric data . the frames with extinction greater than 0.5 mag were discarded . in the case of the ia679 band , where no photometric data were obtained , all the object frames were scaled to the least extinct frame . therefore the photometry in the ia679 band should be used with caution . the flux - matched frames were then smoothed to the same psf fwhm using a gaussian kernel . after the resampling onto the final astrometric grid , the psf - matched frames were combined with a weight of the inverse variance of each frame , clipping outlier values at more than @xmath10 from the median value in the calculation of each pixel . in this procedure , we also generated a root mean square ( rms ) map that reflects the true pixel - to - pixel rms , which is the value expected if the effects of the resampling and smoothing do not exist . the rms measured in a given area on this rms map represents the variance that would be measured in the background of the same area if the variance was calculated on the individual images that went into the final mosaiced image . the psf sizes of final images are summarized in table [ optical_imaging ] . in addition to these psf - matched combined images , we also provided the original - psf images for each band from the frames that were not convolved in order to provide a maximum sensitivity for detection of ( compact ) sources . note that the psf varies as a function of position in these original - psf images , and therefore the color measurements with a relatively small aperture can be less reliable . these reduced images were divided into tiles with a dimension of @xmath64 as shown in figure 5 of paper i. for the color measurements and generating the official multi - band photometric catalog , we additionally convolved the psf - matched combined images to match the psf among all the optical nir data from @xmath65 to @xmath66 band . these data were convolved with a gaussian kernel so that the flux ratio between a @xmath36 and @xmath67 aperture for a point source in each band is the same as that of ctio / kpno @xmath66-band data , which have the lowest flux ratio of @xmath68 ( capak et al . the width ( @xmath69-value ) of the gaussian kernel used in the convolution of the ia and @xmath70-band data is shown in the last column of table 4 . using these convolved data , we carried out the multi - band photometry with a @xmath36 diameter aperture and the results are presented in the official photometric catalog . note that the matching of the flux ratio between @xmath36 and @xmath67 apertures for point sources does not necessarily guarantee the same flux ratio for extended sources , because the detailed shapes of the psf are not the same among the different bands . if one needs to measure colors for extended sources with high accuracy , the photometric values in the official catalog should be used with caution . since we used the original intermediate- and narrow - band filters in this project , the theoretical synthetic magnitudes of the spectrophotometric standard stars are necessary for the photometric calibration . the theoretical complete system response in each band was computed by multiplying the filter transmission , the subaru telescope mirror reflectivity , the prime focus unit s transmission , the ccd quantum efficiency , and the atmosphere transmission for an airmass of 1.2 . for the atmosphere , we used the same model as the one used for the broad - band suprime - cam filters ( k. shimasaku , private communication ) . we computed the theoretical synthetic magnitude in the ab system of our standard stars using the available calspec spectra , i.e. , gd50_004.fits for gd 50 , gd108_005.fits for gd 108 , hz4_stis_001.fits for hz 4 , hz21_stis_001.fits for hz 21 and hz44_stis_001.fits for hz 44 ( bohlin 1996 ; bohlin et al . we present these theoretical magnitudes in the appendix [ sec : app - b ] . note that for the hz stars , the calspec spectra have been interpolated between 5150 and 5212 , resulting in a larger overall uncertainty in the calibration of the ia505 and ia527 bands , because this more uncertain region lies on the edges of these two passbands . these standard stars have been observed at various airmass as mentioned above , and we determined the zero point for each band taking account of the airmass dependence . if @xmath71 are the counts measured from the standard star under an airmass @xmath72 , the relation between the standard magnitude @xmath73 and the zero point @xmath74 is : @xmath75 so that the airmass term @xmath76 is the slope of the expected linear fit to the data of the standard stars and its intercept gives the zero point . figure [ fig : standards ] shows typical examples of the calibration for the ia624 and the ia738 bands . for most bands , our zero point determination is accurate within 0.02 mag . in a few cases , there was a lack of standard star observations at high airmass that prevented us to derive the airmass term from the standard star data . in this case , we used the airmass point of the object data for the zero - point determination . note that for the ia679 band , we obtained no photometric data both for the cosmos field and the standard stars , and we scaled to the photometry of the neighboring bands ( i.e. , ia624 and ia709 ) assuming the objects were flat in @xmath77 for the interpolation . after the photometric calibration , all the reduced images are converted to be in units of nanojanskys per pixel ( the zero point of 31.4 mag in the ab magnitude system ) . we also checked the consistency among the zero points in the different bands through the spectral energy distribution fitting for a large number of galaxies with spectroscopic redshift ( ilbert et al . the multi - band photometry from uv to mir wavelength including the ia- and narrow - band data were fitted with population synthesis models , and the systematic difference between the model and observed magnitudes in a certain band is considered to reflect the zero - point offset . the details of the method to determine the zero - point offsets are described in ilbert et al . ( 2006 , 2009 ) . these offsets of the photometric zero points are shown in the second - last column of table [ optical_imaging ] . we note that the offset for the ia679 band is much higher than the other ia bands , which probably reflect the larger uncertainty in the photometric calibration for this band mentioned above . the zero - point offsets in table [ optical_imaging ] are calculated for the upgraded version ( v2.0 ) of the photometric redshift catalog from ilbert et al . ( 2009 ) including the new ultravista data from the dr1 ( mccracken et al . therefore , the offsets in table [ optical_imaging ] are slightly different from those in table 1 of ilbert et al . ( 2009 ) . @xmath42effective wavelength calculated from the filter response curve including the effects of the ccd sensitivity , the atmospheric transmission , and the transmission of the telescope and the instrument shown in figure [ filter_response ] . @xmath43fwhm of the filter response curve mentioned above . @xmath46the target dedicated time . @xmath47the average 3@xmath69 limiting magnitude in the ab system within @xmath36 diameter aperture . @xmath49the standard deviation of @xmath78 measured in the 81 tiles . @xmath50the psf size of the final images . note that the psf of each filter band is finally matched so that the flux ratio between a @xmath36 and @xmath67 apertures is the same as that in the ctio / kpno @xmath66-band data to provide official photometric catalog ( see text ) . @xmath52systematic offset of the photometric zero point for each filter ( see text ) . @xmath79the @xmath69-value of the gaussian kernel used for the psf matching among the different bands ( see text in section [ subsec : datareduction ] ) . note that the magnitudes in the public catalog are not corrected for the galactic extinction . instead , we provide the galactic extinction value , @xmath80 , from schlegel et al . ( 1998 ) for each object in the catalog . the correction for the galactic extinction in each band can be calculated from these values . we estimated the limiting magnitudes using the 81 tiles ( the cosmos _ hst_/acs field ) for each band . for each tile , we set 50,000 random points and performed aperture photometry with a @xmath36 diameter aperture on the psf - matched images which were convolved to the resolution of the cosmos @xmath66-band image . in order to measure the background fluctuation properly , we masked objects on the images . we used sextractor version 2.3.2 ( bertin & arnouts 1996 ) with the detection criteria of 5-pix connection above the @xmath81 significance . then we replaced the masked regions with pseudo noise images , which were provided from randomly - shifted object - masked images . then we evaluated the limiting magnitudes from the standard deviation for the distribution of the random photometry . the average limiting magnitudes of the 81 tiles for the ia and nb711 bands are listed in table [ optical_imaging ] . as shown in table [ optical_imaging ] and figures [ fig : limitmag_ia1 ] and [ fig : limitmag_ia2 ] , the @xmath82 limiting magnitudes are @xmath8325.9 mag in ia427ia827 bands . the nb711 data reach to the limiting magnitude of @xmath84 mag as shown in table [ optical_imaging ] and figure [ fig : limitmag_nb711 ] . the standard deviation of the limiting magnitudes among the 81 tiles for each band is @xmath850.16 mag . as seen in figures [ fig : limitmag_ia1][fig : limitmag_nb711 ] , the limiting magnitudes are brighter in the tiles at the edge of our survey field , because the total exposure time is smaller in these regions . some tiles where very bright stars illuminate surrounding sky region also show brighter limiting magnitudes . , but for @xmath70band data.[fig : limitmag_nb711 ] ] we present deep optical imaging observations made with the suprime - cam on the subaru telescope with 20 filters [ 6 broad - band , 12 intermediate - band ( ia ) , and 2 narrow - band ( nb ) filters ] : subaru cosmos 20 . in this paper , we describe the details of our imaging with the 12 ia filters and nb711 . note that those of the other seven filters are given in paper i. the use of intermediate - band filters has generally a couple of scientific merits : ( 1 ) improvement of the accuracy of photometric redshifts and ( 2 ) selection of very strong emitters . first , we discuss the improvement of the accuracy of photometric redshifts . as described in mobasher et al . ( 2007 ) , our previous accuracy of photometric redshifts based on six subaru broad band , cfht @xmath86 band , acs f814w , nb816 , and ctio / kpno @xmath66 photometric data is @xmath87 where @xmath88 ; note that @xmath89 and @xmath90 are photometric and spectroscopic redshifts , respectively . since both @xmath86 and @xmath66 are used together with optical data , the accuracy of @xmath89 is better than that of typical optical studies ( e.g. , hogg et al . however , in the cosmos project , thanks to its multi - wavelength campaign , 30 band photometric data including subaru cosmos 20 data are accumulated to obtain much more accurate estimates of @xmath89 ( ilbert et al . 2009 ; see also salvato et al . 2009 , 2011 ) . the accuracy of @xmath89 is improved to @xmath91 for @xmath92 . even at fainter magnitudes of @xmath93 , the accuracy is found to be still high as @xmath94 for the galaxies at @xmath95 ( ilbert et al . 2009 ) . there are several similar surveys with the use of intermediate band filters . [ 1 ] combo-17 ( classifying objects by medium - band observations in 17 filters ) : this is a pioneering optical survey with multi - band filters ( wolf et al . 2003 ) . the combo-17 covers three @xmath96 fields , including the extended chandra deep field south ( ecdf - s ) . in this survey , twelve intermediate - band filters were used together with five broad - band ones ( @xmath65 , @xmath0 , @xmath2 , @xmath97 , and @xmath98 ) by using the wide field imager at the mpg / eso 2.2 m telescope on la silla , chile . their intermediate - band filters cover 410 nm to 920 nm . the spectral resolution is not fixed for all the filters but ranges from @xmath99 to 61 ( mostly from 30 to 40 ) . the use of intermediate - band filters improves the accuracy of photometric redshifts to @xmath91 for @xmath100 ( wolf et al . this enables them to construct a large sample of agns at @xmath1015 . [ 2 ] musyc ( the multiwavelength survey by yale - chile ) : in this project , 18 intermediate - band filters in the ia filter system for suprime - cam on the subaru telescope were used together with 14 board - band data from optical to mid - infrared ( seven optical filters from @xmath65 to @xmath102 , three near - infrared filters , @xmath103 , @xmath104 , and @xmath105 , and four spitzer irac bands , 3.6 , 4.5 , 5.8 , and 8.0 @xmath106 m ) ( cardamone et al . 2010 ) . these data cover a @xmath107 field of the ecdf - s , which is one of the musyc fields . the use of ia filters improves the accuracy of photometric redshifts at @xmath108 to 1.2 and @xmath109 : @xmath110 for @xmath111 , see table 8 in cardamone et al . ( 2010 ) in more detail . this is attributed that the balmer break ( 3648 ) or lyman break ( 912 ) falls in wavelength interval covered by the 18 ia filters . according to cardamone et al . ( 2010 ) , _ the use of ia filters not only tightens the accuracy of photometric redshifts but also can help to rule out false redshift solution ( so called catastrophic failures)_. [ 3 ] mahoroba-11 : this survey is a scaled down version of subaru cosmos 20 ( yamada et al . 2005 ) . in this survey seven ia filters are used together with five broad - band filters . these data cover a @xmath112 area in the subaru xmm - newton deep survey field . their main purpose is to search for ly@xmath21 emitters at @xmath113 by using a photometric redshift method . they showed that the fraction of false detection is only 10% . [ 4 ] alhambra ( the advanced large homogeneous area medium - band redshift astronomilca ) : this survey has been carried out by using the wide - field optical camera , large area imager for calar alto ( laica ) on the calar alto 3.5 m telescope with 20 intermediate - band filters with 300 spacing ( moles et al . 2008 ; molino et al . the surveyed area size is 2.79 deg@xmath114 . the accuracy of photometric redshifts is @xmath110 for @xmath115 and @xmath116 for @xmath117 . in this way , a number of optical wide - field deep surveys have been carried out by using their original intermediate - band filter systems . the main reason for this is to obtain more reliable photometric redshifts for large numbers of objects in the individual surveys ; see figure 1b in molino et al . ( 2014 ) for a comprehensive comparison among available optical surveys including surveys with broad - band filters only such as hdf , sdss , and so on . the subaru cosmos 20 is the widest survey among the deep ( @xmath118 ) optical intermediate - band surveys . some efficient multiple - object spectrographs are available on 8 m class telescopes ( e.g. , vimos on the vlts and fmos on the subaru telescope ) . however , imaging surveys with intermediate - band filters are more efficient to obtain redshift information for large numbers of objects . we mention about our future works on study of strong emission - line objects . the wide imaging with the ia filter set of the subaru cosmos 20 enables us to detect very strong emission - line objects ( star forming galaxies and agns ) over a extremely large volume . in our forthcoming papers , we will present a large sample of ia - excess strong emission - line objects ( kajisawa et al . 2015 , in preparation ) and a new population of maestlo ( @xmath119 massive extremely strong ly@xmath21 emitters ) at @xmath120 with rest - frame ly@xmath21 equivalent width of @xmath121 and @xmath122 ( taniguchi et al . 2015 ) . finally , we note that the major cosmos datasets including the subaru images and catalogs are publicly available ( following calibration and validation ) through the web site for ipac / irsa : + * http://irsa.ipac.caltech.edu / data / cosmos/*. the _ hst _ cosmos treasury program was supported through nasa grant hst - go-09822 . we gratefully acknowledge the contributions of the entire cosmos collaboration consisting of more than 70 scientists . more information on the cosmos survey is available at * http://www.astro.caltech.edu/~cosmos*. it is a pleasure the acknowledge the excellent services provided by the nasa ipac / irsa staff ( anastasia laity , anastasia alexov , bruce berriman and john good ) in providing online archive and server capabilities for the cosmos datasets . we are deeply grateful to the referee for his / her useful comments and excellent refereeing , which helped us to improve this paper very much . we would also like to thank the staff at the subaru telescope for their invaluable help . in particular , we would like to thank hisanori furusawa because his professional help as a support scientist made our suprime - cam observations successful . data analysis were in part carried out on common use data analysis computer system at the astronomy data center , adc , of the national astronomical observatory of japan . this work was financially supported in part by jsps ( yt : 15340059 , 17253001 , 19340046 , 23244031 , tn : 23654068 and 25707010 ) and by the yamada science foundation ( tn ) . ajiki , m. , taniguchi , y. , fujita , s. s. , et al . 2004 , , 56 , 597 amorn , r. , grazian , a. , castellano , m. , et al . 2014 , , 788 , l4 amorn , r. , prez - montero , e. , contini , t. , et al . 2015 , , 578 , a105 atek , h. , siana , b.,scarlata , c. , et al . 2011 , , 743 , 121 bertin , e. , & arnouts , s. 1996 , , 117 , 393 bohlin , r. c. 1996 , , 111 , 1743 bohlin , r. c. , dickinson , m. e. , & calzetti , d. 2001 , , 122 , 2118 capak , p. , aussel , h. , ajiki , m. , et al . 2007 , , 172 , 99 cardamone , c. , schawinski , k. , sarzi , m. , et al . 2009 , , 399 , 1191 cardamone , c. n. , van dokkum , p. g. , urry , c. m. , et al . 2010 , , 189 , 270 elvis , m. , civano , f. , vignali , c. , et al . 2009 , , 184 , 158 feruglio , c. , aussel , h. , le floch , e. , et al . 2010 , , 721 , 607 fujita , s. s. , ajiki , m. , shioya , y. , et al . 2003 , , 125 , 13 georgakakis , a. , mountrichas , g. , salvato , m. , et al . 2014 , , 443 , 3327 hasinger , g. , cappelluti , n. , brunner , h. , et al . 2007 , , 172 , 29 hayashino , t. , tamura , h. , matsuda , y. , et al . 2003 , publications of the national astronomical observatory of japan , 7 , 33 hayashino , t. , taniguchi , y. , yamada , t. , et al . 2000 , , 4008 , 397 hogg , d. w. , cohen , j. g. , blandford , r. , et al . 1998 , , 115 , 1418 hopkins , a. m. 2004 , , 615 , 209 hu , e. m. , cowie , l. l. , barger , a. j. , et al . 2010 , , 725 , 394 hu , e. m. , cowie , l. l. , capak , p. , et al . 2004 , , 127 , 563 ideue , y. , nagao , t. , taniguchi , y. , et al . 2009 , , 700 , 971 ideue , y. , taniguchi , y. , nagao , t. , et al . 2012 , , 747 , 42 iye , m. , karoji , h. , ando , h. , et al . 2004 , , 56 , 381 ilbert , o. , arnouts , s. , mccracken , h. j. , et al . 2006 , , 457 , 841 ilbert , o. , capak , p. , salvato , m. , et al . 2009 , , 690 , 1236 kaifu , n. , usuda , t. , hayashi , s. s. , et al . 2000 , , 52 , 1 kajisawa , m. , shioya , y. , aida , y. , et al . 2013 , , 768 , 51 kakazu , y. , cowie , l. l. , & hu , e. m. 2007 , , 668 , 853 kashikawa , n. , shimasaku , k. , yasuda , n. , et al . 2004 , , 56 , 1011 kobayashi , m. a. r. , murata , k. l. , koekemoer , a. m. , et al . 2015 , submitted to kodaira , k. , taniguchi , y. , kashikawa , n. , et al . 2003 , , 55 , l17 koekemoer , a. m. , aussel , h. , calzetti , d. , et al . 2007 , , 172 , 196 kova , k. , lilly , s. j. , knobel , c. , et al . 2014 , , 438 , 717 mccracken , h. j. , milvang - jensen , b. , dunlop , j. , et al . 2012 , , 544 , a156 miyazaki , s. , komiyama , y. , nakaya , h. , et al . 2012 , , 8446 , miyazaki , s. , komiyama , y. , sekiguchi , m. , et al . 2002 , , 54 , 833 mobasher , b. , capak , p. , scoville , n. z. , et al . 2007 , , 172 , 117 moles , m. , bentez , n. , aguerri , j. a. l. , et al . 2008 , , 136 , 1325 molino , a. , bentez , n. , moles , m. , et al . 2014 , , 441 , 2891 monet , d. g. , levine , s. e. , canzian , b. , et al . 2003 , , 125 , 984 murata , k. l. , kajisawa , m. , taniguchi , y. , et al . 2014 , , 786 , 15 murayama , t. , taniguchi , y. , scoville , n. z. , et al . 2007 , , 172 , 523 nagao , t. , murayama , t. , maiolino , r. , et al . 2007 , , 468 , 877 nagao , t. , sasaki , s. s. , maiolino , r. , et al . 2008 , , 680 , 100 radovich , m. , bonnarel , f. , mellier , y. , et al . 2001 , the new era of wide field astronomy , 232 , 297 salvato , m. , hasinger , g. , ilbert , o. , et al . 2009 , , 690 , 1250 salvato , m. , ilbert , o. , hasinger , g. , et al . 2011 , , 742 , 61 sanders , d. b. , salvato , m. , aussel , h. , et al . 2007 , , 172 , 86 schinnerer , e. , carilli , c. l. , scoville , n. z. , et al . 2004 , , 128 , 1974 schinnerer , e. , smoli , v. , carilli , c. l. , et al . 2007 , , 172 , 46 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 scoville , n. , abraham , r. g. , aussel , h. , et al . 2007a , , 172 , 38 scoville , n. , arnouts , s. , aussel , h. , et al . 2013 , , 206 , 3 scoville , n. , aussel , h. , brusa , m. , et al . 2007b , , 172 , 1 shimasaku , k. , ouchi , m. , furusawa , h. , et al . 2005 , , 57 , 447 shioya , y. , taniguchi , y. , ajiki , m. , et al . 2005 , , 57 , 287 shioya , y. , taniguchi , y. , sasaki , s. s. , et al . 2008 , , 175 , 128 shioya , y. , taniguchi , y. , sasaki , s. s. , et al . 2009 , , 696 , 546 smolv ci , v. , aravena , m. , navarrete , f. , et al . 2012 , , 548 , a4 takahashi , m. i. , shioya , y. , taniguchi , y. , et al . 2007 , , 172 , 456 taniguchi , y. 2004 , studies of galaxies in the young universe with new generation telescope , proceedings of japan - german seminar , held in sendai , japan , july 24 - 28 , 2001 , eds . : n. arimoto and w. duschl , 2004 , p. 107 - 111 taniguchi , y. , ajiki , m. , nagao , t. , et al . 2005 , , 57 , 165 taniguchi , y. , kajisawa , m. , kobayashi , m. a. r. , et al . 2015 , , 809 , l7 taniguchi , y. , murayama , t. , scoville , n. z. , et al . 2009 , , 701 , 915 taniguchi , y. , scoville , n. , murayama , t. , et al . 2007 , , 172 , 9 ( paper i ) taniguchi , y. , shioya , y. , & trump , j. r. 2010 , , 724 , 1480 wolf , c. , meisenheimer , k. , kleinheinrich , m. , et al . 2004 , , 421 , 913 wolf , c. , wisotzki , l. , borch , a. , et al . 2003 , , 408 , 499 yamada , s. f. , sasaki , s. s. , sumiya , r. , et al . 2005 , , 57 , 881 zamojski , m. a. , schiminovich , d. , rich , r. m. , et al . 2007 , , 172 , 468 in this section , we present optical properties of the ia filter system for suprime - cam on the subaru telescope . this filter system was developed as a private type of filters by the two authors ( th and yt ) . early short descriptions on this filter system are given in hayashino et al . ( 2000 ) and taniguchi ( 2004 ) . the ia filter system consists of 20 intermediate band filters with a spectral resolution of @xmath12326 , covering 410 nm to 1000 nm . ( table [ iafilters ] ) . the filter response curves are shown in figure [ filter_response_all ] . note that these response curves are those of the filters themselves ; that is , the effects of the ccd sensitivity , the atmospheric transmission , and the transmission of the telescope and the instrument are not included . [ iafilters ] @xmath42center wavelength defined as the center of the two wavelengths at which the filter transmission becomes the half maximum . @xmath43fwhm calculated from the same filter response curve used to evaluate the center wavelength . all the ia filters were manufactured by barr associates co. ltd ( now , materion co. ltd ) . the specifications for the ia filters are summarized in table [ specifications ] . although some of the specifications were found not to be fully satisfied , all the filters are highly useful for scientific observations ( e.g. , fujita et al . 2003 ; ajiki et al . 2004 ; shioya et al . 2005 ; yamada et al . 2005 ; nagao et al . details of measurements of the filter transmission is given in hayashino et al . the measured data are available at http://www.awa.tohoku.ac.jp/astro/filter.html . [ specifications ] in table [ tab : standards ] , we summarize the theoretical ia - band magnitudes of the standard stars computed from the calspec spectra and the complete system response , in ab magnitudes . the uncertainty given here assumes a perfect knowledge of the system response and are based solely on the calspec statistical and systematic uncertainties .

the broadening of exciton emission lines due to electron - exciton collisions was first observed in bulk gaas @xcite . an experimental comparison between exciton - electron and exciton - exciton scattering mechanisms in bulk gaas @xcite and in gaas quantum wells ( qw s ) @xcite showed that the exciton - electron scattering efficiency is an order of magnitude larger than that of the exciton - exciton process . both scattering processes are enhanced for the 2d excitons as compared to bulk excitons . theoretical calculations of elastic and inelastic exciton - electron scattering were reported for bulk semiconductors @xcite and for qw s @xcite . the semiclassical treatment of the latter neglected the exchange term which was shown recently to be the dominant term in qw s @xcite . the existence of the negatively charged exciton @xmath1 ( trion ) was first proposed by lampert @xcite as the semiconductor analogue of the hydrogen ion . due to their confinement in the growth direction , 2d trions have a binding energy which is an order - of - magnitude larger than that of the bulk trions @xcite . this fact facilitated the observation of trions in cdte @xcite and in gaas @xcite qw s containing a low - density two - dimensional electron gas ( 2deg ) . theoretical calculations of the binding energy of both negatively ( @xmath1 ) and positively ( @xmath4 ) charged excitons were performed using various trial functions , showing that only the singlet state is bound @xcite . particular attention was devoted to the modification of the trion s properties in the presence of strong magnetic fields . in this limit , the @xmath1 binding energy is increased , and the electron spin triplet state becomes bound as well @xcite . it was also observed that the dependence on the magnetic field of the @xmath1 and @xmath4 binding energies , as well as their zeeman splitting , differ drastically ( while being nearly identical at zero magnetic field ) @xcite . while these aspects of the trions where extensively studied , there are very few reports on their broadening mechanisms @xcite . in this paper we present a theoretical model for the scattering of both neutral and charged excitons with electrons in qw s . this paper continues our previous work which introduced neutral excitons elastic scattering with electrons in the context of cavity polaritons @xcite . here we extend our model , to incorporate both excitons and trions , taking into account elastic and inelastic scattering processes . the model is based on calculating the exciton - electron direct and exchange interaction matrix elements , from which we derive the exciton scattering rates . these are integrated over all final excitonic states , resulting in the exciton linewidth due to scattering , as a function of its initial momentum ( or energy ) . a major difference between the exciton and trion scattering lies in the charge of the latter , which results in a divergence of its matrix elements in the limit of zero transferred momentum . this divergence , originating from the infinite range of the coulomb potential , is treated by applying the lindhard model for the potential screening . the screening action complicates the trion linewidth dependence on the electron density ; thus , while for electron densities larger than @xmath5 the trion linewidth due to electron scattering in a gaas qw at @xmath3 is comparable to that of the neutral exciton , for a very dilute 2deg ( @xmath6 ) the trion linewidth becomes roughly an order - of - magnitude larger than that of the neutral exciton . inelastic scattering is also much more efficient in the trion s case , due to its smaller binding energy . the calculated reflection spectra are obtained by considering the three electronic excitations of the system , namely , the heavy - hole and light - hole 1s neutral excitons , and the heavy - hole 1s charged exciton ( trion ) , with the appropriate oscillator strengths . a qualitative validation of our calculations is given by considering photoluminesence ( pl ) measurements that were done on a mixed type i type ii gaas / alas qw ( mtqw ) structure @xcite . the paper is organized as follows . in section [ ex - el ] , we present in detail the model for the neutral exciton - electron scattering . in section [ trion ] we consider the trion - electron scattering . first we construct the trion s wave function , which is used in the matrix elements calculations . we then treat the divergences that originate from the infinite range of the coulomb potential , by applying the lindhard model for the potential screening . the effect of the screening is discussed in the context of the trion linewidth dependence on the electron density . in section [ bare - ex ] we use our model to calculate qw reflection spectra and discuss their relevance to the available experimental data . a summary and conclusions are given in section [ conc ] . in appendices a and b , we provide some details of the direct and exchange integrals calculations for the neutral exciton - electron scattering . finally , a derivation of the trion binding energy , using a chandrasekhar - type trial function , is given in appendix c. in this section we present a detailed description of the exciton - electron scattering , considering separately the elastic process , for which the exciton remains bound , and the inelastic process , where the exciton breaks into an unbound electron - hole pair . we first construct the state of a single exciton with an in - plane center - of - mass ( cm ) momentum @xmath7 in the fermionic hilbert space of electron - hole pairs , using the notations of tassone and yamamoto @xcite . it is a superposition of wave functions with different electron momenta @xmath8 and electron and hole @xmath9 coordinates , given by @xmath10 where @xmath11 is the electron ( hole ) creation operator with in - plane momentum @xmath7 and @xmath12 ( @xmath13 ) coordinate , @xmath14 , @xmath15 are the electron and exciton in - plane effective masses , respectively , @xmath16 , and @xmath17^{-3/2}\chi_e(z_e ) \chi_h(z_h)$ ] is the in - plane fourier transform of the exciton wave function ( see appendix a ) . a state comprising an exciton and an unbound electron having @xmath18 and @xmath19 will be written as @xmath20 where @xmath21 is the electron wave function . applying the coulomb interaction operators @xmath22 , @xmath23 to this state , we have for the electron - hole and electron - electron interactions : [ vehvee ] @xmath24 the first term in eq . ( [ veh ] ) represents the coulomb interaction between the constituents of the exciton , thus contributing to its self energy , and can be discarded in the calculation of the scattering matrix elements . in eqs.([veh ] ) and ( [ vee ] ) , @xmath25 is the 2-dimensional fourier transform of the coulomb interaction , where @xmath26 is the qw area , and the plus ( minus ) sign is used in @xmath22 ( @xmath23 ) . using the anticommutation relations for the fermion operators , one can easily find the scattering matrix elements to be [ matel ] @xmath27\end{aligned}\ ] ] @xmath28,\end{aligned}\ ] ] where we denote @xmath29 . the first term in each of eqs . ( [ matel ] ) contributes to the direct ( classical ) coulomb interaction and the second contributes to the exchange matrix element . the direct term is calculated analytically ( the details can be found in appendix a ) , resulting in : @xmath30 , \label{vdir}\ ] ] where @xmath31 is the qw width , @xmath32 is a variational parameter associated with the exciton bohr radius in the qw , and we have defined the dimensionless functions @xmath33^{-3/2 } \nonumber \\ e}^{-lq}- 1+lq + \frac{5(lq)^3}{8\pi^2 } + \frac{3(lq)^5}{32 \pi^4}}{(\lambda q)^3 \left [ 1+(lq/2 \pi)^2 \right]^2}. \label{gh}\end{aligned}\ ] ] it is evident from eq . ( [ vdir ] ) that the direct term is identically zero for equal electron and hole masses ( @xmath34 ) . similarly , we have for the exchange term @xmath35 h(\lambda k_1 ) , \label{vexch}\end{aligned}\ ] ] where we have defined @xmath36 . the exchange term ( [ vexch ] ) is computed numerically , however its angular part can be calculated analytically ( see appendix b ) . it is convenient to transform to dimensionless direct and exchange integrals given by @xmath37 the direct and exchange integrals are plotted in fig . [ dir_exc ] as a function of the transferred momentum @xmath38 , for the case @xmath39 , where the angular dependence of @xmath40 disappears . as @xmath41 the direct integral approaches zero , while the exchange integral has its maximum ( this is also the case for exciton - exciton interaction @xcite ) . in the general case , the exchange integral is a function of the transferred momentum @xmath38 , the momentum difference @xmath42 ( which can be regarded , for convenience , as the in - plane momentum of the colliding electron in the rest frame of the exciton ) , and of the angle @xmath43 . the exchange interaction term has the following features : ( i ) the interaction favors the case @xmath44 . physically , this means that the electron is inclined to transfer as much momentum as possible to the exciton , preferably in the same direction . ( ii ) the interaction retains its strength for quite large values of @xmath38 ( or @xmath42 ) even though the excitonic wave function vanishes much more rapidly with momentum . ( iii ) the differential cross section is largest for @xmath45 and decreases to a minimum for back scattering ( @xmath46 ) . = 0.55 next we calculate the scattering rate of an exciton , with an initial in - plane momentum @xmath7 , to an excitonic state with an in - plane momentum @xmath47 . the exchange matrix element ( [ vexch ] ) depends on five variables : @xmath48 , where we have defined @xmath49 . it is convenient to transform from the angles @xmath50 and @xmath51 to the angles @xmath52 and @xmath53 @xcite : @xmath54 it can be verified that for the special case of @xmath55 , we have @xmath56 and @xmath57 . conservation of energy and momentum for the exciton - electron scattering process reads @xmath58 where @xmath59 ( @xmath60 ) is the exciton ( electron ) kinetic energy . this equation is satisfied with : @xmath61 \ , \ ] ] the scattering rate is then found using fermi s golden rule the scattering rate is then found using fermi s golden rule @xmath62 \nonumber \\ & = & \frac{a m_e}{\pi \hbar^3 } \int_0^{2 \pi } \frac{d \delta } { \left| \cos \delta \right|}\frac{k_{\rm e}^*}{q } \left|v_{\rm dir } + v_{\rm exc } \right|^2 f_{\rm fd } ( k_{\rm e}^*)\left[1-f_{\rm fd } ( |{\bf k}_{\rm e}^*-{\bf q}| ) \right ] , \label{rate_ex}\end{aligned}\ ] ] where @xmath63^{-1}$ ] is the electron fermi - dirac distribution function . in the two dimensional case that we are considering , the chemical potential is @xmath64 , and we assume that the fermi energy is given by the free 2d electrons value : @xmath65 . we note that in the high temperature limit , @xmath66 , the fermi - dirac distribution is practically classical , and that for low electron densities , the chemical potential becomes negative . using the relations ( [ gamdel ] ) , the integrand in ( [ rate_ex ] ) can be expressed as a function of the variables : @xmath67 . the exciton linewidth due to electron scattering is calculated by integrating over all final excitonic states @xmath68 figures [ fgam_ex ] a , b show the heavy - hole exciton ( @xmath69 ) and light - hole exciton ( @xmath70 ) linewidths at @xmath71 and @xmath3 for @xmath72 . the large linewidths obtained for relatively low @xmath73 reflect the high efficiency of the electron scattering mechanism . this should be compared to an exciton linewidth of @xmath74 0.2mev for acoustic phonon scattering at @xmath71 . the larger linewidth obtained at @xmath3 ( fig . [ fgam_ex]b ) is explained by noticing that the exciton - electron interaction matrix elements favor small energy transfer transitions . thus , at high temperatures , the electrons have too high energy to be effective scatterers . decreasing @xmath75 thus increases the scattering rate until the electron gas becomes degenerate , and fewer final states are available for the scattered electrons . in figure [ fgam_ex]c the heavy hole exciton linewidth at @xmath55 is plotted as a function of @xmath73 for @xmath3 and @xmath76 ( the light hole exciton linewidth exhibits a similar behaviour ) . it is seen that while at @xmath71 the linewidth amplitudes scale practically linearly with @xmath77 for a large range of electron densities ( up to @xmath78 ) , the linearity region at @xmath3 is much smaller . this linearity threshold seems to be in accord with the onset of the phase - space filling effect which becomes noticeable at higher densities and effectively enlarges the exciton bohr radius @xcite . as the temperature decreases , the effect of phase - space filling becomes important at a much lower electron density . we note that the functional dependence of the linewidths on the initial exciton momentum is almost unchanged within the linear regime . increasing @xmath73 further results in a shift of the maximum linewidth from @xmath55 to higher momenta . = 0.55 we now consider the case when the electron - exciton scattering results in dissociation . this scattering process is depicted schematically in fig . [ xion_scat ] . ( 150,150 ) ( 0,0)(150,150 ) ( 13,116)x ( 25,118 ) ( 30,115)(1,-1)30 ( 20,135)@xmath7 ( 15,30)e ( 25,32 ) ( 30,35)(1,1)30 ( 20,10)@xmath79 ( 80,90)h ( 90,92 ) ( 95,95)(1,1)30 ( 87,135)@xmath80 ( 80,73)e ( 90,75 ) ( 95,75)(1,0)40 ( 110,80)@xmath81 ( 80,55)e ( 90,58 ) ( 95,55)(1,-1)30 ( 100,10)@xmath82 the initial state of the system is given again by eq . ( [ initial ] ) while the final state is that of three free particles , symbolically written as @xmath83 , where @xmath84 is the momentum transferred from the free electron to the electron - hole pair , and @xmath81 is the second electron momentum . calculating the coulomb interaction matrix elements between the initial and final states of the system , we again obtain direct and exchange interaction terms : @xmath85 \\ v_{{\rm exc}}^{\rm dis}\!\!\!&(&\!\!\!{\bf k}_{\rm x},{\bf q},{\bf k}',{\bf k}_{\rm e } ) = -\frac{4 e^2 \lambda^4}{\epsilon_0 l^2}\left(\frac{2 \pi}{a}\right)^{3/2 } h(\lambda \left|{\bf k}'-{\bf k}_{\rm e}\right| ) \times \nonumber \\ & & \left [ g(\lambda|\alpha { \bf k}_{\rm x}+{\bf q}-{\bf k}'| ) -g(\lambda|\alpha { \bf k}_{\rm x}+{\bf q}-{\bf k}_{\rm e}| ) \right]\end{aligned}\ ] ] where @xmath86 are given in ( [ gh ] ) . we note that contrary to the elastic scattering case , here the direct and exchange matrix elements are comparable . in order to calculate the scattering rate we first introduce energy - momentum conservation @xmath87 where @xmath88 is the exciton binding energy . denoting the angle @xmath89 together with @xmath90 which were defined previously , this equation is satisfied with @xmath91 \right\}\end{aligned}\ ] ] the scattering rate is then calculated to be : @xmath92 \times \\ & & \left [ 1-f_{\rm fd}(k ' ) \right ] \nonumber , \label{xrate_ion}\end{aligned}\ ] ] where the + ( - ) sign between the direct and exchange matrix elements corresponds to the singlet ( triplet ) electron spin configuration . for an unpolarized electron gas , an averaging over the two spin configurations must be performed , in order to take into account the contributions of all the electrons to the exciton linewidth due to dissociating scattering : @xmath93 @xmath94 ( @xmath95 ) denotes the exciton linewidth contribution from the singlet ( triplet ) spin configuration , and both are calculated using eq . ( [ gam_ex ] ) . we note that the spin configuration averaging is irrelevant in the elastic scattering case , since the direct term there is much smaller than the exchange term . this results in similar contributions from both singlet and triplet configurations . = 0.65 figure [ lw_ion]a shows the heavy hole exciton linewidth due to dissociating scattering at @xmath96 , and @xmath71 . although the magnitude of @xmath97 is of the same order as the elastic scattering linewidth ( compare to fig . [ fgam_ex]a ) , its functional dependence on the exciton in - plane momentum is very different . in particular , we note that the maximal linewidth is obtained at a very large momentum ( @xmath98 ) . this is due to the fact that in order for an exciton with initially small @xmath99 to be ionized , it must scatter on an electron with energy large enough to overcome its binding energy . this is less likely as the temperature decreases , as is evident from fig . [ lw_ion]b , where @xmath97 is plotted at @xmath3 . the insets in figure [ lw_ion ] show the maximum values of @xmath97 vs. @xmath73 . as in the elastic scattering case , the linear dependence on @xmath73 holds for much larger electron densities at @xmath71 . at low temperatures and low @xmath73 , charged excitons are formed . in this section we calculate the trion linewidth due to scattering with free electrons . three scattering processes are considered : elastic scattering , where the trion remains bound , capturing scattering , where the heavy - hole exciton captures a free electron to form a trion , and dissociating scattering where the trion dissociates into a heavy - hole exciton and an extra free electron . the latter process is important due to the small binding energy of the trion with respect to the exciton . we neglect the scattering in which both electrons become unbound . in order to facilitate the matrix elements calculation , we consider a simple two parameter chandrasekhar type @xcite trial function for the trion @xmath100 , \label{phitr}\end{aligned}\ ] ] where the + ( - ) sign applies to the singlet ( triplet ) spin configuration . in eq . ( [ phitr ] ) , @xmath101 are the in - plane coordinates of the two electrons with respect to the hole in - plane coordinate , @xmath102 are the electrons @xmath9 coordinates , and @xmath103 , @xmath104 are the confinement functions in the @xmath9 direction , taken to be the same for all constituents ( eq . ( [ chi ] ) ) . the two electron orbitals are given by the normalized functions @xmath105 and the trion wavefunction normalization factor is given by @xmath106 where @xmath107 the trial function ( [ phitr ] ) was used in @xcite for the case of a two dimensional negative - donor center ( i.e. , @xmath108 limit ) . in addition , these authors performed the binding energy calculations using a more complicated version of ( [ phitr ] ) which includes a correlation term ( with an additional variational parameter ) . the variational parameters @xmath109 are calculated by maximizing the trion binding energy for the two spin configurations . the details of the binding energy calculation are given in appendix c , resulting in @xmath110 , @xmath111(@xmath112 , @xmath113 ) for the singlet ( triplet ) spin configuration , where @xmath114 is calculated with respect to the heavy hole 1s neutral exciton energy . we note that the triplet configuration is barely bound , in accordance with the experimental observations , thus it can be disregarded . comparing with the much more elaborate treatment of stb _ et al _ @xcite , our results are fairly accurate , giving us the confidence to proceed in calculating the trion - electron scattering matrix elements using the wave function ( [ phitr ] ) . we construct the state of a single trion with an in - plane cm momentum @xmath115 , similarly to that of a single exciton ( [ ex_st ] ) : @xmath116 where we have added the spin index to the electron creation operators , and denoted @xmath117 , and @xmath118 as the projection of a generic spin configuration of two electrons on the singlet spin configuration . the in - plane fourier transformed trion wave function in eq . ( [ tr_st ] ) is given by @xmath119 where , @xmath120^{3/2 } } \\ \phi'_{{\bf k}}&=&\sqrt{\frac{8 \pi \lambda^{\prime 2}}{a } } \frac{1 } { \left[1+(\lambda ' k)^2 \right]^{3/2}}.\end{aligned}\ ] ] a state comprising a trion and a free electron having @xmath121 , @xmath12 , and @xmath122 will be written as @xmath123 in the following we shall use these wave functions to calculate the scattering matrix elements for the various scattering processes . applying the coulomb interaction operators @xmath124 , @xmath125 , and @xmath23 to the initial state of the system ( [ initr ] ) , and discarding the self energy contributions , we can write a generic expression for the three scattering matrix elements : @xmath126 where we put the @xmath127 prefactor to indicate averaging over the initial electron spin states , and used a numeric index for the electron operators , indicating both spatial and spin degrees of freedom . note that we have omitted the hole operators since they anticommute with the electron operators and do not contribute additional constraints . using the electrons anticommutation relations , the fermi vacuum expectation value of the operators reads : @xmath128 where we have symbolically written @xmath129 . for the 1st electron interaction term we have : @xmath130 the second electron interaction term is found by making the substitutions : @xmath131 , @xmath132 in eqs . ( [ ops ] ) . similarly , the electron - hole interaction term is found by substituting : @xmath131 in ( [ ops ] ) . applying the delta functions of ( [ delspin ] ) to eq . ( [ vei ] ) , it is evident that the contributions of the @xmath9-dependent parts are identical for all three terms , since we have taken identical confinement functions for the electron and the hole ( see eq . [ chi ] ) . assuming that the trion remains in the singlet electron spin configuration , we sum over the spin degrees of freedom , resulting in selection rules for the three electron final spin states , which determine the relevant signs of the various terms . performing the in - plane momentum integrations , and collecting the matrix elements into one direct and two exchange contributions , we finally find : @xmath133- \right . \nonumber \\ & & \left . \left [ g(\lambda \alpha_t q/2 ) g(\lambda ' \alpha_t q/2)+\kappa^2g^2(\tilde{\lambda } \alpha_t q ) \right ] \right\ } \label{trion_dir}\end{aligned}\ ] ] and [ trion_exc ] @xmath134 \times \right . \nonumber \\ & & \left . \left [ g(\lambda ' |{\bf k}_1-\delta { \bf k}_t+{\bf q}|)- g(\lambda ' |\delta { \bf k}_t- { \bf q}| ) \right ] + [ \lambda \leftrightarrow \lambda']\times [ \lambda \leftrightarrow \lambda ' ] \right\ } \\ v^{(2)}_{\rm exc}({\bf q},\delta{\bf k}_t)&= & -\frac{{\cal n}_{\rm es}}{\pi } \int d^2 k_1 h(\lambda k_1)\left\ { \left [ \lambda \lambda ' \kappa g(\tilde{\lambda}|{\bf k}_1-\alpha_t q| ) g(\lambda ' |{\bf q}-\delta { \bf k}_t| ) + \lambda^2 g(\lambda ' k}_t| ) \right ] \times \right . \nonumber \\ & & \left . g(\lambda |{\bf k}_1-\delta { \bf k}_t+\alpha_t{\bf q}| ) + [ \lambda \leftrightarrow \lambda']\times[\lambda \leftrightarrow \lambda ' ] \right\},\end{aligned}\ ] ] where we have denoted @xmath135 , @xmath136 , and @xmath137 . in eqs.([trion_dir])-([trion_exc ] ) , the prefactor , multiplying the dimensionless quantities , is given by @xmath138 we note that the direct term ( [ trion_dir ] ) diverges at @xmath139 . this divergence originates from the infinite range of the coulomb potential , which is manifest whenever a scattering event between two charged particles occurs ( note that for the neutral exciton this divergence cancelled out , due to the equal contributions with opposite signs from the electron and hole constituting the exciton ) . in practice the coulomb potential is screened by the presence of the 2deg . since we are considering a low density 2deg , its fermi wavevector , @xmath140 is small , thus the semiclassical ( thomas - fermi ) approximation , which assumes that @xmath141 , is inadequate @xcite . in particular , the two - dimensional semiclassical analysis results ( at @xmath142 ) in a density - independent screening wavevector @xmath143 , implying that the presence of a single electron is sufficient to screen the external potential , which is clearly unrealistic . we adopt instead the lindhard approach which utilizes a perturbative scheme to evaluate the induced charge density in first order of the total potential ( by total we mean the sum of the external potential , i.e. , that of the trion , and the 2deg potential that is induced by the trion s presence ) . the in - plane fourier transformed dielectric function which results from the effect of the screening is given by @xcite @xmath144 where @xmath145 and @xmath146 is the electrons effective bulk bohr radius . two factors appear in eq . ( [ screene ] ) , which influence the screening . the first is the screening form factor given by @xmath147 which arises from the finite qw width . in the infinite barrier limit we are considering , the confinement functions , @xmath148 , are given by eq . ( [ chi ] ) , and @xmath149 can be calculated analytically , resulting in @xmath150 ^ 2}.\ ] ] this screening form factor is smaller than 1 , thus it reduces the effect of the screening , compared with the strictly two - dimensional case ( @xmath151 ) . the second factor which appears in eq . ( [ screene ] ) is given by @xmath152 we note that @xmath153 does not depend on the particular details of the qw , but only on the two - dimensional nature of the 2deg motion . substituting the screening dielectric function into the coulomb interaction , we find the screened potential to be @xmath154 thus the use of the screened potential implies replacing @xmath155 in eq . ( [ trion_dir ] ) with @xmath156 in general , the effect of the quasi two - dimensional screening saturates in larger electron densities , in contradistinction to the bulk case , resulting in a complicated dependence of the trion linewidth on @xmath73 . we note that since there are no divergences in the exchange terms , replacing @xmath157 with @xmath158 in eqs . ( [ trion_exc ] ) amounts to a very small effect ( less than @xmath159 in the relevant parameter range ) . = 0.6 figure [ matel_fig]a shows the direct and exchange dimensionless matrix elements as a function of transferred momentum @xmath38 , for the case @xmath160 , where the angular dependence of @xmath161 , @xmath162 disappears . the effect of the screening on the total dimensionless matrix element is shown in fig . [ matel_fig]b , where various electron densities are considered , giving rise to a change in the screening action . as expected , the effect of the screening is noticeable in the region of small momenta of the direct matrix element . = 0.57 the elastic scattering rate of the trion , @xmath163 , is calculated by writing the energy - momentum conservation equation and finding its solution @xmath164 , similarly to eqs.([econs])-([rate_ex ] ) . integrating over all final trionic states , we calculate the trion linewidth due to elastic electron scattering , @xmath165 , plotted for two electron densities in fig . [ trion_lw ] . the linewidth dependence on @xmath73 is quite complicated . first , as we are considering @xmath3 , the density range for which the linewidth exhibits a linear dependence on @xmath73 is quite small , similar to the neutral exciton scattering ( cf . [ fgam_ex]c ) . second , the screening potential becomes larger as the density decreases , thus for very small densities ( @xmath166 ) decreasing @xmath73 results in a larger trion linewidth . we note that for such low densities , it is probable that other screening mechanisms become appreciable ( e.g. lattice impurities ) , thus such a behavior will probably not be observed experimentally . we now consider the case in which a heavy - hole exciton captures a free electron thereby forming a trion . the initial and final states of the system are given by eqs . ( [ initial ] ) and ( [ tr_st ] ) , respectively . the calculation of the coulomb interaction matrix elements between these states essentially follows the procedure described in the previous section . in this process there is no exchange term . instead , the two electrons which constitute the trion form a singlet spin configuration . defining @xmath167 , the calculation yields : @xmath168 g\left(\lambda \left|{\bf k}_1+\frac{\delta { \bf k}}{\alpha_{\rm x}}\right| \right ) + \right . \nonumber \\ & & \left . \frac{\lambda \lambda'}{(\lambda_0+\lambda)^2 } \left [ g\left(\tilde{\lambda } |{\bf k}_1+\delta { \bf k}|\right)-g(\tilde{\lambda } \delta k ) \right ] \times \right . \\ & & \left . g\left ( \lambda ' \left|{\bf k}_1+\frac{\delta { \bf k}}{\alpha_{\rm x}}\right| \right ) \right\ } \nonumber\end{aligned}\ ] ] where we have defined the reduced bohr radii : @xmath169 and @xmath170 , and the prefactor : @xmath171 fig . [ fig_extr]a shows the dependence of the dimensionless matrix element on @xmath42 . proceeding as before , we write conservation of energy : @xmath172 which is satisfied for @xmath173 where @xmath174 . we note that for the energy conserving solutions , @xmath175 , the interaction matrix element is constant since @xmath176 . thus , the dependence of the heavy hole exciton linewidth on its initial momentum , shown in fig . [ fig_extr]b , comes solely from energy momentum conservation and the fermi - dirac distribution of the scattered electron : @xmath177 . lastly , we consider the case in which the scattered trion dissociates into a neutral exciton and a free electron . this scattering process is shown schematically in fig.[trex_scat ] . the initial state of the system is given by eq . ( [ initr ] ) while the final state is that of three free particles , symbolically written as @xmath178 , where @xmath179 is the momentum transferred from the free electron to the trion , and @xmath7 is the neutral exciton cm momentum . ( 160,150 ) ( -5,0)(175,150 ) ( 15,115)t ( 25,118 ) ( 30,115)(1,-1)30 ( 0,135)@xmath180 ( 15,30)e ( 25,32 ) ( 30,35)(1,1)30 ( 10,10)@xmath181 ( 80,90)x ( 90,92 ) ( 95,95)(1,1)30 ( 87,135)@xmath182 ( 80,73)e ( 90,75 ) ( 95,75)(1,0)40 ( 100,83)@xmath183 ( 80,55)e ( 90,58 ) ( 95,55)(1,-1)30 ( 80,10)@xmath184 proceeding as before , we calculate the interaction matrix elements between the initial and final states of the system . to this end we denote the final spin state of the three electrons as @xmath185 ( see fig . [ trex_scat ] ) . assuming the interaction is spin independent , this final state will be one of the two states : @xmath186 , where the state with @xmath187 ( @xmath188 ) corresponds to the initial electron spin @xmath122 being up ( down ) . actually , we need not concern ourselves with the exact form of the final electron spin state , since the projection of the initial spin state ( which is a tensor product of a singlet state with the additional free electron spin ) on a generic final spin state will extract the relevant contributions . the proper ( anti)symmetrization of this final spin state is performed inherently by the anticommutation relations of the electron operators , as in eq . ( [ delspin ] ) . as in the elastic scattering case , we find one direct and two exchange contributions . denoting @xmath189 , and @xmath190 as before , these terms are expressed as : @xmath191 \right\ } + \right . \nonumber \\ & & \left . \frac{\lambda \lambda'}{(\lambda_0+\lambda')^2 } \left\ { { \renewcommand{\arraystretch}{0.5 } \begin{array}{l } \lambda \leftrightarrow \lambda ' \\ \tilde{\lambda } \rightarrow \tilde{\lambda } ' \end{array } } \right\ } \right ) \label{trex_dir}\end{aligned}\ ] ] and [ trex_exc ] @xmath192 \times \right . \nonumber \\ & & \left . \left . g(\lambda'| { \bf q}-\delta { \bf k}_{\rm t}| ) + g(\lambda |{\bf k}_1-\delta { \bf k}_{\rm t}+{\bf q}| ) g\left(\lambda ' \left| \frac{\delta { \bf k}}{\alpha_{\rm x}}+{\bf q } \right| \right ) \right\}+\ { \lambda \leftrightarrow \lambda ' \ } \right ) \\ v^{(2)}_{\rm exc}({\bf q},\delta{\bf k}_{\rm t},\delta{\bf k}_{\rm x})&= & -\frac{1}{2}{\cal n}_{\rm dis } h_s\left(\lambda_0 \left| \frac{\delta { \bf k}_{\rm x}}{\alpha_{\rm x}}-\frac{\delta { \bf k}_{\rm t}}{\alpha_{\rm t}}+{\bf q}\right| \right ) \left ( \frac{\lambda \lambda'}{(\lambda_0+\lambda)^2 } \left\ { g(\lambda ' \frac{\delta { \bf k}_{\rm t}-\beta_{\rm x}\delta { \bf k}_{\rm x}}{\alpha_{\rm x}}-{\bf q } \right| \right ) + \right . \nonumber \\ & & \left . g(\tilde{\lambda } \delta k ) \left [ g\left(\lambda ' \frac{\delta k}{\alpha_{\rm x}}\right ) - g(\lambda'| \delta { \bf k}_{\rm t}-{\bf q}| ) \right ] \right\ } + \frac{\lambda \lambda'}{(\lambda_0+\lambda')^2 } \left\ { { \renewcommand{\arraystretch}{0.5 } \begin{array}{l } \lambda \leftrightarrow \lambda ' \\ \tilde{\lambda } \rightarrow \tilde{\lambda } ' \end{array } } \right\ } \right ) , \label{trex_exc2}\end{aligned}\ ] ] where @xmath193 , @xmath194 and @xmath42 were defined in the previous subsection and the prefactor multiplying the dimensionless quantities in eqs.([trex_dir])-([trex_exc ] ) , is given by @xmath195 the direct term ( [ trex_dir ] ) and the 2nd exchange term ( [ trex_exc2 ] ) diverge at @xmath196 and at @xmath197 , respectively , implying the need of using a screened coulomb potential , as was done in the trion elastic scattering . note that the direct term ( [ trex_dir ] ) depends only on @xmath198 , and on the angles @xmath199 and @xmath200 , whereas the exchange terms depend also on the initial electron momentum @xmath201 and on the angle @xmath202 . in practice , we express everything in terms of @xmath203 , and the angles : @xmath204 , @xmath205 , and @xmath206 , given by : @xmath207 the calculation of the scattering rate is performed as before , where conservation of energy now reads : @xmath208 equation ( [ cons_trex ] ) is satisfied for @xmath209 \right\}.\end{aligned}\ ] ] = 0.37 using fermi s golden rule as in eq . ( [ xrate_ion ] ) , we find the scattering rate @xmath210 . integrating over all final trionic states , we find the trion linewidth due to inelastic scattering as a function of its initial momentum @xmath211 , @xmath212 , shown in fig . [ lw_trex ] for @xmath213 . in this section we apply the linewidth calculations for the various scattering processes , presented in the previous sections , in order to calculate the line shapes of neutral excitons and trions . these line shapes can be directly compared with measured reflection spectra , taken from mtqw structures containing a variable density 2deg . without electrons in the qw and for @xmath214 , the two dominant scattering processes are due to acoustic phonons and static disorder . we note that the radiative lifetime of a free qw exciton is @xmath215 , resulting in a radiative homogeneous broadening of @xmath216 . the exciton - acoustic phonon interaction leads to a homogeneous broadening whose lorentzian line shape , @xmath217 , is characterized by @xmath218 for the @xmath219 and @xmath220 for the @xmath221 at @xmath71 , using the parameters of gaas / alas qw s @xcite . the 1s exciton energy , @xmath222 , has a gaussian distribution ( characterized by @xmath223 ) due to the roughness of the qw surface . convolving these two distributions results in the familiar voigt function @xcite @xmath224 where we take @xmath225 . when a low density 2deg is present , @xmath226 is convolved with a lorentzian lineshape that is associated with the homogeneous broadening that is due to the electron scattering : @xmath227 where @xmath228 is given by eq . ( [ gam_ex ] ) , with the exciton energy , @xmath229 , corresponding to @xmath99 . a pictorial description of the convolution in eq . ( [ ls ] ) is given in fig . [ conv ] , for both elastic and dissociating exciton - electron scattering processes ( the trion - electron scattering exhibits a similar behavior ) . it is seen that the electron scattering admixes excitons having @xmath230 with the @xmath55 state . the degree of admixture is determined by the value of the lorentzian that is due to electron scattering , given in the figure by the dashed line . note that since @xmath231 is a decreasing function of @xmath229 , the lorentzian peak is shifted to higher exciton energies . ( this is observed experimentally as an increased shift with increasing @xmath77 @xcite ) . = 0.45 in the calculation , the gaussian width , associated with the inhomogeneity , was taken as a fitting parameter , yielding the value @xmath232 for both @xmath219 and @xmath221 ( the same value was taken also for the trion line ) . the calculated line shapes are identified with the imaginary part of the dielectric function in the qw that is related to the exciton ( trion ) resonance . use of the kramers - kronig relations yields the real part of the dielectric function . examples of calculated exciton line shapes together with the associated real part of the dielectric function are depicted in fig [ line_sh ] for two electron densities . = 0.4 combining the three exciton line shapes ( heavy - hole and light - hole neutral excitons , and heavy - hole trion ) , by using the appropriate oscillator strengths , the reflection spectra of a qw can be obtained . for high temperature ( @xmath71 ) , where the trion is dissociated , these calculated spectra and their dependence on @xmath73 were shown to be in good agreement with the experimental data @xcite . in the low temperature regime ( @xmath233 ) , the vindication of our results with experimental measurements is harder to perform . since the heavy hole exciton and trion lines are very close in energy , and the trion oscillator strength is much lower than that of the exciton , it is practically impossible to discriminate between them by reflection measurements . pl experiments were performed in mtqw with 2deg , at low temperatures , facilitating the observation of the trion line @xcite , but there the detailed balance between the exciton and the trion line intensities is determined by the 2deg density , and the system dynamics should be addressed by solving , e.g. , rate equations . = 0.6 in fig . [ amnon]a the trion linewidth due to elastic electron scattering is shown as a function of @xmath73 for three temperatures . the combined effect of both elastic and dissociating electron scattering on the trion linewidth in shown in fig.[amnon]b . qualitatively , the trion line observed in @xcite indeed broadens considerably at much lower @xmath234 than the exciton line , as indicated by our results . the electron densities stated in @xcite are probably an overestimate , as implied by the fact that the neutral exciton linewidth remains unchanged up to @xmath235 , in clear contradiction to both our calculations and other experimental data ( e.g. , @xcite ) . it should be noted that calibrating the photoexcitation intensity @xmath236 to @xmath234 in mtqws is not an easy task . even if one assumes a linear dependence of @xmath73 on @xmath236 ( which is valid for a limited range ) , the slope depends on the temperature , since the holes tunneling times decrease exponentially with @xmath75 , so a different calibration should be performed not only for different experiments , but also for each temperature . finally , the strong enhancement of the dissociating scattering with increasing temperature , evident in fig . [ amnon]b , implies that the trion line should practically vanish for @xmath237 at very low electron densities . in this paper we presented a theoretical study of the various exciton - electron scattering processes that take place in a qw which contains a low density 2deg . we have demonstrated that even for small densities , of the order of @xmath238 in the case of a gaas / alas qw , exciton - electron scattering is much more efficient than both exciton - phonon and exciton - exciton scattering , thus playing a central role in the system dynamics . the dependence of the exciton linewidth on @xmath73 was shown to be drastically different for the cases of a classical ( @xmath71 ) and a degenerate ( @xmath3 ) 2deg . in general , the trion - electron scattering poses a much harder problem , one of the major difficulties being the evaluation of the screening effects . we showed that the screening model has a crucial effect on the trion - electron scattering matrix elements , and hence on the resulting trion linewidth . this is particularly true for very low densities ( @xmath239 ) , where the 2deg screening is less effective , making other possible mechanisms more dominant . using the calculated scattering rates we showed that dissociating scattering can produce a sizable broadening of the excitonic lines , in particular for the trion , whose binding energy is considerably smaller than that of the neutral exciton . a simple method to produce excitonic line shapes from the calculated linewidths was devised , allowing our results to be easily compared with experimental reflection spectra . a reliable source of experimental data is achievable from reflection measurements taken from a mtqw structure , embedded in a microcavity ( mc ) . in the limit of large mc mode energy , the exciton linewidths are restored , since their coupling with the mc mode is barely noticeable . a major advantage of mc experiments is that the trion line is easily resolved by reflection @xcite , allowing a direct comparison with theory . the calculated interaction matrix elements can be readily used to evaluate polariton - electron scattering rates in mcs with 2deg . in a recent experiment we have demonstrated polariton final - state stimulation , assisted by these scattering mechanisms , which was hitherto observed using polariton - polariton scattering mechanism @xcite . incorporation of the various scattering processes which were considered in this paper , in a consistent explanation for these nonlinear effects , will be the topic of a forthcoming publication . this work was supported by the united states - israel binational science foundation ( bsf ) , jerusalem , israel , by the fund for promotion of research at the technion , and by the technion vpr fund - jewish communities of germany research fund . g.r . gratefully acknowledges the financial help from the technion . in this appendix we provide the details of the direct term calculation . separating the coordinates in the qw plane ( @xmath240 ) from the perpendicular coordinate ( @xmath9 ) , and denoting the electron and hole in - plane momenta by @xmath8 and @xmath241 , respectively , we write the in - plane fourier transform of the exciton wave function @xmath242 $ ] @xmath243 \times \nonumber \\ & & { \rm e}^{-{\rm i}({\bf k}_1 \cdot { \bf r}_{e\| } + { \bf k}_2 \cdot { \bf r}_{h\|})},\end{aligned}\ ] ] where @xmath26 denotes the qw surface area . transforming to center - of - mass ( cm ) and relative coordinates in the qw plane ( @xmath244 , where @xmath245 , @xmath29 , and @xmath246 , @xmath15 are the electron and exciton in - plane effective masses , respectively ) we can decompose the exciton wave function into an envelope function , @xmath247 , and a free motion part related to the in - plane cm coordinate . denoting @xmath7 as the in - plane cm momentum we find @xmath248 } \nonumber \\ & = & \delta({\bf k}_{\rm x}- { \bf k}_1-{\bf k}_2 ) \phi_{{\bf k}_1 -\alpha { \bf k}_{\rm x } } ( z_e , z_h)\end{aligned}\ ] ] we use the simplest exciton wave function for the exciton ground state : @xmath249 , where @xmath250 is a normalization factor and @xmath32 is a variational parameter associated with the exciton bohr radius in the qw , which is fixed by maximizing the binding energy of the exciton @xcite . we note that the use of a wave function separable in @xmath9 and @xmath251 , facilitates considerably the calculation of the scattering matrix elements , although it is strictly justifiable only for narrow well structures . assuming perfect confinement of electrons and holes in the qw , and taking the @xmath9 axis origin in the center of the qw , the confinement functions are : @xmath252 this approximation is quite accurate for almost all practical thicknesses @xmath31 . calculating the binding energy of the 1s exciton in a gaas qw of width @xmath253 , we find @xmath254(@xmath255 ) for the heavy - hole ( light - hole ) exciton . these results , obtained with our single parameter trial function , are quite close to those obtained in @xcite , which took into account finite qw barriers , and anisotropic masses . combining the first terms from eqs . ( [ matel ] ) , the direct term reads @xmath256 g(\lambda k_1 ) \times \nonumber \\ & & \cos^2(\pi z_e / l ) \cos^2(\pi z_h / l ) \cos^2(\pi z_c / l),\end{aligned}\ ] ] where @xmath257 was defined in eq . ( [ gh ] ) . we note that the electron wave functions in the qw plane are absent , as they contribute , together with the in - plane center of mass part of the exciton wave function , a fixed phase factor which is unimportant for our purposes . the momentum integral is a simple convolution and can be readily evaluated @xmath258 g(\lambda k_1 ) \nonumber \\ & = & \frac{1}{4 \pi^2 } \int d^2 \rho \tilde{g}^2(\rho/\lambda ) \left [ \exp ( { \rm i } \mbox{\boldmath $ \rho$ } \cdot \beta { \bf q } ) - \exp ( { \rm i } \mbox{\boldmath $ \rho$ } \cdot \alpha { \bf q } ) \right ] \nonumber \\ & = & \frac{\pi}{2 \lambda^2 } \left [ g(\lambda \beta q/2 ) - g(\lambda \alpha q/2 ) \right]\end{aligned}\ ] ] where @xmath259 is the exciton spatial wave function in the qw plane . in order to evaluate the @xmath9 integrals , we change to the coordinates @xmath260 which implies the following change in the integration limits : @xmath261 and performing the integrations results in the direct term , given by eq . ( [ vdir ] ) in this appendix we present an analytical calculation of the angular part of the exchange integral . this calculation , while quite complicated , is worth the trouble for two reasons . the first is , obviously , a major reduction in the time consumed by the numerical computation , making it feasible . as a subsidiary benefit , the partially analytical calculation removes some convergence problems encountered in the previously used numerical algorithm . using dimensionless momenta ( e.g. , by writing @xmath38 we mean @xmath262 ) the exchange integral reads @xmath263^{-3/2}-\left [ 1+({\bf q}- \delta { \bf k})^2 \right]^{-3/2 } \right\ } \times \nonumber \\ & & \left[1+({\bf k}_1- \delta { \bf k}+\alpha { \bf q})^2 \right]^{-3/2 } \equiv -4 \pi \int dk_1 k_1 h(k_1 ) \left\ { ( i ) + ( ii ) \right\ } , \label{geni}\end{aligned}\ ] ] where @xmath50 is the angle between @xmath179 and @xmath264 and @xmath51 is the angle between @xmath8 and @xmath264 . we shall work out separately the two parts of eq . ( [ geni ] ) . the first part is given by @xmath265^{-3/2},\end{aligned}\ ] ] where we have defined @xmath266 transforming @xmath267 and then to @xmath268 we have @xmath269^{-3/2}.\end{aligned}\ ] ] finding the roots of this 4th order polynomial we have @xmath270^{-3/2 } \label{i1}\end{aligned}\ ] ] where @xmath271 we note for later use that @xmath272 obey the inequalities @xmath273 one can reduce the polynomial in eq . ( [ i1 ] ) to a product of two quadratic binomial forms , by making use of the following transformation @xmath274 with @xmath275 , @xmath38 given by @xmath276 ( note that we assume @xmath277 , deferring for later the treatment in this special case ) . noticing that @xmath275 and @xmath38 always have opposite signs , we choose @xmath278 and divide the integral into @xmath279 where the limit @xmath280 corresponds to @xmath281 and the limit @xmath282 corresponds to @xmath281 with @xmath38 and @xmath275 interchanged . the integral now reads @xmath283 ^ 2 } { \left [ ( y^2+s^2)(y^2+t^2 ) \right]^{3/2 } } - \frac{1}{(st)^3 } \int^{-q / p}_{-1 } dy \frac { \left [ y^2(1+p^2)+2y(1+pq)+1+q^2 \right]^2 } { \left [ ( y^2 + 1/s^2)(y^2 + 1/t^2 ) \right]^{3/2 } } \right\}\ ] ] where @xmath284 and @xmath285^{3/2}}.\ ] ] making the transformation @xmath286 in the first integral and @xmath287 in the second , we find @xmath288 ^ 2}{(1-k^2 \sin^2 \psi)^{3/2 } } - \right . \\ & & \left . st^3 \int_{\tan^{-1 } ( -s)}^{\tan^{-1 } ( -qs / p ) } d \psi \frac{\left [ ( p / s \sin \psi + q \cos \psi)^2+(1/s \sin \psi + \cos \psi)^2 \right]^2}{(1-\tilde{k}^2 \sin^2 \psi)^{3/2 } } \right\},\end{aligned}\ ] ] where we have defined @xmath289 changing @xmath290 in the second integral , and using @xmath291 we find that the second integral exactly matches the first except that its upper limit is larger by @xmath292 ; thus we are left with @xmath293 ^ 2}{(1-\tilde{k}^2 \sin^2 \psi)^{3/2}}.\ ] ] this integral can be evaluated using incomplete elliptic integrals ( see p. 201 in @xcite ) . furthermore , one can exploit the functional relations ( p. 911 in @xcite ) @xmath294 where @xmath295 denote incomplete elliptic integrals of the first and second kind , respectively , and @xmath296 denote complete elliptic integrals of the first and second kind , respectively . the final result is @xmath297\left [ \frac{1}{t^2}(1+p^2)^2+(1+q^2)^2s^2 \right ] - \right . \nonumber \\ & & \left . 2(3 + 3p^2q^2 + 4pq+p^2+q^2 ) \left [ 2s^2 { \bf e}(\tilde{k})-(s^2+t^2 ) { \bf k}(\tilde{k } ) \right ] \right\ } \ . \label{ires}\end{aligned}\ ] ] for imaginary values of @xmath298 , the result is the same as eq . ( [ ires ] ) with the interchanges @xmath299 and @xmath300 , @xmath301 being defined in ( [ k ] ) . we now return to examine two special cases which can not be treated using ( [ ires ] ) . the first case corresponds to @xmath302 . here we have @xmath303 although @xmath304 , @xmath298 is still finite , taking the value @xmath305 thus , enabling the use of ( [ ires ] ) with some modifications . remembering the inequalities ( [ ineq ] ) we note that the sign of @xmath306 is determined by the sign of @xmath307 . we obtain , for @xmath308 @xmath309^{3/2 } ( d - b)^2 } \times \label{ires1 } \\ & & \!\!\ ! \left\ { \left[1 + \frac{(1+a^2)^2}{(b - a^2)(d - a^2 ) } \right ] \left [ ( 1+d - a^2){\bf e}(\tilde{k})- \right . \right . \nonumber \\ & & \!\!\ ! \left . 2 { \bf k}(\tilde{k } ) \right ] \!\!- 2(3a^2 + 1 ) \!\ ! \left [ 2{\bf e}(\tilde{k})-\!\ ! \left ( 1+\frac{b - a^2}{d - a^2 } \right){\bf k}(\tilde{k } ) \right ] \right\}. \nonumber\end{aligned}\ ] ] for @xmath310 the result is the same as eq . ( [ ires1 ] ) with the interchanges @xmath311 and @xmath312 . the second special case corresponds to @xmath313 . here the integral ( [ i1 ] ) reduces to @xmath314^{3/2 } } \int_{-\infty}^{\infty } dz \frac{(1+z^2)^2}{(z^2 + 2az+b)^3}.\ ] ] this integral can be easily evaluated ( see p. 81 in @xcite ) and the result is @xmath315^{3/2}(b - a^2)^{5/2}}.\ ] ] finally , we evaluate the second part of the exchange integral , appearing in eq . ( [ geni ] ) . we have @xmath316 where @xmath317 where defined in ( [ abcd ] ) . using the transformation @xmath318 , where @xmath319 , we have @xmath320 where @xmath321 . this integral can be evaluated using the elliptic integral of second kind , resulting in @xmath322 in this appendix we present the calculation of the trion binding energy , based on the trial wave function given in eq.([phitr ] ) . it is useful to work with the fourier transformed trion wave function ( [ tr_k ] ) , where we have for the two spin configurations , explicitly @xmath323.\end{aligned}\ ] ] the effective mass hamiltonian is given by and we employed natural units of length and energy , namely , bulk effective bohr radius and rydberg . the expectation value of the trion energy is given by @xmath330 where @xmath331 was defined in eq . ( [ s ] ) , and the upper ( lower ) signs apply to the singlet ( triplet ) spin configuration . the various terms appearing in eq . ( [ etr ] ) are calculated in the in - plane fourier space , resulting in simple 2d convolutions , similar to those in appendix a. the results are : @xmath332= \left ( \frac{\pi}{l } \right)^2(2+\sigma_z ) + \nonumber \\ & & \left[\frac{1}{\lambda^2}(1+\sigma)-\lambda^3\left(\frac{2}{l } \right)^2 \int_0^{\infty}dq q h(\lambda q)g(\lambda q/2 ) \right]+[\lambda \rightarrow \lambda ' ] \\ { \cal i}_2&=&\left(\frac{2}{l}\right)^2 \langle \phi \chi_1 \chi_h |h_0| \phi ' \chi_1 \chi_h \rangle = \frac{\kappa}{2}\left ( \frac{\pi}{l}\right)^2 ( 2+\sigma_z)+ \frac{4(1+\sigma)}{(\lambda+\lambda')^2}-(\lambda+\lambda ' ) \left(\frac{\tilde{\lambda}\kappa}{l}\right)^2 \int_0^{\infty}dq q h(\tilde{\lambda}q ) g(\tilde{\lambda}q ) \\ { \cal j}_1&=&\left(\frac{2}{l}\right)^2 \langle \phi_1 \chi_1 \phi'_2 \chi_2|h_{12}|\phi_1 \chi_1 \phi'_2 \chi_2 \rangle = 2\lambda^3 \left(\frac{2}{l}\right)^2 \int_0^{\infty}dq q h(2\lambda q ) g(\lambda q/2)g(\lambda'q/2 ) \\ { \cal j}_2&=&\left(\frac{2}{l}\right)^2 \langle \phi_1 \chi_1 \phi'_2 \chi_2 |h_{12}|\phi'_1 \chi_1 \phi_2 \chi_2 \rangle = 2\kappa^2\tilde{\lambda}^3 \left(\frac{2}{l}\right)^2 \int_0^{\infty}dq q h(2\tilde{\lambda}q ) g^2(\tilde{\lambda}q).\end{aligned}\ ] ] in eqs . ( [ ij ] ) , @xmath333 are the functions defined in eq . ( [ gh ] ) , and we have denoted @xmath334 . we note that in the 2d limit ( @xmath335 ) , eliminating the kinetic energy term in the @xmath9 direction , eqs . ( [ ij ] ) reduce to @xmath336 , & \lambda<\lambda ' \\ \frac{\lambda ' \kappa}{(\lambda'-\lambda)^2}\left [ \left(1 + \frac{\lambda^2}{\lambda^{\prime 2}}\right){\bf e}(1-\lambda^{\prime 2}/ \lambda^2)-2 { \bf k}(1-\lambda^{\prime 2}/ \lambda^2)\right ] , & \lambda>\lambda ' \\ \frac{3 \pi}{4 \lambda } , & \lambda=\lambda ' \end{array } \right . \\ { \cal j}_2&=&\frac{3 \pi \kappa}{2 ( \lambda+\lambda')}\end{aligned}\ ] ] where @xmath337 , @xmath338 are the complete elliptic integrals of the first and second kind , respectively . these closed form expressions coincide with the integral expressions of sandler and proetto @xcite , for @xmath339 and vanishing magnetic field ( note a misprint in eq . a6 of this paper ) . in order to maximize the trion binding energy , we must first write the expectation value of the exciton energy : @xmath340 resulting in the 1s exciton bohr radius and binding energy values stated in appendix a. note that in the 2d limit we have @xmath341 minimizing eq . ( [ ex0 ] ) in the @xmath339 limit , results in @xmath342 , @xmath343 , as expected . the trion binding energy with respect to the exciton is related to the trion and exciton energies through @xmath344 where @xmath345 is the confined electron energy , in our infinite barrier model . we note that the kinetic energy terms in the @xmath9 direction cancel out in the expression for the trion binding energy , so they can be removed from the calculation . maximizing eq . ( [ ebtr ] ) with respect to the two variational parameters @xmath109 , we finally find the results given in section [ trion ] , where we used the gaas bulk values of the effective bohr radius ( @xmath346 ) , and rydberg ( 5.7mev ) . 99 r. c. c. leite , j. shah , and j. p. gordon , phys . * 23 * , 1332 ( 1969 ) . l. schultheis , j. kuhl , a. honold and c. w. tu , phys . * 57 * , 1635 ( 1986 ) . a. honold , l. schultheis , j. kuhl and c. w. tu , phys . b * 40 * , 6442 ( 1989 ) . v. capozi , l. pavesi , and j. l. staehli , phys . b * 47 * , 6340 ( 1993 ) . k. matsuda , m. hirooka , and s. sunakawa , prog . theor . phys . * 54 * , 79 ( 1975 ) ; s. g. elkomoss and g. munschy , j. phys . solids * 38 * , 557 ( 1977 ) ; * 40 * , 431 ( 1979 ) ; * 42 * , 1 ( 1981 ) . y .- feng and h. n. spector , j. phys . solids * 48 * , 593 ( 1987 ) ; superlattices and microstructures * 3 * , 459 ( 1987 ) ; y .- p . feng z. huang , and h. n. spector , j. phys . solids * 50 * , 117 ( 1989 ) ; t. s. koh , y .- p . feng and h. n. spector , j. appl . phys . * 81 * , 2236 ( 1997 ) . g. ramon , r. rapaport , a. qarry , e. cohen , a. mann , arza ron , and l. n. pfeiffer , phys . b * 65 * , 085323 ( 2002 ) . m. a. lampert , phys . lett . * 1 * , 450 ( 1958 ) . b. stb and a. ainane , superlatt . microstruct . * 5 * , 545 ( 1989 ) . k. kheng , r. t. cox , y. merle daubign , f. bassani , k. saminadayar , and s. tatarenko , phys . * 71 * , 1752 ( 1993 ) . g. finkelstein , h. shtrikman , and i. bar - joseph , phys . 74 * , 976 ( 1995 ) ; a. j. shields , m. pepper , d. a. ritchie , m. y. simmons , and g. a. c. jones , phys . b * 51 * , 18049 ( 1995 ) ; arza ron , h. w. yoon , m. d. sturge , a. manassen , e. cohen , and l. n. pfeiffer , sol . state comm . * 97 * , 741 ( 1996 ) . a. i. bobrysheva , m. v. grodetskii , and v. t. z. zyukov , j. phys . c * 16 * , 5723 ( 1983 ) ; a. thilagam , phys . rev . b * 55 * , 7804 ( 1997 ) . b. stb , g. munschy , l. stauffer , f. dujardin , and j. murat , phys . rev . b * 56 * , 12454 ( 1997 ) . j. r. chapman , n. f. johnson , and v. n. nicopoulos , phys . b * 55 * , r10221 ( 1997 ) ; d. m. whittaker and a. j. shields , _ ibid _ * 56 * , 15185 ( 1997 ) ; a. wjs , j. j. quinn , and p. hawrylak , _ ibid _ * 62 * , 4630 ( 2000 ) . s. glasberg , g. finkelstein , h. shtrikman , and i. bar - joseph , phys . b * 59 * , r10425 ( 1999 ) . y. yayon , m. rappaport , v. umansky , and i. bar - joseph , phys . b * 64 * , 081308(r ) ( 2001 ) . a. manassen , e. cohen , arza ron , e. linder , and l. n. pfeiffer , phys . b * 54 * , 10609 ( 1996 ) . f. tassone and y. yamamoto phys . b * 59 * , 10830 ( 1999 ) . c. ciuti , v. savona , c. piermarocchi , a. quattropani and p. schwendimann , phys . b * 58 * , 7926 ( 1998 ) . the angle between @xmath7 and @xmath347 is thus given by @xmath348 , where it is noted that since one can make the transformation @xmath349 , the result of the integration over @xmath350 is indifferent to the sign choice in @xmath351 . d. huang , j .- chyi , and h. morko , phys . b * 42 * , 5147 ( 1990 ) . s. chandrasekhar , astrophys . j. * 100 * , 176 ( 1944 ) ; h. a. bethe and e. e. salpeter , _ quantum mechanics of one- and two- electron atoms _ ( plenum , new york , 1957 ) . n. p. sandler and c. r. proetto , phys . b * 46 * , 7707 ( 1992 ) . g. bastard , _ wave mechanics applied to semiconductor heterostructures _ ( les editions de physique cnrs , paris , 1988 ) . r. rapaport , a. qarry e. cohen , arza ron , e. linder , and l. n. pfeiffer , in _ proceedings of the 25th international conference on the physics of semiconductors _ , edited by n. miura and t. ando ( springer - verlag , 2001 ) , p. 679 . j. christen , and d. bimberg , phys . b * 42 * , 7213 ( 1990 ) . v. savona , l. c. andreani , p. schwendimann and a. quattropani , solid state comm . * 93 * , 733 ( 1995 ) . m. grundmann and d. bimberg , phys . b * 38 * , 13486 ( 1988 ) . i. s. gradshteyn and i. m. ryzhik , _ tables of integrals , series , and products _ ( academic press , san diego , 1994 ) . r. rapaport , r. harel , e. cohen , arza ron , e. linder and l. n. pfeiffer , phys . lett . * 84 * , 1607 ( 2000 ) ; phys . a * 178 * , 481 ( 2000 ) . g. savvidis , j. j. baumberg , r. m. stevenson , m. s. skolnick , d. m. whittaker , and j. s. roberts , phys . lett . * 84 * , 1547 ( 2000 ) ; j. j. baumberg , p. g. savvidis , r. m. stevenson , a. i. tartakovskii , m. s. skolnick , d. m. whittaker , and j. s. roberts , phys . b * 62 * r16247 ( 2000 ) ; p. g. savvidis , c. ciuti , j. j. baumberg , d. m. whittaker , m. s. skolnick , and j. s. roberts , _ ibid . _ * 64 * 075311 ( 2001 ) . le si dang , d. heger , r. andr , f. buf , and r. romestain , phys . lett . * 81 * , 3920 ( 1998 ) ; p. senellart , j. bloch , b. sermage , and j. y. marzin , _ ibid . _ * 62 * , r16263 ( 2000 ) ; r. huang , f. tassone , and y. yamamoto , _ ibid . _ * 61 * , r7854 ( 2000 ) .

reed - solomon ( rs ) codes are one of the most widely used error - correcting codes in digital communication and data storage systems . this is primarily due to the fact that rs codes are maximum distance separable ( mds ) codes , can correct long bursts of errors , and have efficient hard - decision decoding ( hdd ) algorithms , such as the berlekamp - massey ( bm ) algorithm , which can correct up to half the minimum distance ( @xmath0 of the code . an @xmath1 rs code of length @xmath2 and dimension @xmath3 is known to have @xmath4 due to its mds nature . since the arrival of rs codes , people have put a considerable effort into improving the decoding performance at the expense of complexity . a breakthrough result of guruswami and sudan ( gs ) introduces a hard - decision list - decoding algorithm based on algebraic bivariate interpolation and factorization techniques that can correct errors beyond half the minimum distance of the code @xcite . nevertheless , hdd algorithms do not fully exploit the information provided by the channel output . koetter and vardy ( kv ) later extended the gs decoder to an algebraic soft - decision ( asd ) decoding algorithm by converting the probabilities observed at the channel output into algebraic interpolation conditions in terms of a multiplicity matrix @xcite . both of these algorithms however have significant computational complexity . thus , multiple runs of error - and - erasure and error - only decoding with some low complexity algorithm , such as the bm algorithm , has renewed the interest of researchers . these algorithms essentially first construct a set of either erasure patterns @xcite , test patterns @xcite , or patterns combining both @xcite and then attempt to decode using each pattern . there has also been recent interest in lowering the complexity per decoding trial as can be seen in @xcite . in the scope of multiple error - and - erasure decoding , there have been several algorithms using different sets of erasure patterns . after multiple decoding trials , these algorithms produce a list of candidate codewords and then pick the best codeword on this list , whose size is usually small . the nature of multiple error - and - erasure decoding is to erase some of the least reliable symbols since those symbols are more prone to be erroneous . the first algorithm of this type is called generalized minimum distance ( gmd ) @xcite and it repeats error - and - erasure decoding while successively erasing an even number of the least reliable positions ( lrps ) ( assuming that @xmath5 is odd ) . more recent work by lee and kumar @xcite proposes a soft - information successive ( multiple ) error - and - erasure decoding ( sed ) that achieves better performance but also increases the number of decoding attempts . literally , the lee - kumar s sed@xmath6 algorithm runs multiple error - and - erasure decoding trials with every combination of an even number @xmath7 of erasures within the @xmath8 lrps . a natural question that arises is how to construct the `` best '' set of erasure patterns for multiple error - and - erasure decoding . inspired by this , we first design a rate - distortion framework to analyze the asymptotic trade - off between performance and complexity of multiple error - and - erasure decoding of rs codes . the framework is also extended to analyze multiple algebraic soft - decision decoding ( asd ) . next , we proposed a group of multiple - decoding algorithms based on this approach that achieve better performance - versus - complexity trade - off than other algorithms . the multiple - decoding algorithm that achieves the best trade - off turns out to be a multiple error - only decoding using the set of patterns generated by random codes combining with covering codes . these are the main results of this paper . the paper is organized as follows . in section [ sec : multiplebma ] , we design an appropriate distortion measure and present a rate - distortion framework to analyze the performance - versus - complexity trade - off of multiple error - and - erasure decoding of rs codes . also in this section , we propose a general multiple - decoding algorithm that can be applied to error - and - erasure decoding . then , in section [ sec : computing - rd ] , we discuss a numerical computation of r - d function which is needed for the proposed algorithm . in section [ sec : multiple asd ] , we analyze both bit - level and symbol - level asd decoding and design distortion measures so that they can fit into the general algorithm . in section [ sec : ext - and - gen ] , we offer some extensions that help the algorithm achieve better performance and running time . simulation results are presented in section [ sec : simulation - results ] and finally , conclusion is provided in section [ sec : conclusion ] . in this section , we set up a rate - distortion framework to analyze multiple attempts of conventional hard decision error - and - erasure decoding . let @xmath9 be the galois field with @xmath10 elements denoted as @xmath11 . we consider an @xmath1 rs code of length @xmath2 , dimension @xmath3 over @xmath9 . assume that we transmit a codeword @xmath12 over some channel and receive a vector @xmath13 where @xmath14 is the receive alphabet for a single rs symbol . in this paper , we assume that @xmath15 and all simulations are based on transmitting each of the @xmath16 bits in a symbol using binary phase - shift keying ( bpsk ) on an additive white gaussian noise ( awgn ) channel . for each codeword index @xmath17 , let @xmath18 be the permutation given by sorting @xmath19 in decreasing order so that @xmath20 . then , we can specify @xmath21 as the @xmath22-th most reliable symbol for @xmath23 at codeword index @xmath17 . to obtain the reliability of the codeword positions ( indices ) , we construct the permutation @xmath24 given by sorting the probabilities @xmath25 of the most likely symbols in increasing order . thus , codeword position @xmath26 is the @xmath17-th lrp . these above notations will be used throughout this paper . consider @xmath27 and @xmath28 . assume that we have the probability @xmath29 written in a matrix form as follows@xmath30_{j , i}\ ] ] then @xmath31 and @xmath32 . [ con : bmaerr - n - era](classical decoding threshold , see @xcite ) : if @xmath33 symbols are erased , a conventional hard - decision error - and - erasure decoder such as the bm algorithm is able to correct @xmath34 errors in unerased positions if @xmath35 [ def:(conv . patterns)](conventional error and erasure patterns ) we define @xmath36 and @xmath37 as an error pattern and an erasure pattern respectively , where @xmath38 means that an error occurs ( i.e. the most likely symbol is incorrect ) and @xmath39 means that an erasure occurs at index @xmath17 . if @xmath5 is odd then @xmath40 is the set of erasure patterns for the gmd algorithm . for the sed@xmath41 algorithm , the set of erasure patterns has the form @xmath42 . here , in each erasure pattern the letters are written in increasing reliability order of the codeword positions . let us revisit the question how to construct the best set of erasure patterns for multiple error - and - erasure decoding . first , it can be seen that a multiple error - and - erasure decoding succeeds if the condition ( [ eq : scbma_0 ] ) is satisfied during at least one round of decoding . thus , our approach is to design a distortion measure that converts the condition ( [ eq : scbma_0 ] ) into a form where the distortion between an error pattern @xmath43 and an erasure pattern @xmath44 , denoted as @xmath45 , is less than a fixed threshold . given a _ letter - by - letter _ distortion measure @xmath46 , the distortion between an error pattern @xmath43 and an erasure pattern @xmath44 is defined by@xmath47 [ prop : bma1]if we choose the _ letter - by - letter _ distortion measure @xmath48 as follows@xmath49 then the condition ( [ eq : scbma_0 ] ) for a successful error - and - erasure decoding then reduces to the form where the distortion is less than a fixed threshold@xmath50 first , we define @xmath51 to count the number of @xmath52 pairs equal to @xmath53 for every @xmath54 . noticing that @xmath55 and @xmath56 , the condition ( [ eq : scbma_0 ] ) for one error - and - erasure decoding attempt to succeed becomes @xmath57 . by seeing that @xmath58 we conclude the proof . next , we try to maximize the chance that this successful decoding condition is satisfied by at least one of the decoding attempts ( i.e. @xmath59 for at least one erasure patterns @xmath44 ) . mathematically , we want to build a set @xmath60 of no more than @xmath61 erasure patterns @xmath44 in order to @xmath62 the exact answer to this problem is difficult to find . however , one can see it as a covering problem where one wants to cover the space of error patterns using a minimum number of balls centered at the chosen erasure patterns . this view leads to an asymptotic solution of the problem based on rate - distortion theory . more precisely , we view the error pattern @xmath43 as a source sequence and the erasure pattern @xmath44 as a reproduction sequence . r7.2 cm = [ draw = black , fill = blue!05,rectangle , rounded corners ] ( 0.5,0.2 ) circle ( 15pt ) ; ( 0,1 ) circle ( 15pt ) ; ( 0.9,0.9 ) circle ( 15pt ) ; ( 1.4,0.2 ) circle ( 15pt ) ; ( 1.8,1 ) circle ( 15pt ) ; ( 0.5,0.2 ) circle ( 1pt ) ; ( 0.5,0.2 ) circle ( 15pt ) ; ( 0,1 ) circle ( 1pt ) ; ( 0,1 ) circle ( 15pt ) ; ( 0.9,0.9 ) circle ( 1pt ) ; ( 0.9,0.9 ) circle ( 15pt ) ; ( 1.4,0.2 ) circle ( 1pt ) ; ( 1.4,0.2 ) circle ( 15pt ) ; ( 1.8,1 ) circle ( 1pt ) ; ( 1.8,1 ) circle ( 15pt ) ; ( 1,-0.05 ) rectangle ( 1.05,0 ) ; ( 0.05,-0.05 ) rectangle ( 0.1,0 ) ; ( -0.35,1.35 ) rectangle ( -0.3,1.4 ) ; ( 2.2,1.2 ) rectangle ( 2.25,1.25 ) ; ( 1.7,-0.2 ) rectangle ( 1.75,-0.15 ) ; ( 1,1.35 ) rectangle ( 1.05,1.4 ) ; ( 2,1.25 ) rectangle ( 2.05,1.3 ) ; ( 0.7,0.55 ) rectangle ( 0.75,0.6 ) ; ( 1.25,0.65 ) rectangle ( 1.3,0.7 ) ; ( 1.05,0.5 ) rectangle ( 1.1,0.55 ) ; ( -0.15,0.55 ) rectangle ( -0.1,0.6 ) ; ( 0.3,1.325 ) rectangle ( 0.35,1.375 ) ; ( 0.45,1.05 ) rectangle ( 0.5,1.1 ) ; ( 0.7,-0.25 ) rectangle ( 0.75,-0.2 ) ; ( 1.85,0.1 ) rectangle ( 1.9,0.15 ) ; ( 2.1,0.7 ) rectangle ( 2.15,0.75 ) ; ( box ) at ( 3.2,0 ) error pattern + erasure pattern ; ( 3.875,0.125 ) rectangle ( 3.925,0.175 ) ; ( 3.9,-0.125 ) circle ( 1pt ) ; rate - distortion theory shows that the set @xmath60 of @xmath61 reproduction sequences can be generated randomly so that@xmath63\leq d\ ] ] where the distortion @xmath64 is minimized for a given rate @xmath65 . thus , for large enough @xmath2 , we have@xmath66 with high probability . here , @xmath65 and @xmath64 are closely related to the complexity and the performance , respectively , of the decoding algorithm . therefore , we characterize the trade - off between those two aspects using the relationship between @xmath65 and @xmath64 . in this subsection , we consider a generalization of the conventional error and erasure patterns under the same framework to make better use of the soft information . at each index of the rs codeword , beside erasing the symbol we can try to decode using not only the most likely symbol but also other ones as the hard decision ( hd ) symbol . to handle up to the @xmath8 most likely symbols at each index @xmath17 , we let @xmath67 and consider the following definition . [ def:(patternsbma)](generalized error patterns and erasure patterns ) consider a positive integer @xmath68 . let us define @xmath69 as the generalized error pattern where , at index @xmath17 , @xmath70 implies that the @xmath22-th most likely symbol is correct for @xmath71 , and @xmath38 implies none of the first @xmath8 most likely symbols is correct . let @xmath72 be the generalized erasure pattern used for decoding where , at index @xmath17 , @xmath73 implies that the @xmath22-th most likely symbol is used as the hard - decision symbol for @xmath74 , and @xmath39 implies that an erasure is used at that index . for simplicity , we will refer to @xmath43 as the error pattern and @xmath44 as the erasure pattern like in the conventional case . next , we also want to convert the condition ( [ eq : scbma_0 ] ) to the form where @xmath45 is less than a fixed threshold . proposition [ prop : bma1 ] _ _ is thereby generalized into the following theorem . [ thm : extbma-1]we choose the _ letter - by - letter _ distortion measure @xmath75 defined by @xmath76_{x,\hat{x}}$ ] in terms of the @xmath77 ) matrix@xmath78 using this , the condition ( [ eq : scbma_0 ] ) for a successful error - and - erasure decoding becomes@xmath50 the reasoning is similar to proposition [ prop : bma1 ] using the fact that @xmath79 and @xmath80 where @xmath81 for every @xmath82 . for each @xmath83 , we will refer to this generalized case as mbm-@xmath8 decoding . we consider the case mbm-2 decoding where @xmath84 . the distortion measure is given by following the matrix@xmath85 here , at each codeword position , we consider the first and second most likely symbols as the two hard - decision choices like in the chase - type decoding method proposed by bellorado and kavcic @xcite . in this section , we propose a general multiple - decoding algorithm for rs codes based on the rate - distortion approach . this general algorithm applies to not only multiple error - and - erasure decoding but also multiple - decoding of other decoding schemes that we will discuss later . the first step is designing a distortion measure that converts the condition for a single decoding to succeed to the form where distortion is less than a fixed threshold . after that , decoding proceeds as described below . * _ phase i : compute rate - distortion function . _ _ step 1 : _ transmit @xmath86 ( say @xmath87 ) arbitrary test rs codewords , indexed by time @xmath88 , over the channel and compute a set of @xmath86 @xmath89 matrices @xmath90 where @xmath91_{j , i}=p_{i,\pi_{i}^{(t)}(j)}^{(t)}$ ] is the probability of the @xmath22-th most likely symbol at position @xmath17 during time @xmath92 . _ step 2 : _ for each time @xmath92 , obtain the matrix @xmath93 from @xmath90 through a permutation @xmath94 that sorts the probabilities @xmath95 in increasing order to indicate the reliability order of codeword positions . take the entry - wise average of all @xmath86 matrices @xmath93 to get an average matrix @xmath96 . _ step 3 : _ compute the r - d function of a source sequence ( error pattern ) with probability of source letters derived from @xmath96 and the designed distortion measure ( see section [ sec : computing - rd ] and section [ sub : analyticalrd ] ) . determine the point on the r - d curve that corresponds to a designated rate @xmath65 along with the test - channel input - probability distribution vector @xmath97 that achieves that point . * _ phase ii : run actual decoder . _ _ step 4 : _ based on the actual received signal sequence , compute @xmath25 and determine the permutation @xmath98 that gives the reliability order of codeword positions by sorting @xmath25 in increasing order . _ step 5 : _ randomly generate a set of @xmath61 erasure patterns using the test - channel input - probability distribution vector @xmath97 and permute the indices of each erasure pattern by the permutation @xmath99 _ step 6 : _ run multiple attempts of the corresponding decoding scheme ( e.g. error - and erasure decoding ) using the set of erasure patterns in step 5 to produce a list of candidate codewords . _ step 7 : _ use maximum - likelihood ( ml ) decoding to pick the best codeword on the list . in this section , we will present a numerical method to compute the r - d function and the test - channel input - probability distribution that achieves a specific point in the r - d curve . this probability distribution will be needed to randomly generate the set of erasure patterns in the general multiple - decoding algorithm that we have proposed . for an arbitrary discrete distortion measure , it can be difficult to compute the r - d function analytically . fortunately , the blahut - arimoto ( b - a ) algorithm ( see details in @xcite ) gives an alternating minimization technique that efficiently computes the r - d function of a single discrete source . more precisely , given a parameter @xmath100 which represents the slope of the @xmath101 curve at a specific point and an arbitrary all - positive initial test - channel input - probability distribution vector @xmath102 , the b - a algorithm shows us how to compute the rate - distortion point @xmath103 by means of computing the test - channel input - probability distribution vector @xmath104 and the test - channel transition probability matrix @xmath105 that achieves that point . however , it is not straightforward to apply the b - a algorithm to compute the r - d for a discrete source sequence @xmath43 ( an error pattern in our context ) of @xmath2 independent but non identical source components @xmath106 . in order to do that , we consider the group of source letters @xmath107 where @xmath108 as a super - source letter @xmath109 , the group of reproduction letters @xmath110 where @xmath111 as a super - reproduction letter @xmath112 , and the source sequence @xmath43 as a single source . for each super - source letter @xmath113 , @xmath114 follows from the independence of source components . while we could apply the b - a algorithm to this source directly , the complexity is a problem because the alphabet sizes for @xmath113 and @xmath115 become the super - alphabet sizes @xmath116 and @xmath117 respectively . instead , we avoid this computational challenge by choosing the initial test - channel input - probability distribution so that it can be factorized into a product of @xmath2 initial test - channel input - probability components , i.e. @xmath118 . then , we see that this factorization rule still applies after every step of the iterative process . by doing this , for each parameter @xmath119 we only need to compute the rate - distortion pair for each component ( or index @xmath17 ) separately and sum them together . this is captured into the following theorem . [ thm:(factored - blahut)](factored blahut - arimoto algorithm ) consider a discrete source sequence @xmath43 of @xmath2 independent but non identical source components @xmath106 . given a parameter @xmath100 , the rate and the distortion for this source sequence are given by@xmath120 where the components @xmath121 and @xmath122 are computed by the b - a algorithm with the parameter @xmath119 . this pair of rate and distortion can be achieved by the corresponding test - channel input - probability distribution @xmath123 where the component probability distribution @xmath124 . see appendix [ sec : appblahut ] . in this section , we analyze and design a distortion measure to convert the condition for successful asd decoding to a suitable form so that we can apply the general multiple - decoding algorithm to asd decoding . first , let us give a brief review on asd decoding of rs codes . given a set @xmath125 of @xmath2 distinct elements in @xmath126 from each message polynomial @xmath127 , we can have a codeword @xmath128 by evaluating the message polynomial at @xmath129 , i.e. @xmath130 for @xmath131 . consider a received vector @xmath132 , we can compute the _ a posteriori _ probability ( app ) matrix @xmath133 as follows.@xmath134_{j , i}=p_{i , j}=\pr(c_{i}=\alpha_{j}|\mathbf{r})\,\,\,\mbox{for}\,\,1\leq i\leq n,1\leq j\leq2^{q}.\ ] ] the asd decoding as in @xcite has the following main steps . 1 . _ multiplicity assignment _ : use a particular multiplicity assignment scheme ( mas ) to derive a @xmath89 multiplicity matrix , denoted as @xmath135 , of non - negative integer entries @xmath136 from the app matrix @xmath133 . _ interpolation _ : construct a bivariate polynomial @xmath137 of minimum @xmath138 weighted degree that passes through each of the point @xmath139 with multiplicity @xmath140 for @xmath141 and @xmath142 . factorization _ : find all polynomials @xmath143 of degree less than @xmath3 such that @xmath144 is a factor of @xmath137 and re - evaluate these polynomials to form a list of candidate codewords . in this paper , we denote @xmath145 as the maximum multiplicity . intuitively , higher multiplicity should be put on more likely symbols . increasing @xmath146 generally gives rise to the performance of asd decoding . however , one of the drawbacks of asd decoding is that its decoding complexity is roughly @xmath147 which sharply increases with @xmath146 . thus , in this section we will work with small @xmath146 to keep the complexity affordable . one of the main contributions of @xcite is to offer a condition for successful asd decoding represented in terms of two quantities specified as the score and the cost as follows . the score @xmath148 with respect to a codeword @xmath149 and a multiplicity matrix @xmath135 is defined as @xmath150,j}$ ] + where @xmath151=i$ ] such that @xmath152 . the cost @xmath153 of a multiplicity matrix @xmath135 is defined as @xmath154 ( asd decoding threshold , see @xcite ) . the transmitted codeword will be on the list if@xmath155\end{aligned}\ ] ] @xmath156 to match the general framework , the asd decoding threshold ( or condition for successful asd decoding ) should be converted to the form where the distortion is smaller than a fixed threshold . in this subsection , we consider multiple trials of asd decoding using bit - level erasure patterns . a bit - level error pattern @xmath157 and a bit - level erasure pattern @xmath158 has length @xmath159 since each symbol has @xmath16 bits . similar to definition [ def:(conv . patterns ) ] of a conventional error pattern and a conventional erasure pattern , @xmath160 in a bit - level error pattern implies a bit - level error occurs and @xmath161 in a bit - level erasure pattern implies that a bit - level erasure occurs . from each bit - level erasure pattern we can specify entries of the multiplicity matrix @xmath135 using the bit - level mas proposed in @xcite as follows : for each codeword position , assign multiplicity 2 to the symbol with no bit erased , assign multiplicity 1 to each of the two candidate symbols if there is 1 bit erased , and assign multiplicity zero to all the symbols if there are @xmath162 bits erased . all the other entries are zeros by default . this mas has a larger decoding region compared to the conventional error - and - erasure decoding scheme . ( bit - level asd decoding threshold , see @xcite ) for rs codes of rate @xmath163 , asd decoding using the proposed bit - level mas will succeed ( i.e. the transmitted codeword is on the list ) if@xmath164 where @xmath165 is the number of bit - level erasures and @xmath166 is the number of bit - level errors in unerased locations . we can choose an appropriate distortion measure according to the following proposition which is a natural extension of proposition [ prop : bma1 ] in the symbol level . [ prop : bitasd]if we choose the bit - level _ letter - by - letter _ distortion measure @xmath48 as follows@xmath167 then the condition ( [ eq : bgmdsc ] ) becomes @xmath168 the proof uses the same reasoning as the proof of _ _ proposition _ _ [ prop : bma1 ] . we refer the the multiple - decoding of bit - level asd as m - b - asd . in this subsection , we try to convert the condition for successful asd decoding in general to the form that suits our goal . we will also determine which multiplicity assignment schemes allow us to do so . ( multiplicity type ) for some codeword position , let us assign multiplicity @xmath169 to the @xmath22-th most likely symbol for @xmath170 where @xmath171 . the remaining entries in the column are zeros by default . we call the sequence , @xmath172 , the column multiplicity type for `` top-@xmath8 '' decoding . first , we notice that a choice of multiplicity types in asd decoding at each codeword position has the similar meaning to a choice of erasure decisions in the conventional error - and - erasure decoding . however , in asd decoding we are more flexible and may have more types of erasures . for example , assigning multiplicity zero to all the symbols ( all - zero multiplicity type ) at codeword position @xmath17 corresponds to the case when we have a complete erasure at that position . assigning the maximum multiplicity @xmath146 to one symbol corresponds to the case when we choose that symbol as the hard - decision one . hence with some abuse of terminology , we also use the term ( generalized ) erasure pattern @xmath44 for the multiplicity assignment scheme in the asd context . each erasure - letter @xmath106 gives the multiplicity type for the corresponding column of the multiplicity matrix @xmath135 . ( error and erasure patterns for asd decoding ) consider a mas with @xmath173 multiplicity types . let @xmath174 be an erasure pattern where , at index @xmath17 , @xmath70 implies that multiplicity type @xmath22 is used at column @xmath17 of the multiplicity matrix @xmath135 . notice that the definition of an error pattern @xmath69 in definition [ def:(patternsbma ) ] applies unchanged here . rate - distortion theory gives us the intuition that in general the more multiplicity types ( erasure choices ) we have , the better performance of multiple asd decoding we achieve as @xmath2 becomes large . thus , we want to find as many as possible multiplicity types for `` top-@xmath8 '' that allow us to convert condition for successful asd decoding to the correct form . choosing @xmath175 , for example , gives four column multiplicity types for `` top-2 '' decoding as follows : the first is @xmath176 where we assign multiplicity 2 to the most likely symbol @xmath177 , the second is @xmath178 where we assign equal multiplicity 1 to the first and second most likely symbols @xmath177 and @xmath179 , the third is @xmath180 where we assign multiplicity 2 to the second most likely symbol @xmath179 , and the fourth is @xmath181 where we assign multiplicity zero to all the symbols at index @xmath17 ( i.e. the @xmath17-th column of @xmath135 is an all - zero column ) . as a corollary of theorem [ thm : genasd ] below , the distortion matrix that converts ( [ eq : asdsc ] ) to the correct form for this case is@xmath182 the following definition and theorem provide a set of allowable multiplicity types that converts the condition for successful asd decoding into the form where distortion is less than a fixed threshold . the set of allowable multiplicity types for `` top-@xmath8 '' decoding with maximum multiplicity @xmath146 is defined to be if @xmath183.]@xmath184 taking the elements of this set in an arbitrary order , we let the @xmath22-th multiplicity type in the allowable set be @xmath185 . @xmath186 consists of all permutations of @xmath187 . meanwhile , @xmath188 comprises all the permutations of @xmath189 and we refer to the multiple asd decoding algorithm using this set of multiplicity types as masd-2 . @xmath190 consists of all the permutations of @xmath191 and this case is referred as masd-3 . we also consider another case called masd-2a that uses the set of multiplicity types @xmath192 . [ thm : genasd ] let @xmath193 be the number of multiplicity types in a mas for `` top-@xmath8 '' decoding with maximum multiplicity @xmath146 . let @xmath194 be a _ letter - by - letter _ distortion measure defined by @xmath76_{x,\hat{x}}$ ] , where @xmath195is the @xmath196 matrix@xmath197 with @xmath198 . then , the condition ( [ eq : asdsc ] ) for successful asd decoding of a rs code with rate @xmath199 is equivalent to @xmath200 [ sketch of proof ] ( see details in @xcite ) let @xmath201 and @xmath202 be the score and cost of the multiplicity assignment . first , we show that @xmath203 in ( [ eq : asdsc ] ) implies that @xmath204 . combining this inequality with the high - rate constraint in theorem [ thm : genasd ] implies that @xmath205 . from ( [ eq : asdsc ] ) , we also know that @xmath206 and this implies that @xmath207 . but , the conditions of the theorem can also be used to show that @xmath208 . combining this with @xmath207 gives a contradiction unless @xmath209 . thus , we conclude that @xmath210 . therefore , the condition in ( [ eq : asdsc ] ) is equivalent to @xmath211 because @xmath212 is a consequence of @xmath210 and @xmath213 is satisfied by the high - rate constraint . finally , one can show that @xmath211 is equivalent to @xmath59 with the chosen distortion matrix . for a fixed @xmath146 , the size of @xmath214 is maximized when @xmath215 . multiplicity types @xmath216 and any permutation of @xmath217@xmath218 are always in the allowable set @xmath219 . the r - d framework we use is most suitable when @xmath220 . for a finite @xmath2 , the random coding approach may have problems with only a few lrps . we can instead use good covering codes to handle these lrps . in the scope of covering problems , one can use an @xmath8-ary @xmath221-covering code ( e.g. a perfect hamming or golay code ) with covering radius @xmath221 to cover the whole space of @xmath8-ary vectors of the same length . the covering may still work well if the distortion measure is close to , but not exactly equal to the hamming distortion . in order take care of up to the @xmath8 most likely symbols at each of the @xmath222 lrps of an @xmath1 rs , we consider an @xmath223 @xmath8-ary @xmath221-covering code whose codeword alphabet is @xmath224 then , we give a definition of the ( generalized ) error patterns and erasure patterns for this case . in order to draw similarities between this case and the previous cases , we still use the terminology `` generalized erasure pattern '' and shorten it to erasure pattern even if error - only decoding is used . for error - only decoding , condition [ con : bmaerr - n - era ] _ _ for successful decoding becomes@xmath225 ( error and erasure patterns for error - only decoding ) let us define @xmath226 as an error pattern where , at index @xmath17 , @xmath70 implies that the @xmath22-th most likely symbol is correct for @xmath71 , and @xmath38 implies none of the first @xmath8 most likely symbols is correct . let @xmath227 be an erasure pattern where , at index @xmath17 , @xmath73 implies that the @xmath22-th most likely symbol is chosen as the hard - decision symbol for @xmath74 . if we choose the _ letter - by - letter _ distortion measure @xmath228 defined by @xmath76_{x,\hat{x}}$ ] in terms of the @xmath229 matrix@xmath230 then the condition for successful error - only decoding then becomes@xmath231 it follows directly from @xmath232 . if we delete the first row which corresponds to the case where none of the first @xmath8 most likely symbols is correct then the distortion measure is exactly the hamming distortion . [ [ split - covering - approach ] ] split covering approach : + + + + + + + + + + + + + + + + + + + + + + + + we can break an error pattern @xmath43 into two sub - error patterns @xmath233 of @xmath234 least reliable positions and @xmath235 of @xmath236 most reliable positions . similarly , we can break an erasure pattern @xmath44 into two sub - erasure patterns @xmath237 and @xmath238 . let @xmath239 be the number of positions in the @xmath234 lrps where none of the first @xmath8 most likely symbols is correct , or @xmath240 . if we assign the set of all sub - error patterns @xmath241 to be an @xmath223 @xmath221-covering code then @xmath242 because this covering code has covering radius @xmath221 . since @xmath243 , in order to increase the probability that the condition ( [ eq : scpf ] ) is satisfied we want to make @xmath244 as small as possible by the use of the r - d approach . the following proposition summarizes how to generate a set of @xmath61 erasure patterns for multiple runs of error - only decoding . in each erasure pattern , the letter sequence at @xmath234 lrps is set to be a codeword of an @xmath223 @xmath8-ary @xmath245covering code . the letter sequence of the remaining @xmath246 is generated randomly by the r - d method ( see section [ sec : proposed - algorithm ] ) with rate @xmath247 and the distortion measure in ( [ eq : pfdst ] ) . since this covering code has @xmath248 codewords , the total rate is @xmath249 for a ( 7,4,3 ) binary hamming code which has covering radius @xmath250 , we take care of the @xmath251 most likely symbols at each of the 7 lrps . we see that @xmath252 is a codeword of this hamming code and then form erasure patterns @xmath253 with assumption that the positions are written in increasing reliability order . the @xmath254 sub - erasure patterns @xmath255 are generated randomly using the r - d approach with rate @xmath256 . while it also makes sense to use a covering codes for the @xmath234 lrps of the erasure patterns and set the the rest to be letter @xmath257 ( i.e. chose the most likely symbol as the hard - decision ) , our simulation results shows that the performance can be improved by using a combination of covering codes and random ( i.e. , generated by the r - d approach ) codes . for some simple distortion measures , we can compute the r - d functions analytically in closed form . first , we observe an error pattern as a sequence of independent but non - identical random sources . then , we compute the component r - d functions at each index of the sequence and use convex optimization techniques to allocate the total rate and distortion to various components . this method converges to the solution faster than the numerical method in section [ sec : computing - rd ] . the following two theorems describe how to compute the r - d functions for the simple distortion measures of proposition [ prop : bma1 ] and [ prop : bitasd ] . [ thm:(bma - rd)](conventional error - and - erasure decoding ) let @xmath258 , the overall rate - distortion function is given by @xmath259^{+}$ ] where @xmath260 and @xmath261 can be found be a reverse water - filling procedure:@xmath262 where @xmath263 should be chosen so that @xmath264 . the @xmath265 function can be achieved by the test - channel input - probability distribution@xmath266 [ sketch of proof ] ( see @xcite for details ) with the distortion measure in ( [ eq : dstfnbma ] ) , we follow the method in @xcite to compute the rate - distortion function component @xmath267^{+}$ ] and the test - channel input - probability distribution @xmath268 and @xmath269 for each index @xmath17 . then , one can show that the optimal allocation of rate and distortion to the various components is given by a reverse - water filling procedure like in @xcite . [ thm:(basd - rd)](bit - level asd case in proposition [ prop : bitasd ] ) the overall rate - distortion function in this case is given by @xmath270^{+}$ ] where @xmath271 and the distortion component @xmath272 is given by@xmath273 where @xmath274 should be chosen so that @xmath275 . the @xmath265 function can be achieved by the test - channel input - probability distribution@xmath276 [ sketch of proof ] ( see @xcite for details ) with the distortion measure in ( [ eq : dstfnbgmd ] ) , using the method in @xcite we can compute the rate - distortion function component @xmath277 where @xmath278 is a lagrange multiplier such that @xmath279 for each index @xmath17 . then , the kuhn - tucker conditions define the the overall rate allocation . performance of various decoding algorithms for the ( 255,239 ) rs code over an awgn channel . ] performance of various decoding algorithms for the ( 255,239 ) rs code over an awgn channel . ] using simulations , we consider the performance of the ( 255,239 ) rs code over an awgn channel with bpsk as the modulation format . the mbm-1 curve corresponds to standard error - and - erasure bm decoding with multiple erasure patterns . for @xmath280 , the mbm-@xmath8 curves correspond to error - and - erasure bm decoding with multiple decoding trials using both erasures and top-@xmath8 symbols . the masd-@xmath146 curves correspond to multiple asd decoding trials with maximum multiplicity @xmath146 . the number of trial decoding patterns is @xmath61 where @xmath65 is denoted in parentheses in each algorithm s acronym ( e.g. , m - bm-2(11 ) uses @xmath281 ) . [ fig : rdcurve ] shows the r - d curves for various algorithms at @xmath282 db . for this code , the fixed threshold for decoding is @xmath283 . therefore , one might expect that algorithms whose average distortion is less than 17 should have a frame error rate ( fer ) less than @xmath284 . the r - d curve allows one to estimate the number of decoding patterns required to achieve this fer . conventional bm decoding is very similar to mbm-1 decoding at rate 0 . notice that the mbm-1 algorithm at rate 0 , which is very similar to conventional bm decoding , has an expected distortion of roughly 24 . for this reason , the fer on conventional decoding is close to 1 . the r - d curve tells us that trying roughly @xmath285 ( i.e. , @xmath286 ) erasure patterns would reduce the fer to roughly @xmath284 because this is where the distortion drops down to 17 . likewise , the mbm-2(11 ) algorithm has an expected distortion of less than 14 . so we expect ( and our simulations confirm ) that the fer should be less than @xmath284 . one weakness of this approach is that the r - d describes only the average distortion and does not directly consider the probability that the distortion is greater than 17 . still , we can make the following observations from the r - d curve . even at low rates ( e.g. , @xmath287 ) , we see that the distortion @xmath64 achieved by mbm-2 is roughly the same as mbm-3 , masd-2 , and masd-3 but smaller than masd-2a and mbm-1 . this implies that mbm-2 is no worse than the more complicated asd based approaches for a wide range of rates ( i.e. , @xmath288 ) . the fer of various algorithms can be seen in fig . [ fig : simulation ] . the focus on @xmath281 allows us to make fair comparisons with sed(12,12 ) . with the same number of decoding trials , mbm-2(11 ) outperforms sed(12,12 ) by 0.3 db at an fer@xmath289 . even mbm-2(7 ) , with many fewer decoding trials , outperforms both sed(12,12 ) and the kv algorithm with @xmath290 . among all our proposed algorithms with rate @xmath281 , the mbm - hm74(11 ) achieves the best performance . this algorithm uses the hamming ( 7,4 ) covering code for the 7 lrps and the r - d approach for the remaining codeword positions . meanwhile , small differences in the performance between mbm-2(11 ) , mbm-3(11 ) , masd-2(11 ) , and masd-3(11 ) suggest that : ( i ) taking care of the @xmath251 most likely symbols at each codeword position is good enough for multiple decoding of high - rate rs code and ( ii ) multiple runs of error - and - erasure decoding is almost as good as multiple runs of asd decoding . recall that this result is also correctly predicted by the r - d analysis . moreover , it is quite reasonable since we know that the gain of gs decoding , with infinite multiplicity , over the bm algorithm is negligible for high - rate rs codes . to compare with the lcc(@xmath291 ) chase - type approach used in @xcite , we also consider the mbm - hm74(4 ) algorithm , which uses the hamming ( 7,4 ) covering codes for the 7 lrps and the hard decision pattern for the remaining codeword positions . this shows that the covering code achieves better performance with the same number ( @xmath292 ) decoding attempts . the comparison is not entirely fair , however , because of their low - complexity approach to multiple decoding . we believe , nevertheless , that their technique can be generalized to covering codes . a rate - distortion approach is proposed as a unified framework to analyze multiple decoding trials , with various algorithms , of rs codes in terms of performance and complexity . a connection is made between the complexity and performance ( in some asymptotic sense ) of these multiple - decoding algorithms and the rate and distortion of an associated r - d problem . covering codes are also combined with the rate - distortion approach to mitigate the suboptimality of random codes when the effective block - length is not large . as part of this analysis , we also present numerical and analytical computations of the rate - distortion function for sequences of independent but non - identical sources . finally , the simulation results show that our proposed algorithms based on the r - d approach achieve a better performance - versus - complexity trade - off than previously proposed algorithms . one key result is that , for high - rate rs codes , multiple - decoding using the standard bm algorithm is as good as multiple - decoding using more complex asd algorithms . in this paper , we only discuss the rate - distortion approach to solve the problem in ( [ eq : probstatement ] ) . however , the performance can be further improved by focusing on the rate - distortion error - exponent . this allows us to approximately solve the covering problem for finite @xmath2 rather than just as @xmath220 . the complexity of multiple decoding can also be decreased by using clever techniques to lower the complexity per decoding trial ( e.g. , @xcite ) . we will address these two improvements in a future paper . first , let us recall that for each source component @xmath106 , the b - a algorithm computes the r - d pair in the following steps : 1 . choose an arbitrary all - positive test - channel input - probability distribution vector @xmath102 . 2 . iterate the following steps at @xmath293@xmath294 where @xmath295 is the transition probability . it is shown by b - a that @xmath296 and @xmath297as @xmath298 . now , we will prove theorem [ thm:(factored - blahut ) ] . since the input - distribution vector of the test channel is an arbitrary all - positive vector , we choose @xmath301 so that it can be factorized as follows @xmath302 . s. -w . lee and b. v. k. v. kumar , `` soft - decision decoding of reed - solomon codes using successive error - and - erasure decoding , '' in _ proc . ieee global telecom . _ , new orleans , la , nov . 2008 , pp . 15 . h. xia , h. wang , and j. r. cruz , `` a chase - gmd algorithm for soft - decision decoding of reed - solomon codes on perpendicular channels , '' in _ proc . _ , beijing , china , may 2008 , pp . 19771981 .

in many applications where high frequency data are observed , we wish to predict the next values of this time series through an online prediction learning algorithm able to process a large amount of data . the classical stationarity assumption on the distribution of the observations has to be weakened to take into account some smooth evolution of the environment . from a statistical modelling point of view , this is described by some time varying parameters . in order to sequentially track them from high - frequency data , the algorithms must require few operations and a low storage capacity to update the parameters estimation and the prediction after each new observation . the most common online methods are least mean squares ( lms ) , normalized least mean squares ( nlms ) , regularized least squares ( rls ) or kalman . all of them rely on the choice of a gradient step , a forgetting factor , or more generally on a tuning parameter corresponding to some a priori knowledge on how smoothly the local statistical distribution of the data evolves along the time . to adapt automatically to this smoothness , usually unknown in practice , we propose to use an exponentially weighted aggregation of several such predictors , with various tuning parameters . we emphasize that to meet the online constraint , we can not use methods that require a large amount of computations ( such as cross validation ) . the exponential weighting technique in aggregation have been developed in parallel in the machine learning community [ see the seminal paper @xcite ] , in the statistical community [ see @xcite , @xcite ( @xcite ) , @xcite , or more recently @xcite ] and in the game theory community for individual sequences prediction [ see @xcite and @xcite for recent surveys ] . in contrast to the classical statistical setting , in the individual sequence setting the observations are not assumed to be generated by an underlying stochastic process . the link between both settings has been analyzed in @xcite for the regression model with fixed and random designs . exponential weighting has also been investigated in the case of weakly dependent stationary data in @xcite . more recently , an approach inspired from individual sequences prediction has been studied in @xcite for bounded arma processes under some specific conditions on the ( constant ) arma coefficients . in this contribution , we consider two possible aggregation schemes based on exponential weights which can be computed recursively . we provide oracle inequalities applying to the aggregated predictor under the following main assumptions that ( 1 ) the observations are sub - linearly with respect to a sequence of random variables with possibly time varying linear coefficients and ( 2 ) the predictors to be aggregated are lipschitz functions of the past . an important feature of our observation model is that it embeds the well - known class of _ locally stationary _ processes . we refer to @xcite and the references therein for a recent general view about statistical inference for locally stationary processes . as an application , we focus on a particular locally stationary model , that of the time varying autoregressive ( tvar ) process . the minimax rate of certain recursive estimators of the tvar coefficients is studied in @xcite . to our knowledge , there is not a well - established method on the automatic choice of the gradient step when the smoothness index is unknown . here , we are interested in the prediction problem which is closely related to the estimation problem . we show that the proposed aggregation methods provide a solution to this question , in the sense that they give rise to recursive adaptive minimax predictors . the paper is organized as follows . in section [ sec : online - aggr - pred ] , we provide oracle inequalities for the aggregated predictors under general conditions applying to nonstationary sub - linear processes . tvar processes are introduced in section [ section : time - varying_autoregressive_model ] in a nonparametric setting based on hlder smoothness assumptions on the tvar coefficients . a lower bound of the prediction risk is given in this setting and this result is used to show that the proposed aggregation methods achieve the minimax adaptive rate . section [ subsection : oracle_inequalities ] contains the proofs of the oracle inequalities . the proof of the lower bound of the minimax prediction risk is presented in section [ section : proof_lower_bound ] . numerical experiments illustrating these results are then described in section [ sec : numer - exper ] . one and one supplementary material [ @xcite ] complete this paper . and [ @xcite ( @xcite ) , section a ] contain some postponed proofs and useful lemmas , [ giraud , roueff and sanchez - perez ( @xcite ) , section b ] explains how to build nonadaptive minimax predictors which can be used in the aggregation step and [ giraud , roueff and sanchez - perez ( @xcite ) , section c ] provides additional results with improved aggregation rates . in this section , we consider a time series @xmath1 admitting the following _ nonstationary _ sub - linear property with respect to the nonnegative process @xmath2 . [ hyp : gen ] the process @xmath3 satisfies @xmath4 where @xmath5 are nonnegative coefficients such that @xmath6 additional assumptions will be required on @xmath2 to deduce useful properties for @xmath1 . note , for instance , that the condition on @xmath7 in ( [ eq : unif - coeff - bound ] ) guarantees that , if @xmath8 has a uniformly bounded @xmath9-norm , the convergence of the infinite sum in ( [ eq : gen - series - representation ] ) holds almost surely and in the @xmath9-sense , with both convergences defining the same limit . it follows that @xmath1 also has uniformly bounded @xmath9 moments . let us give some particular contexts where the representation ( m-1 ) can be used . [ exple : tvlinear ] standard weakly stationary processes such as arma processes [ see @xcite ] admit a wold decomposition of the form @xmath10 where @xmath11 is a weak white noise with , says , unit variance . this model , sometimes referred to as an ma(@xmath12 ) representation , is often extended to a two - sided sum representation @xmath13 and additional assumptions on the existence of higher moments for @xmath11 or on the independence of the @xmath14 s are often used for statistical inference or prediction ; see @xcite , chapters 7 and 8 . because the sequence @xmath15 may vary with @xmath16 in ( m-1 ) , we may extend this standard stationary setting and also consider linear processes with time varying coefficients . in this case , we have @xmath17 where @xmath18 is a sequence of centered independent random variables with unit variance and @xmath19 is supposed to satisfy ( [ eq : unif - coeff - bound ] ) with @xmath20 , so that holds with @xmath21 . for this general class of processes , statistical inference is not easily carried out : each new observation @xmath22 comes with a new unknown sequence @xmath23 . however , additional assumptions on this set of sequences allow to derive and study appropriate statistical inference procedures . a sensible approach in this direction is to consider a _ locally stationary _ model as introduced in @xcite . in this framework , the set of sequences @xmath24 is controlled as @xmath25 by artificially ( but meaningfully ) introducing a dependence in @xmath26 , hence is written as @xmath27 , and by approximating it with a set of sequences rescaled on the time interval @xmath28 $ ] , @xmath29 , @xmath30 $ ] , @xmath31 , for example , in the following way : @xmath32 then various interesting statistical inference problems based on @xmath33 can be tackled by assuming some smoothness on the mapping @xmath34 and , possibly , additional assumptions on the structure of the sequence@xmath35 for each @xmath30 $ ] [ see @xcite and the references therein ] . [ exple : tvar ] a particular instance of example [ exple : tvlinear ] is the _ time varying autoregressive _ ( tvar ) process , which is assumed to satisfy the recursive equation @xmath36 where @xmath37 is a white noise process ; see @xcite . it turns out that , in the framework introduced by @xcite , under suitable assumptions , such processes admit a time varying linear representation of the form ( [ eq : tvlin - rep ] ) ; see @xcite . in section [ section : time - varying_autoregressive_model ] , we focus on such a class of processes and use the aggregation of predictors to derive adaptive minimax predictors under specific smoothness assumptions on the time varying coefficients . [ example : non - linear ] it can also be interesting to consider nonlinear extensions of example [ exple : tvar ] . a simple example is obtained by setting @xmath38 where @xmath37 is an i.i.d . sequence and @xmath39 is a time varying sub - linear sequence of functions satisfying , for all @xmath16 that @xmath40 for some @xmath41 . since @xmath39 is no longer linear but sub - linear , such a model does not enjoy an exact linear representation of the form ( [ eq : tvlin - rep ] ) . nevertheless , since we have @xmath42 and iterating this equation backwards yields assumption ( m-1 ) with @xmath21 and @xmath43 . in the stationary case , where @xmath44 does not depend on @xmath16 , a well - known nonlinear extension is the threshold autoregressive model where @xmath45 is piecewise linear ; see @xcite . our goal in this section is to derive oracle bounds for the aggregation of predictors that hold for the general model ( m-1 ) with one of the two following additional assumptions on @xmath2 . the nonnegative process @xmath46 satisfies @xmath47<\infty . \ ] ] the nonnegative process @xmath46 is a sequence of independent random variables fulfilling @xmath48 < \infty . \ ] ] assumptions and appear to be quite mild . as mentioned in example [ exple : tvlinear ] , basic assumptions in stationary time series usually include moments of sufficiently high order for the innovations and their independence , or rely on the gaussian assumption , which is contained in . we also note that , in the context of locally stationary time series , our assumptions on the innovations are weaker than those used in the recent works @xcite , @xcite ( @xcite ) . precise comparisons between our assumptions and usual ones in the aggregation literature will be given after corollary [ corollary : oracle - bounds ] . let @xmath49 be a real valued sequence . we say that @xmath50 is a predictor of @xmath51 if it is a measurable function of @xmath52 . throughout this paper , the quality of a sequence of predictors @xmath53 is evaluated for some @xmath54 using the @xmath55 loss averaged over the time period @xmath56 @xmath57 now , given a collection of @xmath58 sequences of predictors @xmath59 , we wish to sequentially derive a new predictor which predicts almost as accurately as or more accurately than the best of them . in the present paper and for our purposes , aggregating the predictors amounts to compute a convex combination of them at each time @xmath16 . this corresponds to choosing at each time @xmath16 an element @xmath60 of the simplex @xmath61 and compute @xmath62}=\sum_{i=1}^n \alpha_{i , t } \widehat{x}_t^{(i ) } . \ ] ] we consider two strategies of aggregation , which are studied in the context of bounded sequences in @xcite . more recent contributions and extensions can be found in @xcite . see also @xcite for a pedagogical . these strategies are sequential and online , meaning that : to compute the aggregation weights @xmath60 at time @xmath16 , only the values of @xmath63 and @xmath64 up to time @xmath65 are used , the computation can be done recursively by updating a set of quantities , the number of which does not depend on @xmath16 . these two properties are met in the algorithm [ algo : strat12 ] detailed below . we consider in the remaining of the paper a convex aggregation of predictors @xmath66}= \sum _ { i=1}^{n}\widehat{\alpha}_{i , t}\widehat { x}_{t}^{(i)},\qquad1\leq t\leq t , \ ] ] with some specific weights @xmath67 defined as follows . the first strategy is to define for all @xmath68 and @xmath69 , the weights @xmath70 by @xmath71 with the convention that a sum over no element is zero , so @xmath72 for all @xmath73 . the parameter @xmath74 , usually called the _ learning rate _ , will be specified later . the second strategy is to define for all @xmath68 and @xmath69 , the weights @xmath70 by @xmath75 with again the convention that a sum over no element is zero . both strategies yield the same algorithm up to the line where @xmath76 is computed . for sake of brevity , we write only one algorithm ( see algorithm [ algo : strat12 ] ) and use a switch / case statement to distinguish between the two strategies . note , however , that the choice of the strategy ( @xmath77 or @xmath78 ) holds for the whole sequence of predictions . we establish oracle bounds on the average prediction error of the aggregated predictors . these bounds ensure that the error is equal to that associated with the best convex combination of the predictors or with the best predictor ( depending on the aggregation strategy ) , up to two remaining terms . one remaining term depends on the number @xmath58 of predictors to aggregate and the other one on the _ variability _ of the original process . the learning rate @xmath79 can then be chosen to achieve the best trade - off between these two terms . the second remaining term indirectly depends on the variability of the predictors . we control below this variability in terms of the variability of the original process by using the following lipschitz property . [ def : unif - lip - pred ] let @xmath80 be a sequence of nonnegative numbers . a predictor @xmath50 of @xmath81 from @xmath52 is said to be @xmath82-lipschitz if @xmath83 we more specifically consider a sequence @xmath82 satisfying the following assumption . the sequence @xmath80 satisfies @xmath84 this condition is trivially satisfied by constant linear predictors depending only on a finite number of previous observations , that is , @xmath85 . in @xcite [ ( @xcite ) , section b.1 ] , we extend this case in the context of the tvar process where the coefficients @xmath86 are replaced by estimates of the time varying autoregressive coefficients . more generally , assumption appears to be quite natural in the general context where @xmath87=f_t((x_{t - s})_{s\geq1})$ ] , where @xmath88 is a lipschitz function from @xmath89 to @xmath90 , with lipschitz coefficients satisfying a condition similar to ( [ eq : hyp : lip ] ) ; see , for instance , @xcite in the case of stationary time series . we now state two upper - bounds on the mean quadratic prediction error of the aggregated predictors defined in the previous section , when the process @xmath91 fulfills the sub - linear property ( m-1 ) . [ thm : oracle - bounds ] assume that assumption holds . let @xmath92 , @xmath93 be a collection of sequences of @xmath82-lipschitz predictors with @xmath82 satisfying . assume that the noise @xmath94 fulfills with @xmath95 and let @xmath96 denote the aggregated predictor obtained using the weights ( [ definition : alphas_gradient_quadratic_loss ] ) with any @xmath74 . then we have @xmath97 & \leq&\inf_{\nu\in\mathcal{s}_n } { 1\over t}\sum _ { t=1}^t{\mathbb{e}}\bigl [ \bigl(\widehat x_t^{[\nu]}-x_{t } \bigr)^2 \bigr ] \nonumber\\[-9pt]\\[-9pt]\nonumber & & { } + { \log n\over t\eta}+2\eta ( 1+l_*)^{4 } a_{*}^4 \m_{4 } .\end{aligned}\ ] ] assume that the noise @xmath94 satisfies with a given @xmath98 and let @xmath96 denote the aggregated predictor obtained using the weights ( [ definition : alphas_quadratic_loss ] ) with any @xmath74 . then we have @xmath99 & \leq & \min_{1\leq i\leq n } { 1\over t}\sum _ { t=1}^t{\mathbb{e}}\bigl [ \bigl(\widehat x_{t}^{(i)}-x_{t } \bigr)^2 \bigr ] \nonumber\\[-10pt]\\[-10pt]\nonumber & & { } + \frac{\log n}{t\eta } + ( 2\eta)^{p/2 - 1}a_{*}^{p}(1+l_*)^{p}m_{p } .\end{aligned}\ ] ] assume that the noise @xmath94 fulfills for some positive @xmath100 and let @xmath96 denote the aggregated predictor obtained using the weights ( [ definition : alphas_quadratic_loss ] ) with @xmath74 . then , for any @xmath101\qquad\mbox{with } a^ * : = \sup _ { j\in{\mathbb{z}}}\sup_{t\in{\mathbb{z } } } a_t(j ) \leq a _ { * } , \ ] ] we have @xmath102\nonumber \\ & & \qquad \leq\min_{1\leq i\leq n } { 1\over t}\sum _ { t=1}^t{\mathbb{e}}\bigl [ \bigl(\widehat x_{t}^{(i)}-x_{t } \bigr)^2 \bigr ] \\ & & \quad\qquad { } + \frac{\log n}{t\eta } + \frac{2}{{\mathrm{e}}}\lambda^{-2 } \bigl(2+\lambda(2 \eta)^{-1/2 } \bigr ) { \mathrm{e}}^{- \lambda(2\eta)^{-1/2 } } \bigl(\mphi(\zeta ) \bigr)^{\lambda a_*(1+l_*)/\zeta } .\nonumber\end{aligned}\ ] ] the proof can be found in section [ sec : proof - theor - strat12 ] . [ rem : oracle - bounds - explicites ] the bounds ( [ eq : strat1-final ] ) , ( [ eq : strat2_p - final ] ) and ( [ eq : strat2_exp - final ] ) are explicit in the sense that all the constants appearing in them are directly derived from those appearing in assumptions ( m-1 ) , , and . the following corollary is obtained by choosing @xmath79 [ and @xmath103 in the case ( iii ) ] adequately in the three cases of theorem [ thm : oracle - bounds ] . [ corollary : oracle - bounds ] assume that assumption holds . let @xmath92 , @xmath93 be a collection of sequences of @xmath82-lipschitz predictors with @xmath82 satisfying . assume that the noise @xmath94 fulfills with @xmath95 and let @xmath96 denote the aggregated predictor obtained using the weights ( [ definition : alphas_gradient_quadratic_loss ] ) with @xmath104 this gives @xmath105 \leq\inf_{\nu\in\mathcal{s}_n } { 1\over t}\sum _ { t=1}^t{\mathbb{e}}\bigl [ \bigl(\widehat x_t^{[\nu]}-x_{t } \bigr)^2 \bigr ] + c_1 \biggl(\frac{\log n}{t } \biggr)^{1/2 } , \ ] ] with @xmath106 . assume that the noise @xmath94 satisfies with a given @xmath107 and let @xmath96 denote the aggregated predictor obtained using the weights ( [ definition : alphas_quadratic_loss ] ) with @xmath108 we then have @xmath109 \nonumber\\[-8pt]\\[-8pt]\nonumber & & \qquad \leq\min_{1\leq i\leq n } { 1\over t}\sum _ { t=1}^t{\mathbb{e}}\bigl [ \bigl(\widehat x_{t}^{(i)}-x_{t } \bigr)^2 \bigr ] + c_2 \biggl(\frac{\log n}{t } \biggr)^{1 - 2/p } , \end{aligned}\ ] ] with @xmath110 . assume that the noise @xmath94 fulfills for some positive @xmath100 and let @xmath96 denote the aggregated predictor obtained using the weights ( [ definition : alphas_quadratic_loss ] ) with @xmath111 then we have @xmath112\nonumber \\ & & \qquad \leq\min_{1\leq i\leq n } { 1\over t}\sum _ { t=1}^t{\mathbb{e}}\bigl [ \bigl(\widehat x_{t}^{(i)}-x_{t } \bigr)^2 \bigr]+ \frac{2a_{*}^2(l_*+1)^2}{\zeta^2 } \frac{\log n}{t } \\ & & \quad\qquad { } \times \biggl\ { \biggl(\log\biggl ( \frac{t}{\log n } \biggr ) \biggr)^{2 } + \frac{\phi(\zeta)}{\mathrm{e}}\biggl(2 + \log\biggl(\frac{t}{\log n } \biggr ) \biggr ) \biggr\ } .\nonumber\end{aligned}\ ] ] [ note that when @xmath113 , the term between curly brackets is equivalent to @xmath114 . ] cases ( i ) and ( ii ) in corollary [ corollary : oracle - bounds ] follow directly from theorem [ thm : oracle - bounds ] . case ( iii ) is more delicate since it requires optimizing @xmath103 as well as @xmath115 in the second line of ( [ eq : strat2_exp - final ] ) . the details are postponed to section [ sec : proof - corollary - strat12 ] . we observe that the bound in ( [ eq : strat2_exp - final - opt - eta ] ) improves that in ( [ eq : strat2_p - final - eta - opt ] ) for any @xmath98 . for @xmath116 , the remaining term @xmath117 in ( [ eq : strat2_p - final - eta - opt ] ) is smaller than the remaining term @xmath118 in ( [ eq : strat1-final - eta - opt ] ) . similarly , the remaining term @xmath119 in ( [ eq : strat2_exp - final - opt - eta ] ) is smaller than @xmath118 in ( [ eq : strat1-final - eta - opt ] ) . yet , we emphasize that the oracle inequalities ( [ eq : strat2_p - final - eta - opt ] ) and ( [ eq : strat2_exp - final - opt - eta ] ) compare the prediction risk of @xmath120 to the prediction risk of the _ best predictor _ @xmath121 , while the oracle inequality ( [ eq : strat1-final - eta - opt ] ) compare the prediction risk of @xmath120 to the prediction risk of the _ best convex combination of the predictors _ @xmath122 , so they can not be directly compared . [ improve - with - net ] as explained in @xcite [ ( @xcite ) , section c ] , under the hypotheses of cases ( ii ) and ( iii ) and for certain values of @xmath26 and @xmath58 , using a more involved aggregation step , we can get a new predictor satisfying an oracle inequality better than that in ( [ eq : strat1-final - eta - opt ] ) . for example , under the hypotheses of case ( iii ) , for @xmath123 , the remaining term @xmath118 in ( [ eq : strat1-final - eta - opt ] ) can be replaced by @xmath124 which is smaller ; see @xcite [ ( @xcite ) , inequality ( c.7 ) , page 8 ] . yet , this aggregation has a prohibitive computational cost and seems difficult to implement in practice . [ rem : eta - choice - vs - conditions ] in cases ( ii ) and ( iii ) , which correspond to the weights ( [ definition : alphas_quadratic_loss ] ) , the choice of the optimal @xmath79 depends on the assumptions on the noise , namely or . under a moment condition of order @xmath125 , the optimal @xmath79 is of order @xmath126 and under an exponential condition , it is of order @xmath127 . it is known from @xcite [ ( @xcite ) , proposition 2.2.1 ] and @xcite [ ( @xcite ) , theorem 5 ] that @xmath79 can be chosen as a constant ( provided that it is small enough ) under a bounded noise condition , or under an exponential moment condition on the noise for predictors at a bounded distance from the conditional mean . hence , coarsely speaking , the heavier the tail of the noise , the smallest @xmath79 should be chosen . observing that @xmath79 allows us to tune the influence of the empirical risk on the weights from no influence at all ( @xmath128 yielding uniform weights ) to the selection of the empirical risk minimizer ( @xmath129 ) , the specific choices of @xmath79 can be interpreted as follows : the heavier the tail of the noise , the less we can trust the empirical risk . in the literature , prediction risk bounds of the form ( [ eq : strat1-final - eta - opt ] ) [ case ( i ) of corollary [ corollary : oracle - bounds ] ] are sometimes called _ convex regret bounds _ , and prediction risk bounds of the form ( [ eq : strat2_p - final - eta - opt ] ) and ( [ eq : strat2_exp - final - opt - eta ] ) [ cases ( ii ) and ( iii ) of corollary [ corollary : oracle - bounds ] ] are sometimes called _ best predictor regret bounds_. @xcite exhibits convex regret bounds in a setting close to ours , namely for an online aggregation of predictors for a sequence of possibly dependent random variables . under our moment condition with @xmath95 , @xcite [ ( @xcite ) , theorem 2 ] provides an upper bound similar to ( [ eq : strat1-final - eta - opt ] ) but with our remaining term @xmath130 replaced by @xmath131 . under the exponential condition , @xcite [ ( @xcite ) , theorem 1 ] provides an upper bound similar to ( [ eq : strat1-final - eta - opt ] ) but with a remaining term @xmath132 , which is still larger than our remaining term under moment conditions . best predictor regret bounds can be found in @xcite for some sequences of possibly dependent random variables . the predictors are assumed to remain at a bounded distance to the conditional means and the scaled innovation noise is assumed to have either a known distribution ( satisfying a certain technical condition ) or an exponential moment . the regret bounds are presented in a slightly different fashion from ours but it is easy to see that a similar result as our bound ( [ eq : strat2_exp - final - opt - eta ] ) is obtained in this setting . however , we do not require bounded prediction errors and our conditions on the noise are milder . the i.i.d . setting has received much more attention and , even if the setting is quite different , it is interesting to briefly compare our results to previous works in this case . let us start with the convex regret bound in case ( i ) of corollary [ corollary : oracle - bounds ] . most of the existing results [ see , e.g. , @xcite , @xcite , @xcite or @xcite for recent extensions to @xmath133 aggregation ] assume the predictors to be bounded and various conditions on the noise are considered ( very often the noise is assumed to be gaussian ) . in such settings , the best possible remaining term typically takes the form @xmath118 when @xmath58 is much larger than @xmath134 and of the form @xmath135 if @xmath58 is smaller than @xmath134 ; see @xcite [ ( @xcite ) , theorem 3.1 ] , @xcite [ ( @xcite ) , theorem 6 ] and @xcite [ ( @xcite ) , theorem 2 ] . hence , our bound ( [ eq : strat1-final - eta - opt ] ) is similar only in the case where @xmath58 is much larger than @xmath134 . however , as explained in remark [ improve - with - net ] and [ giraud , roueff and sanchez - perez ( @xcite ) , section c ] , when @xmath26 is larger than @xmath136 and under the moment condition , we can get via a more involved aggregation procedure , a convex regret bound with a remaining term of the same order @xmath135 up to a @xmath137 factor [ see @xcite , inequality ( c.7 ) , page 8 ] . let us now compare our bound ( [ eq : strat2_p - final - eta - opt ] ) in case ( ii ) to optimal bounds in the i.i.d . setting under moment conditions on the noise . corollary 7.2 and theorem 8.6 in @xcite shows that the optimal aggregation rate is @xmath138 in the i.i.d . setting with bounded predictors and moment conditions of order @xmath125 on the noise . our remaining term @xmath117 in ( [ eq : strat2_p - final - eta - opt ] ) is slightly larger , yet an inspection of the proof of @xcite [ ( @xcite ) , corollary 7.2 ] shows that the aggregation rate would also be @xmath117 in this corollary , if the predictors were assumed to have a moment condition of order @xmath125 instead of being uniformly bounded ( we are not aware of any lower bound in this setting matching this rate ) . finally , when the data and the predictors are bounded , the best aggregation rate is known to be @xmath139 in the i.i.d . setting ; see , for example , @xcite , theorem 8.4 . our bound ( [ eq : strat2_exp - final - opt - eta ] ) in case ( iii ) achieves the same rate up to a @xmath140 factor . we introduce some preliminary notation before defining the model . in the remainder of this article , vectors are denoted using boldface symbols and @xmath141 denotes the euclidean norm of @xmath142 , @xmath143 . for @xmath144 $ ] and an interval @xmath145 , the @xmath146-hlder semi - norm of a function @xmath147 is defined by @xmath148 this semi - norm is extended to any @xmath149 as follows . let @xmath150 and @xmath151 $ ] be such that @xmath152 . if @xmath153 is @xmath154 times differentiable on @xmath155 , we define @xmath156 and @xmath157 otherwise . we consider the case @xmath158 $ ] . for @xmath159 and @xmath149 , the @xmath160-hlder ball is denoted by @xmath161\rightarrow{\mathbb{r}}^{d } , \mbox { such that } { \vert}\mathbf{f}{\vert}_{\beta } \leq r \bigr\ } . \ ] ] the idea of using a rescaled time with the sample size @xmath26 for the tvar parameters goes back to @xcite . since then , it has always been a central example of locally stationary linear processes . in this setting , the time varying autoregressive coefficients and variance which generate the observations @xmath162 for @xmath163 are represented by functions from @xmath28 $ ] to @xmath164 and from @xmath28 $ ] to @xmath165 , respectively . the definition sets of these functions are extended to @xmath166 $ ] in the following definition . [ definition : tvar ] let @xmath167 . let @xmath168 and @xmath169 be functions defined on @xmath166 $ ] and @xmath11 be a sequence of i.i.d . random variables with zero mean and unit variance . for any @xmath54 , we say that @xmath170 is a tvar process with time varying parameters @xmath171 sampled at frequency @xmath172 and normalized innovations @xmath18 if the two following assertions hold : the process @xmath91 fulfills the time varying autoregressive equation @xmath173 the sequence @xmath170 is bounded in probability , @xmath174 this definition extends the usual definition of tvar processes , where the time varying parameters @xmath168 and @xmath175 are assumed to be constant on @xmath176 ; see , for example , @xcite [ ( @xcite ) , page 144 ] . the tvar model is generally used for the sample @xmath177 . the definition of the process for negative times @xmath16 can be seen as a way to define initial conditions for @xmath178 , which are then sufficient to compute @xmath177 by iterating ( [ equation : definition_tvar ] ) . however , in the context of prediction , it can be useful to consider predictors @xmath179 which may rely on historical data @xmath180 arbitrarily far away in the past , that is , with @xmath181 tending to @xmath182 . to cope with this situation , our definition of the tvar process @xmath183 holds for all time indices @xmath184 and we use the following definition for predictors . [ def : predictor ] for all @xmath163 , we say that @xmath179 is a predictor of @xmath162 if it is @xmath185-measurable , where @xmath186 is the @xmath169-field generated by @xmath187 . for any @xmath54 , we denote by @xmath188 the set of sequences @xmath189 of predictors for @xmath177 , that is , the set of all processes @xmath189 adapted to the filtration @xmath190 . in this general framework , the time @xmath191 corresponds to the beginning of the aggregation procedure . such a framework applies in two practical situations . in the first one , we start collecting data @xmath162 at @xmath192 and compute several predictors @xmath193 , @xmath194 from them . thus , the resulting aggregated predictor only depends on @xmath195 . a somewhat different situation is when historical data is available beforehand the aggregation step , so that a given predictor @xmath193 is allowed to depend also on data @xmath180 with @xmath196 , while the aggregation step only starts at @xmath192 , and thus depends on the data @xmath197 only through the predictors . it is important to note that , in contrast to the usual stationary situation , having observed the process @xmath180 for infinitely many @xmath181 s in the past ( for all @xmath198 ) is not so decisive for deriving a predictor of @xmath162 , since observations far away in the past may have a completely different statistical behavior . the next proposition proves that under standard stability conditions on the time varying parameters @xmath168 and @xmath175 , condition ( ii ) in definition [ definition : tvar ] ensures the existence and uniqueness of the solution of equation ( [ equation : definition_tvar ] ) for @xmath199 ( and thus for all @xmath200 ) . we define the time varying autoregressive polynomial by @xmath201 let us denote , for any @xmath202 , @xmath203\rightarrow { \mathbb{r}}^{d},{\bolds\theta}(z;u)\neq0 , \forall{\vert}z{\vert } < \delta^{-1 } , u\in[0,1 ] \bigr\ } .\ ] ] define , for @xmath149 , @xmath159 , @xmath204 , @xmath205 $ ] and @xmath206 , the class of parameters @xmath207\to{\mathbb{r}}^d\times[\rho\sigma_+ , \sigma_+]:\bolds{\theta}\in\lambda_d(\beta , r)\cap s_d(\delta ) \bigr\ } .\end{aligned}\ ] ] the definition of the class @xmath208 is very similar to that of @xcite . the domain of definition in their case is @xmath28 $ ] whereas it is @xmath166 $ ] in ours . we have the following stability result . [ proposition : stationary_tvar ] assume that the time varying ar coefficients @xmath209 are uniformly continuous on @xmath166 $ ] and the time varying variance @xmath210 is bounded on @xmath166 $ ] . assume moreover that there exists @xmath204 such that @xmath211 . then there exists @xmath212 such that , for all @xmath213 , there exists a unique process @xmath170 which satisfies and in definition [ definition : tvar ] . this solution admits the linear representation @xmath214 where the coefficients @xmath215 satisfy that for any @xmath216 , @xmath217 moreover , if @xmath218 for some positive constants @xmath146 , @xmath219 and @xmath220 , then the constants @xmath221 and @xmath222 can be chosen only depending on @xmath223 , @xmath224 , @xmath146 and @xmath219 . a proof of proposition [ proposition : stationary_tvar ] is provided in . this kind of result is classical under various smoothness assumptions on the parameters and initial conditions for @xmath225 , @xmath226 . for instance , in @xcite , bounded variations and a constant @xmath227 for negative times are used for the smoothness assumption on @xmath227 and for defining the initial conditions . the linear representation ( [ eq : inf - series - representation ] ) of tvar processes was first obtained in the seminal papers @xcite . we note that an important consequence of proposition [ proposition : stationary_tvar ] is that for any @xmath213 , the process @xmath170 satisfies assumption ( m-1 ) with @xmath21 and @xmath228 for @xmath229 . moreover , the constant @xmath7 in ( [ eq : unif - coeff - bound ] ) is bounded independently of @xmath26 , and we have , for all @xmath218 , @xmath230 where @xmath231 and @xmath232 can be chosen only depending on @xmath224 , @xmath146 and @xmath219 . based on proposition [ proposition : stationary_tvar ] , given an i.i.d . sequence @xmath37 and constants @xmath204 , @xmath205 $ ] , @xmath206 , @xmath149 and @xmath159 , we consider the following assumption . the sequence @xmath170 is a tvar process with time varying standard deviation @xmath169 , time varying ar coefficients @xmath168 and innovations @xmath11 , and @xmath233 . let @xmath234 denote a generic random variable with the same distribution as the @xmath235 s . under assumption ( m-2 ) , the distribution of @xmath236 only depends on that of @xmath237 and on the functions @xmath227 and @xmath169 . for a given distribution @xmath238 on @xmath90 for @xmath234 , we denote by @xmath239 the probability distribution of the whole sequence @xmath170 and by @xmath240 its corresponding expectation . the next two assumptions on the innovations are useful to prove upper bounds of the prediction error . [ hyp : innov - moment ] the innovations @xmath11 satisfy @xmath241<\infty$ ] . the innovations @xmath242 satisfy @xmath243<\infty$ ] . the following one will be used to obtain a lower bound . [ hyp : innov - lowerbound ] the innovations @xmath11 admit a density @xmath244 such that @xmath245 assumption ( i-3 ) is standard for proving lower bounds in nonparametric regression estimation , see @xcite , chapter 2 . it is satisfied by gaussian density with @xmath246 . the setting of definition [ definition : tvar ] and of assumptions derived thereafter is essentially nonparametric , since for given initial distribution @xmath238 , the distribution of the observations @xmath247 are determined by the unknown parameter function @xmath248 . the doubly indexed @xmath162 refers to the fact that this distribution can not be seen as a distribution on @xmath249 marginalized on @xmath250 as the usual time series setting but rather as a sequence of distributions on @xmath250 indexed by @xmath26 . it corresponds to the usual nonparametric approach for studying statistical inference based on this model . in this contribution , we focus on the prediction problem , which is to answer the question : for given smoothness conditions on @xmath248 , what is the mean prediction error for predicting @xmath162 from its past ? the standard nonparametric approach is to answer this question in a minimax sense by determining , for a given sequence of predictors @xmath251 , the maximal risk @xmath252\\[-8pt]\nonumber & & \qquad = \sup_{({\bolds\theta},\sigma ) } { \frac{1}{t } } \sum_{t=1}^{t } \biggl({\mathbb{e}}^{\psi } _ { ( { \bolds\theta},\sigma ) } \bigl [ ( \widehat{x}_{t , t}-x_{t , t } ) ^{2 } \bigr ] - \sigma^{2 } \biggl ( { \frac{t}{t } } \biggr ) \biggr ) , \end{aligned}\ ] ] where : @xmath253 is assumed to belong to @xmath188 as in definition [ def : predictor ] , the sup is taken over @xmath254 within a smoothness class of functions , the expectation @xmath255 is that associated to assumption ( m-2 ) . the reason for subtracting the average @xmath256 over all @xmath163 in this prediction risk is that it corresponds to the best prediction risk , would the parameters @xmath248 be exactly known . we observe that dividing @xmath162 by the class parameter @xmath220 amounts to take @xmath257 . in addition , we have @xmath258 so the prediction problem in the class @xmath259 can be reduced to the prediction problem in the class @xmath260 . accordingly , we define the reduced minimax risk by @xmath261 in section [ section : lower_bound ] , we provide a lower bound of the minimax rate in the case where the smoothness class is of the form @xmath259 . then , in section [ subsection : minimax ] , relying on the aggregation oracle bounds of section [ subsection : general : bound ] , we derive an upper bound with the same rate as the lower bound using the same smoothness class of the parameters . moreover , we exhibit an online predictor which does not require any knowledge about the smoothness class and which is thus minimax adaptive . in other words , it is able to adapt to the unknown smoothness of the parameters from the data . to our knowledge , such theoretical results are new for locally stationary models . a lower bound on the minimax rate for the estimation error of @xmath227 is given by @xcite [ ( @xcite ) , theorem 4 ] . clearly , a predictor @xmath262 can be defined from an estimator @xmath263 , and the resulting prediction rate can be controlled using the estimation rate ( see @xcite [ ( @xcite ) , section b.1 ] for the details ) . the next theorem provides a lower bound of the minimax rate of the risk of _ any _ predictor of the process @xmath177 . combining this result with [ @xcite ( @xcite ) , lemma 9 ] , we show that a predictor obtained by [ giraud , roueff and sanchez - perez ( @xcite ) , equation ( b.1 ) ] from a minimax rate estimator of @xmath227 automatically achieves the minimax prediction rate . [ theorem : lower_bound_quadratic_risk ] let @xmath204 , @xmath149 , @xmath159 and @xmath205 $ ] . suppose that assumption holds and assume on the distribution @xmath238 of the innovations . then we have @xmath264 where @xmath265 is defined in ( [ eq : reduced - minimax - risk ] ) . the proof is postponed to section [ section : proof_lower_bound ] . in @xcite , an adaptive estimator of the autoregressive function of a gaussian tvar process of order 1 is studied . it relies on the lepski s procedure [ see @xcite ] , which seems difficult to implement in an online context . our minimax adaptive predictor is based on the aggregation of sufficiently many predictors , assuming that at least one of them converges at the minimax rate . the oracle bounds found in section [ subsection : general : bound ] imply that the aggregated predictor is minimax rate adaptive under appropriate assumptions . seminal works using the aggregation to adapt to the minimax convergence rate are @xcite ( nonparametric regression ) and @xcite ( density estimation ) ; see also @xcite for a more general presentation . in the tvar model ( m-2 ) , it is natural to consider @xmath82-lipschitz predictors @xmath266 of @xmath177 with a sequence @xmath82 supported on @xmath267 . then @xmath268 in ( [ eq : hyp : lip ] ) corresponds to the maximal @xmath0-norm of the tvar parameters . since for the process itself to be stable , this norm has to be bounded independently of @xmath26 , condition is a quite natural assumption for the tvar model ; see @xcite [ ( @xcite ) , section b.1 ] for the details . a practical advantage of the proposed procedures is that , given a set of predictors that behaves well under specific smoothness assumptions , we obtain an aggregated predictor which performs almost as well as or better than the best of these predictors , hence which behaves well without any prior knowledge on the smoothness of the unknown parameter . such an adaptive property can be formally demonstrated by exhibiting an adaptive minimax rate for the aggregated predictor which coincides with the lower bound given in theorem [ theorem : lower_bound_quadratic_risk ] . the first ingredient that we need is the following . [ def : minimax - rate - expert ] let @xmath238 be a distribution on @xmath90 and @xmath149 . we say that @xmath269 is a @xmath270-minimax - rate sequence of predictors if , for all @xmath54 , @xmath271 and , for all @xmath204 , @xmath159 , @xmath272 $ ] and @xmath206 , @xmath273 where @xmath274 is defined by ( [ eq : max - risk ] ) . the term _ minimax - rate _ in this definition refers to the fact that the maximal rate in ( [ equation : definition_minimax ] ) is equal to the minimax lower bound ( [ eq : lower - bound - minimax ] ) for the class @xmath259 . we explain in @xcite [ ( @xcite ) , section b ] how to build such predictors which are moreover @xmath82-lipschitz for some @xmath82 only depending on @xmath275 . to adapt to an unknown smoothness , we rely on a collection of @xmath276-minimax - rate predictors with @xmath146 within @xmath277 , where @xmath278 is the ( possibly infinite ) maximal smoothness index . [ def : minimax - rate - expert - loc - bounded ] let @xmath238 be a distribution on @xmath90 and @xmath279 $ ] . we say that @xmath280 is a locally bounded set of @xmath238-minimax - rate predictors if for each @xmath146 , @xmath281 is a @xmath276-minimax - rate predictor and if moreover , for all @xmath204 , @xmath159 , @xmath272 $ ] , @xmath282 and for each closed interval @xmath283 , @xmath284 where @xmath274 is defined by ( [ eq : max - risk ] ) . the following lemma shows that , given a locally bounded set of minimax - rate predictors , we can always pick a finite subset of at most @xmath285 predictors among which the best one achieves the minimax rate of any unknown smoothness index . [ lemma : minimax_experts ] let @xmath238 be a distribution on @xmath90 . let @xmath286 $ ] and @xmath287 be a corresponding locally bounded set of @xmath238-minimax - rate predictors . set , for any @xmath288 , @xmath289 suppose moreover , in the case where @xmath290 , that @xmath291 , and , in the case where @xmath292 , that @xmath293 . then we have , for all @xmath294 , @xmath204 , @xmath159 , @xmath295 and @xmath206 , @xmath296 the proof of this lemma is postponed to @xcite [ ( @xcite ) , section a.8 ] . lemma [ lemma : minimax_experts ] says that to obtain a minimax - rate predictor which adapts to an unknown smoothness index @xmath146 , it is sufficient to select it judiciously among @xmath297 or @xmath140 well chosen nonadaptive minimax - rate predictors . as a consequence of theorem [ thm : oracle - bounds ] and lemma [ lemma : minimax_experts ] , we obtain an adaptive predictor by aggregating them ( instead of selecting one of them ) , as stated in the following result . [ theorem : upper_bound_aggregated_forecaster_adaptative ] let @xmath238 be a distribution on @xmath90 . let @xmath286 $ ] and @xmath298 , @xmath299 be a locally bounded set of @xmath238-minimax - rate and @xmath82-lipschitz predictors with @xmath82 satisfying . define @xmath300 as the predictor aggregated from @xmath301 with @xmath58 defined by @xmath302 @xmath303 defined by ( [ eq : betai - def ] ) , and with weights defined according to one of the following setting depending on the assumption on @xmath238 and @xmath278 : if @xmath238 satisfies with @xmath304 and @xmath305 , use the weights ( [ definition : alphas_gradient_quadratic_loss ] ) with @xmath306 . if @xmath238 satisfies with @xmath98 and @xmath307 , use the weights ( [ definition : alphas_quadratic_loss ] ) with @xmath308 . if @xmath238 satisfies , use the weights ( [ definition : alphas_quadratic_loss ] ) with @xmath309 . then we have , for any @xmath310 , @xmath204 , @xmath159 , @xmath272 $ ] and @xmath206 , @xmath311 the proof of this theorem is postponed to giraud , roueff and sanchez - perez [ ( @xcite ) , section a.9 ] . [ remark : theorem_upper_bound_aggregated_forecaster_adaptative ] the limitation to @xmath305 in ( i ) under assumption ( i-1 ) for @xmath238 follows from the factor @xmath118 obtained in the oracle inequality ( [ eq : strat1-final ] ) of theorem [ thm : oracle - bounds ] after optimizing in @xmath79 [ see ( [ eq : strat1-final - eta - opt ] ) ] . if @xmath116 this restriction is weakened to @xmath307 in ( ii ) taking into account the factor @xmath117 obtained in the oracle inequality ( [ eq : strat2_p - final ] ) of theorem [ thm : oracle - bounds ] after optimizing in @xmath79 [ see ( [ eq : strat2_p - final - eta - opt ] ) ] . in the last case , the limitation of @xmath312 drops when applying the oracle inequality ( [ eq : strat2_exp - final ] ) of the same theorem . however , a stronger condition on @xmath238 is then required . [ rem : beta0-eta - minim - adapt - forec ] it may happen that the locally bounded set of @xmath238-minimax - rate predictors is limited to some @xmath313 [ see the example of the nlms predictors in @xcite , section b.2 ] . in this case , the result roughly needs @xmath297 predictors and the computation of the aggregated one requires less operations than if @xmath312 were infinite . for these reasons , we do not consider in general that @xmath314 . on the one hand , a finite @xmath278 yields a restriction on the set of ( unknown ) smoothness indices @xmath146 for which the aggregated predictors are minimax rate adaptive . on the other hand , if @xmath292 , theorem [ theorem : upper_bound_aggregated_forecaster_adaptative ] then requires the stronger assumption ( i-2 ) on the process . [ rem : sigma - eta - minim - adapt - forec ] the constant @xmath315 present in the definitions of @xmath79 in the three cases ( i ) , ( ii ) and ( iii ) corresponds to the homogenization of the remaining terms appearing in theorem [ thm : oracle - bounds ] [ the second lines of ( [ eq : strat1-final ] ) , ( [ eq : strat2_p - final ] ) and ( [ eq : strat2_exp - final ] ) ] . indeed with the proposed choices and in the three cases , the constant @xmath316 factors out in front of the remaining terms [ see the last three displayed equations in @xcite , section a.9 ] . however , the @xmath317 in the definitions of @xmath79 does not impact the convergence rate in the sense that theorem [ theorem : upper_bound_aggregated_forecaster_adaptative ] is still valid using any other constant ( @xmath77 , e.g. ) in these definitions . we start with a lemma which gathers useful adaptations of well - known inequalities applying to the aggregation of deterministic predicting sequences . [ lemma : deterministic_upper_bound_aggregated_forecaster ] let @xmath318 be a real valued sequence and @xmath319 be a collection of predicting sequences . define @xmath320 as the sequence of aggregated predictors obtained from this collection with the weights ( [ definition : alphas_gradient_quadratic_loss ] ) . then , for any @xmath74 , we have @xmath321}-x_{t } \bigr)^2 + \frac{\log n}{t\eta } + \frac{2\eta}{t}\sum _ { t=1}^{t } y_{t}^{4 } , \end{aligned}\ ] ] where @xmath322 . define now @xmath320 as the sequence of aggregated predictors obtained with the weights ( [ definition : alphas_quadratic_loss ] ) . then , for any @xmath323 , we have @xmath324\\[-8pt]\nonumber & & \qquad \leq\min _ { i=1,\dots , n}\frac{1}{t}\sum_{t=1}^t \bigl(\widehat{x}_{t}^{(i)}-x_{t } \bigr)^2 + \frac{\log n}{t\eta } + \frac{1}{t}\sum _ { t=1}^{t } \biggl(y_{t}^{2}- \frac{1}{2\eta } \biggr)_{+ } , \end{aligned}\ ] ] where @xmath322 . with the weights defined by ( [ definition : alphas_gradient_quadratic_loss ] ) , by slightly adapting [ @xcite ( @xcite ) , theorem 1.7 ] , we have that @xmath325}-x_{t } \bigr)^2 & \leq & \frac{\log n}{t\eta } + \frac{\eta}{8 t } s_{t}^ { * } , \end{aligned}\ ] ] where @xmath326 and @xmath327 . the bound ( [ eq : strat1-deterministic ] ) follows by using that @xmath328 is in the simplex @xmath329 defined in ( [ eq : def - simplexn ] ) . we now prove ( [ eq : strat2-deterministic - allcases ] ) . we adapt the proof of @xcite [ ( @xcite ) , proposition 2.2.1 . ] to unbounded sequences by replacing the convexity argument by the following lemma . [ lemma : bound_integral_concave_exponential ] let @xmath330 and @xmath331 a probability distribution supported on @xmath332 $ ] . then we have @xmath333 the proof of lemma [ lemma : bound_integral_concave_exponential ] is postponed to section [ sec : proof - lemma - technical - exp - concave ] in . now , let @xmath74 and @xmath334 . using lemma [ lemma : bound_integral_concave_exponential ] with the probability distribution @xmath331 defined by @xmath335 and @xmath336 , we get that @xmath337 taking the log , multiplying by @xmath338 and re - ordering the terms , we obtain that @xmath339 taking the average over @xmath69 and developing the expression of @xmath67 , we obtain @xmath340\\[-8pt]\nonumber & & { } + \frac{1}{t}\sum_{t=1}^{t } \biggl(y_{t}^{2}-\frac{1}{2\eta } \biggr)_{+}.\end{aligned}\ ] ] using that @xmath341 , we get the bound ( [ eq : strat2-deterministic - allcases ] ) . we prove the cases ( i ) , ( ii ) and ( iii ) successively . we denote @xmath342 . applying ( [ eq : strat1-deterministic ] ) in lemma [ lemma : deterministic_upper_bound_aggregated_forecaster ] with @xmath343\leq\inf{\mathbb{e}}[\cdots]$ ] , we obtain @xmath344 & \leq&\inf_{\bolds{\nu}\in\mathcal{s}_n}\frac{1}{t}\sum _ { t=1}^t{\mathbb{e}}\bigl [ \bigl(\widehat{x}_{t}^{[\bolds\nu]}-x_{t } \bigr)^2 \bigr ] \nonumber\\[-8pt]\\[-8pt]\nonumber & & { } + \frac{\log n}{t\eta } + \frac{2\eta}{t}\sum _ { t=1}^{t } { \mathbb{e}}\bigl[y_{t}^{4 } \bigr ] .\end{aligned}\ ] ] using that the predictors are @xmath82-lipschitz and the process @xmath1 satisfies ( m-1 ) , we have , for all @xmath345 , @xmath346 where @xmath347 applying the minkowski inequality together with ( [ inequality : x - z ] ) , ( [ eq : unif - coeff - bound ] ) and ( [ eq : hyp : lip ] ) , we obtain , for all @xmath163 , @xmath348 \leq{\mathbb{e}}\biggl [ \biggl(\sum _ { j\in{\mathbb{z } } } b_t(j ) z_{t - j } \biggr)^{4 } \biggr ] \leq a_{*}^4(1+l_*)^{4 } \sup_{t\in { \mathbb{z}}}{\mathbb{e}}\bigl[z_t^4 \bigr ] .\end{aligned}\ ] ] since the process @xmath94 fulfills with @xmath95 , plugging this bound in ( [ inequality : expectation_aggregated_forecaster_moment ] ) we obtain ( [ eq : strat1-final ] ) . we use ( [ eq : strat2-deterministic - allcases ] ) in lemma [ lemma : deterministic_upper_bound_aggregated_forecaster ] and the inequality @xmath349 which holds for @xmath350 and @xmath351 . we get , taking the expectation , @xmath352 & \leq&\min_{i=1,\dots , n}\frac{1}{t}\sum _ { t=1}^{t}{\mathbb{e}}\bigl [ \bigl(\widehat{x}_{t , t}^{(i)}-x_{t , t } \bigr)^2 \bigr ] + \frac{\log n}{t\eta } \nonumber\\[-8pt]\\[-8pt]\nonumber & & { } + ( 2\eta)^{p/2 - 1}\max_{t=1,\ldots , t}{\mathbb{e}}\bigl[y_{t}^p \bigr ] .\end{aligned}\ ] ] applying the minkowski inequality , ( [ inequality : x - z ] ) and assumption , @xmath353 \leq\biggl(\sum _ { j\in z}b_{t}(j ) \bigl({\mathbb{e}}\bigl[z_{t - j}^{p } \bigr ] \bigr)^{1/p } \biggr)^{p } \leq a_{*}^{p}(1+l_*)^{p } \sup_{t\in z}{\mathbb{e}}\bigl[z_t^{p } \bigr ] . \ ] ] using this bound which is independent of @xmath16 , with and ( [ inequality : expectation_aggregated_forecaster_moment_1 ] ) , the inequality ( [ eq : strat2_p - final ] ) follows . to obtain ( [ eq : strat2_exp - final ] ) , we again use ( [ eq : strat2-deterministic - allcases ] ) in lemma [ lemma : deterministic_upper_bound_aggregated_forecaster ] but now with an exponential bound for @xmath354 . we note that , or all @xmath355 , @xmath356 it follows that , for all @xmath357 and @xmath355 , @xmath358 applying this bound with @xmath359 and @xmath360 we get @xmath361 plugging this into ( [ eq : strat2-deterministic - allcases ] ) and taking the expectation , we obtain that @xmath362\nonumber \\ & & \qquad \leq\min_{i=1,\dots , n}\frac{1}{t}\sum _ { t=1}^{t}{\mathbb{e}}\bigl [ \bigl(\widehat{x}_{t , t}^{(i)}-x_{t , t } \bigr)^2 \bigr ] + \frac{\log n}{t\eta } \\ & & \quad\qquad { } + 2\lambda^{-2 } \bigl(2+\lambda(2\eta)^{-1/2 } \bigr ) { \mathrm{e}}^{-1- \lambda(2\eta)^{-1/2 } } \max_{t=1,\ldots , t}{\mathbb{e}}\bigl[{\mathrm{e}}^{\lambda y_t } \bigr ] .\nonumber\end{aligned}\ ] ] we now use assumption . since @xmath363 for all @xmath364 and @xmath365 jensen s inequality and ( [ inequality : x - z ] ) gives that , for any @xmath366 , @xmath367&\leq&{\mathbb{e}}\bigl[{\mathrm{e}}^{\lambda ( { \vert}{x}_{t}{\vert}+\max_{1\leq i\leq n}{\vert}\widehat { x}_{t}^{(i)}{\vert } ) } \bigr ] \\ & \leq&\prod_{j\in{\mathbb{z}}}{\mathbb{e}}\bigl[{\mathrm{e}}^{\lambda b_t(j ) z_{t - j } } \bigr ] \\ & \leq&\prod_{j\in{\mathbb{z } } } \bigl(\phi(\zeta ) \bigr)^{\lambda b_t(j)/\zeta } \leq\bigl(\phi(\zeta ) \bigr)^{\lambda a_{*}(1+l_{*})/\zeta}.\end{aligned}\ ] ] the combination of this bound with ( [ inequality : expectation_aggregated_forecaster_exponential_moment ] ) gives ( [ eq : strat2_exp - final ] ) . the proof of theorem [ thm : oracle - bounds ] is complete . minimizing the sum of the two terms appearing in the second line of ( [ eq : strat2_exp - final ] ) is a bit more involved , since it depends both on @xmath79 and @xmath103 . under condition ( [ eq : def - sup - unif - coeff - bound ] ) , the quantity @xmath368 remains between two positive constants while , for any @xmath74 , @xmath369 is decreasing as @xmath103 increases . to simplify @xmath370 into @xmath371 , we simply take @xmath372 which satisfies ( [ eq : def - sup - unif - coeff - bound ] ) . now that @xmath103 is set , it remains to choose a value of @xmath79 which ( almost ) minimizes @xmath373 the @xmath79 defined as in ( [ eq : strat2_exp - choice - eta ] ) is chosen so that @xmath374 , and we get ( [ eq : strat2_exp - final - opt - eta ] ) . we now provide a proof of theorem [ theorem : lower_bound_quadratic_risk ] . we consider an autoregressive equation of order one @xmath375 where @xmath11 is i.i.d . with density @xmath244 as in ( i-3 ) . in this case , if @xmath376 , the representation ( [ eq : inf - series - representation ] ) of the stationary solution reads , for all @xmath377 as @xmath378 with the convention @xmath379 . the class of models so defined with @xmath380 corresponds to assumption ( m-2 ) with @xmath248 in @xmath260 such that only the first component of @xmath227 is nonzero and @xmath169 is constant and equal to one . we write henceforth in this proof @xmath381 for the law of the process @xmath382 and @xmath383 for the corresponding expectation . let @xmath384 be any predictor of @xmath385 in the sense of definition [ def : predictor ] . define @xmath386 by @xmath387 for any vectors @xmath388 , we define @xmath389 by ( [ eq : lb - tvar - eq ] ) , since @xmath162 and @xmath390 are @xmath391-measurable , they are independent of @xmath392 and we have @xmath393 - 1={\mathbb{e}}_{\theta } \bigl[d_{x}^2 \bigl(\widehat\theta,\mathrm{v}_{t}\{\theta\}\bigr ) \bigr ] , \ ] ] where , for any @xmath394\to{\mathbb{r}}$ ] , @xmath395 denotes the @xmath26-sample of @xmath396 on the regular grid @xmath397 , @xmath398 hence , to prove the lower bound of theorem [ theorem : lower_bound_quadratic_risk ] , it is sufficient to show that there exist @xmath399 , @xmath400 and @xmath401 both depending only on @xmath402 , @xmath146 , @xmath219 and the density @xmath244 , such that for any @xmath403 adapted to @xmath404 and @xmath405 , we have @xmath406\geq c t^{-2\beta/ ( 2\beta+1 ) } .\ ] ] we now face the more standard problem of providing a lower bound for the minimax rate of an estimation error , since @xmath407 is an estimator of @xmath408 . the path for deriving such a lower bound is explained in [ @xcite ( @xcite ) , chapter 2 ] . however , we have to deal with a loss function @xmath409 which depends on the observed process @xmath91 . not only the loss function is random , but it is also not independent of the estimator @xmath410 . the proof of the lower bound ( [ eq : objectif : minimax ] ) thus requires nontrivial adaptations . it relies on some intermediate lemmas . [ lemma : birge ] we write @xmath411 for the kullback leibler divergence between @xmath331 and @xmath412 . for any functions @xmath413 from @xmath28 $ ] to @xmath90 such that @xmath414 and any @xmath415 we have @xmath416 \\ & & \qquad \geq{r^2\over4 } \biggl({1\over2{\mathrm{e}}+1}-\max _ { j=0,\ldots , m}{\mathbb{p}}_{\theta _ { j } } \bigl(\min_{i : i\neq j}d_{x , t } ( \theta_{i},\theta_{j})\leq r \bigr ) \biggr ) , \ ] ] where we denote , for any two functions @xmath417 from @xmath166 $ ] to @xmath90 , @xmath418 the proof is postponed to section [ section : proof_lemma_birge ] in . we next construct certain functions @xmath419 fulfilling ( [ eq : kullback - condition ] ) and well spread in terms of the pseudo - distance @xmath420 . consider the infinitely differentiable kernel @xmath421 defined by @xmath422 given any @xmath423 , vershamov gilbert s lemma [ @xcite ( @xcite ) , lemma 2.9 ] ensures the existence of @xmath424 points @xmath425 in the hypercube @xmath426 such that @xmath427\\[-8pt ] \eqntext{\mbox{for all } j\neq i.}\end{aligned}\ ] ] we then define @xmath413 by setting , for all @xmath428 , @xmath429 where @xmath430 since @xmath431 out of @xmath432 , we observe that @xmath433 and @xmath434,\ ] ] where @xmath435 denotes the fractional part of @xmath436 . thus , we have @xmath437 } \bigl{\vert}\theta_{j}(x)\bigr{\vert}\leq{r_0{\mathrm{e}}^{-1}\over m^{\beta}}\leq \delta<1 .\ ] ] we first check that the definition of @xmath438 ensures that the @xmath439 s are in the expected set of parameters . [ lemma : admissible - part1 ] for all @xmath440 , we have @xmath441 . the proof can be found in section [ section : proof_lemma_admissible - part1 ] of . next , we provide a bound to check the required condition ( [ eq : kullback - condition ] ) on the chosen @xmath442 s . [ lemma : admissible - part2 ] for all @xmath443 , we have @xmath444 where @xmath445 is the constant appearing in . we prove it in section [ section : proof_lemma_admissible - part2 ] of . finally , we need a control on the distances @xmath446 . [ lemma : min_distance ] for any @xmath447 , there exists a constant @xmath448 depending only on @xmath449 and the density @xmath244 of @xmath234 such that for all @xmath450 , @xmath451 and @xmath440 , @xmath452 the proof is postponed to section [ section : proof_lemma_min_distance ] of . we can now complete the proof of theorem [ theorem : lower_bound_quadratic_risk ] . proof of theorem [ theorem : lower_bound_quadratic_risk ] recall that @xmath453 in ( [ eq : def - thetaj ] ) are some parameters only depending on @xmath146 and @xmath224 and a certain integer @xmath423 and that , whatever the value of @xmath454 , lemma [ lemma : admissible - part1 ] insures that @xmath453 belongs to @xmath455 . hence , it is now sufficient to show that ( [ eq : objectif : minimax ] ) holds for a correct choice of @xmath454 , relying on lemmas [ lemma : birge ] , [ lemma : admissible - part2 ] and [ lemma : min_distance ] . let us set @xmath456 where @xmath457 is a constant to be chosen . then @xmath458 and , by lemma [ lemma : admissible - part2 ] , we can choose @xmath457 only depending on @xmath146 , @xmath219 , @xmath445 and @xmath224 so that condition ( [ eq : kullback - condition ] ) of lemma [ lemma : birge ] is met . we thus get that , for any @xmath415 , @xmath459 \\ & & \qquad { } \geq{r^2\over4 } \biggl({1\over2{\mathrm{e}}+1}-\max _ { j=0,\ldots , m}{\mathbb{p}}_{\theta _ { j } } \bigl(\min_{i : i\neq j}d_{x , t } ( \theta_{i},\theta_{j})\leq r \bigr ) \biggr ) . \ ] ] applying lemma [ lemma : min_distance ] with @xmath460 and the previous bound with @xmath461 , we get , as soon as @xmath451 , @xmath462\geq{r^2\over4 } \biggl({1\over4{\mathrm{e}}+2}- \frac{2 r_0{\mathrm{e}}^{-1}}{a ( 1-\delta)m^\beta } \biggr ) . \ ] ] the proof is concluded by observing that , as a consequence of ( [ eq : m ] ) , we can choose a constant @xmath221 only depending on @xmath146 , @xmath219 , @xmath445 and @xmath224 such that @xmath213 implies that @xmath451 and that the term between parentheses is bounded by @xmath463 from below . in this section , we test the proposed aggregation methods on data simulated according to a tvar process with @xmath464 . the choice of a smooth parameter function @xmath465 within @xmath466 for some @xmath467 is done by first picking randomly some smoothly time varying partial autocorrelation functions up to the order @xmath275 that are bounded between @xmath468 and @xmath77 and then by relying on the levinson durbin algorithm . we show the three components of the obtained @xmath469 on @xmath470 $ ] in the top parts of figure [ figure : thetas_and_tvar ] . realizations of the tvar process are then obtained from an innovation sequence @xmath37 of i.i.d . centered gaussian process with unit variance as in definition [ definition : tvar ] by sampling @xmath227 at a given rate @xmath54 . figure [ figure : thetas_and_tvar ] displays one realization of such a tvar process for @xmath471 . , @xmath472 and @xmath473 on the interval @xmath28 $ ] . the last plot displays @xmath471 samples of the corresponding tvar process with gaussian innovations . ] the nlms algorithm [ see @xcite , algorithm 1 ] studied in @xcite provides an online estimator of @xmath227 depending on a gradient step size @xmath474 . for any @xmath144 $ ] , choosing @xmath475 yields a @xmath260-minimax - rate online @xmath82-lipschitz predictor as explained in @xcite , section b.1 . hence , proceeding as in lemma [ lemma : minimax_experts ] to define @xmath58 and @xmath303 , @xmath68 , with @xmath476 , we obtain a finite set of nlms predictors corresponding to gradient step sizes @xmath477 . this set of predictors is aggregated in two possible ways according to the online algorithm [ algo : strat12 ] with the specifications on @xmath79 and @xmath58 given in theorem [ theorem : upper_bound_aggregated_forecaster_adaptative ] . the overall running time of @xmath26 iterates of the algorithm leading to the aggregated predictors from the data @xmath33 is then @xmath478 . since the algorithm is recursive , the corresponding required storage capacity is @xmath479 . we evaluate the obtained nlms predictors and their aggregated predictors by running @xmath480 simulations based on equally distributed realizations of the above gaussian tvar process in the case @xmath471 which yields @xmath481 . in figure [ figure : beta_0=5_t=1024 ] , we compare the averaged downward shifted empirical losses defined for any predictor @xmath266 by @xmath482 this empirical averaged loss mimics the risk considered in ( [ eq : max - risk ] ) . of the nlms predictors @xmath483 . the ones on the right of the same line are those associated to the aggregated predictors using the weights ( [ definition : alphas_gradient_quadratic_loss ] ) and ( [ definition : alphas_quadratic_loss ] ) . ] we observe that the best nlms predictor is the third one while the aggregated predictor of strategy 1 enjoys a smaller loss and that of strategy 2 a slightly larger one . this is in accordance with theorem [ thm : oracle - bounds](i ) and ( iii ) where it is shown that the aggregated predictor of the first strategy may outperform the best predictor as it nearly achieves the loss of the best possible convex combination of the original predictors while the aggregated predictor of the second strategy nearly achieves the loss of the best original predictor . [ section : useful_bounds_tvar_postponed_proofs ] the following lemma provides a uniform bound on the norm of a product of matrices sampled from a continuous function defined on an interval @xmath155 and valued in a set of @xmath484 matrices with bounded spectral radius and norm . [ lem : product ] let @xmath167 and @xmath155 an interval of @xmath90 . let @xmath448 be a function defined on @xmath155 taking values in the set of @xmath485 matrices with eigenvalues moduli at most equal to @xmath224 . let @xmath486 be any matrix norm . denote by @xmath487 the corresponding uniform norm of @xmath448 , @xmath488 and , for any @xmath489 , @xmath490 the modulus of continuity of @xmath448 over @xmath155 , @xmath491 let @xmath492 and assume that @xmath493 . then there exist some positive constants @xmath449 , @xmath494 and @xmath421 only depending on @xmath487 , @xmath224 and @xmath495 such that , for any @xmath496 fulfilling @xmath497 , we have , for all @xmath498 in @xmath155 and all integer @xmath499 , @xmath500 denote by @xmath501 the product of matrices appearing in the left - hand side of ( [ eq : product - univ - bound ] ) . the proof goes along the same lines as [ moulines , priouret and roueff ( @xcite ) , proposition 13 ] but we use the modulus of continuity instead of the @xmath146-lipschitz norm to control the local oscillation of matrices . for @xmath502 and any square matrices @xmath503 , adopting the convention @xmath504 if @xmath505 and @xmath506 is the identity matrix if @xmath507 , we have @xmath508\\[-8pt]\nonumber & = & a_{1}^{\ell_{1}}+\sum_{k=1}^{\ell_{1}-1}a_{1}^{\ell_{1}-k } ( a_{\ell_{1}-k+1}-a_{1 } ) \prod_{i=\ell_{1}-k+2}^{\ell _ { 1}}a_{i } . \nonumber\end{aligned}\ ] ] given a positive integer @xmath494 , using the euclidean division of @xmath509 by @xmath494 , @xmath510 , we decompose the product @xmath501 as @xmath511\\[-8pt]\nonumber & & { } \times\prod_{k=1}^{r}a \biggl(t- \frac{(q\ell+k-1)(t - s)}{p } \biggr ) .\end{aligned}\ ] ] using ( [ equation : decomposition_product ] ) , we have for any @xmath512 , @xmath513 and @xmath514 , @xmath515\\[-8pt]\nonumber & & \qquad \leq\biggl{\vert}\biggl(a \biggl(t-\frac{j\ell(t - s)}{p } \biggr ) \biggr)^{\ell _ { 1}}\biggr{\vert}+ ( \ell_{1}-1 ) \bigl(a^ { * } \bigr)^{\ell _ { 1}-1}\omega_{h } ( a;i ) .\end{aligned}\ ] ] take an arbitrary @xmath516 ( say the middle point ) . the eigenvalues of @xmath448 are at most @xmath224 on @xmath155 and @xmath493 . applying [ @xcite ( @xcite ) , lemma 12 ] we obtain that there is a constant @xmath517 only depending on @xmath224 , @xmath518 and @xmath519 such that @xmath520 . from ( [ equation : decomposition_pi_t_s ] ) and ( [ equation : bound_block_size_l_product ] ) , we derive the following inequality : @xmath521 where @xmath522 . we can choose a positive integer @xmath494 and a positive number @xmath523 only depending on @xmath524 , @xmath495 and @xmath525 such that @xmath526 in the following , we set @xmath527 . the previous bound gives that for any @xmath496 such that @xmath497 and @xmath528 , @xmath529 hence , we have the result . we can now provide a proof of proposition [ proposition : stationary_tvar ] . equation ( [ equation : definition_tvar ] ) can be more compactly written as @xmath530 for all @xmath531 , iterating this recursive equation @xmath154 times , we have @xmath532\mathbf x_{t - k-1,t } \nonumber\\[-8pt]\\[-8pt]\nonumber & & { } + \sum_{j=0}^{k } \sigma \biggl ( { \frac{t - j}{t } } \biggr)\mathbf e'_{1 } \biggl[\prod_{i=1}^{j}a \biggl ( { \frac { t - i}{t } } \biggr ) \biggr]\mathbf e_{1}\xi_{t - j } , \end{aligned}\ ] ] where @xmath533'$ ] and @xmath534 . \ ] ] note that the eigenvalues of @xmath535 are the reciprocals of the roots of the local time varying autoregressive polynomial @xmath536 , and thus are at most @xmath537 . moreover , since @xmath227 is bounded by a constant only depending on @xmath275 and is uniformly continuous on @xmath158 $ ] , so is @xmath448 as a function defined on @xmath155 and we can find @xmath496 such that @xmath538 for any positive @xmath449 . if @xmath539 , this @xmath540 can be chosen depending only on @xmath541 and @xmath219 ( and also on the matrix norm @xmath486 ) . consider @xmath216 . lemma [ lem : product ] gives that there exist some positive constant @xmath449 , @xmath494 and @xmath421 only depending on @xmath487 , @xmath224 and @xmath495 such that , for any @xmath496 fulfilling @xmath497 , we have , for all @xmath54 , @xmath200 and @xmath542 so that @xmath543 , @xmath544 we here consider the @xmath545 operator norm which is the maximum absolute row sum of the matrix , in which case @xmath546 . hence , by ( [ equation : tvar_initial_condition_noise ] ) we obtain that @xmath547\\[-8pt ] \eqntext{1\leq t\leq t,}\end{aligned}\ ] ] with , provided that @xmath548 , for all @xmath200 , @xmath549 and @xmath550 , @xmath551 the result follows . denote @xmath552 , so that @xmath553 . the function @xmath554 is concave on @xmath555 $ ] , so introducing @xmath556 and then using jensen s inequality , we get @xmath557 it only remains to show that @xmath558 , with the assumption that @xmath331 has support on @xmath332 $ ] . this is verified if @xmath559 , so we now assume @xmath560 . we write @xmath561 we note that @xmath562 and @xmath563 . we deduce that the product @xmath564 either take nonpositive values or positive values of the form @xmath565 now , for @xmath566 $ ] with @xmath560 , in the first case , we have @xmath567 since @xmath568 , and , in the second case , @xmath569 . the lemma follows . we define @xmath570 as the ( random ) smallest index which minimizes @xmath571 over @xmath572 so that @xmath573 . note that @xmath574 . hence , @xmath575 \\ & & \qquad \geq { 1\over4 } \max_{j=0,\ldots , m}{\mathbb{e}}_{\theta_{j } } \bigl[d_{x , t}^2(\theta_{\hat{{\mathrm{j } } } } , \theta_{j } ) \bigr ] \\ & & \qquad \geq{r^2\over4}\max_{j=0,\ldots , m}{\mathbb{p}}_{\theta_{j } } \bigl ( \ { \hat{{\mathrm{j}}}\neq j \ } \cap \bigl\{\min_{i : i\neq j}d_{x , t } ( \theta _ { i},\theta_{j } ) > r \bigr\ } \bigr ) \\ & & \qquad \geq{r^2\over4 } \bigl(1-\min_{j=0,\ldots , m } { \mathbb{p}}_{\theta_{j } } ( \hat{{\mathrm{j}}}= j ) -\max_{j=0,\ldots , m}{\mathbb{p}}_{\theta_{j } } \bigl(\min_{i : i\neq j}d_{x , t}(\theta_{i } , \theta_{j})\leq r \bigr ) \bigr).\end{aligned}\ ] ] birg s lemma [ @xcite ( @xcite ) , corollary 2.18 ] implies that @xmath576 so the lemma follows from condition ( [ eq : kullback - condition ] ) . by ( [ eq : thetastar ] ) , we have @xmath577 for all @xmath578 . decompose the hlder - exponent @xmath152 where @xmath154 is an integer and @xmath151 $ ] . differentiating ( [ eq : def - thetaj ] ) @xmath154 times , we have , as in ( [ eq : thetaj ] ) , @xmath579 . \ ] ] thus , for @xmath580 in the same interval @xmath581 $ ] with @xmath582 , we get @xmath583 the same inequality then follows with @xmath584 replaced by @xmath585 for @xmath580 in two such consecutive intervals . now , if @xmath580 are separated by at least one such interval , we have @xmath586 and , using that @xmath421 has support in @xmath432 , we have that @xmath587 is bounded by @xmath588 . we thus get in this case that @xmath589 the last two displays and ( [ eq : l0-def ] ) then yields @xmath590 . let @xmath591 . recall that @xmath592 by ( [ eq : vershamov ] ) and ( [ eq : def - thetaj ] ) . by ( [ eq : thetaj - zero ] ) and ( [ eq : lb - tvar - eq ] ) , we have that @xmath197 has the same distribution under @xmath593 and @xmath594 [ which is the distribution of @xmath595 . hence , the likelihood ratio @xmath596 of @xmath597 is given by the corresponding conditional likelihood ratio of @xmath598 given @xmath197 . hence , under ( i-3 ) , we obtain that @xmath599 where , in the second equality , we used again that @xmath592 . now , under @xmath600 , we have @xmath601 . thus , we get @xmath602 \\ & = & \sum_{t=1}^t { \mathbb{e}}_{\theta_{j } } \biggl[\log\frac{f(\xi _ { t})}{f(\theta_{j}((t-1)/t)x_{t-1,t}+\xi_{t } ) } \biggr ] \\ & = & \sum_{t=1}^t { \mathbb{e}}_{\theta_{j}}\int \log \biggl(\frac{f(u)}{f(\theta_{j}((t-1)/t)x_{t-1,t}+u ) } \biggr ) f(u ) \,{\mathrm{d}}u.\end{aligned}\ ] ] using assumption ( i-3 ) yields @xmath603 \leq\kappa\theta^{*2}\sum _ { t=1}^{t}{\mathbb{e}}_{\theta_{j } } \bigl[x_{t-1,t}^{2 } \bigr].\end{aligned}\ ] ] the series representation ( [ eq : inf - series - representation - d1 ] ) , the fact that @xmath234 is centered with unit variance and ( [ eq : thetastar ] ) imply that for all @xmath604 @xmath605 \leq \bigl(1- \theta ^{*2 } \bigr)^{-1 } . \ ] ] using this bound and ( [ eq : thetastar ] ) in ( [ eq : etap1:admissible ] ) , we obtain @xmath606 the proof of lemma [ lemma : admissible - part2 ] now follows by applying the first bound in ( [ eq : vershamov ] ) . the proof relies on an upper bound of @xmath607 involving the noise @xmath608 . by the expression of @xmath442 in ( [ eq : thetaj ] ) , we have @xmath609 where we denoted @xmath610 and @xmath611 . using ( [ eq : inf - series - representation - d1 ] ) and ( [ eq : thetastar ] ) , we have , for all @xmath612 , @xmath613 which implies @xmath614 inserting this bound in ( [ eq : d2xtdef - thetaj ] ) , we get @xmath615 where @xmath616 thus , with ( [ eq : bound - d2-remainder ] ) , the left - hand side of inequality ( [ eq : min - distance ] ) is upper bounded by @xmath617 using that @xmath234 is centered with unit variance and then ( [ eq : thetastar ] ) , we easily get that @xmath618\leq\frac{2{\mathrm{e}}^{-2}}t \sum_{t=0}^{t-1}\sum _ { j=1}^{\infty}\theta^{*j } \leq \frac{2{\mathrm{e}}^{-2}\theta^*}{1-\theta^ * } \leq\frac{2 r_0{\mathrm{e}}^{-3}}{(1-\delta)m^\beta } .\ ] ] hence , by markov s inequality , to conclude the proof , it now suffices to show that , for @xmath448 well chosen , @xmath619 for @xmath620 we define @xmath621 . we observe that the cardinality of @xmath622 is @xmath623 where the lower bound is a consequence of the assumption @xmath624 in the lemma . moreover , it is easy to check that we have @xmath625 for all index @xmath626 and that , for each @xmath627 , the set @xmath628 is included in the set @xmath629 ( so that , in particular , @xmath630 for @xmath631 ) . it follows that random variables @xmath632 are i.i.d . by the monotonicity of @xmath421 in @xmath633 and its symmetry , we have @xmath634 from ( [ eq : vershamov ] ) , for any @xmath635 there exist at least @xmath636 values of @xmath154 for which @xmath637 equals one in the above sum . hence , using the order statistics @xmath638 , we thus obtain that @xmath639 where we used @xmath640 for @xmath641 in the last inequality . let us denote by @xmath642 the cumulative distribution function of @xmath643 , which only depends on @xmath644 and on the distribution of @xmath645 . for @xmath646 , we have @xmath647 gathering the last two bounds , we get that @xmath648 recall that @xmath649 and note that @xmath650 admits a density , since @xmath234 does . by the strong law of large numbers , we further have that the random variable @xmath643 converges to @xmath77 almost surely when @xmath644 goes to infinity , so there exists @xmath651 depending only on the density of @xmath234 such that @xmath652 whatever the value of . therefore , there exists some @xmath653 , depending only on the distribution of @xmath234 , such that ( [ eq : bound - d2-remainderrt ] ) holds , which achieves the proof . we gratefully acknowledge the fruitful comments of the referees .

the connection between core collapse supernovae and neutron stars @xcite has had a solid observational footing for almost forty years , due largely to the discovery of young radio pulsars in supernova remnants ( snrs ) like vela @xcite and in the crab nebula @xcite . energetic young pulsars like these are strong radio and x - ray sources , and often power synchrotron nebulae called pulsar wind nebulae or pwne @xcite that are indirect markers of pulsars ( e.g. , * ? ? ? the idea that young neutron stars resemble the crab pulsar came to dominate the search for the products of supernovae ( e.g. , * ? ? ? * ) . recently , though , young neutron stars have been revealed in a wide variety of manifestations , from anomalous x - ray pulsars ( axps ) and soft @xmath2-ray repeaters ( sgrs ) , to nearby thermal and radio quiet neutron stars , to long period radio pulsars with high inferred magnetic fields . as exemplified by the identification of the central compact object ( cco ) in the snr @xcite , much of this diversity has come from x - ray observations . while this diversity is clearly demonstrated observationally , theory and simulation can not yet constrain the fundamental birth properties of neutron stars ( e.g. , * ? ? ? * ; * ? ? ? models still have difficulties achieving explosions , much less following the activity in the post - collapse object in any detail . * hereafter ) have attempted to address our lack of understanding of stellar death and neutron star cooling by defining a volume - limited ( @xmath3 kpc ) sample of supernova remnants ( snrs ) , examining the neutron stars that they contain , and outlining a survey designed to detect or significantly constrain neutron stars in the remaining remnants . the primary subsample discussed in is one where the snr diameter is @xmath4 , so that the _ chandra x - ray observatory _ can observe a significant fraction of the snr interior with its acis - i detector and hence cover the area where neutron stars would be with a reasonable range of velocities ( @xmath5 , where @xmath6 is the velocity perpendicular to the line of sight ) . also discuss two other subsamples of snrs : one with diameters @xmath7 for which _ xmm - newton _ is suitable ( and which we will present in a forthcoming paper ) , and one with @xmath8 . it is this subsample of the six largest snrs from that we consider here . the organization of the paper is as follows . in [ sec : snrs ] we give brief summaries of the six snrs discussed here . in [ sec : bsc ] we describe our identification of candidate x - ray sources in and around the snrs . in [ sec : cpt ] we detail the initial identification of optical / ir counterparts to the x - ray sources using available sky surveys : as discussed in , optical / ir observations are a powerful way to reject x - ray sources that are not neutron stars ( see also e.g. , @xcite ) . with the sky surveys we were able to identify most of the x - ray sources with high confidence : those for which we were not certain were selected for additional _ chandra _ observations and optical / ir observations ( [ sec : cxo ] ) . finally , we give our discussion and conclusions in [ sec : disc ] . all coordinates are j2000.0 . l l c c l c c & g65.2 + 5.7 & @xmath9 & 0.8 & optical velocity & 1.4 & 6.7 + & cygnus loop & @xmath10 & 0.44 & optical proper motion & 0.8 & 1.7 + & & 110 & 1.3 & nei fits & 3.5 & 29 + & hb9 & @xmath11 & @xmath12 & hi , optical velocity & 1 & @xmath13 + & monoceros & 220 & 1.2 & optical velocity & 0.8 & 13 + & lupus loop & 180 & 1.2 & nei fits & 0.5 & 12.0 + + we list the snrs for this paper , along with relevant parameters , in table [ tab : snrs ] . each snr has a distance determined from a more reliable method than the @xmath14-@xmath15 method ( e.g. , * ? ? ? * ; * ? ? ? * ) , but they are not all of the same quality . in the best cases , the distances are from kinematic observations of optical or radio lines . in the worst cases , the distances are from fitting shock models to the x - ray data ( e.g. , * ? ? ? these distances involve many uncertainties beyond the kinematic distances , including the assumption of a sedov - phase remnant , the state of equilibration in the system , the non - sphericity of the explosion , and the unknown total explosion energy . for kinematic distances , the uncertainties are probably @xmath16% , but for distances from x - ray fitting they could exceed 50% @xcite . the ages tend to be derived from x - ray fits for all sources , although having an independently determined distance for some sources makes for better constraints . below we discuss each snr in more detail . snr g65.3 + 5.7 ( also known as g65.2 + 5.7 ) was identified as a snr by @xcite by its filamentary line emission . it has major axes of @xmath17 . according to @xcite , the age is 2025 kyr , and the distance is @xmath181.0 kpc ( a kinematic distance derived from the velocity of optical emission lines ; also see @xcite ) . @xcite show data from pointed _ rosat_observations @xcite but do not discuss point sources ; the _ rosat _ data detect emission from much of the interior at @xmath19 , with @xmath20 pixels . @xcite do mention an extended ( @xmath21 radius ) soft source at @xmath22 @xmath23 but come to no conclusion as to its identity . the inner @xmath24 were searched for radio pulsars by @xcite at 1410 mhz down to a limit of 0.1 mjy , but given the size of snr g65.3 + 5.7 a transverse velocity of only @xmath25 would have moved a neutron star outside the search region so the lack of detection was not very constraining , the 0.58-s radio pulsar lies @xmath26 from the center of the remnant . however , with no estimate of the spin - down rate ( and hence no spin - down age ) , the period seems rather large to be associated with an snr ( it would require an unusual but not unheard of magnetic field of @xmath27 g ) , and no definite claim of an association can be made @xcite . snr g74.0@xmath288.5 , also known as the cygnus loop , is a @xmath29 radio and x - ray shell . the distance , estimated from measurements of the shock velocity and proper motion is 0.44 kpc @xcite and the age is 8 kyr @xcite . @xcite reported @xmath30 in the interior over the @xmath31 field of the sis for their _ asca _ observation , and used these data to conclude that snr g74.0@xmath288.5 was likely the result of a type ii supernova on the basis of elemental abundances . @xcite searched snr g74.0@xmath288.5 for promising x - ray point sources that might be compact objects , and identified two , one of which they later concluded was an active galactic nucleus ( agn ) on the basis of its long - term variability , x - ray spectrum , and radio counterpart @xcite , and the other of which they conclude may be a neutron star . the inner @xmath32 were searched for radio pulsars by @xcite down to a 430 mhz flux limit of 0.3 mjy , but given the size of snr g74.0@xmath288.5 the lack of detection was not very constraining ( @xmath33 ) . the inner @xmath34 were also searched by @xcite for pulsars down to a 400-mhz flux of 3 mjy . assuming an average radio spectrum for radio pulsars of @xmath35 , this translates to a 1400-mhz luminosity limit of @xmath36 , which is considerably fainter than the very low luminosity ( @xmath37 ; * ? ? ? snr g156.2 + 5.7 was discovered in the _ rosat _ all - sky survey by @xcite . it has a faint @xmath38 shell in both x - rays and radio , and non - equilibrium fits to the x - ray data place it at a distance of @xmath39 kpc with an age of 15 kyr @xcite . @xcite searched snr g74.0@xmath288.5 for radio pulsars , tiling seven pointings of the 76-m lovell telescope at jodrell bank , each of which covered @xmath40 . the search did not find any pulsars , down to a flux limit of @xmath41 mjy at 606 mhz , or a 1400-mhz luminosity limit of @xmath42 . snr g160.9 + 2.6 , also known as hb 9 , is a @xmath43 radio shell with bright x - rays in the interior . @xcite use x - ray fitting to estimate a distance of 1.5 kpc and an age of 820 kyr . this distance is consistent with the upper limit of 4 kpc derived from other measurements @xcite . the inner @xmath34 were also searched by @xcite for pulsars down to a 610-mhz flux of 15 mjy . @xcite discovered an old radio pulsar ( ) in the interior of the snr , although the association between the pulsar and the snr is generally considered to be false ( e.g. , * ? ? ? * ) based on the large spin - down age and low spin - down energy loss rate for the pulsar ( @xmath44 yr and @xmath45 , respectively ) . snr g205.5 + 0.5 , also known as the monoceros nebula , is a @xmath46 radio shell . the systemic velocity of optical line emission puts the snr at a distance of 0.8 kpc @xcite , although distances up to 1.6 kpc are preferred by low - frequency radio data that show the snr within the mon ob2 association @xcite . the age is likely @xmath47 kyr , as inferred from fits to x - ray data @xcite . snr g330.0 + 15.0 , the lupus loop , is a low surface brightness radio shell approximately @xmath48 in diameter . non - equilibrium fits to the x - ray data and comparison with the column density of the nearby remnant of sn 1006 suggest a distance of 1.01.2 kpc and an age of 50 kyr @xcite . _ chandra _ or _ xmm - newton _ imaging of the entire fields of the large - diameter snrs listed in table [ tab : snrs ] is impractical because of their sizes . their proximities ( @xmath49 kpc ) , though , means that we do not need the high sensitivities of _ chandra _ or _ xmm - newton _ to achieve the same luminosity limit as in . we therefore used the _ rosat _ all - sky survey bright source catalog ( hereafter bsc ; @xcite ) for our source selection . this was a survey of the entire sky with the position - sensitive proportional counter ( pspc ) aboard _ rosat_. the positional accuracy of the pspc does not approach that of _ chandra _ or even _ xmm - newton _ ( typical uncertainties are @xmath50 ) , and the observations are not as deep as the pointed _ chandra _ and _ xmm - newton_observations used for the other snrs . nonetheless , the bsc is useful . as seen in table [ tab : snrs ] , its limit of @xmath51 in the pspc is actually of roughly comparable depth to our _ chandra _ observations in @xmath52 when the smaller distances and column densities of the snrs in this paper are taken into account . while the x - ray positions do not in all cases allow unambiguous optical identifications , the relative brightness and softness of the x - ray sources compared to those in means that very often stars from the digitized sky survey ( dss ) or 2mass @xcite can be identified as counterparts . we selected the bsc sources within twice the nominal radii ( for elliptical sources , we took the semi - major axes ) of the snrs in table [ tab : snrs ] ( as defined by their positions and sizes given by @xcite ) that had @xmath53 and were listed as unextended ( a value of 0 in the ` ext ` column of the bsc catalog ) . searching outside the remnants allowed us to find neutron stars that have overtaken the snr shocks not an uncommon occurrence @xcite in snrs of the ages considered here ( 1030 kyr ) . this gave us all of the x - ray sources listed in table [ tab : srcs ] . for the sake of comparison between sources , we plot the distribution of hardness ratios in figure [ fig : hard ] . in our analysis , we rejected those bsc sources that were identified as extended . this was for several reasons : we eliminated peaks in diffuse background emission that may have been identified as discrete sources , and we eliminated large extended objects such as galaxy clusters . practically , point sources offer much better astrometry and are better suited to counterpart identification . however , in some sense our selection was less than ideal . we would have eliminated any bright pwne , although these might have been identified by previous searches . also , source confusion makes our resulting luminosity limits less constraining than they might otherwise be , as two nearby point - sources could have been identified as a single extended source and hence been rejected . given the relatively low space density of bsc sources ( fig . [ fig : dens ] ) this should not be a major effect , but it should still be noted . in contrast , our _ chandra _ observations do not suffer from any confusion limitations . one might ask if the diffuse emission from the snrs themselves will limit the depth of the bsc in the snr interiors . we have found in general that this is not the case . figure [ fig : dens ] shows the density of bsc sources ( both point - like and of all sizes ) within different radii from the snr centers . while the inner reaches of the snrs have few sources and therefore poor statistics , in no case is there a statistically significant deficit of point sources inside the snr . there might be a slight deficit inside snr g156.2 + 5.7 or snr g160.9 + 2.6 , but these are also the smallest of the snrs and therefore have the fewest total sources . similarly , in figure [ fig : bg ] we show the average background count - rates determined when extracting the sources plotted in figure [ fig : dens ] , with the same binning . two of the snrs ( g65.3 + 5.7 and g74.0@xmath288.5 ) do show background increases in the interiors , two do not ( g205.5 + 0.5 and g330.0 + 15.0 ) , and two are uncertain due to few counts ( g156.2 + 5.7 and g160.9 + 2.6 ) , but even an increase of a factor of three above the mean background rate ( @xmath54 ) would give only @xmath55 within the 90% confidence radius of a pspc source ( for 0.3 kev ) which is a factor of 10 less than the minimum source count - rate for the bsc . therefore the diffuse snr emission should not have significantly affected the bsc source detection , and it is unlikely that there were any point sources that were missed . besides the bsc , we took advantage of x - ray observations in the literature to identify additional sources for _ chandra _ followup . these were : for snr g74.0@xmath288.5 , ax j2049.6 + 2939 ( from _ asca _ ; * ? ? ? * ; * ? ? ? * ) ; and for snr g205.5 + 0.5 : _ einstein _ sources 1 , 3 , and 6 from @xcite , known as 1e 0627.4 + 0537 , 1e 0630.9 + 0611 , and 1e 0636.8 + 0517 , respectively . once we had assembled the list of x - ray sources , we then examined the publicly available surveys ( dss , 2mass , and nrao vla sky survey limit of the nvss . ] [ nvss ; @xcite ] ) , as well as examination of simbad and the relevant literature . with these sources of information , we were able to identify likely counterparts to 41 of the 50 sources in table [ tab : srcs ] . we list the relevant data ( x - ray and optical ) of the identifications in table [ tab : srcs ] , with a summary of all identifications and additional notes in table [ tab : id ] . the separations between the nominal x - ray and optical positions were consistent with the predicted x - ray position uncertainties ( fig . [ fig : offset ] ) . there are a number of cases where there were multiple stars within the x - ray error circles , some of which were known to be physically associated with each other ( as noted in simbad ) . in these cases we list multiple possible counterparts in table [ tab : srcs ] . the true source of the x - ray emission may be any one of the stars , or may in fact come from the interactions between them . while the identifications were made only on the basis of positional coincidence with bright stars , in many cases we can be additionally confident . this is because the stars are so bright ( @xmath56 mag ) that the chances of a false association are negligible or the stars are of types known to have x - ray emission ( e.g. , t tauri stars ) . to aid in the evaluation of our identifications , we plot the cumulative number density of 2mass sources for each snr in figure [ fig : stars ] . for a typical position uncertainty of @xmath50 , we expect that all identifications with 2mass sources brighter than @xmath57 mag will be real ( chance probability @xmath0% ) , and even for sources brighter than @xmath58 mag the identifications will be probable ( chance rates @xmath59% ) . we note , though , that while 1rxs j205812.8 + 292037 is consistent with having an association with one of the identified 2mass sources , the association is not secure . chandra _ followup observation likely would have been definitive as it was for the majority of the ambiguous sources discussed in section [ sec : cxo ] , but due to an oversight on our part this source was not selected for followup _ chandra _ observations . we therefore discuss 1rxs j205812.8 + 292037 in extra detail in section [ sec : rxsd ] . given the uncertainties in spectrum and foreground column density , virtually all of the x - ray sources are consistent with being stars ( as opposed to active galaxies ; fig . [ fig : fxk ] ) . we show 2mass images of the x - ray sources with optical counterparts indicated in figures [ fig : opt1][fig : opt12 ] . the nine bsc sources that had no obvious optical counterparts ( excluding 1rxs j205812.8 + 292037 , as mentioned above ) , plus the four sources from [ sec : addl ] , were selected for _ chandra _ followup observations . here , as in , we selected the exposure times ( 36 ksec ) based on the known column densities to the snrs ( tab . [ tab : snrs ] ) and a blackbody spectrum with @xmath60 kev . the positions are known to sufficient accuracy to allow use of the acis - s3 ccd @xcite . depending on the source brightnesses , however , we were concerned about photon pileup for some of the sources , so we used the @xmath61- or @xmath62-subarray modes ( which also provides improved timing information ) , depending on the positional uncertainties . a log of the observations is in table [ tab : cxo ] . in most of the cases , the _ chandra _ observations revealed nothing extraordinary . in the case of the bsc sources , the _ chandra _ data typically showed that the bsc position was significantly off and/or the counterpart was faint ( figs . [ fig : cxoopt1 ] and [ fig : cxoopt2 ] ) . the additional four sources from the literature were all coincident with stellar sources , once we had _ chandra _ positions . of the 13 sources with _ chandra _ followup , nine had point - like _ chandra _ sources with obvious ir counterparts ( [ sec : notescxo ] ) . of the other four sources : two show extended x - ray emission with _ chandra _ , one has no obvious 2mass counterpart , and one source was not detected in the _ chandra _ observation . we discuss these sources in more detail in section [ sec : remain ] . in the cases where _ chandra _ point - sources were detected , the counterpart identifications are essentially secure . this is due to the very small positional uncertainty of the _ chandra _ positions ( @xmath63 ) coupled with the brightnesses of the counterparts ( see fig . [ fig : stars ] and ) . figures [ fig : cxoopt1 ] and [ fig : cxoopt2 ] contain images of those sources with counterpart identifications . with these identifications we can eliminate these sources as candidate compact objects using the x - ray - to - optical flux ratio ( see fig . [ fig : fxk ] and ) . here we comment on all of the sources observed with _ chandra_. 1rxs j193458.1 + 335301 : : the _ chandra _ source is extended . [ sec : rxsa ] . 1rxs j193228.6 + 345318 : : the _ chandra _ source is point - like , and is at @xmath64 , @xmath65 ( @xmath66 away from the bsc position ) . it is coincident with the @xmath67 mag source 2mass j19322722 + 3453148 , the @xmath68 mag source usno 1248 - 0333432 , and with the nvss source identified in table [ tab : id ] . it was identified as a flare star by @xcite after our initial source selection . 1rxs j205042.9 + 284643 : : there is no _ chandra _ source in the followup observation . [ sec : rxsb ] . ax j2049.6 + 2939 : : this _ asca _ source was identified as a possible neutron star by @xcite . the _ chandra _ source is point - like , and is at @xmath69 , @xmath70 . it is coincident with the @xmath71 mag source 2mass j20493540 + 2938509 ( also usno 1196@xmath280518650 ) , which is presumably the @xmath72 mag g star discussed by @xcite . our observed count - rate for this source ( @xmath73 in the 0.32.0 kev band ) is roughly comparable with that predicted from the latest _ spectroscopy , although as noted in @xcite the source appears to be variable . given the variability of this source and the extremely tight coincidence with a g star ( in this region , there are @xmath74 stars with @xmath75 mag in 2mass , and to find one @xmath76 away from the x - ray source has a chance rate of @xmath77 ; see fig . [ fig : stars ] ) , the x - ray emission is very likely from an active star . @xcite had dismissed this possibility because of the high x - ray intensity , but it is in fact consistent with the majority of the stars that we detect here ( fig . [ fig : fxk ] ; the x - ray flux is @xmath78 ) . 1rxs j045707.4 + 452751 : : the _ chandra _ source is point - like , and is at @xmath79 , @xmath80 ( @xmath50 away from the bsc position ) . it is coincident with the @xmath81 mag source 2mass j04570832 + 4527499 . 1rxs j050339.8 + 451715 : : the _ chandra _ source is point - like , and is at @xmath82 , @xmath83 ( @xmath84 away from the bsc position ) . it is coincident with the @xmath85 mag source 2mass j05033958 + 4516594 and with the nvss source identified in table [ tab : id ] . there is no usno counterpart . 1rxs j062740.3 + 073103 : : the _ chandra _ source is point - like , and is at @xmath86 , @xmath87 ( @xmath88 from the bsc position ) . it is coincident with the @xmath89 mag source + 2mass j06274012 + 0731006 . 1e 0627.4 + 0537 : : the _ chandra _ source is point - like , and is at @xmath90 , @xmath91 . it is coincident with the @xmath71 mag source 2mass j06300529 + 0545407 . 1e 0630.9 + 0611 : : the _ chandra _ source is point - like , and is at @xmath92 , @xmath93 . it is coincident with the @xmath94 mag source 2mass j06333322 + 0608396 . 1e 0636.8 + 0517 : : the _ chandra _ source is point - like , and is at @xmath95 , @xmath96 . it is coincident with the @xmath97 mag source 2mass j06392566 + 0514301 . 1rxs j150818.8@xmath28401730 : : the _ chandra _ source is point - like , and is at @xmath98 , @xmath99 ( @xmath100 away from the bsc position ) . it is coincident with the @xmath101 mag source 2mass j15081819@xmath284017261 . 1rxs j150139.6@xmath28403815 : : the _ chandra _ source is extended . see [ sec : rxsc ] . 1rxs j151942.8@xmath28375255 : : the _ chandra _ source is point - like , and is at @xmath102 , @xmath103 ( @xmath88 away from the bsc position ) . there is no 2mass counterpart , but we do identify a counterpart on our magellan magic optical and panic infrared observations ( [ sec : oir ] and fig . [ fig : mag ] ) . the source has @xmath104 mag and @xmath105 mag . -band images of the source in snr g156.2 + 5.7 from table [ tab : srcs ] . the images are @xmath106 , with north up and east to the left . the x - ray position uncertainties are indicated by the circles , and the proposed optical counterparts are shown by the crosses.[fig : opt5],scaledwidth=50.0% ] -band images of the sources in snr g330.0 + 15.0 from table [ tab : srcs ] ( cont . ) . the images are @xmath106 , with north up and east to the left . the x - ray position uncertainties are indicated by the circles , and the proposed optical counterparts are shown by the crosses.[fig : opt12],scaledwidth=50.0% ] we obtained optical and infrared followup observations of three of the four sources where the _ chandra _ followup did not immediately identify a counterpart , namely 1rxs j193458.1 + 335301 , 1rxs j151942.8@xmath28375255 , and 1rxs j150139.6@xmath28403815 . for 1rxs j193458.1 + 335301 , we got a 300-s @xmath107-band exposure with the large format camera ( lfc ) on the palomar 200-inch telescope and 1200-s exposures in h@xmath108 and a narrowband filter located away from major emission lines with the ccd camera ( p60ccd ) on the palomar 60-inch telescope . for 1rxs j151942.8@xmath28375255 , we obtained a 930-s @xmath109-band exposure with the raymond and beverly sackler magellan instant camera ( magic ) on the 6.5-m clay ( magellan ii ) telescope , and a 200-s @xmath110-band exposure with persson s auxiliary nasmyth infrared camera ( panic ; @xcite ) , also on the 6.5-m clay ( magellan ii ) telescope . for 1rxs j150139.6@xmath28403815 , we got a 160-s @xmath111-band exposure and a 60-s @xmath109-band exposure with magic . the log of the observations is given in table [ tab : optobs ] . reduction and calibration followed standard procedures . the magellan data were taken during the same observing runs as data presented in , and details can be found there . for the reduction of the remaining data , we used standard ` iraf ` routines to subtract the bias , flat - field , and then combine separate exposures . significant focal - plane distortion prevented simple addition of the lfc data , so we used the ` iraf ` ` mscred ` package to flatten each image with custom distortion maps prior to addition . we then performed absolute astrometry , solving for plate - scale , rotation , and central position relative to stars the 2mass catalog , and getting residuals in each coordinate of @xmath112 ( 2100 stars ) and @xmath113 ( 4000 stars ) for p60ccd and lfc , respectively . we find that after investigating 50 bsc sources plus four sources from the literature and obtaining _ chandra _ followup of 13 of these sources , there remain four x - ray sources that do not have very likely optical counterparts and are therefore worthy of extended discussion . as noted in [ sec : cpt ] and [ sec : notescxo ] , these sources are : 1rxs j193458.1 + 335301 , 1rxs j205042.9 + 284643 , 1rxs j205812.8 + 292037 , and 1rxs j150139.6@xmath28403815 . the first and fourth have extended x - ray emission , while the second has no apparent _ chandra _ counterpart . the third has a probable but not definite association with optical / ir sources . we now discuss all of these sources in more detail . the bsc lists 1rxs j193458.1 + 335301 as having @xmath114 in the pspc , with hardness ratios of @xmath115 and @xmath116 ( see @xcite for definitions of bands and hardness ratios ) . the corresponding _ chandra _ source is clearly extended , as shown in figure [ fig:1934 ] . there are no other x - ray sources nearby , indicating that the _ chandra _ source is very likely 1rxs j193458.1 + 335301 despite the @xmath117 distance between the two ( this is somewhat larger than the separations between the optical and x - ray sources in fig . [ fig : offset ] , but given the extended nature of the x - ray source that is not that surprising ) . the peak of the _ chandra_emission is at @xmath118 , @xmath119 , and has an extent of @xmath120 . the overall source is larger and asymmetric , with a maximum visible distance of @xmath121 from the peak to the north - east and a minimum distance of @xmath122 from the peak to the south - west , although there is some diffuse emission that extends further . averaged over azimuth , the half - power radius is @xmath123 , and 95% of the power is within @xmath124 . fitting the spatial profile to a @xmath125-model ( surface brightness @xmath126 , typical for galaxy clusters ) in ` sherpa ` was successful , with core radius @xmath127 , @xmath128 , an amplitude of @xmath129 , and @xmath130 for 12 degrees of freedom . we extracted photon events from a @xmath131 region and created source and background response files using the ` ciao ` task ` acisspec ` . we used versions of 3.0.2 of ` ciao ` and 2.26 ` caldb ` that compensate for low - energy degradation of the acis detector . we then fit the data in ` sherpa ` . the events were binned so that each bin had @xmath132 counts . while there are not very many counts ( 601 source counts before background subtraction , with 77.2 background counts ) , the data are well fit ( fig . [ fig:1934spec ] ) by an absorbed power - law model , with @xmath133 , @xmath134 , and an amplitude of @xmath135 at 1 kev ( giving @xmath136 for 16 degrees of freedom ; all uncertainties are 1-@xmath137 ) . the observed flux from this model is @xmath138 ( 0.38.0 kev ) , while the unabsorbed flux is @xmath139 ( 0.38.0 kev ) . @xcite plasma models do not fit , giving @xmath140/16 . the extended morphology and the hard spectrum of the source suggest several models : ( 1 ) a very hot nebula of galactic origin , ( 2 ) a very hot nebula but of extra - galactic origin ( gas from a cluster or an early - type galaxy ; @xcite ) , ( 3 ) the superposition of many bright lmxbs ( with power - law spectra ) in an early - type galaxy @xcite , or ( 4 ) a pulsar wind nebula radiating via synchrotron emission . here we discuss each interpretation in some detail . galactic nebula : : this would require a very bright central source ( i.e. an ob star ) to heat the nebula , which would also be visible as an extended optical / ir source and should show h@xmath108 emission . while the northern 2mass source ( 2mass j19345557 + 3353136 ) in figure [ fig:1934 ha ] is moderately bright ( with about a 5% chance of a star this bright occurring randomly within @xmath141 of a position ) , it does not appear to be an ob star . the colors are fairly common for the field , they are more consistent with a star of type @xmath142g0 , and the extinction that would be required of an ob star is higher than expected for this line of sight ( @xmath143 mag , compared to a maximum of @xmath144 as determined by @xcite and w3colden ) . the southern 2mass source ( 2mass j19345569 + 3353063 ) is fainter and redder consistent with being a late - type star but its position near the peak of the x - ray emission ( chance probability of @xmath0% ) suggests that there might be an actual association between it and the x - ray emission . since there is no diffuse broadband optical , broadband ir , or h@xmath108 emission , we do not believe that the extended x - ray emission is powered by any of the stars , although it may be related to one or both of the 2mass sources . extragalactic nebula : : the x - ray emission is much more compact than is typical for galaxy groups or clusters ( even clusters at @xmath1451 have @xmath146 ; @xcite ) , and the spectrum is wrong ( thermal plasma models do not fit ) . early - type galaxy : : the size is similar to what is often seen for early - type ( e and s0 ) galaxies . in those galaxies the x - ray emission comes from a combination of hot gas ( plasma with @xmath1471.0 kev ; @xcite ) and the superposition of many hard x - ray point sources ( whose spectra are power - laws with @xmath148 ) reasonably compatible with the observed spectrum of 1rxs j193458.1 + 335301 . + for these galaxies , the hard x - ray luminosity scales reasonably well with the integrated @xmath111 or @xmath149 band luminosities as @xmath150 or @xmath151 , where the x - ray luminosity is in the 0.38.0 kev band and all luminosities are corrected for extinction @xcite ; the scatter in this ( from a sample of 14 galaxies ) is a factor of 23 ( there is more scatter in @xcite , but their work concerns the soft emission more than the hard emission , and the scatter is still only a factor of @xmath152 ) . we can convert the relations from @xcite into relations for fluxes and magnitudes ( i.e. observables ) , giving @xmath153 where @xmath154 is the absorption - corrected x - ray flux in the 0.38.0 kev band in units of @xmath155 , @xmath156 ( @xmath157 ) is the observed @xmath111-band ( @xmath149-band ) magnitude , @xmath158 is the galactic column density along the line of sight in units of @xmath159 , and @xmath160 mag ( @xmath161 mag ) is the @xmath111-band ( @xmath149-band ) absolute magnitude of the sun ( we have assumed @xmath162 following @xcite ) . + the relation in equation [ eqn : flux ] implies that for @xmath163 and @xmath164 , @xmath165 mag and @xmath166 mag ( this indeed matches what is seen for sources in @xcite ) . therefore , such sources should be readily visible on even shallow images . examining 2mass and palomar ( fig . [ fig:1934 ha ] ) images of 1rxs j193458.1 + 335301 we see that there are two optical / ir sources near the peak of the x - ray emission : the northern source appears stellar ( fwhm@xmath167 ) , while the southern source may have some extended emission mag ) m star @xmath168 to the south . ] to the north - east ( although this could be a superposition of point sources ) . however , neither of these sources is a great candidate for the origin of the x - ray emission , as they are too faint by several orders of magnitude ( @xmath169 mag and @xmath170 mag for the northern and southern sources , respectively ) and not extended enough . therefore , while it is not impossible that 1rxs j193458.1 + 335301 is an early - type galaxy , we consider it unlikely . deeper x - ray observations should be conclusive : if 1rxs j193458.1 + 335301 is a galaxy , it should resolve into discrete point sources . optical spectroscopy would also be useful in determining the natures of the optical / ir sources . + there are some early - type galaxies with significant excesses of x - ray emission @xcite , largely due to increases in the amounts of hot gas that give roughly the same x - ray - to - optical ratio as would be necessary here . however , the optical / ir sources in fig . [ fig:1934 ha ] do not look like bright galaxies ( unlike the galaxies from @xcite which are typically @xmath171 in the optical ) and the spectrum of 1rxs j193458.1 + 335301 is wrong : again , thermal plasma models do not fit . pwn : : a pulsar wind nebula ( i.e. , a nebula excited by a pulsar or pwn ; for a review , see @xcite ) is consistent with the size and spectrum of 1rxs j193458.1 + 335301 , although the source is slightly softer toward the center ( fig . [ fig : radprof ] ) , contrary to what is expected for pwne . there is no obvious h@xmath108 emission from 1rxs j193458.1 + 335301 in figure [ fig:1934 ha ] as there can be near pwne ( associated either with sn ejecta or with the passage of the pulsar through the interstellar medium ; * ? ? ? * ; * ? ? ? * ) , but this could be because the conditions are not favorable . + 1rxs j193458.1 + 335301 is outside snr g65.3 + 5.7 . if it were a bow - shock nebula that originated in the interior of the snr and then moved outside the shell ( and not a static pwn inflated by the wind of its central pulsar ) , one would expect h@xmath108 emission and for the x - ray nebula to trail away from the direction of motion / toward the snr center ( e.g. , @xcite , although this is not always the case ) . since the x - ray emission trails _ away _ from the snr center ( suggesting motion toward the snr instead of out of it ) , we see no h@xmath108 , the fitted value of @xmath172 is just about at the maximum predicted for this line of sight by w3colden and is somewhat higher than the nominal value for snr g65.3 + 5.7 ( suggesting that the x - ray source may be more distant and highly absorbed than snr g65.3 + 5.7 ) , we believe that an association between the two is unlikely . however , this is not entirely unexpected , as there are a number of young , newly - discovered pwne that have no definitively associated snrs ( similar to the the crab nebula ; @xcite ) . the bsc lists 1rxs j205042.9 + 284643 as having @xmath173 in the pspc , with hardness ratios of @xmath174 and @xmath175 . the _ chandra _ observation of 1rxs j205042.9 + 284643 had a total exposure time of 3.7 ksec , and which should give @xmath176 acis - s counts depending on the source spectrum . however , as shown in figure [ fig : rxsb ] there are no point sources detected anywhere within three times the nominal position uncertainty ( a conservative limit , as seen in fig . [ fig : offset ] ) : the only significant source detected ( using ` wavdetect ` on scales from 132 pixels ; see for the detection method ) in the data - set is at @xmath177 , @xmath178 ( with @xmath179 counts ) , which is @xmath180 or 6-@xmath137 away from 1rxs j205042.9 + 284643 . this x - ray source is almost certainly not related to 1rxs j205042.9 + 284643 . we can then set a limit of @xmath181 counts to any point source . there are no obvious extended sources , but such limits are more difficult to quantify : overall , there are 1047 counts in the 0.35.0 kev energy range over the whole @xmath182 image , so the average background rate is @xmath183 . then , in a region @xmath184 in area , the 3-@xmath137 limits will be @xmath185 counts . there are no regions in the event list with such concentrations , so no extended sources are present . one obvious explanation of the disappearance of 1rxs j205042.9 + 284643 is variability . this is not atypical among the most common class of soft x - ray sources in the galactic plane : active stars . flares and other chromospheric / magnetospheric events often lead to dramatic changes in the fluxes of these sources . while other sources , such as x - ray binaries , active galaxies , and some anomalous x - ray pulsars , do exhibit variability , these sources have hard x - ray spectra generally inconsistent with the bsc emission . we therefore think it likely that 1rxs j205042.9 + 284643 is an active star , but of course this can not be confirmed without additional data . it is also possible that the source is extended , and therefore too diffuse to have been detected by _ the softness of the bsc emission make this unlikely , though , as most known types of extended sources are relatively hard ( e.g. , 1rxs j193458.1 + 335301 and 1rxs j150139.6@xmath28403815 ) . it is possible , but unlikely , that 1rxs j205042.9 + 284643 is a neutron star . as discussed above , most of the neutron stars considered in have stable x - ray emission : only some of the axps vary significantly . however , the spectrum of 1rxs j205042.9 + 284643 is unlike those of axps ( typically a power - law with @xmath186 ) . the bsc lists 1rxs j205812.8 + 292037 as having @xmath187 in the pspc , with hardness ratios of @xmath188 and @xmath189 . this is moderately hard compared to the other sources in fig . [ fig : hard ] , but is not too extreme . unlike the rest of the sources without _ chandra _ followup , the counterpart(s ) shown in figure [ fig : opt4 ] are not entirely secure . within @xmath32 of 1rxs j205812.8 + 292037 , there are 1627 2mass sources , for an average density of @xmath190 . to find a source within @xmath191 ( as in the case of 1rxs j205812.8 + 292037 ) has a chance probability of 37% , and the chance probability for two sources is 13% . these are not small enough for a definite association . 1rxs j205812.8 + 292037is similar , in both hardness ratio and optical brightness , to other sources like 1rxs j193228.6 + 345318 , 1rxs j205042.9 + 284643 , 1rxs j045707.4 + 452751 , or 1rxs j151942.8@xmath28375255 . these sources did not have certain associations based on _ rosat _ alone , but the _ chandra _ data are unambiguous . these sources may represent a population of x - ray sources that are somewhat fainter than the majority of the sources in table [ tab : srcs ] . this faintness , together with the hardness of the x - ray spectrum , likely reflects extragalactic origins of the sources ( i.e. they are active galaxies ) : in figure [ fig : fxk ] these sources are largely those with the highest x - ray - to - ir flux ratios most similar to the extragalactic sample . overall , 1rxs j205812.8 + 292037 is consistent with having an association with one of the identified 2mass sources . chandra _ followup observation would have made the case secure , but it was not selected for _ chandra _ due to an oversight . as with 1rxs j205042.9 + 284643 , we do not believe that 1rxs j205812.8 + 292037 is a neutron star , but we can not rule out this possibility . + the bsc lists 1rxs j150139.6@xmath28403815 as having @xmath192 in the pspc , with hardness ratios of @xmath193 and @xmath194 . the _ chandra _ source is fainter than that of 1rxs j193458.1 + 335301 , but nonetheless it appears extended , as shown in figure [ fig:1501 ] . since this source is more diffuse than 1rxs j193458.1 + 335301 , the spatial measurements are not as precise , but the center is at approximately @xmath195 , @xmath196 . the total extent of the source is @xmath197 in radius . as with 1rxs j193458.1 + 335301 , while there is some offset between the _ rosat _ and _ chandra_positions , this does not appear inconsistent with the position uncertainties for such extended sources . again , we can be quite confident that the _ chandra _ source is 1rxs j150139.6@xmath28403815 , since there are no other sources nearby . similar to our analysis of 1rxs j193458.1 + 335301 ( [ sec : rxsa ] ) , we extracted photon events from a @xmath198 region and created source and background response files using the ` ciao ` task ` acisspec ` . we then fit the data in ` sherpa ` , where the events were binned so that each bin had @xmath132 counts . there are 1305 source counts and 478.5 background counts . the data are well fit ( fig . [ fig:1501spec ] ) by an absorbed power - law model , with @xmath199 , @xmath200 , and an amplitude of @xmath201 at 1 kev ( giving @xmath202 for 32 degrees of freedom ; all uncertainties are 1-@xmath137 ) . the observed flux from this model is @xmath203 ( 0.38.0 kev ) , and the unabsorbed flux is @xmath204 ( 0.38.0 kev ) . the column density is higher than but consistent ( given the uncertainties ) with both the column density of snr g330.0 + 15.0 and the total expected along this line of sight ( @xmath205 ) . as with 1rxs j193458.1 + 335301 , we considered different models for 1rxs j150139.6@xmath28403815 . figure [ fig:15012mass ] does not identify a single hot source , so a thermal galactic nebula is unlikely . our first idea was that 1rxs j150139.6@xmath28403815 is a pwn . the size is about right and the spectrum is typical for pwne . however , as with 1rxs j193458.1 + 335301 there is a problem : 1rxs j150139.6@xmath28403815is outside of snr g156.2 + 5.7 , and the largely symmetric morphology rules out an association between 1rxs j150139.6@xmath28403815 and snr g330.0 + 15.0 ( i.e. 1rxs j150139.6@xmath28403815 can not be a bow - shock nebula ) . 1rxs j150139.6@xmath28403815 could instead be a pressure - confined bubble pwn related to another supernova ; we searched the sydney university molongolo sky survey ( summs ; @xcite ) for evidence of radio emission from or another supernova shell surrounding 1rxs j150139.6@xmath28403815 , but there is no extended or point - like emission present at the position of 1rxs j150139.6@xmath28403815 nor is there any sign of a new snr around it . sumss is particularly sensitive to extended emission , and would almost certainly have identified any snr around 1rxs j150139.6@xmath28403815 . like 1rxs j193458.1 + 335301 , the lack of a clear snr shell does not mean that 1rxs j150139.6@xmath28403815 is not a pwn . we then examined possible extragalactic classifications for 1rxs j150139.6@xmath28403815 . this source is larger than 1rxs j193458.1 + 335301 , and is more compatible with the sizes of typical galaxy clusters ( @xmath206 ) : a fit to a @xmath125 model has @xmath207 and @xmath208 . the spectral data are reasonably well fit by a @xcite plasma model , having @xmath209 , @xmath210 kev , and a normalization , where @xmath211 is the angular - size distance ( in cm ) , @xmath212 is the electron density ( in @xmath213 ) , and @xmath214 is the hydrogen density ( in @xmath213 ) . ] of @xmath215 ( giving @xmath216 for 32 degrees of freedom ) , such as what one would expect for a cluster @xcite . with this model the observed flux is @xmath217 ( 0.38.0 kev ) , and the unabsorbed flux is @xmath218 . examining the 2mass image we see an extended elliptical source , 2masx j15014110@xmath284038093 , near the center of the x - ray emission ( fig . [ fig:15012mass ] ) . this source has a radius of @xmath219 ( @xmath220 isophotal radius ) , a @xmath110 magnitude of @xmath221 mag within that radius , and @xmath222 mag . higher - resolution optical images of 1rxs j150139.6@xmath28403815 ( fig . [ fig:15012mass ] ) show that + 2masx j15014110@xmath284038093 is partially decomposed into two sources : an extended source labeled a that is at the exact position of 2masx j15014110@xmath284038093 ( within uncertainties ) , and a source labeled b @xmath223 to the east . there is also another extended source labeled c @xmath224 to the north - east , but this is a separate 2mass source ( 2mass j15014145@xmath284038068 ) . we performed a rough photometric calibration using 80 stars from the usno - b1.0 catalog and then ran ` sextractor ` @xcite on the images : the results for sources a c are given in table [ tab : rxscopt ] . source a is very clearly extended , although it is not as large as 2masx j15014110@xmath284038093 ( the fwhm of the ir emission is @xmath225 ) . source b is very likely unresolved ( within uncertainties ) , and source c is extended . source a is very red ( @xmath226 mag ) , consistent with the 2mass data . we believe that the 2mass source is primarily due to source a , given the position coincidence and the extreme redness of a compared to b or c. if this is the case , though , then a has the relatively blue color of @xmath227 mag , but this could be partly due to the difficulties of measuring an extended source from images with drastically different seeing ( 2mass versus magic @xmath109-band ) . the ir colors of 2masx j15014110@xmath284038093 are similar to the brighter galaxies in known clusters ( e.g. , * ? ? ? therefore , 2masx j15014110@xmath284038093 could be the central galaxy of an unknown cluster . the x - ray temperature is reasonably high , implying a high luminosity ( @xmath228 ; @xcite ) , so this source can not be part of a nearby , low-@xmath229 group . however , the value of @xmath125 is lower than those of most known clusters @xcite , and is more similar to those of low-@xmath229 systems @xcite . while 1rxs j150139.6@xmath28403815 is compatible with the sizes and spectra of early - type galaxies , and there is an extended optical / ir source near the peak of the x - ray emission , the scenario is not entirely consistent . the optical / ir source is , like those in 1rxs j193458.1 + 335301 , about 7 magnitudes fainter than expected ( the predicted magnitude following eqn . [ eqn : flux ] is @xmath230 mag ) . this is far greater than the variation seen among galaxies . we do not believe that the difference can be due to an excess of soft emission in 1rxs j193458.1 + 335301 or 1rxs j150139.6@xmath28403815 ( eqn . [ eqn : flux ] refers only to the contribution of hard point sources ) , as the spectra of 1rxs j193458.1 + 335301 and 1rxs j150139.6@xmath28403815 are hard and similar to the prototypical sources assumed in @xcite and when one fits primarily for the soft emission ( as in @xcite ) one finds a similar relation to that of @xcite . it is possible that 1rxs j150139.6@xmath28403815 is an over - luminous elliptical galaxy , such as those discussed in @xcite , as the size , optical / x - ray flux ratio , and luminosity are similar to these sources ( @xmath231 ; @xmath232 assuming @xmath233 ) , but again there are difficulties : the temperature of 1rxs j150139.6@xmath28403815 is considerably higher than those of @xcite , and the value of @xmath125 is too low . we see that no scenario is entirely consistent for 1rxs j150139.6@xmath28403815 . pwne , isolated elliptical galaxies , and galaxy clusters all have problems . we believe it likely that 1rxs j150139.6@xmath28403815 does have an extragalactic origin , as 2masx j15014110@xmath284038093 looks like an elliptical galaxy and it is probably associated with the x - ray emission : there are 39 extended 2mass sources within @xmath234 of 1rxs j150139.6@xmath28403815 , giving a false - association rate of 0.005% for a source within @xmath235 . however , it is not clear exactly what 1rxs j150139.6@xmath28403815 is . as with 1rxs j193458.1 + 335301 , deeper x - ray observations and optical spectroscopy should be conclusive for 1rxs j150139.6@xmath28403815 . we have fully investigated the population of _ rosat _ bsc point sources in six large - diameter snrs . our identifications of counterparts to 50 of the 54 sources are quite secure : in most cases the positional coincidence between the x - ray and optical / ir sources has been augmented by identification of an abnormal stellar type ( variable , t tau , binary , etc . ) , by the extreme brightness ( and hence rarity ) of the optical source , or by a previous classification in the literature . this conclusion echoes that of @xcite , who searched for older neutron stars using _ rosat _ and found only previously identified neutron stars , along with 17 sources that are definitely not neutron stars and 13 that are probably not . the remaining sources , as discussed in section [ sec : remain ] , are more intriguing . however , none of them is likely to be a neutron star associated with one of the snrs in table [ tab : snrs ] . to begin with , all are outside their snrs . while this is not impossible for older sources and high velocity neutron stars ( e.g. , * ? ? ? * ) , it lessens the chance of association . for 1rxs j193458.1 + 335301 and 1rxs j150139.6@xmath28403815 the x - ray morphologies rule out associations , since any pwne outside the snrs would likely have elongated bow - shock appearances , in contrast to what we see ( of course , 1rxs j193458.1 + 335301 and/or 1rxs j150139.6@xmath28403815 could be extragalactic ) . 1rxs j205042.9 + 284643 and 1rxs j205812.8 + 292037 , neither of which has a _ chandra _ detection , are more uncertain . 1rxs j205042.9 + 284643 is likely a flare star . 1rxs j205812.8 + 292037 does not have a provisional classification but is probably extragalactic in origin . since we have ruled out ( to some degree of certainty ) neutron stars in all six snrs considered here , we can then follow and draw the x - ray luminosity limits on a cooling diagram . this is shown in figure [ fig : cool ] . to account for the uncertainties of 1rxs j205042.9 + 284643and 1rxs j205812.8 + 292037 , both in snr g74.0@xmath288.5 , we have adjusted the luminosity of that snr from table [ tab : snrs ] to correspond to @xmath236 above the count - rates of both of the uncertain sources and therefore a more secure limit . further x - ray observations of 1rxs j205812.8 + 292037 would very likely detect counterparts ( for 1rxs j205042.9 + 284643 , it might have only been included in the bsc due to a flare , and therefore significantly deeper x - ray observations may be necessary ) . with secure counterparts , the limit for snr g74.0@xmath288.5 would decrease by a factor of 3 . the limits in figure [ fig : cool ] are not as uniform or as constraining as those from . the lack of uniformity is due to the sample construction : the different distances and column densities of the snrs make the bsc limit of @xmath237 translate into different luminosities . so , snrs g74.0@xmath288.5 , g330.0 + 15.0 , and g65.3 + 5.7 all have reasonably tight limits ( and those of snr g74.0@xmath288.5could get better ) . snrs g160.9 + 2.6 and g205.5 + 0.5 have loose limits primarily due to uncertain distances : we have used the upper limit of 4 kpc for snr g160.9 + 2.6 and the full range of 0.81.6 kpc for snr g205.5 + 0.5 in figure [ fig : cool ] . finally , snr g156.2 + 5.7 is more highly absorbed than the other snrs . while all of the limits are below the luminosities of central sources in cas a , , and ( and are therefore in concordance with our original survey design from ) , some are further below than others . the utility of these limits is somewhat lessened , though , as the snrs are all reasonably large and are older ( 1030 kyr ) than the sources in ( 310 kyr ) . therefore the cooling curves have descended , and there are other snrs that have similar or even lower neutron star luminosities ( , , for the snrs with tighter limits , and , snrs , , and for the remaining snrs ) , although 5/7 of these sources have x - ray pwne that increase their luminosities by a factor of @xmath152 . in one sense , though , the limits here are tighter than those of . by using the bsc to go to twice the snr radii , we have virtually eliminated the possibility that there are high - velocity neutron stars in these snrs ( as discussed in [ sec : ext ] confusion is most likely not a limiting factor in detecting x - ray sources ) , while in we only searched a portion of the snr interiors . it is of course possible that the sn explosions were type ia or produced black holes , but as discussed in these alternate scenarios are not very likely for an ensemble . therefore , while not as tight as those of ( or e.g. , @xcite ; @xcite ) , our limits are still useful . they are not below all detected neutron stars , so do not require appeals to exotic physics or cooling processes , but they conclusively demonstrate that there is a significant range in the observed luminosities of neutron stars , even including experimental uncertainties . it is clear that the neutron stars of a single age must be able to produce a luminosities differing by a factor of @xmath238 . whether the unknown parameter that controls the luminosity is one of the usual culprits ( mass , rotation , composition , magnetic field ) or something entirely different is not known . it is also clear that there is a significant number of objects that do not show non - thermal emission and would therefore not go on to evolve as traditional radio pulsars . we thank an anonymous referee for helpful comments . d. l. k. was partially supported by a fellowship from the fannie and john hertz foundation . b. m. g. and p. o. s. acknowledge support from nasa contract nas8 - 39073 and grant g02 - 3090 . b. m. g. is supported by nasa ltsa grant nag5 - 13032 . s. r. k. is supported by grants from nsf and nasa . support for this work was provided by the national aeronautics and space administration through chandra award go3 - 4088x issued by the _ chandra x - ray observatory _ center , which is operated by the smithsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 39073 . the nrao is a facility of the national science foundation operated under cooperative agreement by associated universities , inc . the digitized sky surveys were produced at the space telescope science institute under u.s . government grant nag w-2166 . we have made extensive use of the simbad database , and we are grateful to the astronomers at the centre de donnes astronomiques de strasbourg for maintaining this database . we would like to thank e. persson and c. rakowski for assistance with the panic observing , and d. fox for assistance with the p60 observing . green , d. a. 2000 , a catalogue of galactic supernova remnants ( 2000 august version ) , mullard radio astronomy observatory , cavendish laboratory , cambridge , uk ( available on the world - wide - web at http://www.mrao.cam.ac.uk/surveys/snrs/ ) , r. f. 2004 , in clusters of galaxies : probes of cosmological structure and galaxy evolution , ed . j. s. mulchaey , a. dressler , & a. oemler ( cambridge : cambridge university press ) , 123 , ( astro - ph/0311105 ) r c c r r c c l c r c r + 1 & & 0.066 & 45.1 & 11 & @xmath239 & @xmath240 & & 19344524 + 3030590 & 5.0 & 10.41 & 8.108 + 2 & & 0.083 & 82.7 & 9 & @xmath241 & @xmath242 & & 19384012 + 3030284 & 7.1 & 12.6 & 11.150 + 3 & & 0.054 & 101.8 & 12 & @xmath243 & @xmath244 & & 19392261 + 3009119 & 9.9 & 4.685 & 2.499 + 4 & & 0.256 & 155.8 & 8 & @xmath245 & @xmath246 & & 19433674 + 3225206 & 5.6 & 9.9 & 8.082 + & & & & & @xmath247 & @xmath248 & & 19433790 + 3225124 & 13.6 & 10.7 & 7.179 + 5 & & 0.051 & 165.0 & 14 & @xmath249 & @xmath250 & & & & & + 6 & & 0.110 & 194.4 & 15 & @xmath251 & @xmath252 & usno 1181@xmath280406270 & 19272197 + 2809452 & 9.8 & 11.620 & 8.426 + 7 & & 0.138 & 202.7 & 8 & @xmath253 & @xmath254 & & 19440138 + 2845099 & 14.0 & & 8.691 + 8 & & 0.357 & 219.9 & 7 & @xmath255 & @xmath256 & & 19490298 + 2952582 & 1.2 & 8.32 & 5.734 + 9 & & 0.091 & 223.4 & 9 & @xmath257 & @xmath258 & & & & & + 10 & & 0.057 & 237.0 & 14 & @xmath259 & @xmath260 & & 19145022 + 3151371 & 16.5 & 6.618 & 4.312 + 11 & & 0.290 & 255.7 & 8 & @xmath261 & @xmath262 & & 19385576 + 3514132 & 8.0 & 8.29 & 6.858 + 12 & & 0.074 & 285.6 & 8 & @xmath263 & @xmath264 & & 19393666 + 2637169 & 3.0 & 11.1 & 7.345 + 13 & & 0.153 & 298.4 & 7 & @xmath265 & @xmath266 & & 19311257 + 3607300 & 13.1 & & 8.839 + 14 & & 0.052 & 305.3 & 11 & @xmath267 & @xmath268 & usno 1188@xmath280330651 & 19115080 + 2850075 & 5.7 & 12.510 & 8.898 + + 1 & & 0.111 & 113.3 & 12 & @xmath269 & @xmath270 & & & & & + 2 & & 0.104 & 114.4 & 10 & @xmath271 & @xmath272 & & 20445809 + 2916211 & 11.9 & 10.8 & 8.746 + 3 & & 0.129 & 122.3 & 8 & @xmath273 & @xmath274 & usno 1193@xmath280519643 & 20581282 + 2920282 & 8.9 & 15.980 & 14.258 + & & & & & @xmath275 & @xmath276 & & 20581257 + 2920454 & 8.5 & & 13.405 + 4 & & 0.225 & 214.7 & 8 & @xmath277 & @xmath278 & & 20520768 + 2705491 & 11.4 & 4.576 & 2.722 + + 1 & & 0.051 & 75.3 & 20 & @xmath279 & @xmath280 & & 05064067 + 5135519 & 20.8 & 4.980 & 4.124 + + 1 & & 0.061 & 82.5 & 9 & @xmath281 & @xmath282 & & & & & + 2 & & 0.061 & 87.2 & 11 & @xmath283 & @xmath284 & & & & & + 3 & & 0.052 & 99.0 & 10 & @xmath285 & @xmath286 & & 04522151 + 4556236 & 8.6 & 6.695 & 5.407 + + 1 & & 0.086 & 74.3 & 11 & @xmath287 & @xmath288 & usno 0155@xmath2801104@xmath281 & 06410807 + 0522522 & 4.0 & 10.780 & 8.905 + & & & & & @xmath289 & @xmath290 & & 06410796 + 0522438 & 8.0 & & 12.121 + 2 & & 0.054 & 100.1 & 11 & @xmath291 & @xmath292 & & 06413601 + 0802055 & 13.0 & 10.3 & 8.712 + 3 & & 0.076 & 107.4 & 17 & @xmath293 & @xmath294 & usno 0947@xmath280100759 & 06410953 + 0447354 & 5.4 & 17.650 & 15.084 + & & & & & @xmath295 & @xmath296 & usno 0947@xmath280100763 & 06410988 + 0447350 & 8.1 & 15.840 & 14.636 + & & & & & @xmath297 & @xmath298 & & 06410900 + 0447239 & 10.1 & & 12.182 + & & & & & @xmath299 & @xmath300 & usno 0155@xmath2802167@xmath281 & 06410924 + 0447187 & 14.3 & 11.100 & 8.688 + 4 & & 0.092 & 160.1 & 10 & @xmath301 & @xmath302 & & 06464073 + 0821471 & 7.6 & 7.5 & 6.080 + 5 & & 0.082 & 179.5 & 8 & @xmath303 & @xmath304 & & & & & + 6 & & 0.088 & 183.9 & 9 & @xmath305 & @xmath306 & & 06293689 + 0829327 & 5.1 & 7.62 & 6.306 + 7 & & 0.109 & 191.7 & 10 & @xmath307 & @xmath308 & usno 0150@xmath2800332@xmath281 & 06371522 + 0320081 & 7.6 & 10.640 & 9.323 + & & & & & @xmath309 & @xmath310 & & 06371585 + 0320043 & 2.5 & & 14.508 + & & & & & @xmath311 & @xmath312 & & 06371555 + 0320026 & 3.7 & & 11.813 + 8 & & 0.106 & 196.7 & 11 & @xmath313 & @xmath314 & usno 0145@xmath2801717@xmath281 & 06255524 + 0655386 & 7.8 & 9.660 & 8.019 + + 1 & & 0.077 & 26.1 & 10 & @xmath315 & @xmath316 & & & & & + 2 & & 0.082 & 35.0 & 12 & @xmath317 & @xmath318 & usno 0505@xmath280350285 & 15105908@xmath283926590 & 7.2 & & 15.066 + & & & & & @xmath319 & @xmath320 & usno 7826@xmath2800179@xmath281 & 15105821@xmath283926499 & 16.4 & 10.600 & 8.374 + & & & & & @xmath321 & @xmath322 & usno 0505@xmath280350290 & & 2.4 & & + 3 & & 0.069 & 40.3 & 16 & @xmath323 & @xmath324 & & 15081213@xmath284035022 & 27.1 & 5.762 & 5.934 + 4 & & 0.051 & 72.7 & 11 & @xmath325 & @xmath326 & & 15042865@xmath283924261 & 3.9 & 10.7 & 8.293 + 5 & & 0.073 & 81.8 & 8 & @xmath327 & @xmath328 & & 15052586@xmath283857031 & 9.4 & 12.55 & 9.124 + 6 & & 0.125 & 103.2 & 14 & @xmath329 & @xmath330 & & & & & + 7 & & 0.107 & 113.6 & 17 & @xmath331 & @xmath332 & & 15185282@xmath284050528 & 38.4 & 11.01 & 8.547 + 8 & & 0.085 & 117.1 & 11 & @xmath333 & @xmath334 & & 14595244@xmath284011594 & 8.6 & 9.90 & 8.401 + 9 & & 0.065 & 123.0 & 16 & @xmath335 & @xmath336 & & 15165935@xmath283826514 & 3.4 & 6.876 & 5.789 + 10 & & 0.074 & 135.1 & 10 & @xmath337 & @xmath338 & & 15204697@xmath284053526 & 15.1 & 7.83 & 6.518 + 11 & & 0.084 & 138.2 & 10 & @xmath339 & @xmath340 & usno 0507@xmath280344267 & 14583744@xmath283915033 & 4.9 & 11.850 & 8.648 + 12 & & 0.080 & 140.1 & 11 & @xmath341 & @xmath342 & & 15221162@xmath283959509 & 8.4 & 12.02 & 9.100 + 13 & & 0.099 & 145.9 & 9 & @xmath343 & @xmath344 & & 14572207@xmath284013586 & 4.2 & & 13.333 + 14 & & 0.115 & 148.5 & 12 & @xmath345 & @xmath346 & & 15180715@xmath283804238 & 5.3 & & 10.600 + & & & & & @xmath347 & @xmath348 & usno 7822@xmath2800433@xmath281 & 15180762@xmath283804237 & 10.7 & 10.650 & 7.999 + 15 & & 0.054 & 150.7 & 13 & @xmath349 & @xmath350 & & 15144748@xmath284220149 & 14.2 & & 9.011 + 16 & & 0.153 & 152.8 & 8 & @xmath351 & @xmath352 & & 15201253@xmath283821579 & 4.2 & & 8.454 + 17 & & 0.061 & 169.1 & 10 & @xmath353 & @xmath354 & & & & & + 18 & & 0.105 & 170.4 & 11 & @xmath355 & @xmath356 & & 15244501@xmath283942367 & 4.7 & 5.367 & 5.495 + 19 & & 0.176 & 172.5 & 8 & @xmath357 & @xmath358 & & 14561404@xmath283851200 & 5.2 & 9.15 & 7.323 + 20 & & 0.051 & 173.7 & 15 & @xmath359 & @xmath360 & usno 0483@xmath280366208 & 14574495@xmath284141394 & 6.9 & 11.790 & 9.351 + & & & & & @xmath361 & @xmath362 & & 14574414@xmath284141408 & 1.8 & & 14.698 + l c p1.25 in c c p2.5 in + & 1 & & [ fig : opt1 ] & n & planetary nebula , associated with 235.2 mjy nvss source at @xmath363 @xmath364 @xcite + & 2 & & [ fig : opt1 ] & n & dwarf nova ; x - ray emission reported by @xcite and @xcite + & 3 & & [ fig : opt1 ] & n & spectroscopic binary ; type g8iii + & 4 & / & [ fig : opt1 ] & n & binary system , late - type + & 5&&[fig : opt1],[fig:1934 ha ] & y & extended _ chandra _ source ; see [ sec : rxsa ] + & 6 & usno 1181@xmath280406270 & [ fig : opt1 ] & n & + & 7 & & [ fig : opt2 ] & n & late - type + & 8 & & [ fig : opt2 ] & n & pulsating variable star of type k2ii - iii + & 9&2mass j19322722 + 3453148&[fig : opt2],[fig : cxoopt1 ] & y & 35.4 mjy nvss counterpart at @xmath365 , @xmath366 ; flare star @xcite + & 10 & & [ fig : opt2 ] & n & late - type + & 11 & & [ fig : opt2 ] & n & double star system ; late - type + & 12 & & [ fig : opt2 ] & n & late - type + & 13 & & [ fig : opt3 ] & n + & 14 & usno 1188@xmath280330651 & [ fig : opt3 ] & n + + & 1 & & [ fig : opt4 ] & y & flare star ? see [ sec : rxsb ] + & 2 & & [ fig : opt4 ] & n & late - type + & 3 & usno 1193@xmath280519643/ 2mass j20581257 + 2920454&[fig : opt4 ] & n & association may be somewhat questionable ; see [ sec : rxsd ] + & 4 & & [ fig : opt4 ] & n & variable type g7iii star + & & 2mass j20493540 + 2938509 & [ fig : cxoopt1 ] & y & late - type + + & 1 & & [ fig : opt5 ] & n & variable star of the @xmath2-dor type + + & 1 & 2mass j04570832 + 4527499 & [ fig : opt6],[fig : cxoopt1 ] & y + & 2 & 2mass j05033958 + 4516594 & [ fig : opt6],[fig : cxoopt1 ] & y & associated with the 34.3 mjy nvss source at @xmath82 , @xmath367 ; flare star ? + & 3 & & [ fig : opt6 ] & n & late - type + + & 1 & usno 0155@xmath2801104@xmath281/ 2mass j06410796 + 0522438 & [ fig : opt7 ] & n & binary system ? + & 2 & & [ fig : opt7 ] & n & late - type + & 3 & usno 0947@xmath280100759/ usno 0947@xmath280100763/ 2mass j06410900 + 0447239/ usno 0155@xmath2802167@xmath281 & [ fig : opt7 ] & n & multiple - star system + & 4 & & [ fig : opt7 ] & n & variable star of the @xmath2-dor type + & 5 & 2mass j06274012 + 0731006&[fig : opt7],[fig : cxoopt1 ] & y & + & 6 & & [ fig : opt7 ] & n & late - type + & 7 & usno 0150@xmath2800332@xmath281/ 2mass j06371585 + 0320043/ 2mass j06371555 + 0320026&[fig : opt8 ] & n & multiple - star system ? + & 8 & usno 0145@xmath2801717@xmath281&[fig : opt8 ] & n + 1e 0627.4 + 0537 & & 2mass j06300529 + 0545407 & [ fig : cxoopt1 ] & y & + 1e 0630.9 + 0611 & & 2mass j06333322 + 0608396 & [ fig : cxoopt2 ] & y & + 1e 0636.8 + 0517 & & 2mass j06392566 + 0514301 & [ fig : cxoopt2 ] & y & + + & 1 & 2mass j15081819@xmath284017261 & [ fig : opt9],[fig : cxoopt2 ] & y & also 1axg j150818@xmath284016 @xcite + & 2 & usno 0505@xmath280350285/ usno 7826@xmath2800179@xmath281/ usno 0505@xmath280350290 & [ fig : opt9 ] & n & multiple - star system ? + & 3 & & [ fig : opt9 ] & n & variable star of the @xmath368-cvn type ; has late - type companion @xcite + & 4 & & [ fig : opt9 ] & n + & 5 & & [ fig : opt9 ] & n & t tauri star + & 6 & & [ fig : opt9],[fig:15012mass ] & y & extended _ chandra _ source ; see [ sec : rxsc ] + & 7 & & [ fig : opt10 ] & n & t tauri star + & 8 & & [ fig : opt10 ] & n & late - type + & 9 & & [ fig : opt10 ] & n & late - type + & 10 & & [ fig : opt10 ] & n & late - type + & 11 & usno 0507@xmath280344267 & [ fig : opt10 ] & n + & 12 & & [ fig : opt10 ] & n & t tauri star + & 13 & 2mass j14572207@xmath284013586 & [ fig : opt11 ] & n & extended on the dss/2mass images ( 2masx j14572207@xmath284013588 ) ; likely a galaxy + & 14 & usno 7822@xmath2800433@xmath281/ 2mass j15180762@xmath283804237&[fig : opt11 ] & n & multiple - star system ? + & 15 & & [ fig : opt11 ] & n + & 16 & & [ fig : opt11 ] & n & multiple - star system ? + & 17 & & [ fig : mag ] & y & star detected in magellan data + & 18 & & [ fig : opt11 ] & n & double star ; type a0sp@xmath369 + & 19 & & [ fig : opt12 ] & n & late - type + & 20 & usno 0483@xmath280366208/ 2mass j14574414@xmath284141408 & [ fig : opt12 ] & n & multiple - star system ? + l c c c + & & 3.7 & 1/4 + & & 3.5 & 1/4 + + & & 3.2 & 1/2 + & & 3.9 & 1/4 + + & & 5.2 & 1/4 + + & & 5.4 & 1/4 + & & 3.4 & 1/4 + 1e 0627.4 + 0537 & & 3.5 & 1/2 + 1e 0630.9 + 0611 & & 3.6 & 1/2 + 1e 0636.8 + 0517 & & 3.8 & 1/2 + + & & 2.9 & 1/4 + & & 3.7 & 1/4 + & & 2.9 & 1/4 + l l l l c r 1rxs j193458.1 + 335301 & 2003-jul-03 & p200 & lfc & @xmath370 & 300 + 1rxs j193458.1 + 335301 & 2003-jul-24 & p60 & p60ccd & h@xmath108/offband & 1200 + 1rxs j151942.8@xmath28375255 & 2003-apr-04 & magellan ii / clay & magic & @xmath109 & 930 + & 2003-apr-18 & magellan ii / clay & panic & @xmath110 & 200 + 1rxs j150139.6@xmath28403815 & 2004-feb-16 & magellan ii / clay & magic & @xmath109 & 60 + & & & & @xmath111 & 160 +

a computer is a machine which processes data according to a set of instructions . every computer is a composition of hardware on which information is processed and software by which information is processed . hardware is the physical part of a computer , while software is a collection of computer programs ( algorithms ) designed to perform a required task . a quantum computer ( qc ) emerges when the computation is executed under the framework of quantum mechanics . there are several models for quantum computation . the first conceptual model of a qc , the quantum turing machine ( qtm ) , was given by deutsch @xcite . it is rather an abstract model and useful for investigating the computational complexity . " the qtm is a quantum version of its classical analogue which allows the superposition of different computational paths . later on , a rather practical model of a qc the unitary - evolution - based quantum computation model ( uqcm)was established @xcite . the uqcm is a quantum edition of the reversible classical circuit model . " in the step from classical to quantum , the bits are replaced by the qubits , and the logic gates are replaced by the quantum gates ( coherent unitary evolution ) . unlike the bits , the qubits can exist in a superposition of different computational states . unlike the logic gates , the quantum gates are able to create and destroy a superposition as well as an entanglement . the computation in uqcm is run by a sequence of unitary gates and represented by its circuit diagram , where the connecting wires stand for the logical qubits or bits which carry the information , and the information is processed by the sequence of quantum gates . in the end , the result of the computation is read out by the projective measurements on the qubits . in the well - known textbook by nielsen and chuang @xcite , many popular algorithms such as deutsch s algorithm @xcite , grover s search algorithm @xcite , and shor s factoring algorithm @xcite are narrated in terms of the uqcm . the measurement - based quantum computation model ( mqcm ) is another well recognized model of a qc @xcite . here , a multi - qubit entangled state known as a cluster state @xcite , or more generally a graph state @xcite is the main ingredient , and it provides all the entanglement beforehand for the subsequent computation . the computation in this model is run by a sequence of single - qubit ( adaptive ) projective measurements on the graph state . the methods of mqcm enable one to simulate any quantum circuit on a sufficiently large two - dimensional graph state by arranging the spatial pattern of measurement bases for the graph qubits according to the temporal order of quantum gates in the circuit . in the mqcm , the measurements on graph qubits are performed in a certain temporal order for the purpose of running the computation deterministically . furthermore , the measurement outcomes are recorded classically and are used for setting the measurement bases for the subsequent measurements @xcite . by contrast , in the uqcm there is no such temporal order of measurements , but an order in which the unitary gates are executed . the task of a qc is to `` simulate a quantum circuit . '' both the uqcm and the mqcm are universal : they can simulate any quantum circuit . where the uqcm uses the unitary gates , the mqcm uses the measurements for simulating a circuit . in this paper our focus will be on constructing a hybrid " model of the uqcm and the mqcm , which we call hybrid quantum computation model " ( hqcm ) , that combines elements of both the uqcm and the mqcm with the aim of exploiting the strengths of both models . of course , the hqcm is universal as well . there are two main objectives of the present investigation . the first objective is to develop a theoretical understanding of the hqcm , where part of a quantum circuit is simulated by unitary gates and the rest by measurements on small graph states . the second objective is the investigation of the experimental optimization . both the uqcm and the mqcm possess their own advantages along with some similarities . for example , where the implementation of an arbitrary single - qubit operation in the mqcm costs a chain of five qubits graph state @xcite , it can be implemented rather simply by a unitary evolution in the uqcm . in return , certain multi - qubit gates that are complicated in the uqcm can be realized in one shot " in the mqcm . here is an overview of the paper . in sec . [ sec : mqcm ] we give a brief review of the mqcm , it has two main parts . the first part contains the methodology for the computation in mqcm with an example and the second part is the classical information processing in the mqcm . we use a portion of sec . [ sec : mqcm ] in sec . [ sec : hqcm ] . section [ sec : hqcm ] is reserved for a detailed discussion on the hqcm , it has three main parts . the first part is the methodology for the computation in hqcm and the second part is the classical information processing in the hqcm . the third part is about the implementation of multi - qubit controlled rotations with the hqcm , and this we are going to use in sec . [ sec : ga ] , which is about grover s search algorithm within the hqcm . finally , we conclude our report in sec . [ sec : conc ] with a summary and outlook . three appendices deal with technical details . we start this section with a short introduction about the preparation of graph states @xcite , and then we proceed towards the methods for the computation in mqcm @xcite . we conclude this section with an example . throughout the text we represent the pauli vector operator @xmath2 by @xmath3 and the identity operator by @xmath4 . graph states can be realized in many physical systems by first preparing all the qubits of graph @xmath5 in an eigenstate of their respective pauli operator @xmath6 . in other words , every qubit ` @xmath7 ' of @xmath5 is initialized in the state @xmath8 , where @xmath9 . then , entanglement between each pair of nearest - neighbor qubits is established by the controlled-@xmath0 gate @xmath10 here , the indices @xmath7 and @xmath11 stand for the qubits at the lattice site ` @xmath7 ' and its nearest - neighbor lattice site ` @xmath11 ' of the graph @xmath5 , respectively . a unitary gate of this kind can be generated by turning on the ( controlled ) ising - type nearest - neighbor interaction for an appropriately chosen time period . experimentally , graph states have been generated by using controlled collisions between cold atoms in optical lattices @xcite and by using linear optics @xcite . mathematically , quantum correlations among the qubits of a graph are specified by correlation operators @xmath12 s , which are given below . the resultant graph state @xmath13 is an eigenstate of these operators , and it is completely specified by the set of eigenvalue equations @xmath14 with the set of eigenvalues @xmath15 . here , @xmath16 stands for the set of all nearest - neighbor qubits which are entangled ( connected ) to the qubit ` @xmath7 ' by the @xmath17 operations . for every qubit ` @xmath7 ' of the graph state @xmath13 , there exists a correlation operator @xmath12 and an eigenvalue @xmath9 . the physical meaning of eq . ( [ eigenvalue ] ) is this : there exists either a correlation ( @xmath18 ) or an anti - correlation ( @xmath19 ) between the outcome of the measurement on qubit ` @xmath7 ' in the @xmath6 eigenbasis and the outcomes of the measurements on all the qubits of @xmath16 in the @xmath0 eigenbasis . these quantum correlations provide the framework for the computation in mqcm . once the resource graph state is ready , then the logical qubits holding the input information are attached to the resource via the same entangling operations given by eq . ( [ cz ] ) . unlike the quantum error - correction , here , one logical qubit stands for one physical qubit . now , the computation is carried out by a sequence of single - qubit ( adaptive ) projective measurements in a certain direction of the bloch sphere and in a certain temporal order . the bloch sphere offers an adequate geometrical description for the direction of single - qubit projective measurement , where the direction of measurement is completely specified by the bloch vector @xmath20 and the corresponding projector is given by @xmath21 the measurement outcomes @xmath22 and @xmath23 mean that the measured qubit is projected onto the states with the kets @xmath24 and @xmath25 respectively . in other words , the choice of measurement basis is characterized by the direction of measurement @xmath26 in the bloch sphere . there are three kinds of measurements in the mqcm @xcite . measurements along the @xmath1-axis effectively detach the measured ( redundant ) qubits from the graph state . measurements along the bloch vector @xmath27it lies in @xmath28,@xmath29 plane of the bloch sphere process the information as well as teleport it from one place to another on the graph . the significance of this kind of measurements is revealed by the example given in appendix [ sec : sqr ] . measurements along the bloch vector @xmath30it lies in @xmath1,@xmath29 plane of the bloch sphere only process the information . this remark will be illustrated by the example given below . the two measurement outcomes @xmath31 for every qubit of the graph state are equally probable because the reduced density matrix for each qubit is the completely mixed state @xmath32 . in the process of getting the desired operations on the logical qubits , one also gets some additional operations . these additional operations are called `` byproduct operators , '' and they belong to the pauli group . these byproduct operators depend on the random measurement outcomes and the eigenvalues of the graph state @xmath13 . the measurement outcome @xmath33 for every graph qubit and the eigenvalues @xmath34 are binary numbers , so one can record them classically in order to take care of the byproduct operators . the classical information processing of this data makes the computation deterministic and helps to set the measurement bases for the subsequent measurements . section [ sec : ci - mqcm ] contains a comprehensive discussion about this matter . right now we are content with illustrating the procedure of mqcm with an example . @xmath35 the unitary operation for the @xmath36-qubit rotation around the @xmath1-axis is @xmath37 where the superscripts @xmath38 symbolize the logical qubits on which this operation will be carried out @xcite . one can accomplish this operation by performing a single measurement on a @xmath39-qubits star graph state , the associated star graph is shown in fig . [ fig:1](i ) . in fig . [ fig:1](i ) , the input quantum register of @xmath36 qubits is displayed by the circles and the ancilla qubit ` @xmath7 ' by the diamond . the input register is in a @xmath36-qubit input state @xmath40 , and the ancilla qubit is prepared in the state @xmath8 . then we perform @xmath36 @xmath17 operations , represented by the bonds in the figure and given by eq . ( [ cz ] ) , between the qubit ` @xmath7 ' and every logical qubit . in principle , all the @xmath17 operations can be performed in one shot , " because they all commute with each other . this series of steps leads us to the resultant star graph state @xmath41 . \label{g2}\end{aligned}\ ] ] the subscript @xmath42 reveals that the final graph state is of one ancilla qubit and @xmath36 logical qubits . a measurement on the ancilla qubit ` @xmath7 ' in the basis @xmath43 transforms the input state of the quantum register into the output state @xmath44 here , the direction of measurement lies in @xmath1,@xmath29 plane of the bloch sphere , and @xmath45 is the measurement outcome . @xmath46 is the byproduct operator , which is represented by the dotted - boxes on all the logical qubits in fig . [ fig:1](ii ) . after the measurement , all bonds ( illustrated in fig . [ fig:1](i ) ) are broken and the qubit ` @xmath7 ' gets projected either onto the state @xmath47 ( if @xmath48 ) or onto the state @xmath49 ( if @xmath50 ) . we shall use this kind of multi - qubit rotations for the hqcm in sec . [ sec : hqcm ] . in contrast to the example @xmath51 given in appendix [ sec : sqr ] where the qubits used for input and output are different , in the case of @xmath52 the input and output states reside in the same @xmath36 logical qubits . in other words , here the information gets processed , but does not get transferred from one place to another . as a side remark , the resultant byproduct operator @xmath53 in the case of @xmath51 ( see eq . ( [ out1 ] ) ) and the measurement basis @xmath43 in the case of @xmath52 depend on the eigenvalue @xmath54 . every quantum gate from the generating set of the clifford group the @xmath55 gate @xmath56 ( the labels @xmath7 and @xmath11 are for the control and target qubits , respectively ) , the hadamard gate @xmath57 the @xmath58-phase gate @xmath59can be executed in a single time step in the mqcm @xcite . this holds because every measurement in these cases is performed either in the @xmath6 eigenbasis or in the @xmath60 eigenbasis and is not influenced by the result of any other measurement . hence all the measurements can be performed simultaneously . a @xmath55 gate can be achieved with a 15-qubit graph state , and both the hadamard gate and the @xmath58-phase gate can be implemented with a chain of five qubits graph state @xcite . single - qubit and @xmath55 gates together constitute a universal set of gates , and they are realizable in the mqcm . in this sense , the mqcm is also universal like the uqcm . in order to simulate a complex unitary gate in the mqcm , it is customary to first decompose it efficiently into a sequence of elementary gates from the universal gate set . then , the temporal order of gates is transformed into the spatial pattern of measurement bases for the graph qubits . afterwards , the measurements are performed in the required order . up to now , we were dealing with the individual gate simulations only , where we need not to worry about the byproduct operators . but in the next section , our focus shall be on the simulation of a sequence of gates , where the study of classical information processing and the temporal order of measurements become necessary . classical information processing is needed for taking care of the byproduct operators . a comprehensive discussion about it is provided in the following section . this section serves as a summary of the results which were discussed in ref . @xcite . when a sequence of gates is simulated in the mqcm , the byproduct operator which originates from the implementation of gates `` passes through '' the sequence . the propagation of the byproduct operator either transforms the next gates in the sequence or the byproduct operator in itself gets transformed . the first part of this section is about propagation relations for some elementary gates , and in the second part we shall define an information flow vector . the third part which concludes this section is reserved for the propagation matrices for some elementary gate based on their propagation relations . we shall use a portion of this section for the hqcm in sec . [ sec : hqcm ] . the structure of the byproduct operator on the logical qubit @xmath61 is @xmath62 , where @xmath63 and @xmath64 are non - negative integers . both @xmath63 and @xmath64 depend on the outcomes of measured qubits and the eigenvalues @xmath34 @xcite . their dependence on @xmath34 is in our control . for example , the @xmath34 dependent parts disappear from the calculation by preparing a graph state with @xmath65 for all the graph qubits . but we can not control the dependence of the byproduct operators on the measurement outcomes which are intrinsically random . in ref . @xcite , the authors took @xmath66 , but we find it simpler to take both @xmath63 and @xmath64 as non - negative integers . this is permissible because in @xmath67 and @xmath68 only the modulo-2 values of @xmath63 and @xmath64 matter . throughout the paper , we reserve the sign ` + ' for the ordinary addition and the sign ` @xmath69 ' for the modulo-2 addition . in principle , we can correct the byproduct operators step by step after completing each gate of the sequence under simulation . but it is more convenient to choose not to correct them and let them pass through the gates , and just keep track of the measurement outcomes in a systematic way using simple classical information processing . at the end of the computation , either we set the measurement bases for the final readout depending on the history of outcomes or we just perform the final measurements in the computational basis and interpret the result with the help of the record of measurement outcomes . the propagation of the byproduct operator through a gate is given by the propagation relation . the euler decomposition of an arbitrary single - qubit rotation @xmath70 is @xmath71 and the propagation relation for @xmath72 is @xmath73 an arbitrary single - qubit rotation @xmath72 gets transformed to @xmath74 , but the byproduct operator stays as it is . we can take eq . ( [ pror ] ) as an illustration of the importance of `` the temporal order of the measurements . '' this is because , when @xmath72 is a part of a circuit , the superscripts @xmath28 and @xmath1 are functions of the earlier measurement outcomes . in order to determine the right sign for the measurement angles @xmath75 , @xmath76 , and @xmath77 , we have to wait until the necessary measurements are completed @xcite . equation ( [ pror ] ) also justifies the following points . the measurement directions for these qubits lie in @xmath28,@xmath29 plane of the bloch sphere , @xmath78 with @xmath79 , their measurement bases depend on the results of previous measurements ( see appendix [ sec : sqr ] and ref . when @xmath80 , then the directions for @xmath81 and @xmath82 coincide and do not get influenced by the outcomes of other measurements . measurements of this kind are either in the @xmath6 ( @xmath83 ) or the @xmath60 ( @xmath84 ) eigenbasis . the gates from the generating set of the clifford group ( given by eqs . ( [ cnot])([rz ] ) ) are realized by such measurements . the propagation relation for the gate @xmath85 is @xmath86 where @xmath87 and @xmath88 in case of the @xmath85 gate , eq . ( [ procnot ] ) , the gate stays as it is , but the byproduct operator @xmath89 gets transformed to @xmath90 . this is also the case for the other two gates from the generating set of the clifford group . the propagation relation for the hadamard gate @xmath91 is @xmath92 and for the @xmath58-phase gate @xmath93 it is ( up to a phase factor @xmath94 ) @xmath95 the propagation relations ( [ procnot ] ) , ( [ proh ] ) , and ( [ prophase ] ) can also be understood from the definition of the clifford group which maps the pauli group into itself under conjugation . now , let us define an information flow vector . at every stage of the computation , the accumulated byproduct operator @xmath96 upon the logical qubits @xmath97 is of the form @xmath98 . after the implementation of a gate , only the values @xmath99 and @xmath100 get changed , and the new values determine the measurement bases for the subsequent gates . these values are processed by a classical computer . there is a one - to - one correspondence between the byproduct operator @xmath96 ( ignoring the global phase @xmath101 ) and a @xmath102-component `` information flow vector '' @xmath103 which is given as follows : @xmath104 where @xmath105 here , the multiplication of byproduct operators ( up to a phase factor @xmath101 ) corresponds to the component - wise addition of information flow vectors . the information flow vector @xmath106 keeps track of the sign(s ) of the measurement angle(s ) for a gate . in accordance with eq . ( [ pror ] ) , the signs of the measurement angles for the operation @xmath107 on the qubit @xmath108 are determined by the current value of @xmath63 and @xmath64 in @xmath106 . the propagation relations ( [ procnot ] ) , ( [ proh ] ) , and ( [ prophase ] ) suggest that none of the gates from the generating set of the clifford group gets altered under the propagation of the byproduct operator . so , the measurement angles for these gates are independent of the values stored in @xmath106 . we can also define a @xmath109 `` propagation matrix '' @xmath110 for a gate @xmath111 , representing the transformation in the information flow vector when the corresponding byproduct operator passes through the gate @xmath111 . the propagation matrices given below are derived from the propagation relations ( [ pror ] ) , ( [ procnot ] ) , ( [ proh ] ) , and ( [ prophase ] ) with the help of the one - to - one correspondence given by eq . ( [ i ] ) , and the entries in the information flow vectors and the propagation matrices are given only for relevant qubits . for the case of @xmath36 logical qubits , the propagation matrices for the @xmath70- , @xmath55- , @xmath91- , and @xmath93-gate are given in appendix [ sec : pm ] . the byproduct operator passes through an arbitrary single - qubit rotation @xmath72 without getting transformed . so the information flow vector stays as it is , @xmath112 the information flow vector gets transformed when the associated byproduct operator passes through the @xmath85 gate in the following way : @xmath113 under the one - to - one correspondence given by eq . ( [ i ] ) , the propagation relation ( [ proh ] ) for the hadamard gate @xmath91 becomes @xmath114 and the propagation relation ( [ prophase ] ) for the @xmath58-phase gate @xmath93 becomes @xmath115 the information flow vector and the propagation matrices provide a simple description , and they are easily handled by a classical computer . as a side remark , the temporal order of the measurements does not typically follow the temporal order of gates in a circuit which we want to simulate with the mqcm . in the mqcm , there exists an efficient measurement scheme where measurements are performed round by round , and in each round all the measurements are executed at the same time @xcite . the information flow vector is updated after every round . after the final round , the result of the computation is interpreted from the @xmath28-part of the information flow vector @xmath116 . an extended discussion of this efficient measurement scheme is provided in appendix [ sec : ems ] . in this section , we formulate the methodology of the hybrid quantum computation model ( hqcm ) by combining the advantages of the uqcm and the mqcm . before going into the details , let us first focus on what benefits we can get from each of the models in different situations . here , we consider the preparation of a graph state , the set of elementary gates for the hqcm , and the simulation of a quantum circuit with the hqcm one by one . the very first experimental step in the mqcm is the preparation of a resource graph state , whereas in the uqcm no such preparation of a resource is needed . while preparing a graph state , in principle , the initialization of every graph qubit in the @xmath6 eigenbasis can be completed in one shot . to this end , we have to `` talk '' to every graph qubit simultaneously . consequently , this requires a lot of experimental resources , and the very many interactions are difficult to control . likewise , the subsequent two - qubit entangling operations ( @xmath117 s given by eq . ( [ cz ] ) ) to create the resource graph state can be performed in one step , because they commute with each other . and , the larger the graph state , the more difficult it is to prepare and control the state and to protect it against decoherence . so , for the hqcm , we choose not to prepare the whole two - dimensional universal graph state at once , but instead prepare small ( non - universal ) graph states step - by - step as we need them when the computation progresses . only the star graph states , such as @xmath118 given in eq . ( [ g2 ] ) , are required for the hqcm . we choose single - qubit operations , the @xmath17 gate and the multi - qubit rotation around the @xmath1-axis @xmath52 for an arbitrary value of @xmath119 given by eq . ( [ nr ] ) as the elementary gates for hqcm . in analogy to the procedure for the uqcm , first we `` efficiently decompose '' any big unitary gate we are trying to simulate into a sequence of elementary gates in such a way that the number of elementary gates grows polynomially with the number of logical qubits , and then every elementary gate is implemented one after another . every single - qubit operation and the @xmath17 gate are carried out by the unitary evolution under the formalism of uqcm . the rotation @xmath52 is implemented by the method given in sec . [ sec : methodology1qc ] ( see the example ) under the formalism of mqcm . the motivation behind these choices is explained in the following . the implementation of an arbitrary single - qubit rotation in the mqcm costs us at least a chain of five qubits graph state and four measurements @xcite . but it can be realized quite simply by the unitary evolution of the respective single qubit . furthermore , the euler decomposition for an arbitrary single - qubit rotation @xmath72 given in eq . ( [ ro ] ) is not needed . the @xmath17 operations themselves are part of the experimental setup for constructing the graph states , and for this we have to execute them by the unitary evolution . that is why we consider the @xmath17 gate as an elementary gate for the hqcm . furthermore , it is more economical to implement @xmath85 by the unitary evolution @xmath120 instead of first preparing a 15-qubit graph state and then implement it with the mqcm @xcite . although we already have the universal set of gates ( single - qubit and @xmath17 gates ) , we prefer to include @xmath52 as an elementary gate in the hqcm . this is because of two reasons . the first reason is the optimization . the resource @xmath39-qubit graph state @xmath118 ( given by eq . ( [ g2 ] ) ) needed for the implementation of @xmath52 is relatively easy to create experimentally . it has only one ancilla qubit , and the entanglement can be established in one shot . furthermore , a single measurement on the ancilla qubit is enough to realize @xmath52 all together on @xmath36 logical qubits . while it is also possible to decompose the rotation @xmath52 in terms of the gates from the universal gate set and implement it under the formalism of uqcm , its implementation there will not be so optimal , and we can not regard it as a single unit . generally , one is using either unitary evolution ( uqcm ) or measurements on the graph state ( mqcm ) in order to simulate a quantum circuit . so , the second reason for including @xmath52 as an elementary gate in the hqcm is to investigate a model of qc in which a part of a circuit ( @xmath52 rotations ) is simulated by the measurements and the rest by the unitary evolution , and to see how the classical information processing plays its role in such a model . the importance of looking at the classical information - processing parts is explained in the following . now let us consider the simulation of a sequence of gates with the hqcm in which the classical information processing becomes crucial . the classical processing does not come into the picture of uqcm where the measurements are used only for the readout of the final result of computation . in all those schemes where measurements are needed for the computation ( for example quantum teleportation @xcite ) , the classical information processing in parallel is essential . in the hqcm also , classical information processing is needed , because the rotations @xmath52 are executed by the measurements . but here the classical information - processing parts are rather simple and straightforward , requiring only the information flow vector and the propagation matrices . a comprehensive discussion of this is given in the following section . in this section , our focus will be on the classical information - processing parts of the hqcm . we only need the information flow vector and the propagation matrices for the elementary gates . we first redefine the information flow vector in the context of hqcm , and then discuss the propagation relations as well as the propagation matrices for the elementary gates . at every computation step @xmath121 , the form of the byproduct operator is the same , as given in eqs . ( [ i ] ) and ( [ byprombqcm ] ) , @xmath122 so , the form of the related information flow vector @xmath123 is also the same as given by the @xmath124 column vector in eqs . ( [ i ] ) and ( [ iso ] ) . but in the hqcm , there are some differences in comparison to the efficient measurement scheme of mqcm given in appendix [ sec : ems ] . in that scheme of mqcm , the index ` @xmath125 ' of @xmath126 stands for ` the measurement round . ' but in the hqcm , every elementary gate is taken as a single computational step , and the index ` @xmath121 ' of @xmath127 is the label for them . in the mqcm , @xmath126 gets updated after each round , but in the hqcm it is updated after each gate . in this scheme of mqcm , the initial value of the information flow vector @xmath128 is determined by the set of eigenvalues @xmath34 and by some particular gates . but in the hqcm , just before starting the computation all the entries of @xmath129 are zeros , i.e. , both @xmath130 and @xmath131 for all @xmath132 . that means that the byproduct operator at @xmath133 is the identity operator @xmath4 on every logical qubit . in fact , the first relevant byproduct operator appears in the computation when the first multi - qubit rotation is implemented , and then the information flow vector gets some nonzero entries . in the mqcm , the information flow vector gets updated from @xmath134 to @xmath126 after the @xmath125th measurement round . in the hqcm , the information flow vector gets updated from @xmath135 to @xmath127 after the implementation of @xmath121th gate . @xmath127 influences the @xmath136th gate of a quantum circuit under simulation . similar to the case of the uqcm , the total number of computation steps ( the logical depth ) is denoted by @xmath137 , which is the total number of elementary gates used for the computation . furthermore , @xmath137 is also the total number of steps taken by a classical computer for the classical information processing in the hqcm . let us turn to the issue how one can interpret the final result of the computation in the hqcm . in the case of uqcm , every gate of a circuit is executed by their respective unitary evolution , and the final readout measurements are performed in the computational basis . in this case , the output state @xmath138 gets projected onto the state @xmath139 after the final readout measurements , @xmath140 where @xmath141 are the readout measurement outcomes for the logical qubits @xmath142 . in the case of hqcm , the final state of the output register will be @xmath143 after performing the last gate of the same circuit . without loss of generality , like above , we consider the computational basis for the final readout , where @xmath144 are the readout measurement outcomes for the logical qubits @xmath142 . that means that the output state @xmath143 gets projected onto the state @xmath145 after the readout measurements , @xmath146 we can transform eq . ( [ outh ] ) with the help of eq . ( [ byprohqcm ] ) into @xmath147 the inference we get by comparing eq . ( [ outu ] ) and eq . ( [ out2h ] ) is : the readout measurements on the state @xmath138 with the results @xmath148 give the same circuit - output as the readout measurements on the state @xmath143 with the results @xmath149 , and these sets of results are related by @xmath150 that is how one can interpret the final result of the computation with the help of @xmath151 in the hqcm . let us turn to the propagation relations for the elementary gates . an arbitrary single - qubit rotation around an axis @xmath152 ( defined in eq . ( [ r ] ) ) by an angle @xmath75 is @xmath153 the byproduct operator passes through this gate without any change , but it changes the axis of rotation of the gate from @xmath154 to @xmath155 . the propagation relation for @xmath156 is given by @xmath157 where @xmath158 in other words , the angles @xmath119 , @xmath159 that define the axis of rotation @xmath154 get transformed as @xmath160 and @xmath161 . the byproduct operator passes through @xmath156 without getting transformed , that means that the propagation matrix @xmath162 is the same identity matrix as given in eqs . ( [ cr ] ) and ( [ ccr ] ) . every single - qubit unitary operator in su(2 ) follows this propagation relation , and it becomes the propagation relation ( [ proh ] ) for the hadamard gate when @xmath163 and the propagation relation ( [ prophase ] ) for the @xmath58-phase gate when @xmath164 . so , the hadamard- and the @xmath58-phase - gate remain special cases in the sense that the byproduct operator changes under the propagation but not these gates . they are executed by the unitary evolution like any other single - qubit gate , but for the classical information - processing parts of the hqcm we shall use their propagation matrices given by eqs . ( [ cch ] ) and ( [ ccphase ] ) . the propagation relation for the next elementary gate @xmath117 ( defined in eq . ( [ cz ] ) ) is @xmath165 where @xmath166 and @xmath167 under the one - to - one correspondence given in eq . ( [ i ] ) the propagation relation ( [ hqcprocz ] ) becomes @xmath168 when the control qubit ` @xmath7 ' and the target qubit ` @xmath11 ' belong to the set of @xmath36 logical qubits @xmath169 , then the propagation matrix @xmath170 can be generated by the following relations : @xmath171_{kl}&= \left[\textbf{c}_{zz}(\textsc{cz}(a , b))\right]_{kl}=\delta_{kl},\nonumber\\ \left[\textbf{c}_{xz}(\textsc{cz}(a , b))\right]_{kl}&=\delta_{ka}\delta_{lb}+\delta_{kb}\delta_{la},\nonumber\\ \left[\textbf{c}_{zx}(\textsc{cz}(a , b))\right]_{kl}&= 0.\label{cccz}\end{aligned}\ ] ] note that eqs . ( [ ccz ] ) and ( [ cccz ] ) are different from eqs . ( [ ccnot ] ) and ( [ cccnot ] ) . the @xmath17 and @xmath55 gates are interconvertible by using the hadamard gate , i.e. , @xmath172 and the same is true for their propagation matrices , i.e. , @xmath173 the propagation relation for @xmath52 is @xmath174 where @xmath96 is the same as given in eq . ( [ i ] ) and @xmath175 in this case , the measurement angle @xmath119 gets modified under the propagation , but the byproduct operator stays as it is . so , the propagation matrix @xmath176 will be the identity matrix , which can be defined in the same way as the @xmath162 in eq . ( [ ccr ] ) . the inferences we get here are the following : ( 1 ) the hadamard- , the @xmath58-phase- and the @xmath17-gate remain unchanged under the propagation , while the byproduct operator gets altered . ( 2 ) the single- and multi - qubit rotations ( with nontrivial angles ) get transformed , while the byproduct operator stays unaltered under the propagation . now we have all the basic tools for the hqcm . we shall first describe some important examples , and then proceed to the implementation of grover s search algorithm within the hqcm . in this section we are considering @xmath36-qubit controlled rotations around the @xmath1-axis , which are defined by @xmath177 where the qubits labeled 1 to @xmath178 are the control qubits and the qubits labeled @xmath179 to @xmath36 are the target qubits . only if every control qubit is in the state @xmath180 , then the @xmath181-qubit rotation @xmath182 operates on the target qubits . we discuss the hqcm implementation for three different values of @xmath178 : @xmath178 = 1 ( single - control ) , @xmath178 = 2 ( double - control ) and @xmath183 ( triple - control ) . first , we decompose @xmath184 in terms of multi - qubit rotations like @xmath52 given by eq . ( [ nr ] ) . in order to have the logical qubit 1 as the control and the rest as the target qubits , we express the @xmath36-qubit rotation about the @xmath1-axis as @xmath185 consequently , we get the required decomposition : @xmath186 we observe that @xmath52 is symmetric under permutation of the qubits , so we can take any qubit as the control and the remaining qubits as targets . we consider a ` multi - qubit rotation about the @xmath1-axis ' as a single unit , and then @xmath184 costs only two units of this kind . the circuit representation of eq . ( [ c1u ] ) is given in fig . [ fig:2 ] . in fig . [ fig:2](i ) , the first and the second rectangular boxes depict @xmath52 ( the first rotation ) and @xmath187 ( the second rotation ) , respectively . in practice , both of them are realized by using a single ancilla qubit and a single measurement with the methodology given in sec . [ sec : methodology1qc ] ( see the example ) . moreover , after executing the first rotation we bring back the ancilla qubit into an eigenstate of @xmath6 and use it for the second rotation . the implementation of @xmath52 and @xmath187 require @xmath39-qubit and @xmath36-qubit star graph states , respectively ( see fig . [ fig:1](i ) ) , where the ancilla qubit is connected to the relevant logical qubits . the eigenvalues of the ancilla qubit corresponding to the first and the second rotation are @xmath188 and @xmath189 , and the measurement outcomes are @xmath190 and @xmath191 , respectively . the classical information processing for this case has three main parts . the first part deals with the change in the measurement angles due to the byproduct operator @xmath192 . @xmath192 appears just before implementing the first rotation and is denoted by the dashed vertical line at the input section in fig . [ fig:2 ] . when the gate @xmath184 in itself is a part of a circuit under simulation , then due to the implementation of previous gates the byproduct @xmath192 has emerged prior to the execution of @xmath184 . without loss of generality , we take @xmath193 the same as given in eq . ( [ i ] ) . only the @xmath28-part of the corresponding information flow vector @xmath194 influences the measurement bases @xmath195 for both rotations . according to eq . ( [ hqcprou ] ) , the angles @xmath119 for the first and @xmath196 for the second rotation get altered . @xmath197 where @xmath28 is given by eq . ( [ x ] ) . the second part deals with the eigenvalues @xmath188 and @xmath189 , which influence the azimuthal angle @xmath58 of the measurement bases in the following way : @xmath198 the third part manages the contribution of measurement outcomes @xmath190 and @xmath191 to the byproduct operator @xmath199 . the implementation of both the first and the second rotation cause the byproduct operators @xmath200 and @xmath201 on the relevant logical qubits . furthermore , these byproduct operators update @xmath199 to @xmath202 . @xmath202 is denoted by the dashed vertical line at the output section in fig . [ fig:2 ] . only the @xmath1-part of the information flow vector @xmath203 gets changed , while the @xmath28-part remains as it is , i.e. , @xmath204 , @xmath205 we have to combine two additional units @xmath207 ( the third rotation ) and @xmath208 ( the fourth rotation ) with @xmath184 for the purpose of getting the gate @xmath206 . in other words , @xmath206 with two control qubits 1 and 2 , is made up of four rotations , and its decomposition is given by @xmath209 figure [ fig:3](i ) illustrates the temporal ordering of the multi - qubit rotations given in eq . ( [ c2u ] ) by the rectangular boxes . the treatment for @xmath206 is similar to that of sec . [ sec : c1 ] . all the rotations@xmath52 ( first ) , @xmath187 ( second ) , @xmath207 ( third ) , and @xmath208 ( fourth)are performed one after another under the scheme given in sec . [ sec : methodology1qc ] ( see the example ) . after initializing the ancilla qubit in the @xmath6 eigenbasis , we prepare a necessary star graph state for the first rotation , and then the ancilla qubit is measured in the appropriate basis . the measurement outcome is recorded , and the ancilla qubit is brought back again into an eigenstate of @xmath6 ( recycled ) for executing the next rotation . in this way , we can use the same ancilla qubit for all the rotations . @xmath188 , @xmath189 , @xmath210 , and @xmath211 are the eigenvalues of the ancilla qubit , and @xmath190 , @xmath191 , @xmath212 , and @xmath213 are the measurement outcomes corresponding to the first , second , third , and fourth rotation . as a side remark , one can also choose to perform these four rotations at the same time by using four different ancilla qubits , but this would require more hardware resources . we can take @xmath206 as a single time step , because , in principle , all the four rotations can be executed at the same time . the classical information processing for this case also has three main parts . the first part deals with the modification in the measurement angles because of the byproduct operator @xmath214 , which is represented by the dashed vertical line at the input section in figs . [ fig:3](i ) and [ fig:3](iii ) . here also , only @xmath215 influences the measurement angle @xmath216 for every rotation . @xmath217 where @xmath28 is the same given by eq . ( [ x ] ) . the second part manages the influence of the eigenvalues @xmath218 and @xmath211 on the azimuthal angle @xmath58 of the measurement bases @xmath195 in the following way : @xmath219 the third part handles the random measurement outcomes @xmath190 , @xmath191 , @xmath212 , and @xmath213 ; which cause the byproduct operators @xmath220 , @xmath221 , @xmath222 , and @xmath223 , respectively , on the relevant logical qubits . furthermore , they change @xmath199 into @xmath202 by their contribution . @xmath202 is denoted by the dashed vertical line at the output section in figs . [ fig:3](i ) and [ fig:3](iii ) . so , only the @xmath1-part of the corresponding information flow vector @xmath203 gets changed to @xmath224 , while @xmath204 . @xmath225 where @xmath226 we emphasize two points . first , @xmath206 is symmetric under permutation of the qubits . so , we can take any two qubits as controls and the rest of the qubits as targets by using only four multi - qubit rotations . if we continue along the same direction , then the controlled rotation @xmath227 with @xmath178 control qubits requires @xmath228 units of rotation . the number @xmath228 is independent of the number of target qubits @xmath181 , but when @xmath178 becomes order of @xmath36 , then @xmath228 becomes exponential in @xmath36 . in order to fix this exponential growth problem , we need some extra work qubits @xcite . our next example will justify this remark . second , @xmath206 becomes @xmath229 for @xmath230 . the gate @xmath229 is equivalent to the deutsch s universal gate @xmath231 $ ] @xcite up to single - qubit gates , provided the angle ` @xmath232 ' is incommensurate with @xmath233 . we are going to use @xmath234 with two work qubits for implementing our next gate @xmath235 . we put the bits and pieces of the hqcm together as a summary by taking the gate @xmath235 as an example . the complete scheme about its implementation in terms of its circuit diagram is shown in fig . [ fig:4 ] , and the associated classical information - processing parts are given in table [ table : cip ] . first , we efficiently decompose the gate @xmath235 into a sequence of elementary gates ( single - qubit gates , the @xmath17 gate , and @xmath52 ) . the temporal order of the elementary gates for @xmath235 is depicted by the circuit diagram in fig . [ fig:4 ] , where the qubits 1 , 2 , and 3 act as the control qubits and the qubit 6 as the target qubit , and they all are represented by the black horizontal lines . the work qubits 4 and 5 are initialized in the state @xmath236 , and they are represented by the gray horizontal lines in the figure . the work qubits are used to make the decomposition of @xmath235 economical . the 3-qubit gates @xmath237 ( 1st gate ) , @xmath238 ( 3rd gate ) , @xmath239 ( 7th gate ) , and @xmath240 ( 9th gate ) are represented by rectangular boxes with double - control . every 3-qubit gate is further decomposed into four rotations around the @xmath1-axis according to eq . ( [ c2u ] ) , here @xmath241 and @xmath242 . furthermore , each rotation is executed by preparing a required star graph state , followed by the measurement in the appropriate basis . the detailed methodology is already mentioned in sec . [ sec : c2 ] . the hadamard gates @xmath91 are displayed by rounded rectangles , and the two - qubit gate @xmath243 is shown by the rounded rectangle on the qubit 6 with the qubit 5 as control in fig . [ fig:4 ] . the hadamard and the @xmath243 gates are executed by the unitary evolution . the classical information - processing parts for @xmath235 are handled by a classical computer according to table [ table : cip ] . in this table , the first column is for the computational steps @xmath121 , which are represented by the dashed vertical lines in fig . [ fig:4 ] . there are ten vertical lines in the figure and ten rows in the table for the ten computational steps from @xmath244 to @xmath245 . at each vertical line the information flow vector @xmath127 gets updated . the second and the third columns are reserved for @xmath246 and @xmath247 respectively . if required , the change in the measurement angles for the next gate based on the updated value of @xmath127 is calculated ; they are given in the fourth column . after performing the measurements in the appropriate bases , the measurement outcomes are recorded in the fifth column . @ c c c p5 cm p5 cm @xmath121 & @xmath246 & @xmath247 & angle @xmath216 ( here @xmath248 ) & measurement outcomes + * 0 * & @xmath249 & @xmath249 & @xmath250 & @xmath251 @xmath252 + * 1 * & @xmath249 & @xmath253 & & + * 2 * & @xmath254 & @xmath255 & @xmath256 & @xmath257 @xmath258 + * 3 * & @xmath254 & @xmath259 & & + * 4 * & @xmath260 & @xmath261 & & + * 5 * & @xmath260 & @xmath262 & & + * 6 * & @xmath254 & @xmath263 & @xmath264 & @xmath265 @xmath266 + * 7 * & @xmath254 & @xmath267 & & + * 8 * & @xmath268 & @xmath269 & @xmath270 @xmath271 & @xmath272 @xmath273 + * 9 * & @xmath268 & @xmath274 + [ table : cip ] let us go through the table row by row . before starting the computation ( in the first row @xmath275 ) , all the entries of both @xmath276 and @xmath277 are zeros ( initialization ) . so , there is no change in the measurement angle for each of the four rotations associated with the gate @xmath237 ( 1st gate ) . the measurement outcomes @xmath278 , @xmath279 , @xmath280 , and @xmath281 corresponding to the four rotations @xmath282 , @xmath283 , @xmath284 , and @xmath285 are recorded . these outcomes give some nonzero entries to @xmath286 according to eq . ( [ ivector ] ) . the measurement outcome @xmath287 corresponds to the @xmath288th rotation of the @xmath108th 3-qubit gate . the next gate in the circuit is the hadamard gate @xmath289 , which does not change under the propagation of the byproduct operator ; that is why the forth column of the second row @xmath290 is empty . the @xmath91-gate is realized by the unitary evolution , that is why the fifth column of the second row is also empty . but the @xmath91-gate changes the information flow vector @xmath291 into @xmath292 under the propagation relation given by eq . ( [ proh ] ) , and the propagation matrix for the @xmath91-gate is given by eqs . ( [ ch ] ) and ( [ cch ] ) . the @xmath293rd gate is @xmath238 . the measurement angles @xmath216 only for the rotation @xmath294 and @xmath295 get influenced by @xmath296 according to eq . ( [ thetax ] ) . the measurement outcomes @xmath297 , @xmath298 , @xmath299 , and @xmath300 only transform @xmath301 into @xmath302 . in this way going through table [ table : cip ] along with fig . [ fig:4 ] explains the whole scheme , and the final output result is interpreted according to eq . ( [ sx ] ) with the help of @xmath303 . here , the @xmath1-part of the information flow vector @xmath247 gets the new entries from the implementation of 3-qubit gates only according to eq . ( [ ivector ] ) . the entries of @xmath127 get manipulated under the propagation of the byproduct operator through the hadamard gates and the @xmath243 gate according to the propagation relations ( [ proh ] ) and ( [ hqcprocz ] ) , respectively . but the propagation of the byproduct operator does not change the @xmath91- and @xmath17-gate . the @xmath28-part of the information flow vector @xmath246 influences the measurement angles @xmath216 of the rotations for every 3-qubit gate according to eq . ( [ thetax ] ) . as a side remark , the sign of the azimuthal angle @xmath58 of the measurement bases for the rotations also depends on the eigenvalues of the ancilla qubit according to eq . ( [ phik ] ) , which is not mentioned in the table . we can easily generalize this example up the @xmath36-qubit gate @xmath304 . where the @xmath305 logical qubits @xmath306 are the control qubits and the last one is the target qubit . for implementing this gate we need @xmath307 work qubits , which are initialized in the state @xmath308 . in grover s search algorithm this gate plays a very important role , which we are discussing in the next section . in this section , we illustrate the use of the hqcm in the context of a practical example : grover s search algorithm ( ga ) @xcite . here , we shall see that the @xmath36-qubit gate @xmath304 along with the single - qubit operations is sufficient to perform ga . the best known quantum algorithm , for an unstructured database search , is ga . we have an unstructured database of total @xmath309 items , out of which only one item matches with our query . where a classical computer takes an average @xmath310 steps , a qc with ga takes only @xmath311 steps ( grover s iterations ) to find out the marked item . a `` step '' is a query of the oracle in the current context . a brief description of ga is given as follows : we can recognize each item of our database by a @xmath36-bit string . for simplicity , we choose @xmath312 = the total number of @xmath36-bit strings , and only one string out of @xmath309 is marked . the task of the search problem is to recover the marked @xmath36-bit string in the end of computation . ga begins with the hadamard operation on each qubit , which were initially prepared in the state @xmath313 , and that creates the superposition of the kets for all possible @xmath36-bit strings with equal amplitude . the next step is the implementation of grover s iteration , which is a rotation in effect . it can be decomposed into two reflection operations . the first reflection operator is the oracle @xmath314 , which has the ability to recognize the solution of the search problem . mathematically , @xmath314 gives a conditional phase shift of @xmath233 to the matching string only . the second reflection operator is the diffusion operator @xmath315 , which gives an inversion about the average . like other quantum algorithms , ga is also probabilistic in nature . after @xmath311 steps ( iterations ) the amplitude of the marked string becomes significantly larger than those of the unmarked strings . finally , we read the output by performing measurements on all the qubits . there have been many successful attempts at the implementation of ga , for @xmath316 or @xmath317 , in different physical setups such as with nmr system @xcite , with cavity quantum electrodynamics ( qed ) @xcite , with optics @xcite , and with the mqcm @xcite . now , let us take a look at the structure of these two reflection operators . if only one out of @xmath309 item matches with our query ; then there exist @xmath309 different @xmath314 operations ( one for each item ) . mathematically , the oracle corresponds to the @xmath108th item can be written as @xmath318 the case of @xmath319 , @xmath320 , corresponds to @xmath321 and is nothing but the @xmath36-qubit gate @xmath304 . we already discussed its implementation with the hqcm for @xmath322 in sec . [ sec : c3 ] , and its circuit diagram is shown in fig . [ fig:4 ] . any other oracle can be derived by performing the gate(s ) @xmath323 on the relevant qubit(s ) before and after performing the gate @xmath304 . for example , the oracle associated with the item @xmath324 can be derived as @xmath325x^{\otimes n}. \label{o1}\ ] ] so the gate @xmath304 is used for implementing every oracle . the information , which of the oracle is executed by the black box ? " , is hidden to us . in other words , we do not know on which qubit(s ) the black box is implementing the @xmath6 gate(s ) along with @xmath304 . the mathematical structure of the diffusion operator is @xmath326x^{\otimes n}h^{\otimes n}. \label{d}\ ] ] it can also be constructed by performing the hadamard- and the @xmath6-gate on every logical qubit before and after performing the gate @xmath304 . in summary , the gate @xmath304 together with the hadamard and the @xmath6 gates is sufficient to realize ga in the framework of hqcm . we have established the hqcm , which is a hybrid model of the mqcm and the uqcm , where at first , a big unitary gate under simulation is decomposed into a sequence of elementary gates . the elementary gates in the hqcm are an arbitrary single - qubit gate , the @xmath17 gate and the multi - qubit rotation around the @xmath1-axis . every single - qubit gate and the @xmath17-gate are realized by a unitary evolution . every multi - qubit rotation is executed by preparing a respective star graph state followed by a single measurement . the hqcm is a model where a portion of the quantum circuit is simulated by the unitary evolution , and the rest is by the measurements . the choice of elementary gates is governed by the experimental easiness in terms of resources and computational steps . the implementation of an arbitrary single - qubit gate with the unitary evolution is straightforward in comparison with its implementation with the mqcm . the @xmath17-gate in itself is the part of experimental setup for creating the graph states . the star graph states for the multi - qubit rotations can be realized in one shot , and a single measurement is enough for executing these rotations . the classical information processing in the hqcm is very simple in comparison with the mqcm . in the hqcm , only the @xmath102-component information flow vector and the propagation matrices for the elementary gates are needed for the classical information processing , and the total number of steps are taken by a classical computer in parallel for doing this is the total number of gates in a quantum circuit under simulation . furthermore , no preprocessing and no additional computational steps are required for classical information processing in the hqcm . we also have shown that how one can realize efficiently the multi - control gates ( like @xmath304 ) with the hqcm , which play a very important role in the implementation of ga . the gate @xmath304 together with the hadamard and the @xmath6 gates is sufficient to realize ga in the framework of hqcm . one can carry on the investigation in the following directions . in addition to the multi - qubit rotations in the set of elementary gates for the hqcm , one could include some more gates which can be executed in one shot by the mqcm without adding further complications in the model in the set of elementary gates . further , implementation errors need to be considered for any practical realization . in order to establish the fault - tolerance version of hqcm , one needs to design the elementary gates in a fault - tolerant manner but this is beyond the scope of the present paper . in this appendix , we discuss the implementation of the single - qubit rotation around the @xmath1-axis @xmath327 with the mqcm @xcite . a two - qubit graph state corresponding to the graph depicted in fig . [ fig:5](i ) is sufficient for accomplishing the job . the logical qubit 1 ( represented by the circle ) is in a general input state @xmath328 , and this is the single - qubit state on which we want to apply @xmath51 . in order to generate the required graph state , we prepare the qubit @xmath7 ( represented by the diamond ) in the state @xmath8 . then both the qubits are connected by the operation @xmath17 , which is represented by the bond in fig . [ fig:5](i ) and given by eq . ( [ cz ] ) . the resulting graph state @xmath329 \label{g1}\ ] ] is ready for the computation . here , the subscript 1 + 1 indicates that this graph state is made of two qubits , the logical qubit 1 and the ancilla qubit @xmath7 . in order to generate the desired effect on the input state , we measure the qubit 1 in the basis @xmath330 and the value of @xmath190 is the result of the measurement . after the measurement , the output state ( up to a global phase ) @xmath331 is obtained from the qubit @xmath7 , and the qubit 1 gets projected either onto the state @xmath332 ( if @xmath333 ) or onto the state @xmath334 ( if @xmath335 ) . the net effect on the input state is the required operation @xmath51 followed by the hadamard gate @xmath336 ( represented by the boxes in fig . [ fig:5](ii ) ) and the byproduct operator @xmath53 ( represented by the dotted - box in fig . [ fig:5](ii ) ) . here , the direction of measurement lies in @xmath28,@xmath29 plane of the bloch sphere , and the input information is not only teleported from one lattice site to other but also gets processed by the measurement . as we know , every rotation in the bloch sphere corresponds to a single - qubit operation ( up to a global phase ) . owing to the euler decomposition of an arbitrary rotation @xmath337 , where @xmath338 , one can generate an arbitrary single - qubit operation with a chain of five qubits graph state with the measurement direction for each qubit ( the angles @xmath339 ) in @xmath28,@xmath29 plane of the bloch sphere @xcite . the realization of an arbitrary rotation by such a sequence of two @xmath1 rotations sandwiching an @xmath28 rotation illustrates the importance of the temporal ordering of the measurements in the mqcm . in this appendix , the propagation matrices for the @xmath70- , @xmath55- , @xmath91- , and @xmath93-gate for the case of @xmath36 logical qubits are given . the propagation matrix @xmath340 is a @xmath109 matrix of the form @xmath341 , where @xmath342 and @xmath343 are @xmath344 matrices with binary - valued entries @xcite . one can generate the propagation matrices for an arbitrary single - qubit rotation @xmath345 on the logical qubit ` @xmath108 ' with @xmath346_{kl}&= \left[\textbf{c}_{zz}(r^{(j)})\right]_{kl}=\delta_{kl},\nonumber\\ \left[\textbf{c}_{zx}(r^{(j)})\right]_{kl}&= \left[\textbf{c}_{xz}(r^{(j)})\right]_{kl}=0;\label{ccr}\end{aligned}\ ] ] here , @xmath347_{kl}$ ] stands for the entry in @xmath288th row and @xmath348th column of the matrix @xmath349 corresponding to the @xmath345 gate . the same notation apply for the propagation matrices defined below . the propagation matrix for the @xmath85 gate ( both the control qubit ` @xmath7 ' and the target qubit ` @xmath11 ' belong to the set of @xmath36 logical qubits ; @xmath350 ) is given by @xmath351_{kl}&= \delta_{kl}+\delta_{kb}\delta_{la},\nonumber\\ \left[\textbf{c}_{zz}(\textsc{cnot}(a , b))\right]_{kl}&=\delta_{kl}+\delta_{ka}\delta_{lb},\nonumber\\ \left[\textbf{c}_{zx}(\textsc{cnot}(a , b))\right]_{kl}&=\left[\textbf{c}_{xz}(\textsc{cnot}(a , b))\right]_{kl}= 0;\label{cccnot}\end{aligned}\ ] ] for the hadamard gate @xmath352 on the logical qubit ` @xmath108 ' is given by @xmath353_{kl}&= \left[\textbf{c}_{zz}(h^{(j)})\right]_{kl}=\delta_{kl}\oplus\delta_{kj}\delta_{lj},\nonumber\\ \left[\textbf{c}_{zx}(h^{(j)})\right]_{kl}&= \left[\textbf{c}_{xz}(h^{(j)})\right]_{kl}=\delta_{kj}\delta_{lj};\label{cch}\end{aligned}\ ] ] and for the @xmath58-phase gate @xmath354 on the logical qubit ` @xmath108 ' is given by @xmath355_{kl}&= \left[\textbf{c}_{zz}(r^{(j)}_{z}(\pi/2))\right]_{kl}=\delta_{kl},\nonumber\\ \left[\textbf{c}_{zx}(r^{(j)}_{z}(\pi/2))\right]_{kl}&= 0,\nonumber\\ \left[\textbf{c}_{xz}(r^{(j)}_{z}(\pi/2))\right]_{kl}&=\delta_{kj}\delta_{lj}.\label{ccphase}\end{aligned}\ ] ] it is advantageous to deal with the information flow vector @xmath106 together with the propagation matrices ( eqs . ( [ ccr])([ccphase ] ) ) by a classical computer , than directly dealing with the corresponding byproduct operator @xmath96 together with the propagation relations ( eqs . ( [ pror ] ) , ( [ procnot ] ) , ( [ proh ] ) , and ( [ prophase ] ) ) . this appendix holds a discuss on an efficient measurement scheme of the mqcm where the temporal order of the measurements plays an important role @xcite . on one hand in the uqcm , we can not parallelize two gates of a sequence that do not commute . on the other hand in the mqcm , all the gates from the generating set of the clifford group can be executed in a single time step irrespective of their positions in the circuit . in other words , the temporal order of the measurements in the mqcm is not pre - imposed by the temporal order of the gates . so , the most efficient scheme for the measurements does not necessarily follow the temporal order of the gates in a circuit under simulation . first , the spatial pattern of the measurement bases is assigned to the graph qubits according to the sequence of gates . then the measurements are performed round by round according to the scheme which is given as follows . first , the graph @xmath5 is divided into disjoint subsets of qubits @xmath356 , where index @xmath125 stands for the round of measurements and @xmath357 . mathematically , @xmath358 and @xmath359 for all @xmath360 . the subset @xmath356 is a collection of all those qubits which will be measured simultaneously in @xmath125th round . all the measurements in the @xmath6 , @xmath60 and @xmath0 eigenbasis are put together in the very first round ( @xmath361th round ) , and there is no need for adjusting the measurement bases according to the previous measurement results for the qubits of @xmath362 . in the first measurement round , the `` redundant graph qubits '' are removed by the @xmath0-measurements , the `` readout qubits '' are measured in the @xmath0 eigenbasis , and the clifford part of the circuit is executed by the @xmath6- , @xmath60-measurements @xcite . in the mqcm , the `` readout qubits '' which play the role of `` output register '' are not the last ones to be measured , they are among the first ones . the observables for all subsequent measurement rounds are of the form @xmath363 with @xmath79 , where the measurement bases for the qubits are decided by the measurement outcomes from the previous rounds . all those qubits whose measurement bases depend on the outcomes from the first measurement round belong to the next subset @xmath364 . similarly , the measurement outcomes from the subset @xmath364 together with @xmath362 decide the measurement bases for the qubits in @xmath365 , and so on . these subsets are measured one by one up to the final measurement round @xmath366 . one can think of the total number of measurement rounds ( @xmath367 ) as the logical depth ( temporal complexity ) for the mqcm . parallel to the measurement rounds , the classical data - processing parts are taken care of by a classical computer . after preparing the graph state and just before starting the measurements , the information vector is initialized to @xmath368 . @xmath368 depends on the eigenvalues @xmath34 of the graph and some particular gates ( like @xmath55 , @xmath93 ) which appear in a quantum circuit under simulation @xcite . after executing the first measurement round on the set @xmath362 , @xmath368 gets updated to @xmath369 through the measurement results . @xmath369 then determines the measurement bases for the qubits of @xmath364 . similarly , the measurement outcomes from round @xmath125 update the information flow vector from @xmath134 to @xmath126 . the corresponding byproduct operator is given by @xmath370 then , @xmath371 sets the measurement bases for the @xmath372th round . after the final measurement round @xmath366 , the @xmath28-part of the information flow vector @xmath373 enables us to interpret the result of the computation . in this measurement scheme , the following technical points are worth emphasizing , which are discussed in ref . @xcite : ( 1 ) in order to construct the subsets of graph qubits @xmath374 , a classical computer needs the forward cones for all the graph qubits . the forward cones decide a strict partial ordering among the qubits , and the sets @xmath374 are constructed accordingly . ( 2 ) in order to account for the influence of the measurement outcomes and the set of eigenvalues @xmath34 on @xmath126 , a classical computer needs the byproduct images for all the graph qubits . ( 3 ) @xmath34 , the byproduct images , and @xmath126 are required for setting the measurement bases for the as - yet unmeasured qubits . a classical computer uses the symplectic scalar product for doing this . centre for quantum technologies is a research centre of excellence funded by ministry of education and national research foundation of singapore . d. z. would like to express his gratitude to cqt for its kind hospitality , and further thanks to h. j. briegel and w. dr . s. wishes to thank h. j. briegel and iqoqi for their warm hospitality . a. s. also wishes to express his grateful feelings to w. dr , hui khoon , le huy nguyen , and philippe raynal for enlightening comments and friendly discussions . 17 d. deutsch , proc . london ser . a * 400 * , 97 ( 1985 ) . d. deutsch , proc . london ser . a * 425 * , 73 ( 1989 ) . a. barenco , c. h. bennett , r. cleve , d. p. divincenzo , n. margolus , p. shor , t. sleator , j. a. smolin and , h. weinfurter , phys . rev . a * 52 * , 3457 ( 1995 ) . d. p. divincenzo , science * 270 * , 255 ( 1995 ) . m. a. nielsen and i. l. chuang , quantum computation and quantum information , cambridge university press , cambridge ( 2007 ) . d. deutsch , r. jozsa , proc . london ser . a * 439 * , 553 ( 1992 ) . l. k. grover , phys . lett . * 79 * , 325 ( 1997 ) . p. w. shor , e - print arxiv : quant - ph/9508027v2 . c. h. bennett , g. brassard , c. crpeau , r. jozsa , a. peres , and w. k. wootters , phys . * 70 * , 1895 ( 1993 ) . h. j. briegel and r. raussendorf , phys . lett . * 86 * , 910 ( 2001 ) . m. hein , j. eisert , and h. j. briegel , phys . a * 69 * , 062311 ( 2004 ) . r. raussendorf and h. j. briegel , phys . * 86 * , 5188 ( 2001 ) . r. raussendorf , d. e. browne , and h. j. briegel , phys . a * 68 * , 022312 ( 2003 ) . r. raussendorf and h. j. briegel , quant . inf . comp . * 6 * , 443 ( 2002 ) , eprint arxiv : quant - ph/0108067v2 . d. browne and h. j. briegel , e - print arxiv : quant - ph/0603226v2 . d. jaksch , h. j. briegel , j. i. cirac , c. w. gardiner , and p. zoller , phys . lett . * 82 * , 1975 ( 1999 ) . m. a. nielsen , phys . * 93 * , 040503 ( 2004 ) . d. e. browne and t. rudolph , phys . * 95 * , 010501 ( 2005 ) . t. p. bodiya and l .- m . duan , phys . lett . * 97 * , 143601 ( 2006 ) . c. y. lu , x. q. zhou , o. ghne , w. b. gao , j. zhang , z. s. yuan , a. goebel , t. yang , and j. w. pan , nature physics * 3 * , 91 ( 2007 ) . j. a. jones , m. mosca , and r. h. hansen , nature * 393 * , 344 ( 1998 ) . i. l. chuang , n. gershenfeld , and m. kubinec , phys . lett . * 80 * , 3408 ( 1998 ) . m. s. anwar , d. blazina , h. a. carteret , s. b. duckett , and j. a. jones , chem . . lett . * 400 * , 94 ( 2004 ) . f. yamaguchi , p. milman , m. brune , j. m. raimond , and s. haroche , phys . a * 66 * , 010302(r ) ( 2002 ) . z. j. deng , m. feng , and k. l. gao , phys . a * 72 * , 034306 ( 2005 ) . w. l. yang , c. y. chen , and m. feng , phys . a * 76 * , 054301 ( 2007 ) . p. g. kwait , j. r. mitchell , p. d. d. schwindt , and a. g. white , j. mod . opt . * 47 * , 257 ( 2000 ) . n. bhattacharya , h. b. vanlindenvandenheuvell , and r. j. c. spreeuw , phys . * 88 * , 137901 ( 2002 ) . p. walther , k. j. resch , t. rudolph , e. schenck , h. weinfurter , v. vedral , m. aspelmeyer , and a. zeilinger , nature * 434 * , 169 ( 2005 ) . k. chen , c. m. li , q. zhang , y. a. chen , a. goebel , s. chen , a. mair , and j. w. pan , phys . lett . * 99 * , 120503 ( 2007 ) .

the integrability problem for systems of first order odes is indeed a relevant issue in the theory of dynamical systems ( see , for instance , @xcite and references therein ) , and the same question for coupled systems of odes could be also meaningful from the viewpoint of applications ( for instance , regarding synchronization problems @xcite ) . in this context , the aim of this paper is two fold : on one hand , to present a novel approach to the explicit construction of liouville integrable coupled systems of odes endowed with lie poisson symmetries and , on the other hand , to show how integrable deformations of the previous systems can be systematically obtained . the construction here presented establishes a ( to the best of our knowledge ) novel connection between poisson lie groups and integrable deformations of dynamical systems . such a link is based on the well - known result by drinfeld that establishes the one - to - one correspondence between poisson lie groups and lie bialgebras @xcite , as well as on the general construction of integrable hamiltonian systems from poisson coalgebras that was introduced in @xcite , thus giving rise to the so - called coalgebra symmetry approach " to finite dimensional integrable systems ( see @xcite for applications of this method to hnon heiles and lotka volterra systems , respectively , and @xcite for the coalgebra based construction of nonlinear superposition rules arising in nonautonomous lie hamilton systems @xcite ) . in particular , we will apply the formalism here introduced in order to obtain new integrable deformations of certain lorenz @xcite and rssler @xcite systems , as well as of some coupled versions of them . nevertheless , we stress that the method presented in the paper is completely general and could be applied to any other interesting hamiltonian systems of odes endowed with a lie poisson symmetry . the basics of the poisson coalgebra approach to integrability will be summarized in the next section . in short , we will consider a finite dimensional integrable hamiltonian system of odes that is defined through a lie poisson algebra @xmath3 , where @xmath4 is the algebra of smooth functions on @xmath5 and the lie poisson bracket is @xmath6 . a set of functions in involution will be given by a single hamiltonian function @xmath7 and a number of casimir functions @xmath8 , @xmath9 . then , since lie poisson algebras are endowed with a ( primitive or non deformed ) coalgebra structure , the coalgebra symmetry approach will straightforwardly provide , by following @xcite , a coupled integrable generalization of the system . this approach will be illustrated in section 3 through the explicit construction of an integrable coupled rssler system . a detailed technical presentation of the construction of poisson lie groups as deformations of lie bialgebras will be given in section 4 . if we denote by @xmath10 the abstract lie algebra associated with the lie poisson coalgebra @xmath3 , we will assume that we are able to find a cocommutator map @xmath11 compatible with @xmath10 , so that @xmath12 defines a lie bialgebra structure @xcite . then , it is well known that the dual of the cocommutator @xmath11 will define a second lie algebra structure ( that we will denote by @xmath13 ) , and the dual of the structure tensor for @xmath14 will provide a compatible cocommutator map @xmath15 for @xmath13 , in such a way that @xmath16 is just the dual lie bialgebra to @xmath12 . once a given @xmath17 is identified , the ( connected component containing the identity ) of the lie group @xmath18 with lie algebra @xmath13 can be constructed by exponentiation , and the unique poisson lie structure having @xmath19 as its linearization at the identity can be computed , thus providing the deformed poisson structure @xmath20 we were looking for . simultaneously , the group law for @xmath18 will give us the explicit deformed coproduct map @xmath21 on the algebra of smooth functions @xmath22 . moreover , the deformation nature of this construction arises in a more transparent way if we consider the one - parametric cocommutator @xmath23 , since the parameter @xmath24 will give us information concerning the orders in the deformation process that arises from the ` exponentiation ' of the tangent lie bialgebra structure . it is worth to be stressed that the procedure that we have just sketched is algorithmic and can be made fully explicit by making use of symbolic computation packages . therefore , as a poisson coalgebra , @xmath25 can be interpreted as a smooth @xmath24deformation of @xmath26 on which the initial dynamical system with hamiltonian function @xmath27 is defined . as a consequence , if we consider the same hamiltonian function @xmath27 on @xmath25 , then the associated dynamical system is just an integrable @xmath24deformation of the initial system of odes . moreover , the coalgebra symmetry approach @xcite guarantees that the set of involutive functions for the deformed system will be given by the deformed coproducts of the hamiltonian @xmath27 and of the deformed casimir functions @xmath28 for the poisson structure @xmath20 , with @xmath29 . sections 5 and 6 are devoted to use this poisson - deformation approach in order to obtain integrable deformations of two coupled rssler and lorenz systems , respectively . it is worth stressing that , as a result of the exponentiation process that underlies the full construction , such integrable deformations of dynamical systems are provided by non polynomial vector fields , which is a non standard finding from the viewpoint of integrability . also , the underlying poisson lie groups are responsible of the fact that the coupling between the systems is quite involved , since such coupling is generated by their non abelian group law . a final section including some remarks and open problems closes the paper . let @xmath14 be a finite dimensional lie algebra . as it is well known , the associated lie poisson algebra @xmath4 is the algebra of @xmath30 functions on @xmath5 , equipped with the poisson bracket @xmath31,x \rangle , \qquad x \in g^ * , \quad f , g \in { \cal f}(g^ * ) , \label{lpintrinsic}\ ] ] where @xmath32 is the non - degenerate pairing between @xmath14 and @xmath5 . let us assume that @xmath14 is generated by the set @xmath33 , with commutation relations @xmath34=c_{ij}^k\ , x_k , \label{structurecoeffs}\ ] ] and let us denote with @xmath35 the corresponding dual basis in @xmath19 @xmath36 then the lie poisson bracket ( [ lpintrinsic ] ) for the coordinate functions on @xmath5 is given by the fundamental poisson brackets @xmath37 now , by defining the following coproduct map @xmath38 for the coordinate functions @xmath39 we can say that @xmath26 is endowed with a poisson coalgebra structure @xcite , since the coassociativity condition @xmath40 is fulfilled , and the coproduct map @xmath41 can be straightforwardly shown to define a poisson map between @xmath42 and @xmath43 , namely @xmath44 with respect to the poisson structure on @xmath45 induced from . the coalgebra approach to integrability is based on the fact that if we start from a liouville integrable system defined on @xmath42 , we can use the coproduct map ( [ primitive ] ) in order to extend it to an integrable system defined on @xmath46^n:= { \cal{f}}(g^ * ) \otimes \dots^{n}\otimes { \cal{f}}(g^*)$ ] , with @xmath47 being an arbitrary number of copies of @xmath4 . the keystone for this result is given by the @xmath48-th coproduct maps , which are defined recursively as @xmath49 and , by construction , satisfy the poisson homorphism property ( [ homorphism ] ) @xmath50 the simplest possible case in this framework occurs when the integrable system on @xmath42 is defined by a single hamiltonian function @xmath27 and an arbitrary number of casimir functions @xmath51 , @xmath9 . in this case all the involutive functions defining an integrable system on @xmath46^n$ ] are obtained as images of the hamiltonian and casimir functions under the @xmath52 maps @xmath53 . in this paper we will consider only this case , but we recall that this approach can be generalized for poisson algebras in which integrability involves either more than a single hamiltonian or even non - coassociative poisson maps ( see @xcite ) . in order to illustrate this integrability approach , let us consider the specific lie poisson algebra that underlies the integrability of a dynamical system of rssler type . we recall that the so called rssler system @xmath54 was introduced in @xcite as a prototype of 3d system of odes exhibiting the phenomenon of continuous chaos . when @xmath55 , the system ( [ s1 ] ) is known to be integrable in the liouville sense @xcite , and gives rise to @xmath56 in particular , this system is known to be hamiltonian @xcite with respect to the poisson bracket @xmath57 provided that we consider the hamiltonian function @xmath58 the bracket can be thought of as a particular case of the lie poisson bracket @xmath59 where we have introduced a fourth central " poisson algebra generator @xmath60 . it is immediate to show that this is just the lie poisson algebra corresponding to the 4d real solvable lie algebra @xmath61 ( here we follow the classification of four dimensional real lie algebras given in @xcite ) . the casimir functions for ( [ la ] ) are given by @xmath62 indeed , if the hamiltonian of the system is again given by ( [ rh ] ) , the rssler system ( [ rs ] ) is recovered when @xmath63 . poisson coalgebra @xmath64 is defined through the primitive coproduct map @xmath65 where in the second equality we have identified the coordinate functions on each copy of @xmath42 through a different subscript . now , by using the standard poisson coalgebra structure , we are able to define integrable coupled generalizations of the rssler system ( examples of nonintegrable couplings can be found in @xcite ) . explicitly , if we consider the coproduct of the hamiltonian , we get @xmath66^{2}+[\delta(y)]^{2}\right)+\delta(z)= \dfrac{1}{2}(x_1^{2}+y_1^{2})+z_1+\dfrac{1}{2}(x_2^{2}+y_2^{2})+z_2+x_1x_2+y_1y_2,\ ] ] and the corresponding 6d equations of motion lead to a coupled rssler system @xmath67 \dot{x}_{2}=-w_{2}(y_{1}+y_{2})-z_{2},\\ \dot{y}_{2}=w_{2}(x_{1}+x_{2}),\\ \dot{z}_{2}=z_{2}(x_{1}+x_{2 } ) , \end{array } \label{coro}\ ] ] together with @xmath68 . notice that the coupling between both systems comes from the non linear terms of the hamiltonian , and the restriction of the dynamics to the submanifold @xmath69 ( or , alternatively , to @xmath70 ) leads us to the uncoupled rssler system ( [ rs ] ) . the 6d hamiltonian system is integrable since , due to the underlying coalgebra symmetry , we have three additional ( besides the hamiltonian ) independent integrals of the motion that are in involution . two of them are given by the ( nontrivial ) casimir functions on each of the copies of the lie poisson algebra , namely _ 1=z_1 e^- , _ 2=z_2 e^- , and the the third one is just the coproduct of the casimir @xmath71 , which reads @xmath72 the @xmath47dimensional generalization of this result is straightforward by considering the @xmath47-th coproduct map @xmath73 of the hamiltonian and of the casimir functions , that is obtained as the pull - back through the @xmath47-th coproduct of the coordinate functions @xcite & & ^(n)(x)=x11 ^n-1)1 & & + 1x1 ^n-2 ) 1 + & & + 11 ^n-1)1x , [ fo ] ( and the same for @xmath74 and @xmath75 ) . explicitly , @xmath76^{2}+[\delta^{(n)}(y)]^{2}\right)+\delta^{(n)}(z)= \dfrac{1}{2}\left(\sum_{i=1}^n x_i\right)^{2 } + \dfrac{1}{2}\left(\sum_{i=1}^n y_i \right)^{2}+\sum_{i=1}^n z_i , \label{nr}\ ] ] and the @xmath47-coupled rssler system reads @xmath77 this system of @xmath78 coupled odes has the following ( @xmath79 ) integrals of the motion : _ i = z_i e^- , i=1, ,n @xmath80 by construction , all these functions are in involution and poisson - commute with the hamiltonian @xmath81 . the aim of this paper is to construct integrable deformations of coalgebra symmetric systems of the type for any number @xmath82 of coupled constituents . as we will show in the sequel , this can be achieved by considering the lie poisson algebra @xmath19 as the linearization of a poisson lie group structure on a certain _ different _ lie group @xmath18 . in his way , the construction of the corresponding full poisson lie group structure @xmath25 will provide a poisson coalgebra @xmath24deformation of @xmath64 , and the application of the coalgebra approach to a hamiltonian defined on the poisson lie group @xmath18 will automatically provide an integrable @xmath24deformation of the initial system defined on @xmath42 . given a lie algebra @xmath14 with basis @xmath33 and lie bracket @xmath34=c_{ij}^k\ , x_k , \label{structurecoeffs}\ ] ] a lie bialgebra structure @xmath83 is given by a skew - symmetric cocommutator map @xmath84 @xmath85 such that * \i ) @xmath11 is a 1-cocycle , _ i.e. _ , ( [ x , y])=[(x ) , y1 + 1y ] + [ x1 + 1x , ( y ) ] , x , y . [ 1cocycle ] * \ii ) the dual map @xmath86 is a lie bracket . therefore , the dual @xmath87 of the cocommutator map defines a second ( in general , different ) lie algebra structure , that we will call @xmath13 , and whose lie brackets are @xmath88= f^{ij}_k \ , \hat x^k . \label{geta}\ ] ] here the duality relation is established through the canonical pairing @xmath89 . note that , given the initial lie algebra @xmath14 , the dual lie algebra @xmath13 has to fulfill the 1-cocycle condition , which implies the following compatibility equations between the two lie algebras @xmath14 and @xmath13 : f^ab_k c^k_ij = f^ak_i c^b_kj+f^kb_i c^a_kj + f^ak_j c^b_ik + f^kb_j c^a_ik . [ compatfc ] by taking the components of the structure tensor @xmath90 as initial data , the solutions to the ( now linear ) equations together with the ( quadratic ) equations coming from the jacobi identity for @xmath91 provide all possible lie bialgebra structures for @xmath14 . finally , if the solutions for the tensor @xmath92 are classified into equivalence classes under the action of all possible automorphisms for @xmath14 , the classification of all possible lie bialgebra structures of @xmath14 is obtained . two remarks are in order : * it is immediate to prove that if @xmath83 is a lie bialgebra with structure tensors @xmath93 , then @xmath94 with @xmath95 being the cocommutator map @xmath96 defines a lie bialgebra structure on @xmath13 with structure tensors @xmath97 , that is called the dual lie bialgebra to @xmath83 . in other words , @xmath13 and @xmath11 are defined through the same structure tensor @xmath91 , and we can say that @xmath98 . in the same manner , @xmath14 and @xmath15 are defined by the same structure tensor @xmath90 and we can say that @xmath99 . * for our purposes it will be convenient to multiply all the components @xmath92 of the cocommutator map by a real parameter @xmath24 , that will play the role of a deformation parameter . the resulting cocommutator will be denoted by @xmath100 , and the dual lie algebra structure @xmath101 will be @xmath88=\eta f^{ij}_k \,\hat x^k . \label{geta}\ ] ] obviously , @xmath13 is just @xmath102 and these two lie algebras are isomorphic under the change of basis @xmath103 . nevertheless , as we will see in the sequel , the explicit appearance of @xmath24 will allow us to control the full deformation / exponentiation procedure . a poisson lie ( pl ) group is a lie group @xmath18 endowed with a poisson structure on the algebra @xmath104 of functions on the group manifold , such that the lie group multiplication is a poisson map . therefore , by defining the coproduct @xmath105 as the pull back of a function @xmath106 by the group multiplication , we obtain a natural structure of poisson coalgebra on the algebra of smooth functions on the poisson lie group @xmath22 . in algebraic terms , if @xmath18 is a lie group , the group multiplication on @xmath18 induces a coproduct map : c^(d)c^(d)c^(d ) , which is coassociative ( since the group law is associative ) . therefore , pl groups are instances of poisson coalgebras , since the coproduct map is -by definition- a poisson algebra homomorphism : =( ) a , bc^(d ) , where \ { a b , c d}=\{a , c } bd + ac \ { b , d } , [ pois2 ] is the natural poisson structure on @xmath107 . now , the relevance of lie bialgebras arises due to their one - to - one correspondence with poisson lie groups , a result that can be explicitly stated as follows . * theorem @xcite*. _ let @xmath18 be a lie group with lie algebra @xmath13 : _ \a ) if @xmath108 is a poisson lie group , then @xmath13 has a natural lie bialgebra structure @xmath94 , called the tangent lie bialgebra of @xmath108 . \b ) conversely , if @xmath18 is connected and simply connected , every lie bialgebra structure @xmath94 is the tangent lie bialgebra of a unique poisson structure on @xmath18 which makes @xmath18 into a poisson lie group . explicitly , if @xmath108 is a pl group , the canonical lie algebra structure on @xmath109 is given by _ d^:=(d\{f_1,f_2})_e=^ ( ^1 ^2 ) , ^1,^2 d^where @xmath110 are such that @xmath111 . in other words , the linearization of the pl bracket in terms of the local canonical coordinates @xmath112 on the @xmath18 group manifold is just @xmath113 , the dual of the cocommutator map that defines the associated tangent lie bialgebra structure . a novel construction of coalgebra deformations of lie poisson algebras can be envisaged if drinfeld theorem is revisited by taking the tangent lie bialgebra @xmath94 as the initial data for the construction of the associated pl group . let us assume that we want to construct a poisson coalgebra deformation of a given lie poisson algebra @xmath19 . then , we can proceed as follows : 1 . take the lie algebra @xmath14 , find a non trivial lie bialgebra structure @xmath83 and construct its dual lie bialgebra @xmath94 . now , recall that drinfeld theorem ensures that @xmath94 will be the tangent lie bialgebra structure of a certain poisson lie group @xmath108 such that @xmath114 . 2 . introduce the isomorphic lie algebra @xmath101 in order to have an explicit deformation parameter in the tangent lie bialgebra . 3 . construct the pl group @xmath108 , that will provide the deformation of the poisson coalgebra @xmath64 . in order to obtain it , we have firstly to find a faithful representation @xmath115 of @xmath101 , and use it to construct the matrix element of the ( connected component ) of the lie group @xmath18 through the usual exponentiation : @xmath116 4 . with an abuse of notation , we will make use the same symbol @xmath117 for the coordinate functions on @xmath5 and for the local coordinates of the @xmath18 group in ( [ matrixgroup ] ) . in this way , the pullback of the group multiplication will provide a coproduct map _ : ( d ) ( d ) ( d ) , where , due to the coordinate identification that we have previously made , the coproduct @xmath21 is a @xmath24 deformation of the primitive one for @xmath64 , namely @xmath118 note that in the ` non - deformed case ' @xmath119 we have from that the dual lie algebra @xmath13 is @xmath120=0 $ ] , and then the group @xmath18 is abelian . therefore , its group law is just the addition of group coordinates , which is algebraically expressed through the primitive coproduct . 5 . therefore , by drinfeld theorem , there exists a unique poisson bracket @xmath20 on @xmath22 and such that @xmath121 is a poisson lie group @xmath122 it turns out that the poisson bracket @xmath20 is also an @xmath24 deformation of the lie poisson bracket @xmath6 on @xmath19 , in the sense that @xmath123 indeed , the poisson lie bracket @xmath20 has to be explicitly found by solving simultaneously the equations and , but this problem turns out to be computationally feasible . due to the properties ( [ deltadeformation ] ) and ( [ bracketdeformation ] ) we can conclude that , as a poisson coalgebra , the poisson lie group @xmath121 is an @xmath24deformation of the lie poisson algebra @xmath26 . indeed , there will be as many poisson - coalgebra deformations of @xmath19 as different pl groups @xmath18 such that @xmath19 induces an admissible tangent lie bialgebra cocommutator @xmath15 for @xmath114 . in the sequel we will make use of the fact that the integrable rssler system is an outstanding example of lie poisson coalgebra - symmetric system . this will allow the construction of an @xmath24deformed coproduct map that will give rise to integrable coupled systems of odes of rssler type . let us start by considering the real lie algebra a@xmath124 in the basis given in @xcite ( note that this is exactly the four dimensional algebra ) @xmath125=-w , \quad & [ x , z]=-z , \quad & [ y , z]=0 , \quad&[w,\cdot ] = 0 . \label{laconm } \end{array}\ ] ] it is straightforward to check that the following cocommutator map _ ( x)= x z , _ ( y)= z w , _ ( z)=_(w)= 0 , is compatible with the lie poisson algebra ( [ la ] ) in the sense of , and thus defines a lie bialgebra structure @xmath126 . the dual lie algebra @xmath101 obtained from the dual cocommutator map is @xmath127=0 , \qquad [ \hat x,\hat z]= \eta\ , \hat x , \qquad [ \hat x,\hat w]=0 , \nonumber \\ & & [ \hat y,\hat z]= 0 , \qquad [ \hat y,\hat w]= 0 , \qquad [ \hat z,\hat w]= \eta\ , \hat y. \label{pda}\end{aligned}\ ] ] it can be checked that the dual lie algebra is also isomorphic to the @xmath128 algebra . therefore , we have a self dual lie bialgebra structure . now we have to construct the lie group @xmath18 whose lie algebra is @xmath101 . if we define the @xmath129 matrix @xmath130 as the one with the only nonvanishing entry @xmath131 in the @xmath132th row and @xmath133th column , a faithful representation @xmath115 of the algebra is given by @xmath134 by using @xmath115 , we construct the matrix group element @xmath135 where @xmath136 are the local coordinate functions on the group @xmath18 . at this point it is worth stressing that the parameter @xmath24 appears as as multiplicative factor within all the nonvanishing matrix entries of the fundamental representation of the lie algebra @xmath13 . therefore , when we exponentiate in order to obtain the group element @xmath18 , the powers of @xmath24 will give us the corresponding contributions coming from different powers of the lie algebra generators . by solving the functional equations @xmath137 through the algorithm presented in @xcite , we get the group law , _ i.e. _ , the coproduct maps for the coordinate functions of @xmath18 , which read @xmath138 that are indeed @xmath24-deformations of the primitive coproduct ( [ coproducts ] ) such that @xmath139 . again , by following the approach described in @xcite , the unique poisson lie bracket @xmath140 for which the coproduct @xmath21 is a poisson map is found to be @xmath141 as expected , @xmath142 and we have thus obtained a @xmath24deformation of the lie poisson algebra ( [ la ] ) . finally , the casimir functions for the algebra are found to be @xmath143 and , as expected , they are again @xmath24deformations of the casimir functions . therefore , the poisson lie group @xmath121 has been fully constructed . now , by taking the same hamiltonian of the rssler system , the pl bracket gives rise to the system of odes @xmath144 , \cr \displaystyle \dot{y}=x w e^{-\eta z}=w x-(w x z ) \eta + \left(\frac{1}{2 } w x z^2\right ) \eta ^2+o[\eta^3 ] , \\ \displaystyle \dot{z}=x\frac{\left ( 1-e^{-\eta z } \right ) } { \eta } = x z- \left(\frac{1}{2 } x z^2\right ) \eta + \left(\frac{1}{6 } x z^3\right ) \eta ^2 + o[\eta^3 ] , \cr \displaystyle \dot{w}=0 , \end{array } \label{req}\ ] ] that , for @xmath145 , provides an integrable deformation of the rssler system , with deformed integrals of the motion given by and . the preservation of the closed nature of the trajectories under deformation is clearly appreciated in figure 1 . ( 140,66)(0,0 ) [ figure2 ] once the poisson hopf algebra @xmath121 has been fully determined , the construction of an integrable deformation for the @xmath47coupled system is straightforward . namely , in the @xmath146 case the deformed coproduct @xmath21 on the hamiltonian of the rssler system ( [ rh ] ) will be _ ( ) = ( [ _ ( x)]^2+[_(y)]^2)+_(z ) , and from we get @xmath147 this hamiltonian gives rise to the following integrable deformation of the @xmath146 coupled rssler system @xmath148 the three integrals of the motion are given by the coproduct of the casimir @xmath149 _ ( c^(2)_)= ( - ) , together with the ` one site ' deformed casimirs ^(1)_= ( - ) , ^(2)_= ( - ) . the generalization of this construction to an arbitrary number @xmath47 of integrable coupled rssler systems is straightformard by making use of the @xmath47-th order coproduct map , that can be explicitly obtained through the recurrence . the approach described above can be applied to any other hamiltonian system of odes endowed with lie poisson symmetry . in particular , the celebrated lorenz equations @xcite are given by the system @xmath150 here we will consider a conservative limit of ( [ lorenz ] ) ( see @xcite ) , which is obtained through the following rescaling @xmath151 and by taking the limit @xmath152 . as a consequence of this transformation , the system ( [ lorenz ] ) is mapped into @xmath153 which is a hamiltonian system with respect to the lie poisson algebra @xcite @xmath154 and provided that we have considered the hamiltonian function @xmath155 in order to proceed in a similar way as for the rssler system , let us consider the 4d lie algebra @xmath156=-\dfrac{1}{2 } ( z+w ) , \qquad [ x , z]=\dfrac{1}{2 } y , \qquad [ y , z]=x , \qquad [ w,\cdot]=0 . \label{liel}\ ] ] this lie algebra is isomorphic to the @xmath157 lie algebra , where @xmath158 is a central generator . obviously , the lie poisson algebra @xmath19 arising from is a generalization of the poisson algebra through a new central coordinate function @xmath60 @xmath159 now , if we consider the hamiltonian @xmath160 the equations of motion obtained through are @xmath161 and we recover the conservative lorenz equations when @xmath145 . finally , the conserved quantities for the system are given by the casimir functions for ( [ clt ] ) , namely @xmath162 with this prerrequisites we can proceed with the formalism presented in section 4 . a cocommutator map compatible with the lie algebra ( [ liel ] ) is found to be @xmath163 the dual lie algebra @xmath101 obtained from the dual cocommutator map @xmath164 is given by @xmath127=0 , \qquad [ \hat x,\hat z]= \eta\ , \hat x , \qquad [ \hat x,\hat w]= \eta\ , \hat x + \eta\ , \hat y , \nonumber \\ & & [ \hat y,\hat z]= \eta\ , \hat y , \qquad [ \hat y,\hat w]= \frac{\eta}{2}\ , \hat x + \eta\ , \hat y , \qquad [ \hat z,\hat w]= 0 . \label{pd}\end{aligned}\ ] ] this lie algebra @xmath101 can be shown to be isomorphic to the direct sum of two copies of the two dimensional real lie algebra @xmath165=e_1 $ ] . a four dimensional faithful representation @xmath115 of @xmath101 is given by the matrices @xmath166 by using @xmath115 , we construct the matrix group element @xmath167 as a consequence , the coproduct map arising from the group multiplication of two @xmath18 matrices is found to be @xmath168,\\ \delta_\eta(y)=y_{1}+e^{-\eta\ , ( z_{1}+w_{1})}\left[y_{2}\cosh \left(\tfrac{\eta } { \sqrt{2}}\ , w_{1}\right)-\sqrt{2}x_{2}\sinh \left(\tfrac{\eta}{\sqrt{2}}\ , w_{1}\right)\right ] , \\ \delta_\eta(z)=z_1+z_2,\\ \delta_\eta(w)=w_1+w_2 . \end{array } \label{colo}\ ] ] now , by applying again the algorithmic procedure introduced in @xcite , we get the following unique solution for the poisson lie structure on the group @xmath18 that has @xmath19 as its linearization and is compatible with : @xmath169 this is indeed an @xmath24deformation of the @xmath170 lie poisson algebra , which is recovered in the @xmath171 limit , and the ( deformed ) casimir functions for the poisson algebra ( [ dt ] ) are found to be @xmath172 - 2 e^{z \eta } [ 1+w \eta ( 1+w \eta ) ] \right)}{\eta ^2}. \label{casl}\ ] ] note that , as expected , @xmath173 therefore , an integrable deformation of the integrable lorenz system is obtained by considering the same hamiltonian and the poisson lie brackets , and reads @xmath174 , \\[2pt ] \dot{z}= & x \ , y , \end{array } \label{cl2def}\ ] ] where the only deformed equation occurs for the @xmath74 variable , and the conserved quantities are given by the hamiltonian and the casimir functions . several periodic trajectories for this system are plotted in figure 2 , and are compared with the underformed lorenz ones . ( 140,66)(0,0 ) [ figure2 ] the undeformed lie poisson coalgebra @xmath64 is again defined through the primitive coproduct map . therefore , the @xmath146 coupled lorenz hamiltonian will be given by @xmath175^{2},\ ] ] where the coproduct @xmath41 is the undeformed one . the corresponding equations of motion read @xmath176 \dot{x}_{2}= y_2,\\ \dot{y}_{2}= 2x_2-(x_1+x_2)(w_2+z_2),\\ \dot{z}_{2}= y_2(x_1+x_2 ) , \end{array } \label{couplo}\ ] ] together with @xmath68 ( see @xcite for some instances of nonintegrable coupled lorenz oscillators ) . again , three additional independent integrals of the motion in involution for this system are provided by the coalgebra symmetry : namely , the casimir functions on each of the two copies of the lie poisson algebra _ 1= -2x_1 ^ 2+y_1 ^ 2+z_1(z_1 + 2 w_1 ) , _ 2= -2x_2 ^ 2+y_2 ^ 2+z_2(z_2 + 2 w_2 ) , together with the coproduct of the casimir @xmath71 , which reads @xmath177 the @xmath47dimensional integrable generalization of this system arises from the @xmath47-th coproduct map @xmath73 of the hamiltonian , namely @xmath178 in this way the following @xmath47-coupled lorenz system is obtained as @xmath179 and the ( @xmath79 ) integrals of the motion ( in involution between themselves and with respect to the hamiltonian @xmath81 ) are given by @xcite the ` one site ' integrals _ i=-2x_i^2+y_i^2+z_i(z_i+2 w_i ) , i=1, ,n together with the @xmath48th order coproducts of the casimir functions @xmath180 , \quad k=2,\dots , n.\ ] ] note that the coupling between the 3d dynamical systems is induced by the non linear @xmath181 term in the hamiltonian . by following the very same approach as for the rssler system , the deformed coproduct map @xmath21 acting on the hamiltonian function ( [ h1 ] ) gives rise to _ ( ) = 2(_(z)-_(w))- [ _ ( x)]^2 , and by taking into account the explicit expressions for @xmath21 we get ( ) = 2(z_1+z_2-w_1-w_2)-^2 . this hamiltonian gives rise to the following ( quite involved ) deformation of the @xmath146 coupled lorenz system @xmath182\notag\\ % % % % % % % % % % % % % % % % % % % % & & + y_{1}\dfrac{\eta}{4}e^{-2\eta ( w_{1}+z_{1})}\left [ ( 2x_{2}^{2}+y_{2}^{2})\cosh(\sqrt{2}\,\eta\ , w_{1})-2\sqrt{2}\,x_{2}\,y_{2}\sinh(\sqrt{2}\,\eta\ , w_{1 } ) \right],\notag\\ % % % % % % % % % % % % % % % % % % % % % % % % \notag\\ % % % % % % % % \dot{y}_{1}&= & \dfrac{x_{1}}{2\eta}\left ( -1+\eta ( 4+\eta(2x_{1}^{2}-y_{1}^{2 } ) ) \right)\notag\\ % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % % & & + \dfrac{x_{1}}{2\eta}e^{-2\eta ( w_{1}+z_{1 } ) } \left ( 1+\eta^{2}(2x_{2}^{2}+y_{2}^{2})\cosh(\sqrt{2}w_{1}\eta ) + \eta^{2}(2x_{2}^{2}-y_{2}^{2}-2\sqrt{2}x_{2}y_{2}\sinh(\sqrt{2}w_{1}\eta ) ) \right)\notag\\ % % % % % % % & & + \dfrac{1}{4\eta}e^{-3\eta ( w_{1}+z_{1})}\left [ \left ( 2x_{2}\cosh\left(\tfrac{\eta}{\sqrt{2}}w_1\right)-\sqrt{2}y_{2}\sinh\left(\tfrac{\eta}{\sqrt{2}}w_1\right ) \right)\left(1+e^{2\eta ( w_{1}+z_{1})}(-1+\eta^{2}(6x_{1}^{2}-y_{1}^{2 } ) ) \right ) \right],\notag\\ % % % % % % % % % % % % % % % % % % % % % \notag\\ % % % % % % % % \dot{z}_{1}&= & x_{1}y_{1}+\dfrac{1}{2}y_{1}e^{-\eta(w_{1}+z_{1})}\left ( 2x_{2}\cosh \left(\tfrac{\eta}{\sqrt{2}}w_1\right)-\sqrt{2}y_{2}\sinh\left(\tfrac{\eta}{\sqrt{2}}w_1\right ) \right ) , \notag\\ % % % % % % % \notag\\ % % % % % % % % \dot{x}_{2}&= & y_{2}+\dfrac{1}{8\eta}e^{-2\eta(w_{1}+w_{2}+z_{1}+z_{2})}\left ( 1+e^{2\eta ( w_{2}+z_{2})}(-1+\eta^{2}(2x_{2}^{2}-y_{2}^{2 } ) ) \right)\times \label{deflor2}\\ % % % % % % % & & \times \left ( y_{2}-y_{2}\cosh(\sqrt{2}w_{1}\eta)+2\sqrt{2}x_{1}e^{\eta(w_{1}+z_{1})}\sinh\left(\tfrac{\eta}{\sqrt{2}}w_1\right)+\sqrt{2}x_{2}\sinh(\sqrt{2}w_{1}\eta ) \right),\notag\\ % % % % % % % \notag\\ % % % % % % % % % \dot{y}_{2}&= & 2x_{2}+\dfrac{1}{4\eta}e^{-2\eta ( w_{1}+w_{2}+z_{1}+z_{2})}\cosh\left(\tfrac{\eta}{\sqrt{2}}w_1\right ) \left(1+e^{2\eta(w_{2}+z_{2})}(-1+\eta^{2}(2x_{2}^{2}-y_{2}^{2}))\right)\times\notag\\ % % % % % % % % % % & & \times \left ( 2x_{1}e^{\eta ( w_{1}+z_{1})}+2x_{2}\cosh\left(\tfrac{\eta}{\sqrt{2}}w_1\right)-\sqrt{2}y_{2}\sinh\left(\tfrac{\eta}{\sqrt{2}}w_1\right ) \right),\notag\\ % % % % % % \notag\\ % % % % % % % \dot{z}_{2}&= & e^{-2\eta(w_{1}+z_{1})}\left ( x_{1}y_{2}e^{\eta(w_{1}+z_{1})}\cosh\left(\tfrac{\eta}{\sqrt{2}}w_1\right)+x_{2}y_{2}\cosh(\sqrt{2}\eta w_{1 } ) \right)\notag\\ % % % % % % % % % % % % % % & & -\dfrac{\sqrt{2}}{4}e^{-2\eta(w_{1}+z_{1})}\left ( 4x_{1}x_{2}e^{\eta(w_{1}+z_{1})}\sinh\left(\tfrac{\eta}{\sqrt{2}}w_1\right)+(2x_{2}^{2}+y_{2}^{2})\sinh\left(\sqrt{2}w_{1}\eta\right ) \right ) . \notag\end{aligned}\ ] ] by taking a power series expansion we get a first order deformation in @xmath24 of the coupled lorenz system given by @xmath183,\\[2pt ] \dot{y}_{1}=2x_{1}-(x_{1}+x_{1})(w_{1}+z_{1 } ) \\[2pt ] % % % % % % % % % % \quad\quad+ \left ( ( w_{1}^{2}+z_{1}^{2}+2 w_{1}z_{1})(x_{1}+2x_{2})-\dfrac{y_{1}^{2}}{2}(x_{1}+x_{2})+\dfrac{w_{1}y_{2}}{2}(w_{1}+z_{1})+x_{1}(x_{1}^{2}+3 x_{1}x_{2}+2x_{2}^{2 } ) \right)\eta + o[\eta^2],\\ \dot{z}_{1}=y_1(x_1+x_2)-\dfrac{y_{1}}{2}(w_{1}(2x_{2}+y_{2})+2z_{1}x_{2})\eta + o[\eta^2 ] \end{array}\ ] ] @xmath184,\\[2pt ] % % % % % % % % % % % % % % \dot{y}_{2}=(2 { x_2}-({x_1}+{x_2 } ) ( { w_2}+{z_2 } ) ) \\ % % % % % % % % % \quad\quad+\dfrac{1}{2 } \left ( ( x_{1}+x_{2})(2x_2 ^ 2-y_2 ^ 2 + 2(w_2+z_2)^2)+(w_2+z_2)(2 ( x_1 + 2x_2)(w_1+z_1)+w_1 y_2)\right)\eta + o[\eta^2],\\[2pt ] % % % % % % % % % % % % % % % % \dot{z}_{2}=y_2({x_1}+{x_2})-\dfrac{1}{2 } \left ( 2 y_2 ( w_1+z_1)(x_1 + 2x_2)+w_1 y_2 ^ 2 + 2 w_1 x_2 ( x_1+x_2 ) \right)\eta + o[\eta^2 ] . \end{array}\ ] ] finally , the three integrals of the motion for are given by the coproduct @xmath21 of the casimir @xmath149 given in , namely l _ ( c^(2)_)= e^-_(z ) ( 1+e^2 ( _ ( w)+_(z ) ) - + - 2 e^ _ ( z ) [ 1+_(w ) ( 1+_(w ) ) ] ) , together with the ` one site ' deformed casimirs l ^(1)_= , + ^(2)_= . of course , since _ 0\{_i,_()}_=\{_i , ( ) } , _ i=\{x_i , y_i , z_i } , all these expressions define an integrable @xmath24-deformation of the @xmath146 coupled lorenz system , and the generalization for arbitrary @xmath47 can be straightforwardly obtained through the @xmath47-th coproduct map that arises from and from the recurrence . in this paper we have established a novel ( and somehow unexpected ) link between the theory of pl groups and the integrability problems for ( coupled ) odes . in fact , since poisson lie groups can be thought of as deformations of lie poisson coalgebras , this property can be used in order to construct integrable deformations of ( coupled ) systems of odes endowed with a lie poisson hamiltonian structure . the applicability of this statement is based on the fact that the explicit construction of the pl groups that deform a given lie poisson algebra can be systematically obtained through the exponentiation of the associated lie bialgebra structures ( therefore , by making use of the poisson version of the well known ` quantum duality principle ' for quantum groups @xcite ) . one of the main advantages of the approach here presented relies on its algorithmic nature , that makes it amenable to be computerized in a nearly complete way ( see @xcite ) . in order to illustrate the method , in this paper we have worked out two specific integrable deformations of ( coupled and uncoupled ) rssler and lorenz systems . it is important to emphasize that other integrable deformations for these systems do exist . in fact , there will be as many of them as the number of possible lie bialgebra structures associated to the initial lie poisson algebra that provides the hamiltonian structure for the undeformed dynamical system ( namely , the a@xmath124 algebra for the rsler system and the @xmath185 algebra for the lorenz one ) . indeed , the very same approach can be applied to many other hamiltonian systems of odes . moreover , there also exists the possibility of generalizing this method in order to consider integrable systems with more than one hamiltonian , either by deforming bihamiltonian poisson pencils or by using the loop coproduct approach introduced in @xcite . finally it is worth mentioning that the analytic solution of coupled integrable systems with coalgebra symmetry is facilitated by the use of the so called ` cluster ' variables ( see @xcite ) , and this fact could shed some light on the dynamical significance of the rssler and lorenz couplings here obtained . work on all these lines is in progress and will be presented elsewhere . this work was partially supported by the spanish ministerio de economa y competitividad ( mineco ) under grant mtm2013 - 43820-p , by junta de castilla y len under grant bu278u14 and by infn sezione roma tre .

mixture models are common tools in statistical pattern recognition @xcite . they offer a mathematical basis to explain data in fields as diverse as astronomy , biology , ecology , engineering , and economics , amongst many others @xcite . a mixture model is composed of component probabilistic models ; a component may variously correspond to a subtype , kind , species , or subpopulation of the observed data . these models aid in the identification of hidden patterns in the data through sound probabilistic formalisms . mixture models have been extensively used in machine learning tasks such as classification and unsupervised learning @xcite . formally , mixture modelling involves representing a distribution of data as a weighted sum of individual probability distributions . specifically , the problem we consider here is to model the observed data using a mixture @xmath0 of probability distributions of the form : @xmath1 where @xmath2 is a @xmath3-dimensional datum , @xmath4 is the number of mixture components , @xmath5 and @xmath6 are the weight and probability density of the @xmath7 component respectively ; the weights are positive and sum to one . the problem of modelling some observed data using a mixture distribution involves determining the number of components , @xmath4 , and estimating the mixture parameters . inferring the optimal number of mixture components involves the difficult problem of balancing the trade - off between two conflicting objectives : low _ hypothesis complexity _ as determined by the number of components _ and _ their respective parameters , versus good quality of _ fit _ to the observed data . generally a hypothesis with more free parameters can fit observed data better than a hypothesis with fewer free parameters . a number of strategies have been used to control this balance as discussed in section [ sec : mixture_existing_methods ] . these methods provide varied formulations to assess the mixture components and their ability to explain the data . methods using the minimum message length criterion @xcite , a bayesian method of inductive inference , have been proved to be effective in achieving a reliable balance between these conflicting aims @xcite . although much of the literature concerns the theory and application of gaussian mixtures @xcite , mixture modelling using other probability distributions has been widely used . some examples are : poisson models @xcite , exponential mixtures @xcite , laplace @xcite , t - distribution @xcite , weibull @xcite , kent @xcite , von mises - fisher @xcite , and many more . @xcite provide a comprehensive summary of finite mixture models and their variations . the use of gaussian mixtures in several research disciplines has been partly motivated by its computational tractability @xcite . for datasets where the _ direction _ of the constituent vectors is important , gaussian mixtures are inappropriate and distributions such as the von mises - fisher may be used @xcite . in any case , whatever the kind of distribution used for an individual component , one needs to estimate the parameters of the mixture , and provide a sound justification for selecting the appropriate number of mixture components . software for mixture modelling relies on the following elements : 1 . an _ estimator _ of the parameters of each component of a mixture , 2 . an _ objective function _ , that is a cost or score , that can be used to compare two hypothetical mixtures and decide which is better , and 3 . a _ search strategy _ for the best number of components and their weights . traditionally , parameter estimation is done by strategies such as maximum likelihood ( ml ) or bayesian maximum _ a posteriori _ probability ( map ) estimation . in this work , we use the bayesian minimum message length ( mml ) principle . unlike map , mml estimators are invariant under non - linear transformations of the data @xcite , and unlike ml , mml considers the number and precision of a model s parameters . it has been used in the inference of several probability distributions @xcite . mml - based inference operates by considering the problem as encoding first the parameter estimates and then the data given those estimates . the parameter values that result in the least _ overall _ message length to explain the whole data are taken as the mml estimates for an inference problem . the mml scheme thus incorporates the cost of stating parameters into model selection . it is self evident that a continuous parameter value can only be stated to some finite precision ; the cost of encoding a parameter is determined by its prior and the precision . ml estimation ignores the cost of stating a parameter and map based estimation uses the probability _ density _ of a parameter instead of its probability measure . in contrast , the mml inference process calculates the optimal precision to which parameters should be stated and a probability value of the corresponding parameters is then computed . this is used in the computation of the message length corresponding to the parameter estimates . thus , models with varying parameters are evaluated based on their resultant total message lengths . we use this characteristic property of mml to evaluate mixtures containing different numbers of components . although there have been several attempts to address the challenges of mixture modelling , the existing methods are shown to have some limitations in their formalisms ( see section [ sec : mixture_existing_methods ] ) . in particular , some methods based on mml are incomplete in their formulation . we aim to rectify these drawbacks by proposing a comprehensive mml formulation and develop a search heuristic that selects the number of mixture components based on the proposed formulation . to demonstrate the effectiveness of the proposed search and parameter estimation , we first consider modelling problems using gaussian mixtures and extend the work to include relevant discussion on mixture modelling of directional data using von mises - fisher distributions . the importance of gaussian mixtures in practical applications is well established . for a _ given _ number of components , the conventional method of estimating the parameters of a mixture relies on the expectation - maximization ( em ) algorithm @xcite . the standard em is a local optimization method , is sensitive to initialization , and in certain cases may converge to the boundary of parameter space @xcite . previous attempts to infer gaussian mixtures based on the mml framework have been undertaken using simplifying assumptions , such as the covariance matrices being diagonal @xcite , or coarsely approximating the probabilities of mixture parameters @xcite . further , the search heuristic adopted in some of these methods is to run the em several times for different numbers of components , @xmath4 , and select the @xmath4 with the best em outcome @xcite . a search method based on iteratively _ deleting _ components has been proposed by @xcite . it begins by assuming a very large number of components and selectively eliminates components deemed redundant ; there is no provision for recovering from deleting a component in error . in this work , we propose a search method which selectively _ splits , deletes _ , or _ merges _ components depending on improvement to the mml objective function . the operations , combined with em steps , result in a sensible redistribution of data between the mixture components . as an example , a component may be split into two children , and at a later stage , one of the children may be merged with another component . unlike the method of @xcite , our method starts with a one - component mixture and alters the number of components in subsequent iterations . this avoids the overhead of dealing with a large number of components unless required . the proposed search heuristic can be used with probability distributions for which the mml expressions to calculate message lengths for estimates and for data given those estimates are known . as an instance of this , section [ sec : search_method ] discusses modelling directional data using the von mises - fisher distributions . directional statistics have garnered support recently in real data mining problems where the _ direction _ , not magnitude , of vectors is significant . examples of such scenarios are found in earth sciences , meteorology , physics , biology , and other fields @xcite . the statistical properties of directional data have been studied using several types of distributions @xcite , often described on surfaces of compact manifolds , such as the sphere , ellipsoid , torus _ etc_. the most fundamental of these is the von mises - fisher ( vmf ) distribution which is analogous to a symmetric multivariate gaussian distribution , wrapped around a unit hypersphere @xcite . the probability density function of a vmf distribution with parameters @xmath8 ( mean direction , concentration parameter ) for a random unit vector @xmath9 on a @xmath10- dimensional hypersphere @xmath11 is given by : @xmath12 where @xmath13 is the normalization constant and @xmath14 is a modified bessel function of the first kind and order @xmath15 . the estimation of the parameters of the vmf distribution is often done using maximum likelihood . however , the complex nature of the mathematical form presents difficulty in estimating the concentration parameter @xmath16 . this has lead to researchers using many different approximations , as discussed in section [ sec : vmf_existing_methods ] . most of these methods perform well when the amount of data is large . at smaller sample sizes , they result in inaccurate estimates of @xmath16 , and are thus unreliable . we demonstrate this by the experiments conducted on a range of sample sizes . the problem is particularly evident when the dimensionality of the data is large , also affecting the application in which it is used , such as mixture modelling . we aim to rectify this issue by using mml estimates for @xmath16 . our experiments section demonstrates that the mml estimate of @xmath16 provides a more reliable answer and is an improvement on the current state of the art . these mml estimates are subsequently used in mixture modelling of vmf distributions ( see sections [ sec : search_method ] and [ sec : vmf_applications ] ) . previous studies have established the importance of von mises circular ( two - dimensional ) and von mises - fisher ( three - dimensional and higher ) mixtures , and demonstrated applications to clustering of protein dihedral angles @xcite , large - scale text clustering @xcite , and gene expression analyses @xcite . the merit of using cosine based similarity metrics , which are closely related to the vmf , for clustering high dimensional text data has been investigated in @xcite . for text clustering , there is evidence that vmf mixture models have a superior performance compared to other statistical distributions such as normal , multinomial , and bernoulli @xcite . movmf is a widely used package to perform clustering using vmf distributions @xcite . + * contributions : * the main contributions of this paper are as follows : * we derive the analytical estimates of the parameters of a multivariate gaussian distribution with full covariance matrix , using the mml principle @xcite . * we derive the expression to infer the concentration parameter @xmath16 of a generic @xmath3-dimensional vmf distribution using mml - based estimation . we demonstrate , through a series of experiments , that this estimate outperforms the previous ones , therefore making it a reliable candidate to be used in mixture modelling . * a generalized mml - based search heuristic is proposed to infer the optimal number of mixture components that would best explain the observed data ; it is based on the search used in various versions of the snob classification program @xcite . we compare it with the work of @xcite and demonstrate its effectiveness . * the search implements a generic approach to mixture modelling and allows , in this instance , the use of @xmath3-dimensional gaussian and vmf distributions under the mml framework . it infers the optimal number of mixture components , and their corresponding parameters . * further , we demonstrate the effectiveness of mml mixture modelling through its application to high dimensional text clustering and clustering of directional data that arises out of protein conformations . the rest of the paper is organized as follows : sections [ sec : gaussian_estimates ] and [ sec : vmf_existing_methods ] describe the respective estimators of gaussian and vmf distributions that are commonly used . section [ sec : mml_framework ] introduces the mml framework for parameter estimation . section [ sec : mml_est_derivations ] outlines the derivation of the mml parameter estimates of multivariate gaussian and vmf distributions . section [ sec : mml_mixture_modelling ] describes the formulation of a mixture model using mml and the estimation of the mixture parameters under the framework . section [ sec : mixture_existing_methods ] reviews the existing methods for selecting the mixture components . section [ sec : search_method ] describes our proposed approach to determine the number of mixture components . section [ sec : gaussian_experiments ] depicts the competitive performance of the proposed mml - based search through experiments conducted with gaussian mixtures . section [ sec : vmf_experiments ] presents the results for mml - based vmf parameter estimation and mixture modelling followed by results supporting its applications to text clustering and protein structural data in section [ sec : vmf_applications ] . the probability density function @xmath17 of a @xmath3-variate gaussian distribution is given as @xmath18 where @xmath19 , @xmath20 are the respective mean , ( symmetric ) covariance matrix of the distribution , and @xmath21 is the determinant of the covariance matrix . the traditional method to estimate the parameters of a gaussian distribution is by maximum likelihood . given data @xmath22 , where @xmath23 , the log - likelihood @xmath24 is given by @xmath25 to compute the maximum likelihood estimates , equation needs to be _ maximized_. this is achieved by computing the gradient of the log - likelihood function with respect to the parameters and solving the resultant equations . the _ gradient vector _ of @xmath24 with respect to @xmath19 and the _ gradient matrix _ of @xmath24 with respect to @xmath20 are given below . @xmath26 the maximum likelihood estimates are then computed by solving @xmath27 and @xmath28 and are given as : @xmath29 @xmath30 is known to be a biased estimate of the covariance matrix @xcite and issues related with its use in mixture modelling have been documented in @xcite and @xcite . an unbiased estimator of @xmath20 was proposed by @xcite and is given below . @xmath31 in addition to the maximum likelihood estimates , bayesian inference of gaussian parameters involving conjugate priors over the parameters has also been dealt with in the literature @xcite . however , the unbiased estimate of the covariance matrix , as determined by the sample covariance ( equation ) , is typically used in the analysis of gaussian distributions . for a von mises - fisher ( vmf ) distribution @xmath17 characterized by equation , and given data @xmath22 , such that @xmath32 , the log - likelihood @xmath24 is given by @xmath33 where @xmath34 is the sample size and @xmath35 ( the vector sum ) . let @xmath36 denote the magnitude of the resultant vector @xmath37 and let @xmath38 and @xmath39 be the maximum likelihood estimators of @xmath40 and @xmath16 respectively . under the condition that @xmath38 is a unit vector , the maximum likelihood estimates are obtained by maximizing @xmath24 as follows : @xmath41 solving the non - linear equation : @xmath42 yields the corresponding maximum likelihood estimate where @xmath43 represents the ratio of bessel functions . because of the difficulties in analytically solving equation ( [ eqn : ml_estimates ] ) , there have been several approaches to approximating @xmath39 @xcite . each of these methods is an improvement over their respective predecessors . @xcite is an improvement over the estimate proposed by @xcite . @xcite is an improvement over @xcite and @xcite fares better when compared to @xcite . the methods are summarized below . the approximation given by equation ( [ eqn : banerjee_approx ] ) is due to @xcite and provides an easy to use expression for @xmath39 . the formula is very appealing as it eliminates the need to evaluate complex bessel functions . @xcite demonstrated that this approximation yields better results compared to the ones suggested in @xcite . it is an empirical approximation which can be used as a starting point in estimating the root of equation ( [ eqn : ml_estimates ] ) . @xmath44 the approximation given by equation ( [ eqn : tanabe_approx ] ) is due to @xcite . the method utilizes the properties of bessel functions to determine the lower and upper bounds for @xmath39 and uses a fixed point iteration function in conjunction with linear interpolation to approximate @xmath39 . the bounds for @xmath39 are given by @xmath45 @xcite proposed to use a fixed point iteration function defined as @xmath46 and used this to approximate @xmath39 as @xmath47 this a heuristic approximation provided by @xcite . it involves refining the approximation given by @xcite ( equation ( [ eqn : banerjee_approx ] ) ) by performing two iterations of newton s method . @xcite demonstrate that this approximation fares well when compared to the approximation proposed by @xcite . the following two iterations result in @xmath48 , the approximation proposed by @xcite : @xmath49 @xmath50 this approximation provided by @xcite uses halley s method which is the second order expansion of taylor s series of a given function @xmath51 . the higher order approximation results in a more accurate estimate as demonstrated by @xcite . the iterative halley s method is truncated after iterating through two steps of the root finding algorithm ( similar to that done by @xcite ) . the following two iterations result in @xmath52 , the approximation proposed by @xcite : @xmath53 @xmath54 the common theme in all these methods is that they try to approximate the maximum likelihood estimate governed by equation ( [ eqn : ml_estimates ] ) . it is to be noted that the maximum likelihood estimators ( of concentration parameter @xmath16 ) have considerable bias @xcite . to counter this effect , we explore the minimum message length based estimation procedure . this bayesian method of estimation not only results in an unbiased estimate but also provides a framework to choose from several competing models @xcite . through a series of empirical tests , we demonstrate that the mml estimate is more reliable than any of the contemporary methods . @xcite have demonstrated the superior performance of the mml estimate for a three - dimensional vmf distribution . we extend their work to derive the mml estimators for a generic @xmath3-dimensional vmf distribution and compare its performance with the existing methods . @xcite developed the first practical criterion for model selection to be based on information theory . the resulting framework provides a rigorous means to objectively compare two competing hypotheses and , hence , choose the best one . as per bayes s theorem , @xmath55 where @xmath56 denotes some observed data , and @xmath57 some hypothesis about that data . further , @xmath58 is the joint probability of data @xmath56 and hypothesis @xmath57 , @xmath59 is the prior probability of hypothesis @xmath57 , @xmath60 is the prior probability of data @xmath56 , @xmath61 is the posterior probability of @xmath57 given @xmath56 , and @xmath62 is the likelihood . mml uses the following result from information theory @xcite : given an event or outcome @xmath63 whose probability is @xmath64 , the length of the optimal lossless code @xmath65 to represent that event requires @xmath66 bits . applying shannon s insight to bayes s theorem , @xcite got the following relationship between conditional probabilities in terms of optimal message lengths : @xmath67 as a result , given two competing hypotheses @xmath57 and @xmath68 , @xmath69 @xmath70 gives the posterior log - odds ratio between the two competing hypotheses . equation ( [ eqn : shannon_bayes_msg ] ) can be intrepreted as the _ total _ cost to encode a message comprising the hypothesis @xmath57 and data @xmath56 . this message is composed over into two parts : 1 . _ first part : _ the hypothesis @xmath57 , which takes @xmath71 bits , 2 . _ second part : _ the observed data @xmath56 using knowledge of @xmath57 , which takes @xmath72 bits . clearly , the message length can vary depending on the complexity of @xmath57 and how well it can explain @xmath56 . a more complex @xmath57 may fit ( i.e. , explain ) @xmath56 better but take more bits to be stated itself . the trade - off comes from the fact that ( hypothetically ) transmitting the message requires the encoding of both the hypothesis and the data given the hypothesis , that is , the model complexity @xmath71 and the goodness of fit @xmath72 . our proposed method of parameter estimation uses the mml inference paradigm . it is a bayesian method which has been applied to infer the parameters of several statistical distributions @xcite . we apply it to infer the parameter estimates of multivariate gaussian and vmf distributions . @xcite introduced a generalized scheme to estimate a vector of parameters @xmath73 of any distribution @xmath17 given data @xmath56 . the method involves choosing a reasonable prior @xmath74 on the hypothesis and evaluating the _ determinant _ of the fisher information matrix @xmath75 of the _ expected _ second - order partial derivatives of the negative log - likelihood function , @xmath76 . the parameter vector @xmath73 that minimizes the message length expression ( equation ( [ eqn : two_part_msg ] ) ) is the mml estimate according to @xcite . @xmath77 where @xmath78 is the number of free parameters in the model , and @xmath79 is the lattice quantization constant @xcite in @xmath78-dimensional space . the total message length @xmath80 in mml framework is composed of two parts : 1 . the statement cost of encoding the parameters , @xmath81 and 2 . the cost of encoding the data given the parameters , @xmath82 . a concise description of the mml method is presented in @xcite . we note here a few key differences between mml and ml / map based estimation methods . in maximum likelihood estimation , the statement cost of parameters is ignored , in effect considered constant , and minimizing the message length corresponds to minimizing the negative log - likelihood of the data ( the second part ) . in map based estimation , a probability _ density _ rather than the probability measure is used . continuous parameters can necessarily only be stated only to finite precision . mml incorporates this in the framework by determining the region of uncertainty in which the parameter is located . the value of @xmath83 gives a measure of the volume of the region of uncertainty in which the parameter @xmath73 is centered . this multiplied by the probability density @xmath74 gives the _ probability _ of a particular @xmath73 and is _ proportional _ to @xmath84 . this probability is used to compute the message length associated with encoding the continuous valued parameters ( to a finite precision ) . based on the mml inference process discussed in section [ sec : mml_framework ] , we now proceed to formulate the message length expressions and derive the parameter estimates of gaussian and von mises - fisher distributions . the mml framework requires the statement of parameters to a finite precision . the optimal precision is related to the fisher information and in conjunction with a reasonable prior , the probability of parameters is computed . a flat prior is usually chosen on each of the @xmath3 dimensions of @xmath19 @xcite and a conjugate inverted wishart prior is chosen for the covariance matrix @xmath20 @xcite . the joint prior density of the parameters is then given as @xmath85 the computation of the fisher information requires the evaluation of the second order partial derivatives of @xmath86 . let @xmath87 represent the determinant of the fisher information matrix . this is approximated as the product of @xmath88 and @xmath89 @xcite , where @xmath88 and @xmath89 are the respective determinants of fisher information matrices due to the parameters @xmath19 and @xmath20 . differentiating the gradient vector in equation with respect to @xmath19 , we have : @xmath90 to compute @xmath89 , @xcite derived an analytical expression using the theory of matrix derivatives based on matrix vectorization @xcite . let @xmath91\ , \forall 1 \leq i , j \leq d$ ] where @xmath92 denotes the element corresponding to the @xmath93 row and @xmath94 column of the covariance matrix . let @xmath95 be the vector containing the @xmath96 free parameters that completely describe the symmetric matrix @xmath20 . then , the fisher information due to the vector of parameters @xmath97 is equal to @xmath89 and is given by equation @xcite . @xmath98 multiplying equations and , we have @xmath99 to derive the message length expression to encode data using a certain @xmath100 , substitute equations , , and , in equation using the number of free parameters of the distribution as @xmath101 . hence , @xmath102 to obtain the mml estimates of @xmath19 and @xmath20 , equation needs to be minimized . the mml estimate of @xmath19 is same as the maximum likelihood estimate ( given in equation ) . to compute the mml estimate of @xmath20 , we need to compute the gradient matrix of @xmath103 with respect to @xmath20 and is given by equation @xmath104 the mml estimate of @xmath20 is obtained by solving @xmath105 ( given in equation ) . @xmath106 we observe that the mml estimate @xmath107 is same as the _ unbiased _ estimate of the covariance matrix @xmath20 , thus , lending credibility for its preference over the traditional ml estimate ( equation ) . parameter estimates for two and three - dimensional vmf have been explored previously @xcite . mml estimators of three - dimensional vmf were explored in @xcite , where they demonstrate that the mml - based inference compares favourably against the traditional ml and map based estimation methods . we use the @xcite method to formulate the objective function ( equation ) corresponding to a generic vmf distribution . regarding choosing a reasonable prior ( in the absence of any supporting evidence ) for the parameters @xmath108 of a vmf distribution , @xcite and @xcite suggest the use of the following _ colourless _ prior that is uniform in direction , normalizable and locally uniform at the cartesian origin in @xmath16 " : @xmath109 regarding evaluating the fisher information , @xcite argue that in the general @xmath3-dimensional case , @xmath110 where @xmath111 and @xmath112 are described by equations and respectively . substituting equations , and in equation with number of free parameters @xmath113 , we have the net message length expression : @xmath114 to obtain the mml estimates of @xmath19 and @xmath16 , equation needs to be minimized . the estimate for @xmath19 is same as the maximum likelihood estimate ( equation ) . the resultant equation in @xmath16 that needs to be minimized is then given by : @xmath115 to obtain the mml estimate of @xmath16 , we need to differentiate equation ( [ eqn : i_kappa ] ) and set it to zero . @xmath116 the non - linear equation : @xmath117 does not have a closed form solution . we try both the newton and halley s method to find an approximate solution . we discuss both variants and comment on the effects of the two approximations in the experimental results . to be fair and consistent with @xcite and @xcite , we use the initial guess of the root as @xmath118 ( equation ( [ eqn : banerjee_approx ] ) ) and iterate twice to obtain the mml estimate . 1 . _ approximation using newton s method : _ @xmath119 2 . _ approximation using halley s method : _ @xmath120 the details of evaluating @xmath121 and @xmath122 are discussed in appendix [ subsec : appendix_derivations ] . equation ( [ eqn : mml_newton_approx ] ) gives the mml estimate ( @xmath123 ) using newton s method and equation ( [ eqn : mml_halley_approx ] ) gives the mml estimate ( @xmath124 ) using halley s method . we used these values of mml estimates in mixture modelling using vmf distributions . [ sec : mml_gaussian_est ] mixture modelling involves representing an observed distribution of data as a weighted sum of individual probability density functions . specifically , the problem we consider here is to model the mixture distribution @xmath0 as defined in equation . for some observed data @xmath125 ( @xmath34 is the sample size ) , and a mixture @xmath0 , the log - likelihood using the mixture distribution is as follows : @xmath126 where @xmath127 , @xmath5 and @xmath6 are the weight and probability density of the @xmath7 component respectively . for a fixed @xmath4 , the mixture parameters @xmath128 are traditionally estimated using a standard _ expectation - maximization_(em ) algorithm @xcite . this is briefly discussed below . the standard em algorithm is based on maximizing the log - likelihood function of the data ( equation ) . the maximum likelihood estimates are then given as @xmath129 . because of the absence of a closed form solution for @xmath130 , a gradient descent method is employed where the parameter estimates are iteratively updated until convergence to some local optimum is achieved @xcite . the em method consists of two steps : * _ e - step _ : each datum @xmath131 has fractional membership to each of the mixture components . these partial memberships of the data points to each of the components are defined using the _ responsibility matrix _ @xmath132 where @xmath133 denotes the conditional probability of a datum @xmath131 belonging to the @xmath7 component . the effective membership associated with each component is then given by @xmath134 * _ m - step _ : assuming @xmath135 be the estimates at some iteration @xmath136 , the expectation of the log - likelihood using @xmath135 and the partial memberships is then _ maximized _ which is tantamount to computing @xmath137 , the updated maximum likelihood estimates for the next iteration @xmath138 . the weights are updated as @xmath139 . the above sequence of steps are repeated until a certain convergence criterion is satisfied . at some intermediate iteration @xmath136 , the mixture parameters are updated using the corresponding ml estimates and are given below . * _ gaussian : _ the ml updates of the mean and covariance matrix are @xmath140 * _ von mises - fisher : _ the resultant vector sum is updated as @xmath141 . if @xmath142 represents the magnitude of vector @xmath143 , then the updated mean and concentration parameter are @xmath144 we will first describe the methodology involved in formulating the mml - based objective function . we will then discuss how em is applied in this context . we refer to the discussion in @xcite to briefly describe the intuition behind mixture modelling using mml . encoding of a message using mml requires the encoding of ( 1 ) the model parameters and then ( 2 ) the data using the parameters . the statement costs for encoding the mixture model and the data can be decomposed into : 1 . encoding the _ number of components _ @xmath4 : in order to encode the message losslessly , it is required to initially state the number of components . in the absence of background knowledge , one would like to model the prior belief in such a way that the probability decreases for increasing number of components . if @xmath145 , then @xmath146 . the prior reflects that there is a difference of one bit in encoding the _ numbers _ @xmath4 and @xmath147 . alternatively , one could assume a uniform prior over @xmath4 within some predefined range . the chosen prior has little effect as its contribution is minimal when compared to the magnitude of the total message length @xcite . 2 . encoding the _ weights _ @xmath148 which are treated as parameters of a multinomial distribution with sample size @xmath149 . the length of encoding all the weights is then given by the expression @xcite : + @xmath150 3 . encoding each of the _ component parameters _ @xmath151 as given by @xmath152 ( discussed in section [ subsec : mml_parameter_estimation ] ) . 4 . encoding the _ data _ : each datum @xmath131 can be stated to a finite precision which is dictated by the accuracy of measurement is a constant value and has no effect on the overall inference process . it is used in order to maintain the theoretical validity when making the distinction between _ probability _ and _ probability density_. ] . if the precision to which each element of a @xmath3-dimensional vector can be stated is @xmath153 , then the _ probability _ of a datum @xmath154 is given as @xmath155 where @xmath156 is the _ probability density _ given by equation . hence , the _ total _ length of its encoding is given by @xmath157 the entire data @xmath56 can now be encoded as : @xmath158 thus , the total message length of a @xmath4 component mixture is given by equation . @xmath159 note that the _ constant _ term includes the lattice quantization constant ( resulting from stating all the model parameters ) in a @xmath78-dimensional space , where @xmath78 is equal to the number of free parameters in the mixture model . the parameters of the mixture model are those that _ minimize _ equation . to achieve this we use the standard em algorithm ( section [ subsec : em_ml ] ) , where , iteratively , the parameters are updated using their respective _ mml estimates_. the component weights are obtained by differentiating equation with respect to @xmath5 under the constraint @xmath160 . the derivation of the mml updates of the weights is shown in appendix [ subsec : wts_mml ] and are given as : @xmath161 the parameters of the @xmath7 component are updated using @xmath162 and @xmath163 ( equations , ) , the partial memberships assigned to the @xmath7 component at some intermediate iteration @xmath136 and and are given below . * _ gaussian : _ the mml updates of the mean and covariance matrix are @xmath164 * _ von mises - fisher : _ the resultant vector sum is updated as @xmath165 . if @xmath142 represents the magnitude of vector @xmath143 , then the updated mean is given by equation . @xmath166 the mml update of the concentration parameter @xmath167 is obtained by solving @xmath168 after substituting @xmath169 and @xmath170 in equation . the em is terminated when the change in the total message length ( improvement rate ) between successive iterations is less than some predefined threshold . the difference between the two variants of standard em discussed above is firstly the objective function that is being optimized . in section [ subsec : em_ml ] , the log - likelihood function is _ maximized _ which corresponds to @xmath171 term in section [ subsec : em_mml ] . equation ( [ eqn : mixture_msglen ] ) includes additional terms that correspond to the cost associated with stating the mixture parameters . secondly , in the m - step , in section [ subsec : em_ml ] , the components are updated using their ml estimates whereas in section [ subsec : em_mml ] , the components are updated using their mml estimates . the standard em algorithms outlined above can be used only when the number of mixture components @xmath4 is fixed or known _ a priori_. even when the number of components are fixed , em has potential pitfalls . the method is sensitive to the initialization conditions . to overcome this , some reasonable start state for the em may be determined by initially clustering the data @xcite . another strategy is to run the em a few times and choose the best amongst all the trials . @xcite point out that , in the case of gaussian mixture modelling , em can converge to the boundary of the parameter space when the corresponding covariance matrix is nearly singular or when there are few initial members assigned to that component . inferring the right " number of mixture components for unlabelled data has proven to be a thorny issue @xcite and there have been numerous approaches proposed that attempt to tackle this problem @xcite . given some observed data , there are infinitely many mixtures that one can fit to the data . any method that aims to selectively determine the optimal number of components should be able to factor the cost associated with the mixture parameters . to this end , several methods based on information theory have been proposed where there is some form of penalty associated with choosing a certain parameter value @xcite . we briefly review some of these methods and discuss the state of the art and then proceed to explain our proposed method . aic in the simplest form adds the _ number _ of free parameters @xmath78 to the negative log - likelihood expression . there are some variants of aic suggested @xcite . however , these variants introduce the same penalty constants for each additional parameter : @xmath172 bic , similar to aic , adds a constant multiple of @xmath173 ( @xmath34 being the sample size ) , for each free parameter in the model . @xcite formulated minimum description length ( mdl ) which formally coincides with bic @xcite . aic and bic / mdl serve as scoring functions to evaluate a model and its corresponding fit to the data . the formulations suggest that the parameter cost associated with adopting a model is dependent only on the number of free parameters and _ not _ on the parameter values themselves . in other words , the criteria consider all models of a particular type ( of probability distribution ) to have the same statement cost associated with the parameters . for example , a generic @xmath3-dimensional gaussian distribution has @xmath175 free parameters . all such distributions will have the same parameter costs regardless of their characterizing means and covariance matrices , which is an oversimplifying assumption which can hinder proper inference . the criteria can be interpreted under the mml framework wherein the first part of the message is a constant multiplied by the number of free parameters . aic and bic formulations can be obtained as approximations to the two - part mml formulation governed by equation @xcite . it has been argued that for tasks such as mixture modelling , where the number of free parameters potentially grows in proportion to the data , mml is known in theory to give consistent results as compared to aic and bic @xcite . to determine the optimal number of mixture components @xmath4 , the aic or bic scores are computed for mixtures with varying values of @xmath4 . the mixture model with the least score is selected as per these criteria . a @xmath3-variate gaussian mixture with @xmath4 number of components has @xmath176 free parameters . all mixtures with a set number of components have the same cost associated with their parameters using these criteria . the mixture complexity is therefore treated as independent of the constituent mixture parameters . in contrast , the mml formulation incorporates the statement cost of losslessly encoding mixture parameters by calculating their relevant probabilities as discussed in section [ sec : mml_mixture_modelling ] . a mml - based scoring function akin to the one shown in equation was used to model gaussian mixtures . however , the authors only consider the specific case of gaussians with diagonal covariance matrices , and fail to provide a general method dealing with full covariance matrices . a rigorous treatment on the selection of number of mixture components @xmath4 is lacking . @xcite experiment with different values of @xmath4 and choose the one which results in the minimum message length . for each @xmath4 , the standard em algorithm ( section [ subsec : em_ml ] ) was used to attain local convergence . the method , also referred to as _ laplace - empirical criterion _ ( lec ) @xcite , uses a scoring function derived using bayesian inference and serves to provide a tradeoff between model complexity and the quality of fit . the parameter estimates @xmath128 are those that result in the minimum value of the following scoring function . @xmath177 where @xmath56 is the dataset , @xmath178 is the negative log - likelihood given the mixture parameters @xmath128 , @xmath4 is the number of mixture components , @xmath3 the dimensionality of the data , @xmath179 are hyperparameters ( which are set to 1 in their experiments ) , @xmath180 is a pre - defined constant or is pre - computed using the entire data , @xmath181 is the hessian matrix which is equivalent to the empirical fisher matrix for the set of component parameters , and @xmath78 is the number of free parameters in the model . the formulation in equation can be obtained as an approximation to the message length expression in equation by identifying the following related terms in both equations . 1 . @xmath182 2 . for a @xmath3-variate gaussian with mean @xmath19 and covariance matrix @xmath20 , the joint prior @xmath183 is calculated as follows : * _ prior on @xmath19 : _ each of the @xmath3 parameters of the mean direction are assumed to be have uniform priors in the range @xmath184 , so that the prior density of the mean is @xmath185 . * _ prior on @xmath20 : _ it is assumed that the prior density is dependent only on the diagonal elements in @xmath20 . each diagonal covariance element is assumed to have a prior in the range @xmath186 so that the prior on @xmath20 is considered to be @xmath187 + the joint prior , is therefore , assumed to be @xmath188 . + thus , @xmath189 + 3 . @xmath190 + 4 . @xmath191 although the formulation is an improvement over the previously discussed methods , there are some limitations due to the assumptions made while proposing the scoring function : * while computing the prior density of the covariance matrix , the off - diagonal elements are ignored . * the computation of the determinant of the fisher matrix is approximated by computing the hessain @xmath192 . it is to be noted that while the hessian is the _ observed information _ ( data dependent ) , the fisher information is the _ expectation _ of the observed information . mml formulation requires the use of the expected value . * further , the approximated hessian was derived for gaussians with diagonal covariances . for gaussians with full covariance matrices , the hessian was approximated by replacing the diagonal elements with the corresponding eigen values in the hessian expression . the empirical fisher computed in this form does not guarantee the charactersitic invariance property of the classic mml method @xcite . the search method used to select the optimal number of components is rudimentary . the optimal number of mixture components is chosen by running the em 10 times for every value of @xmath4 within a given range . an optimal @xmath4 is selected as the one for which the best of the 10 trials results in the least value of the scoring function . the icl criterion _ maximizes _ the _ complete log - likelihood _ ( cl ) given by @xmath193 where @xmath194 is the log - likelihood ( equation ) , @xmath133 is the responsibility term ( equation ) , and @xmath195 if @xmath131 arises from component @xmath196 and zero otherwise . the term @xmath197 is explained as the estimated mean entropy . the icl criterion is then defined as : @xmath198where @xmath78 is the number of free parameters in the model . we observe that similar to bic , the icl scoring function penalizes each free parameter by a constant value and does not account for the model parameters . the search method adopted in this work is similar to the one used by @xcite . the em algorithm is initiated 20 times for each value of @xmath4 with random starting points and the best amongst those is chosen . the method uses the mml criterion to formulate the scoring function given by equation . the formulation can be intrepreted as a two - part message for encoding the model parameters and the observed data . @xmath199 where @xmath200 is the _ number _ of free parameters per component and @xmath5 is the component weight . the scoring function is derived from equation by assuming the prior density of the component parameters to be a jeffreys prior . if @xmath151 is the vector of parameters describing the @xmath7 component , then the prior density @xmath201 @xcite . similarly , a prior for weights would result in @xmath202 . these assumptions are used in the encoding of the parameters which correspond to the first part of the message . we note that the scoring function is consistent with the mml scheme of encoding parameters and the data using those parameters . however , the formulation can be improved by amending the assumptions as detailed in in section [ sec : mml_est_derivations ] . further , the assumptions made in @xcite have the following side effects : * the value of @xmath203 gives the cost of encoding the component parameters . by assuming @xmath201 , the message length associated with using any vector of parameters @xmath151 is essentially treated the same . to avoid this , the use of independent uniform priors over non - informative jeffreys s priors was advocated previously @xcite . the use of jeffreys prior offers certain advantages , for example , not having to compute the fisher information @xcite . however , this is crucial and can not be ignored as it dictates the _ precision of encoding the parameter vector_. @xcite states that jeffreys , while noting the interesting properties of the prior formulation did not advocate its use as a genuine expression of prior knowledge . " + by making this assumption , @xcite _ sidestep _ " the difficulty associated with explicitly computing the fisher information associated with the component parameters . hence , for encoding the parameters of the entire mixture , _ only _ the cost associated with encoding the component weights is considered . * the code length to state each @xmath151 is , therefore , greatly simplified as @xmath204 ( notice the sole dependence on weight @xmath5 ) . @xcite interpret this as being similar to a mdl formulation because @xmath205 gives the expected number of data points generated by the @xmath7 component . this is equivalent to the bic criterion discussed earlier . we note that mdl / bic are highly simplified versions of mml formulation and therefore , equation does not capture the entire essence of complexity and goodness of fit accurately . the method begins by assuming a large number of components and updates the weights iteratively in the em steps as @xmath206 where @xmath207 is the effective membership of data points in @xmath7 component ( equation ) . a component is annihilated when its weight becomes zero and consequently the number of mixture components decreases . we note that the search method proposed by @xcite using the mml criterion is an improvement over the methods they compare against . however , we make the following remarks about their search method . * the method updates the weights as given by equation . during any iteration , if the amount of data allocated to a component is less than @xmath208 , its weight is updated as zero and this component is ignored in subsequent iterations . this imposes a lower bound on the amount of data that can be assigned to each component . as an example , for a gaussian mixture in 10-dimensions , the number of free parameters per component is @xmath209 , and hence the lower bound is 33 . hence , in this exampe , if a component has @xmath210 data , the mixture size is reduced and these data are assigned to some other component(s ) . consider a scenario where there are 50 observed 10 dimensional data points originally generated by a mixture with two components with equal mixing proportions . the method would always infer that there is only one component regardless of the separation between the two components . this is clearly a wrong inference ! ( see section [ subsec : fj_weight_updates_exp2b ] for the relevant experiments ) . * once a component is discarded , the mixture size decreases by one , and it can not be recovered . because the memberships @xmath207 are updated iteratively using an em algorithm and because em might not always lead to global optimum , it is conceivable that the updated values need not always be optimal . this might lead to situations where a component is deleted owing to its low prominence . there is no provision to increase the mixture size in the subsequent stages of the algorithm to account for such behaviour . * the method assumes a large number of initial components in an attempt to be robust with respect to em initialization . however , this places a significant overhead on the computation due to handling several components . _ summary : _ we observe that while all these methods ( and many more ) work well within their defined scope , they are incomplete in achieving the true objective that is to rigorously score models and their ability to fit the data . the methods discussed above can be seen as different approximations to the mml framework . they adopted various simplifying assumptions and approximations . to avoid such limitations , we developed a classic mml formulation , giving the complete message length formulations for gaussian and von mises - fisher distributions in section [ sec : mml_est_derivations ] . secondly , in most of these methods , the search for the optimal number of mixture components is achieved by selecting the mixture that results in the best em outcome out of many trials @xcite . this is not an elegant solution and @xcite proposed a search heuristic which integrates estimation and model selection . a comparative study of these methods is presented in @xcite . their analysis suggested the superior performance of icl @xcite and lec @xcite . later , @xcite demonstrated that their proposed method outperforms the contemporary methods based on icl and lec and is regarded as the current state of the art . we , therefore , compare our method against that of @xcite and demonstrate its effectiveness . with this background , we formulate an alternate search heuristic to infer the optimal number of mixture components which aims to address the above limitations . the space of candidate mixture models to explain the given data is infinitely large . as per the mml criterion ( equation ) , the goal is to search for the mixture that has the smallest overall message length . we have seen in section [ subsec : em_mml ] that if the number of mixture components are fixed , then the em algorithm can be used to estimate the mixture parameters , namely the component weights and the parameters of each component . however , here it is required to search for the optimal _ number _ of mixture components along with the corresponding mixture parameters . our proposed search heuristic extends the mml - based snob program @xcite for unsupervised learning . we define three operations , namely _ split , delete , _ and _ merge _ that can be applied to any component in the mixture . [ algm ] @xmath211 the pseudocode of our search method is presented in algorithm [ algm ] . the basic idea behind the search strategy is to _ perturb _ a mixture from its current suboptimal state to obtain a new state ( if the perturbed mixture results in a smaller message length ) . in general , if a ( current ) mixture has @xmath4 components , it is perturbed using a series of _ split , delete _ , and _ merge _ operations to check for improvement . each component is split and the new @xmath212-component mixture is re - estimated . if there is an improvement ( _ i.e. , _ if there is a decrease in message length with respect to the current mixture ) , the new @xmath212-component mixture is retained . there are @xmath4 splits possible and the one that results in the greatest improvement is recorded ( lines 5 7 in algorithm [ algm ] ) . a component is first split into two sub - components ( children ) which are locally optimized by the em algorithm on the data that belongs to that sole component . the child components are then integrated with the others and the mixture is then optimized to generate a @xmath147 component mixture . the reason for this is , rather than use random initial values for the em , it is better if we start from some already optimized state to reach to a better state . similarly , each of the components is then deleted , one after the other , and the @xmath213-component mixture is compared against the current mixture . there are @xmath4 possible deletions and the best amongst these is recorded ( lines 8 11 in algorithm [ algm ] ) . finally , the components in the current mixture are merged with their closest matches ( determined by calculating the kl - divergence ) and each of the resultant @xmath213-component mixtures are evaluated against the @xmath4 component mixture . the best among these merged mixtures is then retained ( lines 12 15 in algorithm [ algm ] ) . we initially start by assuming a one component mixture . this component is split into two children which are locally optimized . if the split results in a better model , it is retained . for any given @xmath4-component mixture , there might be improvement due to splitting , deleting and/or merging its components . we select the perturbation that best improves the current mixture . this process is repeated until there is no further improvement possible . and the algorithm is continued . the notion of _ best _ or improved mixture is based on the amount of reduction of message length that the perturbed mixture provides . in the current state , the observed data have partial memberships in each of the @xmath4 components . before the execution of each operation , these memberships need to be adjusted and a em is subsequently carried out to achieve an optimum with a different number of components . we will now examine each operation in detail and see how the memberships are affected after each operation . let @xmath214 $ ] be the @xmath215 responsibility ( membership ) matrix and @xmath5 be the weight of @xmath7 component in mixture @xmath0 . 1 . _ split ( line 6 in algorithm [ algm ] ) : _ as an example , assume a component with index @xmath216 and weight @xmath217 in the current mixture @xmath0 is split to generate two child components . the goal is to find two distinct clusters amongst the data associated with component @xmath218 . it is to be noted that the data have fractional memberships to component @xmath218 . the em is therefore , carried out _ within _ the component @xmath218 assuming a _ two - component sub - mixture _ with the data weighted as per their current memberships @xmath219 . the remaining @xmath220 components are untouched . an em is carried out to optimize the two - component sub - mixture . the initial state and the subsequent updates in the maximization - step are described below . + _ parameter initialization of the two - component sub - mixture : _ the goal is to identify two distinct clusters within the component @xmath218 . for _ gaussian _ mixtures , to provide a reasonable starting point , we compute the direction of maximum variance of the parent component and locate two points which are one standard deviation away on either side of its mean ( along this direction ) . these points serve as the initial means for the two children generated due to splitting the parent component . selecting the initial means in this manner ensures they are reasonably apart from each other and serves as a good starting point for optimizing the two - component sub - mixture . the memberships are initialized by allocating the data points to the closest of the two means . once the means and the memberships are initialized , the covariance matrices of the two child components are computed . + there are conceivably several variations to how the two - component sub - mixture can be initialized . these include random initialization , selecting two data points as the initial component means , and many others . however , the reason for selecting the direction of maximum variance is to utilize the available characteristic of data , _ i.e. , _ the distribution within the component @xmath218 . for _ von mises - fisher _ mixtures , the maximum variance strategy ( as for gaussian mixtures ) can not be easily adopted , as the data is distributed on the hypershpere . hence , in this work , we randomly allocate data memberships and compute the components ( initial ) parameters . + once the parameters of the sub - mixture are initialized , an em algorithm is carried out ( just for the sub - mixture ) with the following maximization - step updates . let @xmath221 $ ] be the @xmath222 responsibility matrix for the two - component sub - mixture . for @xmath223 , let @xmath224 be the effective memberships of data belonging to the two child components , let @xmath225 be the weights of the child components within the sub - mixture , and let @xmath226 be the parameters describing the child components . * the effective memberships are updated as given by equation . @xmath227 * as the sub - mixture comprises of two child components , substitute @xmath228 in equation to obtain the updates for the weights . these are given by equation . @xmath229 * for _ gaussian _ mixtures , the component parameters @xmath230 are updated as follows : @xmath231 * for _ von mises - fisher _ mixtures , the component parameters @xmath232 are updated as follows : @xmath233 @xmath234 represents the magnitude of vector @xmath235 . the update of the concentration parameter @xmath236 is obtained by solving @xmath237 after substituting @xmath238 and @xmath239 in equation . + the difference between the em updates in equations , and equations , is the presence of the coefficient @xmath240 with each @xmath131 . since we are considering the sub - mixture , the original responsibility @xmath241 is multiplied by the responsibility within the sub - mixture @xmath242 to quantify the influence of datum @xmath131 to each of the child components . + after the sub - mixture is locally optimized , it is integrated with the untouched @xmath220 components of @xmath0 to result in a @xmath147 component mixture @xmath243 . an em is finally carried out on the combined @xmath147 components to estimate the parameters of @xmath243 and result in an optimized @xmath212-component mixture as follows . + _ em initialization for @xmath243 : _ usually , the em is started by a random initialization of the members . however , because the two - component sub - mixture is now optimal and the @xmath220 components in @xmath0 are also in an optimal state , we exploit this situation to initialize the em ( for @xmath243 ) with a reasonable starting point . as mentioned above , the component with index @xmath218 with component weight @xmath217 is split . upon integration , the ( child ) components that replaced component @xmath218 will now correspond to indices @xmath244 and @xmath245 in the new mixture @xmath243 . let @xmath246 \,\forall 1 \leq i \leq n , 1 \leq j \leq m+1 $ ] be the responsibility matrix for the new mixture @xmath243 and let @xmath247 be the component weights in @xmath243 . * _ component weights : _ the weights are initialized as follows : @xmath248 * _ memberships : _ the responsibility matrix @xmath249 is initialized for all data @xmath250 as follows : @xmath251 where @xmath252 are the effective memberships of the components in @xmath243 . + with these starting points , the parameters of @xmath243 are estimated using the traditional em algorithm with updates in the maximization - step given by equations , , and . the em results in local convergence of the @xmath212-component mixture . if the resultant message length of encoding data using @xmath243 is lower than that due to @xmath0 , that means the perturbation of @xmath0 because of splitting component @xmath218 resulted in a new mixture @xmath243 that compresses the data better , and hence , is a better mixture model to explain the data . 2 . _ delete ( line 10 in algorithm [ algm ] ) : _ the goal here is to remove a component from the current mixture and check whether it results in a better mixture model to explain the observed data . assume the component with index @xmath218 and the corresponding weight @xmath217 is to be deleted from @xmath0 to generate a @xmath220 component mixture @xmath243 . once deleted , the data memberships of the component need to be redistributed between the remaining components . the redistribution of data results in a good starting point to employ the em algorithm to estimate the parameters of @xmath243 as follows . + _ em initialization for @xmath243 : _ let @xmath246 $ ] be the @xmath253 responsibility matrix for the new mixture @xmath243 and let @xmath247 be the weight of @xmath7 component in @xmath243 . * _ component weights : _ the weights are initialized as follows : @xmath254 it is to be noted that @xmath255 because the mml update expression in the m - step for the component weights always ensures non - zero weights during every iteration of the em algorithm ( see equation ) . * _ memberships : _ the responsibility matrix @xmath249 is initialized for all data @xmath250 as follows : @xmath256 where @xmath252 are the effective memberships of the components in @xmath243 . it is possible for a datum @xmath131 to have complete membership in component @xmath218 ( _ i.e. , _ @xmath257 ) , in which case , its membership is equally distributed among the other @xmath220 components ( _ i.e. , _ @xmath258 ) . + with these readjusted weights and memberships , and the constituent @xmath220 components , the traditional em algorithm is used to estimate the parameters of the new mixture @xmath243 . if the resultant message length of encoding data using @xmath243 is lower than that due to @xmath0 , that means the perturbation of @xmath0 because of deleting component @xmath218 resulted in a new mixture @xmath243 with better explanatory power , which is an improvement over the current mixture . 3 . _ merge ( line 14 in algorithm [ algm ] ) : _ the idea is to join a pair of components of @xmath0 and determine whether the resulting @xmath213-component mixture @xmath243 is any better than the current mixture @xmath0 . one strategy to identify an improved mixture model would be to consider merging all possible pairs of components and choose the one which results in the greatest improvement . this would , however , lead to a runtime complexity of @xmath259 , which could be significant for large values of @xmath4 . another strategy is to consider merging components which are close " to each other . for a given component , we identify its _ closest _ component by computing the kullback - leibler ( kl ) distance with all others and selecting the one with the least value . this would result in a linear runtime complexity of @xmath260 as computation of kl - divergence is a constant time operation . for every component in @xmath0 , its closest match is identified and they are merged to obtain a @xmath220 component mixture @xmath243 . merging the pair involves reassigning the component weights and the memberships . an em algorithm is then employed to optimize @xmath243 . + assume components with indices @xmath218 and @xmath261 are merged . let their weights be @xmath217 and @xmath262 ; and their responsibility terms be @xmath219 and @xmath263 respectively . the component that is formed by merging the pair is determined first . it is then integrated with the @xmath264 remaining components of @xmath0 to produce a @xmath213-component mixture @xmath243 . + _ em initialization for @xmath243 : _ let @xmath265 and @xmath266 be the weight and responsibility vector of the merged component @xmath267 respectively . they are given as follows : @xmath268 the parameters of this merged component are estimated as follows : * _ gaussian : _ the parameters @xmath269 are : @xmath270 * _ von mises - fisher : _ the parameters @xmath271 are : @xmath272 the concentration parameter @xmath273 is obtained by solving @xmath274 after substituting @xmath275 and @xmath276 in equation . + the merged component @xmath267 with weight @xmath265 , responsibility vector @xmath277 , and parameters @xmath278 is then integrated with the @xmath264 components . the merged component and its associated memberships along with the @xmath264 other components serve as the starting point for optimizing the new mixture @xmath243 . if @xmath243 results in a lower message length compared to @xmath0 that means the perturbation of @xmath0 because of merging the pair of components resulted in an improvement to the current mixture . r0.35 we explain the proposed inference of mixture components through the following example that was also considered by @xcite . consider a bivariate gaussian mixture shown in fig . [ fig : mix1 ] . the mixture has three components with equal weights of 1/3 each and their means at ( -2,0 ) , ( 0,0 ) , and ( 2,0 ) . the covariance matrices of the three components are the same and are equal to @xmath279 . we simulate 900 data points from this mixture ( as done by @xcite ) and employ the proposed search strategy . the progression of the search method using various operations is detailed below . _ search for the optimal mixture model : _ the method begins by inferring a one - component mixture @xmath280 ( see fig . [ fig : mix1_iter_1_splits](a ) ) . it then splits this component ( as described in _ split _ step of section [ subsec : search_operations ] ) and checks whether there is an improvement in explanation . the red ellipse in fig [ fig : mix1_iter_1_splits](b ) depicts the component being split . the direction of maximum variance ( dotted black line ) is first identified , and the means ( shown by black dots at the end of the dotted line ) are initialized . an em algorithm is then used to optimize the two children and this results in a mixture @xmath281 shown in fig [ fig : mix1_iter_1_splits](c ) . since the new mixture has a lower message length , the current is updated as @xmath281 . in the second iteration , each component in @xmath281 is iteratively split , deleted , and merged . [ fig : mix1_iter_2_splits ] shows the splitting ( red ) of the first component . on splitting , the new mixture @xmath282 results in a lower message length . deletion of the first component is shown in fig . [ fig : mix1_iter_2_deletions ] . before merging the first component , we identify its closest component ( the one with the least kl - divergence ) ( see fig . [ fig : mix1_iter_2_merges ] ) . deletion and merging operations , in this case , do not result in an improvement . these two operations have different intermediate em initializations ( figures [ fig : mix1_iter_2_deletions](b ) and [ fig : mix1_iter_2_merges](b ) ) but result in the same optimized one - component mixture . the same set of operations are performed on the second component in @xmath281 . in this particular case , splitting results in an improved mixture ( same as @xmath282 ) . @xmath282 is updated as the new parent and the series of split , delete , and merge operations are carried out on all components in @xmath282 . [ fig : mix1_iter_3 ] shows these operations on the first component . we see that splitting the first component in @xmath282 results in @xmath283 ( see fig . [ fig : mix1_iter_3](c ) ) . however , @xmath283 is not an improvement over @xmath282 as seen by the message lengths and is , therfore , discarded . similarly , deletion and merging of the components do not yield improvements to @xmath282 . the operations are carried out on the remaining two components in @xmath282 ( not shown in the figure ) too . these perturbations do not produce improved mixtures in terms of the total message length . since the third iteration does not result in any further improvement , the search terminates and the parent @xmath282 is considered to be the best mixture . + + in different stages of the search method , we have different intermediate mixtures . em is a gradient descent technique and it can get trapped in a local optimum . by employing the suggested search , we are exhaustively considering the possible options , and aiming to reduce the possibility of the em getting stuck in a local optimum . the proposed method infers a mixture by balancing the tradeoff due to model complexity and the fit to the data . this is particularly useful when there is no prior knowledge pertaining to the nature of the data . in such a case , this method provides an objective way to infer a mixture with suitable components that best explains the data through lossless compression . another example is shown in appendix [ subsec : appendix_mix2 ] , where the evolution of the inferred mixture is explored in the case of a mixture with overlapping components . + _ variation of the two - part message length : _ the search method infers three components and terminates . in order to demonstrate that the inferred number of components is the optimum number , we infer mixtures with increasing number of components ( until @xmath284 as an example ) and plot their resultant message lengths . for each @xmath285 , the standard em algorithm ( section [ subsec : em_mml ] ) is employed to infer the mixture parameters . r0.5 fig . [ fig : individual_msglens_gaussian ] shows the total message lengths to which the em algorithm converges for varying number of components @xmath4 . as expected , the total message length ( green curve ) drastically decreases initially until @xmath286 components are inferred . starting from @xmath287 , the total message length gradually increases , clearly suggesting that the inferred models are over - fitting the data with increasing statement cost to encode the additional parameters of these ( more complex ) models . we further elaborate on the reason for the initial decrease and subsequent increase in the total message length . as per mml evaluation criterion , the message length comprises of two parts statement cost for the parameters and the cost for stating the data using those parameters . the model complexity ( which corresponds to the mixture parameters ) increases with increasing @xmath4 . therefore , the first part of the message to encode parameters increases with an increase in the number of parameters . this behaviour is illustrated by the red curve in fig . [ fig : individual_msglens_gaussian ] . the first part message lengths are shown in red on the right side y - axis in the figure . as the mixture model becomes increasingly more complex , the error of fitting the data decreases . this corresponds to the second part of the message in the mml encoding framework . this behaviour is consistent with what is observed in fig . [ fig : individual_msglens_gaussian ] ( blue curve ) . there is a sharp fall until @xmath286 ; then onwards increasing the model complexity does not lower the error significantly . the error saturates and there is minimal gain with regards to encoding the data ( the case of overfitting ) . however , the model complexity dominates after @xmath286 . the optimal balance is achieved when @xmath286 . in summary , the message length at @xmath286 components was rightly observed to be the optimum for this example . we note that for a fixed number of mixture components , the em algorithm for the mml metric is monotonically decreasing . however , while searching for the number of components , mml continues to decrease until some optimum is found and then steadily increases as illustrated through this example . we compare our proposed inference methodology against the widely cited method of @xcite . the performance of their method is compared against that of bayesian information criterion ( bic ) , integrated complete likelihood ( icl ) , and approximate bayesian ( lec ) methods ( discussed in section [ sec : mixture_existing_methods ] ) . it was shown that the method of @xcite was far superior than bic , icl and lec ( using gaussian mixtures ) . in the following sections , we demonstrate through a series of experiments that our proposed approach to infer mixtures fares better when compared to that of @xcite . the experimental setup is as follows : we use a gaussian mixture @xmath288 ( true distribution ) , generate a random sample from it , and infer the mixture using the data . this is repeated 50 times and we compare the performance of our method against that of @xcite . as part of our analysis , we compare the number of inferred mixture components as well as the quality of mixtures . _ comparing message lengths : _ the mml framework allows us to objectively compare mixture models by computing the total message length used to encode the data . the difference in message lengths gives the log - odds posterior ratio of any two mixtures ( equation ) . given some observed data , and any two mixtures , one can determine which of the two best explains the data . our search methodology uses the scoring function ( @xmath289 ) defined in equation . as elaborated in section [ subsec : fj ] , @xcite use an approximated mml - like scoring function ( @xmath290 ) given by equation . we employ our search method and the method of @xcite to infer the mixtures using the same data ; let the inferred mixtures be @xmath291 and @xmath292 respectively . we compute two quantities : @xmath293 we use the two different scoring functions to compute the differences in message lengths of the resulting mixtures @xmath292 and @xmath291 . since the search method used to obtain @xmath291 optimizes the scoring function @xmath289 , it is expected that @xmath294 and consequently @xmath295 . this implies that our method is performing better using our defined objective function . however , if @xmath296 , this indicates that our inferred mixture @xmath291 results in a lower value of the scoring function that is defined by @xcite . such an evaluation not only demonstrates the superior performance of our search ( leading to @xmath291 ) using our defined scoring function but also proves it is better using the scoring function as defined by @xcite . _ kullback leibler ( kl ) divergence : _ in addition to using message length based evaluation criterion , we also compare the mixtures using kl - divergence @xcite . the metric gives a measure of the similarity between two distributions ( the lower the value , the more similar the distributions ) . for a mixture probability distribution , there is no analytical form to compute the metric . however , one can calculate its empirical value ( which asymptotically converges to the kl - divergence ) . in experiments relating to mixture simulations , we know the true mixture @xmath288 from which the data @xmath297 is being sampled . the kl - divergence is given by the following expression : @xmath298 \approx \frac{1}{n } \sum_{i=1}^n\log\frac{\pr(\mathbf{x}_i,{\mathcal{m}}^t)}{\pr(\mathbf{x}_i,{\mathcal{m}})}\end{aligned}\ ] ] where @xmath0 is a mixture distribution ( @xmath291 or @xmath292 ) whose _ closeness _ to the true mixture @xmath288 is to be determined . an experiment conducted by @xcite was to randomly generate @xmath299 data points from a two - component ( with equal mixing proportions ) bivariate mixture @xmath288 whose means are at @xmath300 and @xmath301 , and equal covariance matrices : @xmath302 ( the identity matrix ) , and compare the number of inferred components . we repeat the same experiment here and compare with the results of @xcite . the separation @xmath303 between the means is gradually increased and the percentage of the correct selections ( over 50 simulations ) as determined by the two search methods is plotted . [ fig : gaussian_2d_mixture_inference](a ) shows the results of this experiment . as the separation between the component means is increased , the number of correctly inferred components increases . we conducted another experiment where we fix the separation between the two components and increase the amount of data being sampled from the mixture . [ fig : gaussian_2d_mixture_inference](b ) illustrates the results for a separation of @xmath304 . as expected , increasing the sample size results in an increase in the number of correct selections . both the search methods eventually infer the true number of components at sample size @xmath305 . we note that in both these cases , the differences in message lengths @xmath306 and @xmath307 are close to zero . the kl - divergences for the mixtures inferred by the two search methods are also the same . therefore , for this experimental setup , the performance of both the methods is roughly similar . as the difference between the two search methods is not apparent from these experiments , we wanted to investigate the behaviour of the methods with smaller sample sizes . we repeated the experiment similar to that shown in fig . [ fig : gaussian_2d_mixture_inference](a ) but with a sample size of @xmath308 . our search method results in a mean value close to 1 for different values of @xmath303 ( see table [ tab : gaussian_2d_mixture_inference_boxplot ] ) . the mean value of the number of inferred components using the search method of @xcite fluctuates between 2 and 3 . however , there is significant variance in the number of inferred components ( see table [ tab : gaussian_2d_mixture_inference_boxplot ] ) . these results are also depicted through a boxplot ( fig . [ fig : gaussian_2d_mixture_inference_boxplot ] ) . there are many instances where the number of inferred components is more than 3 . the results indicate that the search method ( fj ) is overfitting the data . r0.5 further , we evaluate the correctness of the mixtures inferred by the two search methods by comparisons using the message length formulations and kl-divergence.fig . [ fig : gaussian_2d_msglen_comparisons ] shows the boxplot of the difference in message lengths of the mixtures @xmath291 inferred using our proposed search method and the mixtures @xmath292 inferred using that of @xcite . @xmath295 across all values of @xmath303 for the 50 simulations . as per equation , we have @xmath294 . this implies that @xmath291 has a lower message length compared to @xmath292 when evaluated using our scoring function . similarly , we have @xmath309 , _ i.e. , _ @xmath310 . this implies that @xmath292 has a lower message length compared to @xmath291 when evaluated using fj s scoring function . these results are not surprising as @xmath291 and @xmath292 are obtained using the search methods which optimize the respective mml and mml - like scoring functions . we then analyzed the kl - divergence of @xmath291 and @xmath292 with respect to the true bivariate mixture @xmath288 over all 50 simulations and across all values of @xmath303 . ideally , the kl - divergence should be close to zero . [ fig : gaussian_2d_kldiv_comparisons](a ) shows the kl - divergence of the mixtures inferred using the two search methods with respect to @xmath288 when the separation is @xmath304 . the proposed search method infers mixtures whose kl - divergence ( denoted by red lines ) is close to zero , and more importantly less than the kl - divergence of mixtures inferred by the search method of @xcite ( denoted by blue lines ) . the same type of behaviour is noticed with other values of @xmath303 . [ fig : gaussian_2d_kldiv_comparisons](b ) compares the kl - divergence for varying values of @xmath303 . the median value of the kl - divergence due to the proposed search method is close to zero with not much variation . the search method of @xcite always result in kl - divergence higher than that of ours . the results suggest that , in this case , mixtures @xmath292 inferred by employing the search method of @xcite deviate significantly from the true mixture distribution @xmath288 . this can also be explained by the fact that there is a wide spectrum of the number of inferred components ( fig . [ fig : gaussian_2d_mixture_inference_boxplot ] ) . this suggests that the mml - like scoring function is failing to control the tradeoff between complexity and quality of fit , and hence , is selecting overly complex mixture models . along the lines of the previous experiment , @xcite conducted another experiment for a 10-variate two - component mixture @xmath288 with equal mixing proportions . the means are at @xmath311 and @xmath312 , so that the euclidean distance between them is @xmath313 . the covariances of the two components are @xmath302 ( the identity matrix ) . random samples of size @xmath299 were generated from the mixture and the number of inferred components are plotted . the experiment is repeated for different values of @xmath303 and over 50 simulations . [ fig : gaussian_10d_mixture_inference](a ) shows the number of inferred components using the two search methods . at lower values of @xmath303 , the components are close to each other , and hence , it is relatively more difficult to correctly infer the true number of components . we observe that our proposed method performs clearly better than that of @xcite across all values of @xmath303 . we also compared the quality of these inferred mixtures by calculating the difference in message lengths using the two scoring functions and the kl - divergence with respect to @xmath288 . for all values of @xmath303 , @xmath295 , _ i.e. , _ our inferred mixtures @xmath291 have a lower message length compared to @xmath292 when evaluated using our scoring function . more interestingly , we also note that @xmath314 ( see fig . [ fig : gaussian_10d_comparisons](a ) ) . this reflects that @xmath291 have a lower message length compared to @xmath292 when evaluated using the scoring function of @xcite . this suggests that their search method results in a sub - optimal mixture @xmath292 and fails to infer the better @xmath291 . in addition to the message lengths , we analyze the mixtures using kl - divergence . similar to the bivariate example in fig . [ fig : gaussian_2d_kldiv_comparisons](a ) , the kl - divergence of our inferred mixtures @xmath291 is lower than @xmath292 , the mixtures inferred by @xcite . [ fig : gaussian_10d_comparisons](b ) shows the boxplot of kl - divergence of the inferred mixtures @xmath291 and @xmath292 . at higher values of @xmath315 , the median value of kl - divergence is close to zero , as the number of correctly inferred components ( fig . [ fig : gaussian_10d_mixture_inference](a ) ) is more than 90% . however , our method always infers mixtures @xmath291 with lower kl - divergence compared to @xmath292 . these experimental results demonstrate the superior performance of our proposed search method . another experiment was carried out where the @xmath316 was held constant ( extremely close components ) , gradually increased the sample size @xmath34 , and plotted the average number of inferred components by running 50 simulations for each @xmath34 . [ fig : gaussian_10d_mixture_inference](b ) shows the results for the average number of inferred components as the amount of data increases . our search method , on average , infers the true mixture when the sample size is @xmath317 . however , the search method of @xcite requires larger amounts of data ; even with a sample size of 2000 , the average number of inferred components is @xmath318 . in fig . [ fig : gaussian_10d_mixture_inference](b ) , the red curve reaches the true number of 2 and saturates more rapidly than the blue curve . one of the drawbacks associated with the search method of @xcite is due to the form of the updating expression for the component weights ( equation ) . as discussed in section [ subsubsec : fj_search_drawbacks ] , a particular instance of wrong inference is bound to happen when the net membership of a ( valid ) component is less than @xmath208 , where @xmath200 is the number of free parameters per component . in such a case , the component weight is updated as zero , and it is eliminated , effectively reducing the mixture size by one . we conducted the following experiment to demonstrate this behaviour : we considered the two - component 10-variate mixture @xmath288 as before and randomly generate samples of size 50 from the mixture . since the constituent components of @xmath288 have equal weights , on average , each component has a membership of 25 . we used @xmath319 , so that the two components are well apart from each other . for each @xmath303 , we run 50 simulations and analyze the number of inferred components . as expected , the search method of @xcite always infer a mixture with one component regardless of the separation @xmath303 . our method always infers the correct number of components . in order to test the validity of mixtures inferred by our proposed method , we analyze the resultant mixtures by comparing the message lengths as discussed in section [ subsec : comparison_mixtures ] . [ fig : gaussian_10d_msglens_comparisons_exp2b](a ) shows the difference in message lengths @xmath306 given in equation . we observe that @xmath295 for all @xmath303 . this demonstrates that our search based mixtures @xmath291 have lower message lengths compared to mixtures @xmath292 using our scoring function . the same phenomenon is observed when using the mml - like scoring function of @xcite . in fig . [ fig : gaussian_10d_msglens_comparisons_exp2b](b ) , we observe that @xmath314 , which means our search based mixtures @xmath291 have lower message lengths compared to mixtures @xmath292 when evaluated using their scoring function . r0.45 this demonstrates that @xmath291 is a better mixture as compared to @xmath292 and their search method is unable to infer it . we also note that the differences in message lengths increases with increasing @xmath303 . this is because for the one - component inferred mixture @xmath292 , the second part of the message ( equation ) which corresponds to the negative log - likelihood term increases because of poorer fit to the data . the two modes in the data become increasingly pronounced as the separation between constituent components of the true mixture increases , and hence , modelling such a distribution using a one - component mixture results in a poorer fit . this is clearly an incorrect inference . we further strengthen our case by comparing the kl - divergence of the inferred mixtures @xmath291 and @xmath292 with respect to the true mixture . [ fig : gaussian_10d_kldiv_comparisons_exp2b ] illustrates the results . as @xmath303 increases , the blue coloured plots shift higher . these correspond to mixtures @xmath292 inferred by @xcite . our search method , however , infers mixtures @xmath291 which have lower kl - divergence . the figure indicates that the inferred mixtures @xmath291 are more similar to the true distribution as compared to mixtures @xmath292 . these experiments demonstrate the ability of our search method to perform better than the widely used method of @xcite . we compared the resulting mixtures using our proposed mml formulation and the mml - like formulation of @xcite , showing the advantages of the former over the latter . we also used a neutral metric , kl - divergence , to establish the closeness of our inferred mixtures to the true distributions . we will now illustrate the behaviour of our search method on two real world datasets . at any intermediate stage of the search procedure , current _ mixture with @xmath4 components requires @xmath4 number of split , delete , and merge operations before it is updated . each of the perturbations involve performing an em to optimize the corresponding mixture parameters . to determine the convergence of em , we used a threshold of @xmath320 which was the same as used by @xcite . fj s method also requires to start from an initial large number of components . we used 25 as an initial number based on what was suggested in @xcite . we investigate the number of times the em routine is called and compare it with that of @xcite . we examine with respect to the simulations that were carried out previously . for the bivariate mixture discussed in section [ subsec : bivariate_mix_simulation ] , the number of resulting em iterations when the sample sizes are @xmath299 and @xmath308 are compared in fig . [ fig : em_iterations](a ) , ( b ) respectively . as per the discussion in section [ subsec : bivariate_mix_simulation ] , at @xmath299 , the average number of components inferred by the two methods are about the same ( fig . [ fig : gaussian_2d_mixture_inference](a ) ) . however , the number of em iterations required by fj s method is greater than 200 across all values of @xmath303 ( fig . [ fig : em_iterations](a ) ) . in contrast , the proposed method , on average , requires fewer than 50 iterations . in this case , both methods produce a similar result with fj s method requiring more number of em iterations . when the bivariate mixture simulation is carried out using @xmath308 , the number of em iterations required by fj s method , on average , is greater than 100 , while the proposed method requires fewer than 40 iterations ( fig . [ fig : em_iterations](b ) ) . in this case , the proposed method not only infers better mixtures ( as discussed in section [ subsec : bivariate_mix_simulation ] ) but is also conservative with respect to computational cost . for the simulation results corresponding to the 10-variate mixtures in section [ subsec:10d_mix_simulation ] , the proposed method requires close to 50 iterations on average , while fj s method requires about 20 ( fig . [ fig : em_iterations](c ) ) . however , the mixtures inferred by the proposed method fare better when compared to that of fj ( figs . [ fig : gaussian_10d_mixture_inference ] , [ fig : gaussian_10d_comparisons ] ) . furthermore , for the simulation results explained in section [ subsec : fj_weight_updates_exp2b ] , fj s method stops after 3 em iterations . this is because their program does not accommodate components when the memberships are less than @xmath208 . the proposed method requires 18 em iterations on average and infers the correct mixture components . in these two cases , our method infers better quality mixtures , with no significant overhead with regard to the computational cost . the first example is the univariate _ acidity _ data set which contains 155 points . our proposed search method infers a mixture @xmath291 with 2 components whereas the search method of @xcite infers a mixture @xmath292 with 3 components . the inferred mixtures are shown in fig . [ fig : acidity ] and their corresponding parameter estimates are given in table [ tab : acidity_mixtures_inference ] . in order to compare the mixtures inferred by the two search methods , we compute the message lengths of the inferred mixtures using our complete mml and the approximated mml - like scoring functions . when evaluated using our mml scoring function , our inferred mixture results in a gain of @xmath321 bits ( see table [ tab : acidity_comparison ] ) . based on the mml framework , our two - component mixture @xmath291 is @xmath322 times more likely than the three - component mixture @xmath292 ( as per equation ) . furthermore , when the inferred mixtures are evaluated as per the mml - like scoring function , @xmath291 is still considered better ( @xmath323 bits ) than @xmath292 ( @xmath324 bits ) . thus , using both forms of scoring function , @xmath291 is the better mixture model of this data set . [ tab : acidity_mixtures_inference ] .message lengths ( measured in bits ) of the mixtures ( in fig . [ fig : acidity ] ) as evaluated using the mml and mml - like scoring functions . [ cols="^,^,^ " , ] [ tab : null_models_comparison ] the per residue statistic is calculated by dividing the total message length by the sample size ( the number of @xmath325 pairs ) . this statistic shows that close to 2.5 bits can be saved ( on average ) if the protein data is encoded using the vmf null model . the vmf null model thus supercedes the naive model of encoding . this can potentially improve the accuracy of statistical inference that is central to the various protein modelling tasks briefly introduced above . we presented a statistically robust approach for inferring mixtures of ( i ) multivariate gaussian distributions , and ( ii ) von mises - fisher distributions for @xmath3-dimensional directional data . it is based on the general information - theoretic framework of minimum message length inference . this provides an objective tradeoff between the hypothesis complexity and the quality of fit to the data . an associated search procedure for an optimal mixture model of given data chooses the number of component distributions , @xmath4 , by minimizing the total message length . we established the better performance of the proposed search algorithm by comparing with a popularly used search method @xcite . we demonstrated the effectiveness of our approach through extensive experimentation and validation of our results . we also applied our method to real - world high dimensonal text data and to directional data that arises from protein chain conformations . the experimental results demonstrate that our proposed method fares better when compared with the current state of the art techniques . the authors would like to thank arun konagurthu for discussions pertaining to protein data clustering , and maria garcia de la banda for stimulating discussions and providing interesting insights . the authors would also like to acknowledge wray buntine s inputs with regard to text clustering . 91 [ 1]#1 [ 1]#1 urlstyle [ 1]doi # 1 [ 2][]#2 agusta y , dowe dl ( 2003 ) unsupervised learning of correlated multivariate gaussian mixture models using mml . in : ai 2003 : advances in artificial intelligence , springer , berlin , heidelberg , pp 477489 akaike h ( 1974 ) a new look at the statistical model identification . ieee transactions on automatic control , 19(6):716723 anderson e ( 1935 ) the irises of the gasp peninsula . bulletin of the american iris society 59:25 banerjee a , dhillon i , ghosh j , sra s ( 2003 ) generative model - based clustering of directional data . in : proceedings of the 9th international conference on knowledge discovery and data mining , acm , new york , pp 1928 banerjee a , dhillon is , ghosh j , sra s ( 2005 ) clustering on the unit hypersphere using von mises - fisher distributions . journal of machine learning research 6:13451382 barton de ( 1961 ) unbiased estimation of a set of probabilities . biometrika 48(1 - 2):227229 basu ap ( 1964 ) estimates of reliability for some distributions useful in life testing . technometrics 6(2):215219 best d , fisher n ( 1981 ) the bias of the maximum likelihood estimators of the von mises - fisher concentration parameters . communications in statistics - simulation and computation 10(5):493502 biernacki c , celeux g , govaert g ( 2000 ) assessing a mixture model for clustering with the integrated completed likelihood . ieee transactions on pattern analysis and machine intelligence 22(7):719725 bishop cm ( 2006 ) pattern recognition and machine learning , vol 1 . springer , new york boulton d , wallace c ( 1969 ) the information content of a multistate distribution . journal of theoretical biology 23:269278 bozdogan h ( 1983 ) determining the number of component clusters in the standard multivariate normal mixture model using model - selection criteria . tech . rep . , dtic document bozdogan h ( 1990 ) on the information - based measure of covariance complexity and its application to the evaluation of multivariate linear models . communications in statistics - theory and methods 19(1):221278 bozdogan h ( 1993 ) choosing the number of component clusters in the mixture - model using a new informational complexity criterion of the inverse - fisher information matrix . springer , berlin , heidelberg burnham kp , anderson dr ( 2002 ) model selection and multimodel inference : a practical information - theoretic approach . springer , new york collier jh , allison l , lesk am , de la banda mg , konagurthu as ( 2014 ) a new statistical framework to assess structural alignment quality using information compression . bioinformatics 30(17):i512i518 conway jh , sloane nja ( 1984 ) on the voronoi regions of certain lattices . siam journal on algebraic and discrete methods 5:294305 cordeiro gm , vasconcellos kl ( 1999 ) theory & methods : second - order biases of the maximum likelihood estimates in von mises regression models . australian & new zealand journal of statistics 41(2):189198 dempster ap , laird nm , rubin db ( 1977 ) maximum likelihood from incomplete data via the em algorithm . journal of the royal statistical society : series b ( methodological ) 39(1):138 dowe dl , allison l , dix ti , hunter l , wallace cs , edgoose t ( 1996 ) circular clustering of protein dihedral angles by minimum message length . in : pacific symposium on biocomputing , vol 96 , pp 242255 dowe dl , oliver jj , baxter ra , wallace cs ( 1996 ) bayesian estimation of the von mises concentration parameter . in : maximum entropy and bayesian methods , springer , netherlands , pp 5160 dowe dl , oliver jj , wallace cs ( 1996 ) mml estimation of the parameters of the spherical fisher distribution . in : algorithmic learning theory , springer , berlin , heidelberg , pp 213227 drton m , sturmfels b , sullivant s ( 2009 ) lectures on algebraic statistics . in : oberwolfach seminars ( 39 ) , birkhuser , basel dwyer ps ( 1967 ) some applications of matrix derivatives in multivariate analysis . journal of the american statistical association 62(318):607625 eaton ml , morris cn ( 1970 ) the application of invariance to unbiased estimation . the annals of mathematical statistics 41(5):17081716 figueiredo ams ( 2012 ) goodness - of - fit for a concentrated von mises - fisher distribution . computational statistics 27(1):6982 figueiredo ma , jain ak ( 2002 ) unsupervised learning of finite mixture models . ieee transactions on pattern analysis and machine intelligence 24(3):381396 fisher ni ( 1993 ) statistical analysis of spherical data . cambridge university press , cambridge fisher r ( 1953 ) dispersion on a sphere . proceedings of the royal society of london : series a ( mathematical and physical sciences ) 217(1130):295305 fisher ra ( 1936 ) the use of multiple measurements in taxonomic problems . annals of eugenics 7(2):179188 gauvain j , lee ch ( 1994 ) maximum a posteriori estimation for multivariate gaussian mixture observations of markov chains . ieee transactions on speech and audio processing 2(2):291298 gray g ( 1994 ) bias in misspecified mixtures . biometrics 50(2):457470 hornik k , grn b ( 2013 ) movmf : an r package for fitting mixtures of von mises - fisher distributions . r package version 01 - 2 jain ak , dubes rc ( 1988 ) algorithms for clustering data . prentice - hall , inc . , upper saddle river , nj , usa jain ak , duin rpw , mao j ( 2000 ) statistical pattern recognition : a review . ieee transactions on pattern analysis and machine intelligence 22(1):437 jeffreys h ( 1946 ) an invariant form for the prior probability in estimation problems . proceedings of the royal society of london : series a ( mathematical and physical sciences ) 186(1007):453461 jones p , mclachlan g ( 1990 ) laplace - normal mixtures fitted to wind shear data . journal of applied statistics 17(2):271276 jorgensen ma , mclachlan gj ( 2008 ) wallace s approach to unsupervised learning : the snob program . the computer journal 51(5):571578 kent jt ( 1982 ) the fisher - bingham distribution on the sphere . journal of the royal statistical society : series b ( methodological ) 44(1):7180 konagurthu as , lesk am , allison l ( 2012 ) minimum message length inference of secondary structure from protein coordinate data . bioinformatics 28(12):i97i105 konagurthu as , allison l , abramson d , stuckey pj , lesk am ( 2013 ) statistical inference of protein lego bricks " . in : 2013 ieee 13th international conference on data mining ( icdm ) , ieee , pp 10911096 krishnan t , mclachlan g ( 1997 ) the em algorithm and extensions . wiley , new york kullback s , leibler ra ( 1951 ) on information and sufficiency . the annals of mathematical statistics pp 7986 lee p ( 1997 ) bayesian statistics : an introduction . arnold , london . lo y ( 2011 ) bias from misspecification of the component variances in a normal mixture . computational statistics and data anaysis 55(9):27392747 magnus jr , neudecker h ( 1988 ) matrix differential calculus with applications in statistics and econometrics . wiley , new york mardia k , jupp p ( 2000 ) directional statistics . wiley , hoboken , nj , usa mardia k , holmes d , kent j ( 1984 ) a goodness - of - fit test for the von mises - fisher distribution . journal of the royal statistical society : series b ( methodological ) 46(1):7278 mardia kv , kent jt , bibby jm ( 1979 ) multivariate analysis . academic press , london mardia kv , taylor cc , subramaniam gk ( 2007 ) protein bioinformatics and mixtures of bivariate von mises distributions for angular data . biometrics 63(2):505512 mclachlan g , peel d ( 1997 ) contribution to the discussion of paper by s. richardson and p.j . journal of the royal statistical society b 59:779780 mclachlan g , peel d ( 2000 ) finite mixture models . wiley , new york mclachlan gj , basford ke ( 1988 ) mixture models : inference and applications to clustering ( statistics : textbooks and monographs ) . dekker , new york murzin a , brenner s , hubbard t , chothia c , et al . ( 1995 ) scop : a structural classification of proteins database for the investigation of sequences and structures . journal of molecular biology 247(4):536540 oliver j , baxter r ( 1994 ) mml and bayesianism : similarities and differences . dept comput sci monash univ , clayton , victoria , australia , tech rep 206 oliver jj , baxter ra , wallace cs ( 1996 ) unsupervised learning using mml . in : machine learning : proceedings of the 13th international conference , pp 364372 patra k , dey dk ( 1999 ) a multivariate mixture of weibull distributions in reliability modeling . statistics & probability letters 45(3):225235 peel d , mclachlan gj ( 2000 ) robust mixture modelling using the t - distribution . statistics and computing 10(4):339348 peel d , whiten wj , mclachlan gj ( 2001 ) fitting mixtures of kent distributions to aid in joint set identification . journal of the american statistical association 96(453):5663 rao cr ( 1973 ) linear statistical inference and its applications . wiley , new york richardson s , green pj ( 1997 ) on bayesian analysis of mixtures with an unknown number of components . journal of the royal statistical society : series b ( methodological ) 59(4):731792 rissanen j ( 1978 ) modeling by shortest data description . automatica 14(5):465471 rissanen j ( 1989 ) stochastic complexity in statistical inquiry theory . world scientific publishing co. , inc . , river edge , nj , usa roberts s , husmeier d , rezek i , penny w ( 1998 ) bayesian approaches to gaussian mixture modeling . ieee transactions on pattern analysis and machine intelligence 20(11):11331142 robertson s , zaragoza h ( 2009 ) the probabilistic relevance framework : bm25 and beyond . now publishers inc . , hanover , ma , usa salton g , buckley c ( 1988 ) term - weighting approaches in automatic text retrieval . information processing & management 24(5):513523 salton g , mcgill mj ( 1986 ) introduction to modern information retrieval . mcgraw - hill , inc . , new york , ny , usa schou g ( 1978 ) estimation of the concentration parameter in von mises fisher distributions . biometrika 65(2):369377 schwarz g , et al . ( 1978 ) estimating the dimension of a model . the annals of statistics 6(2):461464 seidel w , mosler k , alker m ( 2000 ) a cautionary note on likelihood ratio tests in mixture models . annals of the institute of statistical mathematics 52(3):481487 shannon ce ( 1948 ) a mathematical theory of communication . the bell system technical journal 27:379423 song h , liu j , wang g ( 2012 ) high - order parameter approximation for von mises fisher distributions . applied mathematics and computation 218(24):11,88011,890 sra s ( 2012 ) a short note on parameter approximation for von mises - fisher distributions : and a fast implementation of @xmath326 . computational statistics 27(1):177190 strehl a , ghosh j , mooney r ( 2000 ) impact of similarity measures on web - page clustering . in : workshop on artificial intelligence for web search ( aaai 2000 ) , pp 5864 taboga m ( 2012 ) lectures on probability theory and mathematical statistics . createspace independent pub . tanabe a , fukumizu k , oba s , takenouchi t , ishii s ( 2007 ) parameter estimation for von mises fisher distributions . computational statistics 22(1):145157 titterington dm , smith af , makov ue , et al . ( 1985 ) statistical analysis of finite mixture distributions . wiley , new york wallace c ( 1986 ) an improved program for classification . in : proceedings of the 9th australian computer science conference , pp 357366 wallace c ( 1990 ) classification by minimum - message - length inference . lecture notes in computer science , vol 468 , springer berlin heidelberg , pp 7281 wallace c , dowe d ( 1994 ) estimation of the von mises concentration parameter using minimum message length . in : proceedings of the 12th australian statistical society conference , monash university , australia wallace cs ( 2005 ) statistical and inductive inference using minimum message length . information science and statistics , springer - verlag , secaucus , nj , usa wallace cs , boulton dm ( 1968 ) an information measure for classification . computer journal 11(2):185194 wallace cs , dowe dl ( 1999 ) minimum message length and kolmogorov complexity . computer journal 42:270283 wallace cs , freeman pr ( 1987 ) estimation and inference by compact coding . journal of the royal statistical society : series b ( methodological ) 49(3):240265 wang p , puterman ml , cockburn i , le n ( 1996 ) mixed poisson regression models with covariate dependent rates . biometrics 52(2):381400 watson g , williams e ( 1956 ) on the construction of significance tests on the circle and the sphere . biometrika 43(3 - 4):344352 wedel m , desarbo ws , bult jr , ramaswamy v ( 1993 ) a latent class poisson regression model for heterogeneous count data . journal of applied econometrics 8(4):397411 white h ( 1982 ) maximum likelihood estimation of misspecified models . econometrica 50(1):125 wood at ( 1994 ) simulation of the von mises fisher distribution . communications in statistics - simulation and computation 23(1):157164 xu l , jordan mi ( 1996 ) on convergence properties of the em algorithm for gaussian mixtures . neural computation 8(1):129151 zhong s , ghosh j ( 2003 ) a comparative study of generative models for document clustering . in : proceedings of the workshop on clustering high dimensional data and its applications in siam data mining conference for brevity , we represent @xmath327 and @xmath328 as @xmath329 and @xmath330 respectively . expressions to evaluate @xmath331 and @xmath332 are given in equations ( [ eqn : ratio_bessels ] ) , ( [ eqn : ratio_first_derivative ] ) , and ( [ eqn : ratio_second_derivative ] ) respectively . we require @xmath330 for its use in the remainder of the derivation and we provide its expression below : @xmath333 now we discuss the derivation of @xmath121 and @xmath122 that are required for computing the mml estimate @xmath334 ( equations ( [ eqn : mml_newton_approx ] ) and ( [ eqn : mml_halley_approx ] ) ) . differentiating equation [ eqn : i_first_derivative ] , we have @xmath335 @xmath336 equations ( [ eqn : tmp1 ] ) and ( [ eqn : tmp2 ] ) can be used to evaluate @xmath121 and @xmath122 which can then be used to approximate the mml estimates @xmath123 and @xmath124 ( equations ( [ eqn : mml_newton_approx ] ) and ( [ eqn : mml_halley_approx ] ) respectively ) . to obtain the optimal weights under the constraint @xmath160 , the above equation is optimized using the _ lagrangian _ objective function defined below using some _ lagrangian multiplier _ @xmath341 . @xmath342 for some @xmath343 , the equation resulting from computing the partial derivative of @xmath344 with respect to @xmath345 and equating it to zero gives the optimal weight @xmath345 . @xmath346 @xmath347 where @xmath348 and @xmath349 are the responsibility and effective membership terms given as per equations and respectively . substituting the above value in equation , we have @xmath350 there are @xmath4 equations similar to equation for values of @xmath343 . adding all these equations together , we get @xmath351 substituting the above value of @xmath341 in equation , we get @xmath352 [ [ subsec : vmf_kldiv ] ] derivation of the kullback - leibler ( kl ) distance between two von mises - fisher distributions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ the closed form expression to calculate the kl divergence between two vmf distributions is presented below . let @xmath353 and @xmath354 be two von mises - fisher distributions with mean directions @xmath355 and concentration parameters @xmath356 . the kl distance between any two distributions is given by @xmath357 \label{eqn : kldiv}\ ] ] where @xmath358 $ ] is the expectation of the quantity @xmath359 $ ] using the probability density function @xmath17 . @xmath360 using the fact that @xmath361 = a_d(\kappa_1)\boldsymbol{\mu_1}$ ] @xcite , we have the following expression : @xmath362 & = \log\frac{c_d(\kappa_1)}{c_d(\kappa_2 ) } + ( \kappa_1\boldsymbol{\mu_1 } -\kappa_2\boldsymbol{\mu_2})^t a_d(\kappa_1 ) \boldsymbol{\mu_1 } \notag\\ d_{kl}(f||g ) & = \log\frac{c_d(\kappa_1)}{c_d(\kappa_2 ) } + a_d(\kappa_1 ) ( \kappa_1 -\kappa_2\boldsymbol{\mu_1}^t\boldsymbol{\mu_2 } ) \label{eqn : vmf_kldiv}\end{aligned}\ ] ] we consider an example of a mixture with overlapping components and employ the proposed search method to determine the number of mixture components . @xcite considered the mixture shown in fig . [ fig : mix2 ] which is described as follows : let @xmath363 be the weight , mean , and covariance matrix of the @xmath364 component respectively . then , the mixture parameters are given as @xmath365 0.5 & 1 \end{bmatrix } , \mathbf{c}_2 = \begin{bmatrix } 6 & -2 \\[0.3em ] -2 & 6 \end{bmatrix},\ , \mathbf{c}_3 = \begin{bmatrix } 2 & -1 \\[0.3em ] -1 & 2 \end{bmatrix},\ , \mathbf{c}_4 = \begin{bmatrix } 0.125 & 0 \\[0.3em ] 0 & 0.125 \end{bmatrix}\end{gathered}\ ] ] we generate a random sample of size 1000 ( similar to @xcite ) and infer mixtures using the proposed heuristic . below are diagrams which portray the splitting of the individual components . we show a selection of operations which result in improved mixtures .

the abelian sandpile model and rotor - routing model are combinatorial models of various dynamics on graphs that were rediscovered several times in several different communities , ranging from combinatorics to self - organized criticality , arithmetic geometry , and algorithms . ( for example , @xcite ; see @xcite for more details and references . ) the stable configurations in the abelian sandpile model on a graph @xmath1 form a group called the sandpile group or the critical group of @xmath1 ; it is also known as the picard group of the graph , which we will denote @xmath0 in analogy with the picard group of a riemann surface . it has long been known that the order of @xmath0 is equal to the number of spanning trees of @xmath1 ( see , e.g. , @xcite and references therein ) . indeed , this is a form of kirchoff s matrix tree theorem . a natural first question to ask , then , is whether there is a canonical bijection between the elements of @xmath0 and the spanning trees of @xmath1 . what does `` canonical '' mean here ? at the very least , we should require that the bijection be invariant under automorphisms of @xmath1 , for certainly it should nt depend on what we name the vertices and edges . but there is clearly no such bijection : many graphs , e.g. a complete graph or an @xmath2-cycle , do nt even have a distinguished spanning tree that could correspond to the identity element of @xmath0 . so it is natural to ask instead for the next best thing : _ is there a canonical free , transitive action of the sandpile group on the set of spanning trees ? _ in other words , is there a canonical @xmath0-torsor structure on the set of spanning trees of @xmath1 ? by `` canonical , '' we still mean an action that is invariant under automorphisms of @xmath1 . such an action would give a bijection between the sandpile group and the set of spanning trees , once we decide on a spanning tree that corresponds to the identity element . but this is still too much to ask for . for example , consider the graph @xmath1 with two vertices and @xmath2 parallel edges between them . the sandpile group @xmath0 is cyclic of order @xmath2 , and each edge of @xmath1 is itself a spanning tree . so an automorphism - invariant action of @xmath3 amounts to an automorphism - invariant cyclic ordering of the edges , which clearly does not exist . this example , and others like it , suggest that we might hope for a free transitive action that is canonically defined on @xmath1 once we ve fixed the additional structure of a cyclic ordering of the edges incident to each vertex . a graph together with such a system of cyclic orderings is called a _ ribbon graph _ ( or sometimes `` combinatorial embedding '' ) . we therefore arrive at the following question . [ q : intro ] is there a free and transitive action of @xmath0 on the spanning trees of @xmath1 that is invariant under automorphisms preserving a ribbon graph structure on @xmath1 ? holroyd _ et al . _ @xcite gave one possible way to answer question [ q : intro ] . given a ribbon graph _ and _ a choice of a basepoint vertex , they used rotor - routing to define a free and transitive action of @xmath0 on the spanning trees of @xmath1 . this action is explained in section 2 . therefore an affirmative answer to question [ q : intro ] would follow if one could prove that the basepoint in the construction of holroyd _ et al _ is unnecessary . in other words , we have the following question , which to our knowledge was first asked by ellenberg @xcite : does the rotor - routing action of @xmath0 on the spanning trees of @xmath1 depend on the choice of basepoint ? in this paper we prove that if @xmath1 is a _ planar _ ribbon graph , then miraculously , the action defined by rotor routing is independent of the basepoint and hence canonical . furthermore , we show that this characterization is tight : only planar ribbon graphs have this property . our main theorem is thus : [ theorem : main ] let @xmath1 be a connected ribbon graph . the action of the sandpile group @xmath0 on the set @xmath4 of spanning trees is independent of the choice of basepoint vertex if and only if @xmath1 is a planar ribbon graph . the proof is based on three key ideas . first , the rotor - routing action of the sandpile group on spanning trees can be partially modeled via _ rotor - routing on unicycles _ ( @xcite ) . this is a related dynamical system with the property that rotor - routing becomes periodic , rather than terminating after finitely many steps . rotor - routing on unicycles is described in section [ sec : reversibility ] . the second main idea is that the independence of the sandpile action on spanning trees can be described in terms of _ reversibility of cycles_. we introduce the notion of reversibility ( previously considered in @xcite only for planar graphs ) , and prove in proposition [ prop : reversibility2 ] that reversibility is a well - defined property of cycles in a ribbon graph . we establish a relation between reversibility and basepoint - independence in section [ sec : spanningtrees ] . third , reversibility is closely related to whether a cycle separates the surface corresponding to the ribbon graph into two components . we prove in lemma [ lemma : almostseparating ] that these conditions are almost equivalent . moreover , although they are not equivalent for individual cycles , we prove in proposition [ prop : planar_is_reversible ] that _ all _ cycles are reversible if and only if all cycles are separating , in which case the ribbon graph is planar . by a _ graph _ we mean a connected and undirected graph , with multiple edges allowed but no self - loops . we write @xmath5 and @xmath6 for the vertex set and edge set of @xmath1 , respectively . throughout , @xmath1 denotes a _ ribbon graph _ : a graph together with a cyclic ordering of the edges incident to each vertex . ribbon graphs are equivalent to graphs embedded into orientable surfaces ( hence their alternate name `` combinatorial embeddings '' ) : from any ribbon graph one can reconstruct a closed orientable surface in which it is naturally embedded , and conversely the local orientation of the surface at a vertex gives the cyclic ordering ( see , e.g. , ( * ? ? ? * theorem 3.7 ) ) . we denote by @xmath4 the set of spanning trees of @xmath1 . when the graph is clear from context , we sometimes write @xmath7 for @xmath4 . * the rotor - routing model . * a _ rotor configuration _ @xmath8 on a directed graph is an assignment to each vertex @xmath9 of an outgoing edge @xmath10 $ ] based at @xmath9 . when we speak about rotor - routing or rotor configurations on a ( ribbon ) graph @xmath1 , we will always mean rotor configurations on the underlying eulerian directed graph @xmath11 obtained by replacing each undirected edge of @xmath1 with two oppositely oriented edges . fix a ribbon graph @xmath1 . the _ rotor - routing model _ is a deterministic process on pairs @xmath12 consisting of a rotor configuration @xmath8 and a vertex @xmath13 . we think of @xmath14 as the position of a _ chip _ that will move around the graph as the process is iterated . one step of the rotor - routing process consists of replacing @xmath12 by a new pair @xmath15 defined as follows . first , the rotor configuration @xmath16 coincides with @xmath8 at every vertex except at @xmath14 , where @xmath17 $ ] is the edge leaving @xmath14 that follows @xmath18 $ ] in the cyclic order . second , the new position of the chip @xmath9 is the vertex at the other end of the edge @xmath17 $ ] from @xmath14 . in short , the rotor - routing process rotates the rotor @xmath18 $ ] at the current position of the chip to the next position , and then moves the chip along this edge in the new direction of the rotor . in this paper we will frequently iterate the rotor - routing process until a certain `` stopping condition '' is reached , and reason about the resulting configuration and the steps taken to reach it . * the sandpile group . * the group @xmath19 of _ divisors _ on the graph @xmath1 is the free abelian group of formal @xmath20-linear combinations @xmath21 of the vertices of @xmath1 . in this paper we will be mostly interested in the subgroup @xmath22 of `` degree-0 '' divisors : @xmath23 if we fix a vertex @xmath24 ( for `` root '' ) , @xmath22 is freely generated by the basis @xmath25 as @xmath26 ranges over the other vertices of @xmath1 . the _ graph laplacian _ of a graph @xmath1 is an operator @xmath27 closely related to the adjacency matrix of @xmath1 . with respect to the basis of vertices @xmath26 , the operator @xmath28 corresponds to the symmetric matrix whose diagonal entries are @xmath29 = degree of @xmath26 and whose off - diagonal entries @xmath30 . the sum @xmath31 of each row and column is 0 , which shows both that the image of @xmath28 is contained in @xmath22 , and that @xmath32 . the image of @xmath28 is a finite - index subgroup of @xmath22 , and the _ sandpile group _ @xmath0 is the quotient : @xmath33 for any graph @xmath1 the group @xmath0 is a finite abelian group whose order is , up to sign , the determinant @xmath34 of any @xmath35 submatrix of the graph laplacian . people who study sandpile groups usually prefer to work with the distinguished representatives of each divisor class that are recurrent under a certain dynamical process . we will not describe the beautiful story of sandpile dynamics here , but that viewpoint does implicitly play a key role in the rotor - routing action that we now define . * action of @xmath0 on @xmath4 . * given a choice of root @xmath36 , holroyd _ et al . _ @xcite defined a natural action of @xmath22 on @xmath4 , which we now describe . this action is trivial on any divisor in @xmath37 and thus descends to an action of @xmath0 on @xmath4 . . the rotor - routing process in the bottom row shows that @xmath38 and thus @xmath39 so the action of @xmath40 on @xmath41 is independent of our choice of @xmath14 or @xmath9 as basepoint . here , the white circle indicates the position of the chip , and the rotors are all oriented clockwise relative to the page.,height=192 ] for a fixed root @xmath24 , given a spanning tree @xmath42 and a divisor @xmath43 , we will write @xmath44 for the image of @xmath41 under the rotor - routing action ( based at @xmath24 ) . we define the action of @xmath22 by describing the action @xmath45 of the generators @xmath46 for each @xmath47 , as follows . given a spanning tree @xmath41 , orient the edges of @xmath41 towards @xmath24 . the resulting collection of directed edges defines a rotor configuration on @xmath1 , except that there is no rotor at the root @xmath24 ; we think of this as a rotor configuration @xmath8 on the graph obtained from @xmath11 by removing the edges leaving @xmath24 . place a chip on the vertex @xmath14 , and iterate the rotor - routing process starting with @xmath12 . stop when the chip first reaches @xmath24 ( which it necessarily must ( * ? ? ? * lemma 3.6 ) ) , and call the resulting state @xmath48 . the rotor configuration @xmath16 is in fact another spanning tree @xmath49 oriented towards @xmath24 , and we define @xmath50 to be the spanning tree @xmath49 . see figure [ f : planar ] . since @xmath22 is an abelian group , the action of different divisors @xmath51 will commute : @xmath52 . however , we emphasize that the actions corresponding to _ different _ root vertices @xmath24 and @xmath53 may _ not _ commute , so it is important to distinguish between @xmath54 and @xmath55 . the action of @xmath22 descends to the sandpile group @xmath0 . in fact , for any choice of root vertex @xmath24 , the resulting action of @xmath0 on @xmath4 is free and transitive ( * ? ? ? * lemmas 3.17 and 3.19 ) . ( this property has been vastly generalized to abelian networks by bond and levine in their recent preprint @xcite . ) we can now restate ellenberg s question @xcite as : if @xmath56 is the permutation defined by this action , does @xmath57 for all @xmath58 and @xmath59 ? in other words , is the rotor - routing @xmath0-torsor structure on the set of spanning trees @xmath4 independent of basepoint ? theorem [ theorem : main ] completely answers this question . a _ path _ @xmath60 in the graph @xmath1 is an edge - injective map @xmath61 from the length-@xmath62 path @xmath63 ; it may or may not be vertex - injective . a _ cycle _ @xmath64 in @xmath1 is the image of a vertex - injective and edge - injective map @xmath65 from the length-@xmath62 cycle @xmath66 ( a circle subdivided into @xmath62 edges ) . we consider both paths @xmath60 and cycles @xmath64 to be oriented , but we identify maps @xmath67 that differ by cyclic permutation ( in other words , we do not specify a basepoint for cycles ) . we denote by @xmath68 or @xmath69 the same path or cycle with the opposite orientation . given an oriented cycle @xmath64 in a ribbon graph and a vertex @xmath70 , the edges incident to @xmath14 naturally fall into three classes : the two edges involved in @xmath64 , the edges ( if any ) on the left of @xmath64 , and the edges ( if any ) on the right of @xmath64 . the edges to the right of @xmath64 , for example , are the ones that occur after the out - edge at @xmath14 in @xmath64 and before the in - edge at @xmath14 in @xmath64 , in the cyclic order at @xmath14 . these three classes are disjoint , since our graphs have no self - loops ; however , if an edge @xmath71 has both endpoints @xmath14 and @xmath9 on @xmath64 , it is possible that @xmath71 is on the left of @xmath64 at @xmath14 but on the right of @xmath64 at @xmath9 . a cycle @xmath64 is _ nonseparating _ if there exists a path @xmath60 in @xmath1 whose endpoints lie on @xmath64 and is disjoint from @xmath64 otherwise , with its first edge on the left of @xmath64 and its last edge on the right of @xmath64 . such a path @xmath60 is a _ witness _ that @xmath64 is nonseparating . the cycle @xmath64 is _ separating _ if no such witness exists ; this is equivalent to saying that @xmath64 is a separating curve on the surface associated to the ribbon graph @xmath1 . the ribbon graph @xmath1 is _ planar _ if every cycle is separating ( equivalently , if its associated surface is a sphere ) . * unicycles and rotor - routing . * a _ unicycle _ @xmath72 consists of a rotor configuration @xmath8 which contains exactly one directed cycle @xmath73 , together with a vertex @xmath26 lying on the cycle @xmath73 . it is not difficult to see that applying rotor - routing to the configuration @xmath72 preserves these conditions , and thus takes unicycles to unicycles ( * ? ? ? * lemma 3.4 ) . in fact , the following is is a key observation that we will use many times : [ lemma : periodic ] let @xmath72 be a unicycle on a graph with @xmath74 edges . iterating the rotor routing process @xmath75 times starting at @xmath72 , the chip traverses each edge of @xmath1 exactly once in each direction , each rotor makes exactly one full turn , and the stopping state is @xmath76 . we write @xmath77 if the configuration @xmath15 can be obtained from @xmath12 by iterating the rotor - routing process some positive number of times . by lemma [ lemma : periodic ] , the relation @xmath77 is in fact an equivalence relation on unicycles . the following lemma shows that each equivalence class is naturally in bijection with the directed edges of @xmath1 . we write @xmath78 if @xmath77 and @xmath79=e$ ] . [ lemma : therotorknows ] let @xmath12 be a unicycle . for any vertex @xmath9 and any directed edge @xmath71 based at @xmath9 , there exists a unique rotor configuration @xmath16 such that @xmath80 . run the rotor - routing process starting at @xmath12 until the next occurrence of @xmath12 . according to lemma [ lemma : periodic ] , each rotor makes precisely one full rotation in this process . since the rotor at @xmath9 only advances when the chip is at @xmath9 , this implies that at some intermediate stage @xmath15 we had @xmath79=e$ ] . the next rotor - routing step then advances the rotor at @xmath9 , so @xmath16 is unique . the _ first _ time the chip arrives at @xmath9 when rotor - routing is iterated starting with @xmath12 , the current rotor configuration @xmath16 satisfies @xmath79=\rho[y]$ ] , since @xmath9 has not previously been visited . accordingly we write @xmath81 as an abbreviation for @xmath82}(\sigma , y)$ ] . finally , when @xmath15 has already been defined , we write @xmath83 as shorthand for the rotor routing process beginning with @xmath12 and ending at the first occurrence of @xmath84 . [ def : reversible ] given a unicycle @xmath72 with directed cycle @xmath85 , let @xmath86 denote the configuration obtained from @xmath8 by reversing the edges of @xmath64 , and keeping all other rotors unchanged . we say that @xmath64 is _ reversible _ if @xmath87 . before using this terminology , we need to verify that it is well - defined : in other words , that reversibility really is a property of the _ cycle _ @xmath64 , and does not depend on the choice of @xmath8 nor the choice of @xmath26 . note that by definition , @xmath64 is reversible if and only if @xmath69 is reversible . we will prove in proposition [ prop : reversibility2 ] that reversibility is well - defined , but we first need the following technical result . this result will also be used , along with lemma [ lemma : almostseparating ] , to show that reversible cycles are quite close to being separating . [ prop : lcrc ] let @xmath64 be a directed cycle . if there exists a unicycle @xmath72 with @xmath85 such that @xmath88 , then the vertices @xmath89 can be partitioned into two sets @xmath90 such that : 1 . if @xmath89 is adjacent to @xmath91 along an edge lying on the left of @xmath64 at @xmath14 , then @xmath92 . 2 . if @xmath89 is adjacent to @xmath91 along an edge lying on the right of @xmath64 at @xmath14 , then @xmath93 . if @xmath89 is adjacent to @xmath94 , then @xmath9 and @xmath95 are either both in @xmath96 or both in @xmath97 . any separating curve @xmath64 separates the surface into two parts @xmath96 and @xmath97 as in this proposition . the key utility of proposition [ prop : lcrc ] is that it allows us to define the sets @xmath96 and @xmath97 knowing only that @xmath64 is _ reversible _ , without knowing whether or not @xmath64 is separating . for any vertex @xmath9 in @xmath1 , let @xmath98 be its degree . consider the rotor - routing process that takes @xmath72 to @xmath99 . for any vertex @xmath9 _ not _ lying on @xmath64 , the rotor ends at its initial position @xmath10={\overline}{\rho}[y]$ ] , so by lemma [ lemma : periodic ] , the vertex @xmath9 was visited either 0 or @xmath98 times . we define @xmath100 and @xmath101 . since @xmath1 has no isolated vertices ( by our convention , all graphs are connected ) , these two sets are disjoint . since no directed edge is traversed more than once , in the latter case @xmath93 the chip must arrive at @xmath9 along each of the @xmath98 adjacent edges exactly once and leave @xmath9 along each edge exactly once . now consider a vertex @xmath14 lying on @xmath64 . since the cycle @xmath64 has been reversed , the rotor at @xmath14 begins at @xmath18=c[x]$ ] and ends at @xmath102={\overline}{c}[x]$ ] . therefore the rotor rotates @xmath103 times , where @xmath104 is 1 plus the number of edges at @xmath14 on the _ right _ of @xmath64 , and so the chip visits @xmath14 precisely @xmath103 times . the first @xmath105 times the chip leaves @xmath14 , it leaves along the @xmath105 edges lying on the right of @xmath64 ; the final time it leaves @xmath14 , it leaves along the edge @xmath102={\overline}{c}[x]$ ] of the reversed cycle @xmath64 ( since this is the final state of the rotor ) . if @xmath89 is adjacent to @xmath91 along an edge lying on the right of @xmath64 at @xmath14 , the chip will leave @xmath14 along this edge at some point , so @xmath93 . similarly , if @xmath89 is adjacent to @xmath91 along an edge lying on the left of @xmath64 at @xmath14 , the chip does not leave @xmath14 along this edge ; in other words , the chip does not arrive at @xmath9 along this edge . this implies that @xmath92 , because we noted above a vertex @xmath9 lying in @xmath97 must be visited along every one of its incoming edges , and this would include the edge from @xmath14 . finally , if @xmath89 is adjacent to @xmath94 and the chip traverses this edge in either direction , then both @xmath9 and @xmath95 lie in @xmath97 ; if not , both lie in @xmath96 . the three conditions above uniquely characterize @xmath96 and @xmath97 as those vertices that can be connected by a path to the left side or right side of @xmath64 respectively ; since this description does not depend on @xmath8 or @xmath26 , the sets @xmath96 and @xmath97 depend only on the cycle @xmath64 . this seems at first to prove that @xmath64 is separating . however , it does not rule out the possibility of a witness consisting of a single edge joining the left side of @xmath91 to the right side of @xmath106 . in fact , nonseparating cycles of this form can indeed be reversible , as we will see in the proof of proposition [ prop : planar_is_reversible ] . [ prop : reversibility2 ] let @xmath64 be a directed cycle . if @xmath107 for _ some _ unicycle @xmath108 with @xmath109 , then @xmath110 for _ any _ unicycle @xmath76 with @xmath85 . in other words , the reversibility of a cycle @xmath64 , as in definition [ def : reversible ] , is well - defined : it only depends on @xmath64 , and not on a choice of unicycle . given a unicycle @xmath72 with @xmath85 , define its _ maximal reversal _ @xmath111 to be the first state @xmath111 encountered in the rotor - routing process such that @xmath112 and @xmath113={\overline}{c}[u]$ ] . in other words , the chip is on the cycle @xmath64 at the vertex @xmath114 , and the rotor at @xmath114 is about to rotate _ past _ the reversed cycle @xmath69 to the edges on the left of @xmath64 for the first time . we will prove that as long as @xmath107 for _ some _ @xmath115 and some @xmath116 , the maximal reversal of @xmath72 is @xmath99 . then we will have @xmath117 as desired . ( we can not use lemma [ lemma : therotorknows ] to deduce this from @xmath113={\overline}{c}[u]={\overline}{\rho}[u]$ ] , because we do not yet know that @xmath118 . ) assume that @xmath107 for some @xmath115 with @xmath109 , and let @xmath96 and @xmath97 be the sets defined by proposition [ prop : lcrc ] . the set of directed edges traversed in the process @xmath119 was described in the proof of that proposition : the chip traverses ( 1 ) every directed edge leaving @xmath93 , ( 2 ) each directed edge leaving @xmath91 on the right side of @xmath64 , and ( 3 ) each edge @xmath120 $ ] leaving @xmath91 . this set of directed edges only depends on @xmath64 , not on @xmath115 or @xmath121 , so we will denote it by @xmath122 . for any vertex @xmath123 , let @xmath124 denote the number of edges in @xmath122 leaving @xmath123 . ( specifically , @xmath125 for @xmath91 , @xmath126 for @xmath93 , and @xmath127 for @xmath92 . ) since the chip ends up back at @xmath26 , it leaves each vertex @xmath123 the same number of times that it arrives there , so @xmath124 is also the number of directed edges in @xmath122 _ arriving _ at @xmath123 . now consider an arbitrary rotor configuration @xmath72 with @xmath85 and @xmath128 . the key to this lemma is the observation that if @xmath111 is the maximal reversal of @xmath72 , the process @xmath129 only traverses edges in @xmath122 . assume otherwise , and let @xmath71 be the first directed edge _ not _ in @xmath122 traversed . since the chip begins on @xmath64 , and no edges connect @xmath97 to @xmath96 , this first edge @xmath71 is based at some @xmath91 . by definition of @xmath122 , either @xmath71 lies on the left side of @xmath64 , or @xmath130 $ ] itself . in either case , the rotor at @xmath14 must rotate past @xmath120 $ ] before it can reach the edge @xmath71 . but this is a contradiction , since by definition @xmath111 is the first time the rotor at @xmath91 is @xmath131 $ ] while the chip is there . in the process @xmath132 , the rotor at @xmath112 rotates from @xmath133=c[u]$ ] to @xmath113={\overline}{c}[u]$ ] , so the chip leaves @xmath114 precisely @xmath134 times . however , if @xmath135 , the chip must have arrived at @xmath114 precisely @xmath136 times , since it did not originate there . since the total number of edges in @xmath122 arriving at @xmath114 is only @xmath134 , this is a contradiction . this shows that @xmath137 , and moreover that the chip has arrived at @xmath26 along each of the @xmath138 edges in @xmath122 directed towards @xmath26 . one of these is the edge @xmath139 $ ] , where @xmath140 is the next vertex after @xmath26 in the cycle @xmath64 , so we must have @xmath141={\overline}{c}[w]$ ] as well . but if the rotor at @xmath123 has reversed from @xmath142=c[w]$ ] to @xmath141={\overline}{c}[w]$ ] , the chip must have left @xmath124 times . to arrive @xmath124 times at @xmath123 , the rotor at the _ next _ vertex after @xmath123 must have been reversed , and so on . by induction , we conclude that we have @xmath17={\overline}{c}[x]={\overline}{\rho}[x]$ ] for all @xmath91 . in particular , the directed cycle of @xmath16 is @xmath143 . since the process @xmath144 does not visit @xmath96 , we know that @xmath79=\rho[y]={\overline}{\rho}[y]$ ] for all @xmath92 . it remains to show that @xmath79=\rho[y]$ ] for @xmath93 . to do this , let @xmath145 be the maximal reversal of @xmath146 . since @xmath143 , the process @xmath147 does not visit @xmath148 , so @xmath149=\sigma[y]$ ] for all @xmath93 . but the previous paragraph shows that @xmath150=\overline{\sigma}[v]=\rho[v]$ ] . since @xmath151 , lemma [ lemma : therotorknows ] implies that @xmath152 . therefore @xmath79=\rho[y]$ ] for all @xmath89 , and @xmath17={\overline}{\rho}[x]$ ] for all @xmath91 , demonstrating that @xmath153 as desired . [ lemma : almostseparating ] let @xmath64 be a directed cycle . if @xmath64 is reversible , then any path @xmath60 witnessing that @xmath64 is nonseparating must have length 1 . conversely , if @xmath64 is a separating cycle , then @xmath64 is reversible . let @xmath60 be a path of length @xmath154 , with vertices @xmath155 , with endpoints @xmath156 and @xmath157 . if @xmath60 is a witness that @xmath64 is nonseparating , the second vertex @xmath158 lies in @xmath96 , and the second - to - last vertex @xmath159 lies in @xmath97 . but this contradicts condition ( 3 ) of proposition [ prop : lcrc ] , which says that membership in @xmath96 is locally constant along paths in @xmath160 . therefore any witness @xmath60 must have length 1 . to prove that any separating cycle is reversible , we will go through the proof of proposition [ prop : lcrc ] in reverse ; the resulting argument closely parallels ( * ? ? ? * corollary 4.11 ) , where this was proved for planar ribbon graphs . ( in fact , a special case of reversibility , for cycles in an @xmath161 planar grid graph , was proved even earlier , in ( * ? ? ? * proposition i ) . ) if @xmath64 is a separating cycle , we can partition the vertices @xmath89 into @xmath162 , where @xmath163 ( resp . @xmath164 ) if there exists a path from @xmath9 to @xmath64 ending with an edge on the left side ( resp . right side ) of @xmath64 . every @xmath9 can be connected to @xmath64 by some path , since @xmath1 is connected , so @xmath165 . if @xmath166 could be connected to both sides of @xmath64 , splicing these paths together would give a witness that @xmath64 was nonseparating , so @xmath167 . let @xmath168 be the ribbon graph obtained from @xmath1 by deleting all vertices in @xmath169 , all edges adjacent to @xmath169 , and all edges lying on the right of @xmath64 . because @xmath64 is separating , no edge on the left side of @xmath64 is removed ( every such edge either joins @xmath91 to @xmath163 , or joins @xmath91 to @xmath106 , in which case it lies on the left side of @xmath64 at both endpoints ) . since no rotor in @xmath170 points along a removed edge , we can restrict the rotor configuration @xmath8 to @xmath168 as @xmath171 . it is easy to run the rotor routing process starting at @xmath172 , because at each @xmath91 the two edges of @xmath64 are now adjacent in the cyclic order . thus at each step , the rotor at one vertex on @xmath64 is reversed , and the chip moves to the previous vertex on @xmath64 . after a number of steps equal to the length of @xmath64 , each rotor on @xmath64 has moved once , the cycle @xmath64 has been reversed , and the chip is back at @xmath26 ; in other words , the process takes @xmath173 . since rotor - routing on @xmath168 is periodic , continuing gives a rotor - routing process on @xmath168 taking @xmath174 to @xmath172 . but this segment of the process does not involve any of the deleted edges . therefore if the rotor - routing process is run in parallel on @xmath1 starting with @xmath99 , the two processes will take precisely the same series of steps . when the former terminates at @xmath175 , the latter will therefore be at @xmath76 , demonstrating that @xmath176 as desired . [ prop : planar_is_reversible ] a connected ribbon graph @xmath1 without loops is planar if and only if all cycles on @xmath1 are reversible . if @xmath1 is planar , the jordan curve theorem implies that every cycle is separating , so lemma [ lemma : almostseparating ] implies that every cycle is reversible . for the converse , suppose @xmath1 is not planar , so it has at least one nonseparating cycle , but that every cycle on @xmath1 is reversible . our first claim is that every nonseparating cycle on @xmath1 has length 2 . indeed , let @xmath60 be a path witnessing that the cycle @xmath64 is nonseparating , joining the left side of @xmath91 to the right side of @xmath106 . this path splits @xmath64 into two directed arcs from @xmath14 to @xmath177 . call them @xmath178 and @xmath179 , labeled so that @xmath180 . let @xmath181 be the cycle @xmath182 , and let @xmath183 be the cycle @xmath184 . the arc @xmath178 is now a witness that @xmath181 is nonseparating , since @xmath178 lies on the left of @xmath181 at @xmath14 but on the right of @xmath181 at @xmath177 . similarly @xmath179 witnesses that @xmath183 is nonseparating , since @xmath179 lies on the right of @xmath183 at @xmath14 but on the left of @xmath183 at @xmath177 . since all three cycles @xmath64 , @xmath181 , and @xmath183 are reversible by assumption , lemma [ lemma : almostseparating ] implies that the paths @xmath60 , @xmath179 and @xmath178 each have length 1 . in particular , the original cycle @xmath185 has length 2 , as claimed . let us rename the edge @xmath60 to @xmath186 , and the cycle @xmath64 to @xmath187 , to expose the cyclic symmetry of our notation . then the cyclic order of these edges at @xmath14 and @xmath177 is the same , namely @xmath188 and @xmath189 respectively , and @xmath190 with indices modulo 3 . see figure [ fig : a1a2a3 ] . let us say that an edge at @xmath14 _ lies between _ @xmath191 and @xmath192 , for @xmath193 , @xmath194 , or @xmath195 , if it comes after @xmath191 and before @xmath192 in the cyclic order of edges at @xmath14 . similarly , we will say that an edge at @xmath177 lies between @xmath196 and @xmath197 if it comes after @xmath196 and before @xmath197 in the cyclic ordering of edges at @xmath177 . and @xmath186 in the proof of proposition [ prop : planar_is_reversible ] . the orientation at each vertex is clockwise with respect to the page.,width=144 ] before moving on , we observe that _ any _ path from @xmath14 to @xmath177 not passing through @xmath14 or @xmath177 except at its endpoints will witness that one of the cycles @xmath198 is nonseparating , so by lemma [ lemma : almostseparating ] any such path has length 1 . indeed , each edge at @xmath14 other than the @xmath191 lies on the right side of exactly one cycle @xmath198 , and on the left side of the other two ; similarly , each edge at @xmath177 lies on the left of exactly one @xmath198 and on the right of the other two . thus the pigeonhole principle implies that no path from @xmath14 to @xmath177 can lie on the same side of @xmath198 for all three simultaneously . we may also assume that no edge @xmath71 connects @xmath14 to @xmath177 and lies between @xmath179 and @xmath186 at @xmath14 , after repeating the following reduction step . suppose @xmath71 is such an edge . if @xmath71 lies between @xmath199 and @xmath200 at @xmath177 , then replace @xmath179 with @xmath71 . if @xmath71 lies between @xmath199 and @xmath201 at @xmath177 , then replace @xmath186 with @xmath71 . if @xmath71 lies between @xmath200 and @xmath201 at @xmath177 , then replace either @xmath179 or @xmath186 with @xmath71 . in each case , this reduces the number of edges at @xmath14 between @xmath179 and @xmath186 while preserving the cyclic orderings of @xmath188 and @xmath202 at @xmath14 and @xmath177 , so we may repeat this step until there are no edges to @xmath177 lying between @xmath179 and @xmath186 at @xmath14 . let @xmath12 be any unicycle with @xmath203 ( e.g. by adding @xmath178 to a spanning tree that uses @xmath179 ; see [ sec : spanningtrees ] for details ) . we have @xmath18=a_1 $ ] and @xmath204={\overline}{a}_2 $ ] by definition . define @xmath16 by @xmath205 . we had @xmath102=a_2 $ ] and @xmath206={\overline}{a}_1 $ ] , and the latter implies @xmath207={\overline}{a}_1 $ ] . but we know moreover that @xmath17=a_3 $ ] . this is because in the rotor routing process @xmath208 , starting from the last time the chip leaves @xmath14 and ending at its arrival at @xmath177 , the chip traces out a path from @xmath14 to @xmath177 . that path has length 1 by the argument above , which means that the chip arrives for the first time at @xmath177 along some edge from @xmath14 . but we have assumed that @xmath186 is the first such edge that the rotor at @xmath14 encounters when it starts at @xmath179 . finally , define @xmath209 by @xmath210 . we had @xmath211=a_1 $ ] and @xmath212={\overline}{a_3}$ ] , and the former implies @xmath213=a_1 $ ] . we have assumed that all cycles are reversible , so @xmath214 and @xmath215 . by transitivity this implies @xmath216 ; since @xmath213=\rho[x]=a_1 $ ] , it must be that @xmath152 . however , this would mean that the rotor @xmath217 $ ] ends up at @xmath218 , which is impossible : this rotor began at @xmath212={\overline}{a}_3 $ ] , so the first edge to @xmath14 that it encounters can not be @xmath218 ( it would encounter @xmath199 first , if not some other edge ) . this contradiction completes the proof that not all cycles in @xmath1 can be reversible if @xmath1 has nonseparating cycles . * spanning trees and unicycles . * the reason that we studied unicycles so carefully in the previous section is that they are closely related to spanning trees , as we now explain . let @xmath42 be a spanning tree . any two vertices @xmath14 and @xmath9 are connected by a unique geodesic path in @xmath41 ; we write @xmath219 for that path , oriented from @xmath14 to @xmath9 . therefore if @xmath71 is an edge of @xmath1 connecting @xmath9 to @xmath14 , the undirected graph @xmath220 contains a unique cycle . now given a directed edge @xmath71 from @xmath9 to @xmath14 and a spanning tree @xmath41 , define a rotor configuration @xmath221 as follows . let @xmath222 denote the collection of directed edges obtained from @xmath41 by directing each edge towards the `` root '' @xmath9 . then the collection of directed edges @xmath223 forms a rotor configuration @xmath221 with unique directed cycle @xmath224 . we will denote this directed cycle by @xmath225 . if @xmath41 already contains @xmath71 , then @xmath225 is the length 2 cycle consisting only of @xmath226 ; note that this cycle is trivially reversible in any ribbon graph . if we took instead the edge @xmath227 from @xmath14 to @xmath9 , the rotor configuration @xmath228 comes from the directed graph @xmath229 , with directed cycle @xmath230 . the following observation will be fundamental for us . for @xmath232 this follows because @xmath233 for vertices @xmath234 , we have @xmath235=t_y[z]$ ] , since the geodesics @xmath236 and @xmath237 share an initial segment . then @xmath238=\rho_{{\overline}e}(t)[z]$ ] , and @xmath238 = { \overline}{\rho_e(t)}[z]$ ] since @xmath95 is not on the cycle . given a spanning tree @xmath239 and an edge @xmath71 from @xmath9 to @xmath14 , the rotor configuration @xmath240 is a unicycle since @xmath241 . conversely , given a unicycle @xmath72 , deleting the directed edge @xmath242 $ ] yields a spanning tree @xmath243 with all edges directed towards the root @xmath26 . this correspondence between spanning trees and rotor configurations lets us partially model the action of the sandpile group on spanning trees . [ prop : rotorscompute ] given a spanning tree @xmath41 in a ribbon graph @xmath1 , let @xmath71 be an edge not in @xmath41 , and @xmath53 any vertex on the @xmath41-geodesic between the endpoints @xmath244 of @xmath71 ( including @xmath14 or @xmath9 ) . let @xmath245 ( so that @xmath246 ) . then @xmath247 is a unicycle , and if @xmath16 is the rotor configuration defined by @xmath248 , then @xmath249 the rotor configuration @xmath8 is equal to @xmath223 , where @xmath222 is obtained by directing each edge of @xmath41 towards @xmath9 . the configuration @xmath8 then has a unique cycle , namely @xmath250 , and by assumption @xmath53 is on that cycle , so @xmath247 is a unicycle . next , by definition , @xmath251 is obtained by beginning with @xmath222 , placing the chip at @xmath53 , and running the rotor - routing process until the chip arrives at @xmath9 for the first time . the presence of the rotor @xmath10=e$ ] does not affect this process , since @xmath9 is not reached until the last step . therefore the rotor - routing process @xmath252 takes precisely the same steps as the process @xmath253 . we conclude that @xmath254 . since @xmath255 by definition , this proves the lemma . * basepoint - independence and reversible cycles . * before giving our last few lemmas and proving our main theorem , let us give some intuition as to why reversibility should be at all related to basepoint - independence . consider a spanning tree @xmath256 , and let @xmath14 and @xmath9 be adjacent vertices connected by an edge @xmath71 from @xmath14 to @xmath9 . suppose that we wish to compute @xmath257 for some divisors @xmath51 . the first step is to compute @xmath258 . by proposition [ prop : rotorscompute ] , the computation of @xmath259 can be modeled ( at least for some divisors ) by a rotor - routing process ; this process terminates with the rotor configuration @xmath260 , which is the spanning tree @xmath41 oriented towards the basepoint @xmath14 . the next step is to compute @xmath261 , and to do this we first need to re - orient the tree @xmath41 towards the basepoint @xmath9 . this does not affect any edges except those on the @xmath41-geodesic between @xmath14 and @xmath9 , which are reversed . if we add the edge @xmath71 to our rotor configuration @xmath260 to obtain @xmath262 , the reversal of the @xmath41-geodesic @xmath219 amounts to the reversal of the cycle @xmath224 in the rotor configuration @xmath221 . if this cycle is reversible , then we can compute both steps @xmath263 and @xmath264 as part of a _ single rotor - routing process _ , even though these actions involve two _ different _ basepoints . this basic observation is the heart of the relation between reversibility and basepoint - independence , upon which our main theorem rests . the next lemma capitalizes on this observation . [ lemma : ongeodesic ] given a spanning tree @xmath41 in a ribbon graph @xmath1 and an edge @xmath71 from @xmath9 to @xmath14 , let @xmath53 be any vertex on the geodesic @xmath219 ( including @xmath265 or @xmath266 ) , and set @xmath267 . let @xmath245 ; as in proposition [ prop : rotorscompute ] , @xmath247 is a unicycle . define rotor configurations @xmath270 , @xmath16 , and @xmath209 by @xmath271 note that @xmath272={\overline}{\rho}[x]={\overline}{e}$ ] , and similarly @xmath79=\rho[y]=e$ ] implies that @xmath213={\overline}{\sigma}[x]={\overline}{e}$ ] . in fact , by lemma [ lemma : ovrho ] and proposition [ prop : rotorscompute ] , we have thus we see that equality is exactly the condition @xmath276 . moreover , since @xmath272=\tau[x]={\overline}{e}$ ] , lemma [ lemma : therotorknows ] implies that this is equivalent to @xmath277 . consider the following diagram relating these six rotor configurations : @xmath278\ar_{\txt{\large ? } } @{2{<}~2{>}}[dl]&&(\sigma , y)\ar^{\txt{\large ? } } @{2{<}~2{>}}[dr]\\ ( { \overline}{\rho},s)\ar@{2{<}~2{>}}[dr]&&&&({\overline}{\sigma},y)\ar@{2{<}~2{>}}[dl]\\ & ( \psi , x)\ar_{\txt{\large ? } } @{2{<}~2{>}}[rr]&&(\tau , x ) } \ ] ] three of these edges hold by definition , namely @xmath279 , @xmath280 , and @xmath281 . by the results of section [ sec : reversibility ] , the upper left edge @xmath282 is equivalent to the cycle @xmath283 being reversible . similarly the upper right edge is equivalent to @xmath284 being reversible . finally , we have already shown that the bottom edge is equivalent to the assertion . therefore parts ( a ) and ( b ) of the lemma each assert that if two of the remaining three edges hold , the third does as well . but this follows immediately from the transitivity of this equivalence relation . by lemma [ lemma : almostseparating ] , the hypothesis of lemma [ lemma : ongeodesic](a ) is always satisfied if @xmath1 is planar , so @xmath286 now applying @xmath287 to both sides , we conclude that @xmath288 . we emphasize that we are _ not _ yet claiming that @xmath289 , since the condition of corollary [ corollary : first ] that @xmath53 lies on the @xmath41-geodesic from @xmath14 to @xmath9 need not hold for all trees @xmath41 . [ lemma : rununtil ] let @xmath290 be a spanning tree of an arbitrary ribbon graph @xmath1 . let @xmath14 , @xmath95 , and @xmath53 be arbitrary vertices such that @xmath53 is the first vertex on @xmath291 visited by the rotor - routing process starting with @xmath292 . then @xmath293 define @xmath294 by @xmath295 ; by definition , the underlying spanning tree @xmath49 of @xmath294 is equal to @xmath296 . similarly , define @xmath297 by @xmath298 , with underlying spanning tree @xmath299 . as we noted in the proof of lemma [ lemma : ovrho ] , the configurations @xmath260 and @xmath300 coincide for all vertices @xmath301 , while on this geodesic , the difference between @xmath260 and @xmath300 is that @xmath291 is reversed and thus replaced by @xmath302 . therefore if the rotor - routing processes are run in parallel starting with @xmath292 and @xmath303 , the same steps will be taken until the chip first arrives at @xmath291 . by our assumption on @xmath53 , this occurs when the chip reaches @xmath53 , i.e. when the second process reaches @xmath304 . the first process coincides except that the geodesic between @xmath14 and @xmath53 is directed towards @xmath14 instead . but by the first sentence of this paragraph , this is the directed tree @xmath305 obtained from @xmath49 by directing its edges towards @xmath14 . to sum up , we have observed that @xmath306 . by definition , @xmath307 can be computed by starting with @xmath308 and running the rotor - routing process until the chip reaches @xmath14 . the above shows that the first process reaches @xmath308 as an intermediate stage ; therefore continuing this process eventually yields @xmath309 , since @xmath298 by definition . this shows that @xmath310 . expanding out our notation , this says that @xmath311 applying @xmath312 to both sides yields @xmath313 , as desired . let @xmath53 be the first vertex on @xmath219 that is reached by the rotor - routing process starting with @xmath292 . note that @xmath53 is _ a fortiori _ the first vertex visited on @xmath302 or on @xmath315 as well . we use this to apply lemma [ lemma : rununtil ] twice , giving @xmath316 write @xmath49 for the tree @xmath296 appearing in . notice that by definition of @xmath53 , in the rotor routing process that computes @xmath296 , the chip never touches any vertex on @xmath219 until it reaches @xmath53 , so @xmath317 . so @xmath53 lies on @xmath318 by corollary [ corollary : first ] applied to @xmath49 ( with the roles of @xmath14 and @xmath9 reversed ) , we have @xmath319 the desired result follows once this equation is combined with , by applying @xmath320 to the left - hand side of and applying @xmath321 to the right - hand side of . ( @xmath322 ) first , assume that @xmath1 is a planar ribbon graph . our goal is to prove for any @xmath43 that @xmath323 as elements of @xmath324 for arbitrary vertices @xmath14 and @xmath9 . since @xmath1 is connected , it suffices to prove this when @xmath14 and @xmath9 are adjacent , with the general case following by induction . fix adjacent vertices @xmath14 and @xmath9 . the group @xmath22 is generated by divisors of the form @xmath325 as @xmath95 ranges over the vertices of @xmath1 . but we proved in corollary [ corollary : fulladjacent ] that @xmath326 when @xmath14 and @xmath9 are adjacent . it follows that @xmath323 for any @xmath43 . this completes the proof that if @xmath1 is a planar ribbon graph , the action of @xmath0 on @xmath7 is independent of the basepoint . ( @xmath327 ) conversely , assume that the action of @xmath0 on @xmath7 is independent of the basepoint . choose any edge @xmath71 of @xmath1 with endpoints @xmath14 and @xmath9 . we claim that for any @xmath239 and any @xmath43 , the cycle @xmath225 is reversible if and only if the cycle @xmath328 is reversible . ( we have dropped the subscripted basepoints , since by assumption the action of @xmath0 does not depend on them . ) we first prove the claim in the case when @xmath329 for some @xmath95 . as in the proof of corollary [ corollary : fulladjacent ] , let @xmath53 be the first vertex on @xmath219 reached by the rotor - routing process starting with @xmath292 . since we have assumed independence of basepoint , this process @xmath330 computes @xmath331 . since this process does not reach @xmath219 until the last step , the cycle @xmath225 remains unchanged , i.e. @xmath332 . we may now apply lemma [ lemma : ongeodesic ] to @xmath333 , since @xmath53 lies on the cycle @xmath332 . the condition is always satisfied by our assumption that the action is independent of the basepoint , so applying lemma [ lemma : ongeodesic](b ) implies that @xmath334 completing the proof of the claim for @xmath329 . since @xmath22 is generated by such elements , the claim follows . we now show that every cycle @xmath64 in @xmath1 is reversible . given a cycle @xmath64 , let @xmath71 be any edge of @xmath64 . extend the edge @xmath71 to a spanning tree @xmath41 , and separately extend the path @xmath335 to a spanning tree @xmath49 . since @xmath0 acts transitively on @xmath7 , there exists @xmath43 for which @xmath336 . thus the claim above shows that @xmath225 is reversible if and only if @xmath268 is reversible . but @xmath225 is the trivial cycle @xmath226 , which is always reversible in any ribbon graph , while @xmath268 is our original cycle @xmath64 . we conclude that every cycle in @xmath1 is reversible . by proposition [ prop : planar_is_reversible ] , @xmath1 is a planar ribbon graph . * acknowledgements . * we are very grateful to jordan ellenberg for calling our attention to the question considered in this paper , and we thank math overflow for providing a venue for the question . the first and third authors would also like to thank the second author for suggesting that the three of us get together to work on this problem , which was great fun ! a. holroyd , l. levine , k. mszros , y. peres , j. propp and d. wilson , chip - firing and rotor - routing on directed graphs , in _ in and out of equilibrium 2 _ , eds . v. sidoravicius and m. e. vares , _ progress in probability _ 60 , birkhuser ( 2008 ) , 331364 . ` arxiv:0801.3306 ` .

spatially periodic non - centrosymmetric systems are able to transport nonequilibrium particles in the absence of an average macroscopic force resulting in the ratchet effect @xcite . ratchet effects whose prerequisites are simultaneous breaking of both thermal equilibrium and spatial inversion symmetry can be realized in a great variety of forms and in particular , as electric transport in semiconductor systems @xcite . recent experiments demonstrated that terahertz radiation induced ratchet effects , can be efficiently excited in semiconductor quantum wells ( qw ) @xcite and graphene @xcite with lateral superlattice structures . these experiments allowed one to explore basic physics of the ratchet effects in low dimensional electron systems , provide information on non - equilibrium transport in such systems and demonstrate that ratchet effects can be applied for room temperature terahertz radiation detection @xcite . the latter , besides high sensitivity and short response times , offer new functionality being a good candidate for all - electric detection of the radiation polarization state including radiation helicity @xcite . here we report on the observation and study of magnetic quantum ratchet effect in ( cd , mn)te/(cd , mg)te diluted magnetic heterostructures and cdte / cdmgte qws superimposed with lateral asymmetric superlattices . applying magnetic field @xmath2 along the growth direction we observe that the ratchet current exhibits sign - alternating @xmath1-periodic oscillations with amplitudes immensely larger than the ratchet signal at zero magnetic field . the results are analyzed in terms of the theory of magnetic ratchet effects in qw structures with a lateral asymmetric periodic potential @xcite . we show that the photocurrent generation is based on the combined action of a spatially periodic in - plane potential and the spatially modulated light due to the near field effects of radiation diffraction . corresponding theoretical analysis describes the experimental results . the paper is organized as follows . first , we present our samples and results of magneto - transport characterization ( sec . [ sample ] ) , as well as briefly describe the experimental technique ( sec . [ experi ] ) . in sec . [ erstes ] we discuss the results on the photocurrents generated at zero magnetic field . section [ magnetic ] contains the main experimental results on the magnetic quantum ratchet effect . it has three subsections describing the basic results , the laser beam scan across the structure and the effect of the electrostatic potential acting on the electron gas . in the following sec . [ discussion ] we present the theory , calculate dependencies of the photocurrent on magnetic field , and compare them with the experimental data . finally , in sec . [ summary ] , we summarize the results and discuss prospectives of future experimental studies of the magnetic quantum ratchet effect . experiments are carried out on ( cd , mn)te / cdmgte and cdte / cdmgte single qw structures grown by molecular beam epitaxy on ( 001)-oriented gaas substrates @xcite . during the growth the fluxes of cd , te , mg and mn have been supplied from elemental sources while iodine flux has been obtained from a zni@xmath3 source . the schematic layout of the ( cd , mn)te / cdmgte layer structure and sketch of the qw are presented in fig . [ structure ] . the thick buffer ( @xmath4 6 @xmath5 m ) , consisting of cdte and cd@xmath6mg@xmath7te layers and cdte / cd@xmath6mg@xmath7te short period superlattices , have been grown to reduce the number of dislocations resulting from the strong lattice mismatch between gaas and cd@xmath8mg@xmath9te . the well width in both kinds of structures is 9.7 nm and the cd@xmath6mg@xmath7te alloy serves as a barrier material . the composition of barriers has been determined from photoluminescence spectra . in order to obtain a two - dimensional electron gas the structures have been modulation doped by iodine donors incorporated into the 5 nm thick region of the top barrier at the distance of either 10 or 15 nm away from the qw , in ( cd , mn)te / cdmgte and cdte / cdmgte structures , respectively . doped regions have been overgrown by an undoped cap layer with thickness of either 50 nm , for ( cd , mn)te / cdmgte qws , or 75 nm , for cdte / cdmgte qws . the density of the two - dimensional electron gas , @xmath10 , and the electron mobility , @xmath11 , determined at liquid helium temperature ( 4.2 k ) for ( cd , mn)te / cdmgte structures are @xmath12 @xmath13 and @xmath14 @xmath15/vs , and for cdte / cdmgte structures @xmath16 @xmath13 and @xmath17 @xmath18/vs . ( cd , mn)te / cdmgte structures contain a single qw made of ( cd , mn)te digital magnetic alloy @xcite , while cdte / cdmgte structures contain a single qw made of non - magnetic cdte . in the former structures two evenly spaced cd@xmath19mn@xmath20te thin layers were inserted during the qw growth , see fig . [ structure](a ) . these two layers were three - monolayer - thick ( ml ) and separated from each other and from the barriers by 8 ml thick cdte ( 1 ml is 0.324 nm thick ) . incorporation of mn@xmath21 ions carrying localized spin @xmath22 into the qw region leads to a strong enhancement of the effective @xmath23 factor of band carriers , and hence to an enhanced zeeman splitting . this has been shown in previous magneto - photoluminescence ( pl ) studies performed on the ( cd , mn)te / cdmgte samples made from the same wafer as used in the current study . magneto - pl showed strong red - shift of the pl line with increasing magnetic field @xcite . from fitting of the modified brillouin function @xcite to the field dependence of the pl line position the effective average concentration of mn in the digital alloy has been estimated to be @xmath24 = 0.015 . the samples have also been characterized by electrical transport measurements . at low temperatures pronounced shubnikov - de - haas ( sdh ) oscillations and very well resolved quantum hall plateaus have been observed . two characteristic dependencies measured in hall bar and van - der - pauw geometries in ( cd , mn)te structures are shown in figs . [ transport](a ) and [ transport](b ) , respectively . while the amplitude of the oscillations increases in cdte and ( cd , mn)te structures at rather high temperatures with raising magnetic field , the amplitude of the oscillations in ( cd , mn)te at low temperatures is a more complex function of the magnetic field @xmath25 and shows a beating like pattern . the latter is clearly seen in fig . [ transport](a ) for @xmath26 k ( @xmath27 t ) and in panel ( b ) for 4.2 k ( @xmath28 t ) . such a behavior in diluted magnetic semiconductor ( dms ) structures is well known to be caused by the exchange interaction of electrons with mn@xmath21 ions resulting in the exchange - enhanced zeeman splitting @xcite . mn@xmath20te layers inserted during qw growth . ( b ) cross - section of the dual - grating gate superlattice formed by metal fingers deposited on top of the qw structure . structure composition is given on the right side of the sketch . the supercell of the grating gate fingers consists of metal stripes having two different widths @xmath29 and @xmath30 separated by spacings @xmath31 and @xmath32 . this asymmetric supercell is repeated @xmath33 times to create a superlattice with period @xmath34 , see table [ tab1 ] . ( c ) photograph of the sample together with schematic top view of the dual - grating gate superlattice . thin red and thick blue lines sketch the top gates having different thicknesses and spacings . ] for the ratchet effect experiments we fabricated a dual grating top gate ( dgg ) superlattices on top of the qw structures . a sketch of the gate fingers and a corresponding optical micrograph are shown in figs . [ structure](b ) and [ structure](c ) , respectively . the grating - gate supercell consists of two metal stripes having widths @xmath29 and @xmath30 and spacings @xmath31 and @xmath32 . the widths of the thin stripes were either @xmath35 or 1.85 @xmath11 m , that of the wide stripes @xmath36 @xmath11 m , see tab . spacings in most of the structures were @xmath37 @xmath11 m and @xmath38 @xmath11 m . to clarify the influence of the superlattice parameters on the ratchet effect , we additionally fabricated one dgg structure with substantially different spacings @xmath39 @xmath11 m and @xmath40 @xmath11 m between the stripes . the supercell is repeated @xmath33 times to produce an asymmetric superlattice with the period @xmath41 , see refs . the two subgrating gates , each formed by stripes of identical width , can be biased independently . for this purpose all thin ( top gate g1 ) and thick ( top gate g2 ) grating stripes have been connected by additional gold stripes , see thick horizontal lines in fig . [ structure](c ) . the benefit of this geometry is that the periodic lateral electrostatic potential of the dgg structure can be varied in a controllable way . several @xmath42 and @xmath43mm@xmath44 size samples of the same batch were prepared . the dgg structure has been fabricated on one half of the samples so that the other half of the sample surface remains unpatterned serving as reference part . it has a total area of @xmath45 where @xmath46 is the overlap length of two gate lattices being of the order of hundreds of micrometer . .sample parameters shown in detail in figs . [ structure](b ) and [ structure](c ) [ cols="^,^,^,^,^,^,^",options="header " , ] . [ tab1 ] in all samples , besides the sample # 4 , the gate fingers have been made by electron beam lithography and subsequent deposition of 25 or 30 nm thick gold films . the dgg structure of sample # 4 has been fabricated by depositing @xmath47 nm thick dy and @xmath48 nm thick au films . pre - magnetizating the hard magnetic dy - based superlattice enables realization of a magnetic ratchet which features an inhomogeneous magnetic field , recently suggested in @xcite . for the photocurrent and magneto - transport measurements several pairs of ohmic contacts have been prepared , see inset in figs.[transport](b ) and [ gatedepb0 ] . contact pads were placed in a way that the photo - induced currents can be measured perpendicularly to the metal fingers ( @xmath49 , contacts 1 and 4 ) and parallel to them ( @xmath50 , contacts 3 and 5 ) . two additional contacts ( 2 and 6 ) were used for detecting the photocurrent signals from the unpatterned area as a reference . magneto - transport data for the ( cd , mn)te dgg structure at liquid helium temperature are shown in fig . [ transport](b ) . note that the application of gate voltages to individual gates does not visibly influence the period of @xmath1-oscillations of the longitudinal resistance , @xmath51 ( not shown ) . this is due to the fact that the area of dgg fingers is a very small fraction of the whole sample area . moreover , the variation of bias voltage applied to gate 1 ( @xmath52 ) also does not substantially affect the value of @xmath51 at a fixed magnetic field . the only detected deviation from the value of @xmath51 at zero gate voltages is an increase of the resistance by @xmath53 % for @xmath54 v. the gate voltage dependencies of the longitudinal resistance measured at one of the maxima of the @xmath51 oscillations are shown in the inset in fig . [ transport](b ) . the dependencies are obtained at fixed magnetic field by sweeping the potential applied to one gate and holding the other one at zero bias . the observed behavior of @xmath51 will be addressed below in the discussion of photocurrent data . note that while the overall characteristic of the data obtained at different cooldowns is the same the onset of the resistance increase can be shifted on the gate voltage scale by @xmath55 v. this is ascribed to cooldown dependent charge trapping in the insulator . and 4.2 k in ( cd , mn)te structures . ( a ) unpatterned hall bar sample . ( b ) square shaped samples with a dgg superlattice . for this measurement a van - der - pauw geometry has been used . inset in ( b ) shows the dependence of the longitudinal resistance , @xmath51 , on the gate voltage @xmath56 ( @xmath57 ) measured for @xmath58 ( @xmath59 ) . measurements are presented for @xmath60 t corresponding to a maximum of the @xmath51-oscillations . bottom inset in ( b ) sketches the dgg sample . ] for thz excitation we applied a continuous wave ( _ cw _ ) molecular optically pumped laser @xcite . the laser operated at the frequency @xmath61 thz ( photon energy @xmath62 mev , wavelength @xmath63 m ) . the incident power about 30 mw was modulated at about @xmath47 or 625 hz by an optical chopper . the radiation at normal incidence was focused onto a spot of about @xmath64 mm diameter . the spatial beam distribution had an almost gaussian profile , measured by a pyroelectric camera @xcite . taking into account the size of the superlattice we obtain that the power irradiating the structure is @xmath65mw . the radiation intensity and electric field strength on the sample are @xmath66mw/@xmath18 and @xmath67v / cm , respectively . note that almost all experimental data , except the scan of the beam across the sample , are obtained for radiation focused on the dgg superlattice . the structures were placed in a temperature variable optical magneto - cryostat . the photocurrents were studied in the temperature range between 2 and 9 k. the laser radiation was linearly polarized along the @xmath68-axis . the photoresponse is measured by the voltage drop @xmath69 across a load resistor @xmath70 ohm @xmath71 using standard lock - in technique . here @xmath72 is the sample resistance . the benefit of using of the small value of @xmath73 is that the detected signal is unaffected by the sample resistance variation and is just proportional to the electric current generated by the thz radiation . the current is calculated via @xmath74 . on the gate voltage @xmath52 ( @xmath57 ) applied to the stripes at zero potential of the other gate @xmath75 ( @xmath59 ) . the polarization independent photocurrent contribution @xmath76 is extracted from the photocurrent polarization dependencies ( not shown ) . the data are obtained for ( cd , mn)te qw sample # 1 at zero magnetic field . here and in the further plots , except for the one showing results on scan across the sample , the data are presented for the laser beam focused on the dgg superlattice . insets show experimental geometries . here and in the further plots , the data are obtained for linearly polarized radiation with the electric field vector @xmath77 directed perpendicularly to the metal gate fingers . ] we begin with briefly introducing the results obtained at zero magnetic field . while the paper is devoted to the observation of magneto - ratchet effect these results are important for further analysis of the data on magnetic field induced photocurrents . illuminating the dgg superlattice at normal incidence we have observed a photocurrent , whose magnitude and direction are sensitive to the electrostatic potentials applied to the first ( g1 ) and the second ( g2 ) gate sublattices . the photoresponse is detected perpendicularly to the gate stripes as well as along them . the fact that the photoresponce is generated by normally incident radiation provides a first indication that it is formed due to the presence of the superlattice . indeed it is well known that in unpatterned ( 001)-oriented qw structures a photocurrent signal is only detectable under oblique incidence @xcite . the photocurrent consists of a contribution independent of the radiation polarization and a substantially smaller one dependent on the orientation of electric field vector @xmath78 of the linearly polarized radiation in respect to the orientation of the gate stripes . this has been demonstrated by measuring the photocurrent as a function the orientation of the @xmath79-field vector in respect to the @xmath68-axis ( not shown ) . the same polarization behavior has been previously observed for the ratchet effect in lateral gaas - based structures @xcite as well as in dgg structures fabricated on top of inalas / ingaas / inalas / inp high electron mobility transistors ( hemt ) @xcite and graphene @xcite . for unbiased gates , a non - zero built - in electrostatic potential is formed due to the presence of metal stripes in the qw vicinity . a direct evidence for the ratchet effect comes from the variation of the photocurrent by application of the bias voltages to the individual gates . indeed the ratchet effect is known to be proportional to the averaged product @xmath80 of the derivative of the coordinate dependent electrostatic potential @xmath81 and the distribution of the electric near - field @xmath82 @xcite . in the following , we call the parameter @xmath83 the _ lateral asymmetry_. the value of @xmath83 may change sign depending on @xmath81 . consequently a variation of individual gate voltages should result in a change of the ratchet current including reversal of its direction , see the discussion in sec . [ discussion ] . exactly this behavior has been observed in the experiment . in order to tune the lateral asymmetry , we applied different bias voltages @xmath84 to the grating gates . figure [ gatedepb0 ] demonstrates the influence of the gate voltage variation on the amplitude and sign of the polarization independent photocurrent @xmath85 @xcite . holding one of the gates at zero bias and varying the gate voltage on the other one we could controllably change the lateral asymmetry . the figure reveals that inversion of this lateral asymmetry obtained either by change of polarity of voltage applied to one gate or by interchange of gate voltage polarities applied to narrow and wide gates results in a change of sign of the photocurrent . this observation agrees well with the signature of ratchet currents : @xmath86 . note that , while the overall characteristics of the data obtained at different cooldowns is the same , the photocurrent sign inversions can be shifted on the gate voltage scale by @xmath55 v. similarly to the results addressed in sec . [ imp ] this is ascribed to cooldown dependent charge trapping in the insulator . we also note that for large negative gate voltages all contributions vanish , which can be attributed to closing the transistor channel beneath the metal stripes of the dgg superlattice . figure [ gatedepb0 ] also reveals that at large positive gate voltages the photocurrent saturates due to the effect of high electron densities . to summarize , experiments at zero magnetic field provide a consistent picture demonstrating that the photocurrents are caused by the ratchet effect . they are ( i ) generated due to the lateral asymmetry , ( ii ) characterized by specific polarization dependencies for directions along and across the metal stripes , and ( iii ) change the direction upon reversing the lateral asymmetry . these results are in full agreement with the theory of ratchet effects excited by the polarized thz electric field in asymmetric lateral superlattices , discussed below in sec . [ discussion ] . the overall behavior of the photocurrent is also in qualitative agreement with that of the electronic ratchet effects observed in semiconductor qw structures and graphene with a lateral superlattice @xcite . as a function of the magnetic field @xmath2 measured in ( cd , mn)te qw sample # 1 for two combinations of the gate voltages . solid blue line show the data for the gate voltages @xmath87 and @xmath88 v and the dashed red line for @xmath89 v and @xmath90 , the setups denoted as @xmath91 v and @xmath92 v. the vertical dashed line indicates zero magnetic field . the inset illustrates the experimental setup . ] now we turn to the main part of the paper devoted to magnetic field induced ratchet effects . figure [ jxjy4k ] shows the normalized photocurrent excited by linearly polarized radiation as a function of magnetic field applied perpendicularly to the qw plane . the principal observation is that with raising magnetic field the photocurrent drastically increases and , at high magnetic fields , exhibits sign - alternating @xmath1-periodic oscillations . moreover , fig . [ jxjy4k ] demonstrates that the reversal of the lateral asymmetry results in an inversion of the magneto - photocurrent direction providing a first clear indication that it is driven by the ratchet effect . measurements of the magneto - photocurrent as a function of radiation power show that it scales linearly with the radiation power , i.e. quadratically with the radiation electric field , @xmath93 , see inset in fig . [ intensity ] . note that the positions of oscillation maxima / minima are independent of the radiation power , fig . [ intensity ] . similar results , particularly magneto - oscillations with current amplitudes much larger than the photocurrent at @xmath94 , are obtained for all studied samples . studying various samples at different experimental conditions we found that the current in the @xmath95-direction is much smaller than @xmath49 . data from the ( cd , mn)te qw sample # 4 with ferromagnetic dy gate are shown in fig . the period and magnitude of the @xmath1-oscillations correlate with transport behavior and are very close to those shown for ( cd , mn)te qw samples with gold dgg structure presented for sample # 1 , fig . [ jxjy4k ] . measured at different levels of power ranging from 0.8 to 4.3 mw in ( cd , mn)te qw sample # 1 . inset shows radiation power dependence of the photocurrent amplitude @xmath96 for @xmath60 t. solid line is a linear fit . ] measured as a function of magnetic field @xmath2 in ( cd , mn)te qw sample # 4 with the dy grating . ] figure [ cdte ] shows another example of magneto - photocurrent measured for nonmagnetic cdte qw sample # 5 with gold dgg . while oscillations are clearly detected the signal is superimposed with a substantial background current . because of the background the photocurrent , in contrast to samples # 1 and # 4 , does nt change sign with varying magnetic field . the background photocurrent has also been obtained for ( cd , mn)te qw samples # 2 and # 3 . we attribute an appearance of the background to imperfections of our large - size superlattices which are present in some structures as confirmed by optical microscope images @xcite . the imperfections locally reduce symmetry and give rise to ratchet unrelated magneto - gyrotropic photogalvanic currents in the `` bulk '' of the qw , see e.g. ref . @xcite . as a function of magnetic field @xmath2 in cdte - based qw sample # 5 at two gate voltage sequences : @xmath97 v@xmath98 and @xmath99@xmath1002.5 v. panels ( b ) and ( c ) show dependencies of the photocurrent magnitude on the gate voltage @xmath52 ( @xmath57 ) obtained for zero potential on the other gate @xmath57 ( @xmath52 ) . the data in panels ( b ) and ( c ) are obtained for the photocurrent maxima / minima at @xmath101 and 4.9 t , respectively . ] to prove that the magneto - ratchet effect stems from the irradiation of the superlattice we scanned the laser spot across the sample along the @xmath68-direction . the photocurrent @xmath49 was measured for @xmath102 t and the gate voltage combination @xmath1030.5 v , i.e. at a maximum of one of the magneto - oscillations , fig . [ jxjy4k ] . the photocurrent generated by linearly polarized radiation with @xmath104 as a function of the radiation spot position and the corresponding experimental geometry are shown in fig . the photocurrent reaches its maximum for the laser spot centered at the superlattice and rapidly decays with the spot shifting away . for the beam spot touching the sample edge the current increases again . this is caused by an edge photocurrent @xcite known for graphene @xcite and semiconductor qws @xcite . comparison with the laser beam spatial distribution measured by a pyroelectric camera ( dashed curve in fig . [ scan ] ) shows that the change of the photocurrent for the scans across the dgg structure only slightly deviates from the gaussian intensity profile . note that the dgg area is smaller than the beam spot . these findings unambiguously demonstrate that the photocurrent is caused by irradiation of the superlattice . measured in ( cd , mn)te qw sample # 1 at a maximum of the photocurrent oscillation at @xmath102 t and @xmath87 , @xmath88 v. inset shows the setup with the laser spot scanned along the @xmath68-axis . positions @xmath105 and @xmath106 correspond to the borders of the superlattice whereas @xmath107 denotes the sample edge position . the dashed line represents the laser beam spatial distribution which is measured by a pyroelectric camera and scaled to the current maximum . ] in order to explore the role of the lateral asymmetry , a prerequisite for the ratchet photocurrents , we have systematically studied magneto - photocurrents for different combinations of the gate voltages . the results are shown for different samples in figs . [ gatedepb0 ] , [ jxjy4k ] , [ cdte ] and [ asymmetry ] . ( @xmath57 ) obtained for zero potential on the other gate @xmath57 ( @xmath52 ) . the data are obtained for ( cd , mn)te qw sample # 1 at @xmath60 t and @xmath108 t , panels ( a ) and ( b ) , respectively . the inset shows the magnetic field dependence of the photocurrent measured for different gate voltage combinations . vertical dashed lines demonstrate that the maxima / minima positions of the magneto - photocurrent are not shifted for all gate voltage combinations . the numbers in the inset correspond to : 1 @xmath109 @xmath110 v ; 2 @xmath109 @xmath111 v ; 3 @xmath109 @xmath112 v ; 4 @xmath109 @xmath113 v ; 5 @xmath109 @xmath114 v and 6 @xmath109 @xmath115 v. ] while the magnitude and the sign of the oscillations are strongly affected by the variation of the gate potentials the maxima / minima positions remain unchanged . to show that this conclusion is valid for the whole range of gate potentials used in our measurements we studied the photocurrent for a large set of different magnitudes and polarities of @xmath52 and @xmath116 . these results are presented in the inset in fig . [ asymmetry ] for @xmath2 ranging from 2.7 to 4.1 t. using the fact that maxima positions do not shift upon variation of @xmath2 we fixed the magnetic field and controllably varied the lateral asymmetry by changing the potential applied to one of the gates and holding the other one at zero bias . figures [ asymmetry](a ) and [ asymmetry](b ) show the results obtained for ( cd , mn)te qw sample # 1 at magnetic field strengths 3.9 and @xmath108 t , respectively . the figures reveal that the polarity of photocurrent changes with the sign of the asymmetry parameter @xmath83 , eq . . this is primarily seen by comparing the curves obtained by variation of the potentials applied either to the thin ( dashed lines ) or thick ( solid lines ) gate stripes . the observed stronger change of the signal for @xmath117 v corresponds to an increase of the longitudinal resistance detected in magneto - transport measurements , see inset in fig . [ transport](b ) . it should be noted that , due to the abovementioned built - in potential , the ratchet current can be detected even for zero gate voltages at both gates . the same behavior upon variation of the lateral asymmetry has been observed for dy - based dgg sample # 4 . studying the photocurrent as a function of the gate voltage in cdte qw sample # 5 and in ( cd , mn)te samples # 2 and # 3 we obtained , that , despite the background , the photocurrent magnitude is also controlled by the lateral asymmetry . corresponding results obtained for cdte qw dgg structure # 5 at @xmath1184.9 t are shown in figs . [ cdte](b ) and ( c ) . as a function of @xmath2 measured in ( cd , mn)te qw dgg sample # 1 for three temperatures and two gate voltage combinations . ] now we turn to the analysis of the @xmath1-oscillations of the photocurrent which are obviously related to the sdh oscillations . comparison with the magneto - transport data reveals that the period of the @xmath1-oscillations is indeed equal to that of the longitudinal magneto - resistance oscillations . this result is confirmed for all samples and temperatures used in our experiments . in the cdte samples and ( cd , mn)te sample at relatively high temperatures the amplitude of oscillations grows with the magnetic field , see e.g. fig . [ temperatur ] . by contrast , in diluted magnetic semiconductor ( cd , mn)te qw dgg samples at liquid helium temperature @xmath119 the rise of the magneto - photocurrent amplitude at low fields is followed by its reduction at higher fields , see fig . [ temperatur](b ) . at the lower temperature @xmath120 k this reduction , now detected at even lower magnetic fields , is followed by a further increase of the photocurrent magnitude at higher @xmath2 , see data in fig . [ temperatur](c ) . similar `` beats '' of the magneto - oscillation amplitudes are detected for the longitudinal resistance , fig . [ transport ] , in agreement with the presence of the exchange enhanced zeeman splitting in dms materials @xcite . the enhanced zeeman splitting is caused by the exchange interaction between electrons and mn@xmath21 ions and described by the modified brillouin function . the spin splitting first grows linearly with @xmath2 and then saturates . the temperature increase leads to a decrease of exchange interaction and shifts the saturation to higher fields . all these features are detected in both transport and magneto - photocurrent experiments . moreover , due to the enhanced spin splitting the oscillations measured at low temperatures and high magnetic fields become spin - split which results in the beats . in low - dimensional semiconductor structures with lateral superlattices , a @xmath0 electric current is generated due to the action of an electromagnetic wave s @xmath121 electric field @xcite . the effect of the superlattice is twofold : it generates a one - dimensional periodic electrostatic potential @xmath81 acting upon the 2d carriers with @xmath68 being the superlattice principal axis , and causes a periodic spatial modulation of the thz electric field due to the near field diffraction . figure [ field ] shows coordinate dependence of the thz electric near - field @xmath122 calculated for the ( cd , mn)te qws based dgg structure for two combinations of the gate voltages . the electric field distribution caused by the near - field diffraction is calculated for radiation with frequency @xmath123 thz applying a self - consistent electromagnetic approach based on the integral equation method described in detail in ref . figure [ absorption ] shows the calculated thz absorption spectrum of the structure . it is seen that the plasmonic resonances in the qw are excited at frequencies well below the operation frequency of 2.54 thz . therefore in the experimental frequency range the absorption follows the drude law , see the inset in fig . [ absorption ] . as a result , the near - field distribution in the qw shown in fig . [ field ] is caused almost solely by the metal dual - grating gate and not by the plasma oscillations . hence one can neglect the dependence of the near field on the voltage applied to the dgg . moreover , since the radiation frequency is much higher than that of plasmonic resonances we consider only electronic ratchet mechanism below . normalized to the electric field of the incident thz wave . shaded regions correspond to the positions of dgg metal fingers . results are obtained for parameters of ( cd , mn)te qw sample # 1 with dielectric constant @xmath124 at the frequency 2.54 thz . the momentum relaxation time @xmath125 is obtained from the electron mobility specified in sec . [ sample ] . ] . ] on the gate voltages @xmath52 ( dashed line ) and @xmath57 ( solid ) fitted to experimental curves shown in fig . [ gatedepb0 ] ( @xmath46 is defined in fig . [ structure ] ) . the lines are obtained after eq . ( [ j_zero_b ] ) . to be specific , we used for calculations @xmath77 in accordance with the experimental conditions . for the calculations we used the ( cd , mn)te dgg qw structure parameters discussed in sec . [ sample ] , electric near - field @xmath122 shown in fig . [ field ] , lattice temperature @xmath126 k , frequency @xmath61 thz and effective mass @xmath127 . the only fitting parameter used here is the relaxation time @xmath128 ns . inset shows the calculated frequency dependence of the photocurrent @xmath49 together with the experimental result for @xmath61 thz ( dot ) . ] the ratchet current is generated by a combined action on carriers of both the radiation near field , transmitted through the grating , and the periodic static potential @xmath81 @xcite . a crucial condition for @xmath0 electric current generation is that the lateral superlattice is asymmetric . applying the voltages @xmath129 and @xmath130 in a controllable way , one can vary the potential @xmath81 and reverse the lateral asymmetry @xmath83 . consequently , the ratchet photocurrent can also reverse its sign as observed in experiment , see fig . [ gatedepb0 ] . the ratchet current has components both perpendicular and parallel to the metal fingers and consists , in general , of polarization - independent , linear- and circular - ratchet contributions . all these individual ratchet currents have been observed and described in various semiconductor - based structures @xcite as well as in graphene with lateral periodical top gates @xcite . the mechanism leading to photocurrent formation can be illustrated on the polarization - independent photocurrent caused by the seebeck ratchet ( thermoratchet ) effect . this contribution dominates at zero magnetic field ( sec . [ erstes ] , fig . [ gatedepb0 ] ) and , as we show below , is most relevant to the observed magnetic ratchet current . the seebeck ratchet effect is caused by inhomogeneous heating of two - dimensional electron gas and subsequent relaxation of the electron temperature . due to the near - field space modulation , the field @xmath82 acting upon electrons depends on the coordinate @xmath68 , fig . [ field ] . the field heats the electron gas changing the effective electron temperature locally to @xmath131 . here we assume that , within a short time , the electron distribution becomes a quasi equilibrium fermi - dirac distribution with the temperature @xmath132 . the space - modulated temperature correction is defined by the energy balance and can be found from the following equation @xcite @xmath133 here @xmath134 is the boltzmann constant , @xmath135 is the temperature relaxation time , @xmath125 is the momentum relaxation time , @xmath136 is the electron charge , and @xmath137 is the electron effective mass . we take into account that , in the present experiment @xmath138 , where @xmath139 is the cyclotron frequency . as a result of the inhomogeneous heating , electrons diffuse from warmer to colder regions , and form a nonequilibrium density profile given by @xmath140 where @xmath33 is the average electron density , @xmath141 is the fermi energy . the ratchet current at zero magnetic field can be represented as a drift current of the electrons in the electric field @xmath142 of the space modulated electrostatic potential @xmath81 . from ohm s law @xmath143 with @xmath144 we have @xmath145 where the bar denotes averaging over the superlattice period . finally , the ratchet current density at zero magnetic field is given by @xcite @xmath146 where @xmath83 is given by eq . . figure [ popovfateev ] shows the photocurrent calculated after eq . together with the data of fig . [ gatedepb0 ] . the result of fig . [ popovfateev ] is obtained for the near - field @xmath122 shown in fig . [ field ] . parameters used for the calculations correspond to the sample # 1 . the only adjustable parameter is the temperature relaxation time taken as @xmath128 ns . this value agrees with the temperature relaxation time of few nanoseconds reported for a degenerate electron gas in gaas - based heterojunctions @xcite . figure [ popovfateev ] shows qualitative and reasonably good quantitative agreement between theory and experiment . the above analysis suggests that , in the absence of magnetic field , the ratchet current is rather weak for degenerate electrons because , according to eq . , the relative concentration correction @xmath147 is much smaller than the temperature correction @xmath148 . below we show that the situation changes drastically in the presence of a quantizing magnetic field @xmath149 . the magnetic field induced ratchet currents in the directions perpendicular , @xmath150 , and parallel , @xmath151 , to the stripes of the superlattice are given by @xcite @xmath152 ^ 2 } { \sinh{z}-z\cosh{z } \over \sinh^2{z } } \delta,\ ] ] @xmath153 here the magneto - oscillations are described by the factor @xmath154 ( @xmath155 ) @xmath156 while the dependence on @xmath157 describes the oscillation suppression with increasing temperature , eqs . - are derived in the first order in @xmath154 . as compared with the theory developed in ref . @xcite we take into account the spin subband splitting @xmath158 due to the zeeman effect . for the dms based ratchet structures the exchange enhanced zeeman splitting is given by @xcite @xmath159 where @xmath160 is the electron land factor in the absence of magnetic impurities , @xmath161 is bohr s magneton , @xmath162 is the mn @xmath23-factor , @xmath163 is the mn - spin system temperature , @xmath164 and @xmath165 account for the mn - mn antiferromagnetic interaction , @xmath166 is the modified brillouin function , and @xmath167 is the exchange integral . equations , predict that the ratchet current components @xmath150 and @xmath168 have opposite parity upon magnetic field reversal due to the photocurrent deflection by the lorentz force @xcite . the ratchet current becomes an oscillating function of the magnetic field as the landau levels move through the fermi level and the photocurrent @xmath1-oscillation period is the same as for sdh oscillations . equation ( [ j ] ) explains the reason for the ratchet current oscillations with giantly enhanced magnitude as compared to the ratchet current at zero magnetic field . the crucial issue is that the ratchet current in a quantized magnetic field arises due to heating - induced corrections to the conductivity rather than a variation of the electron density , and the conductivity in this regime is extremely sensitive to electron temperature variations . indeed , even a weak change of the electron temperature near the dingle temperature results in exponentially strong changes of the conductivity @xcite . therefore , the sensitivity of the conductivity to electron gas heating results in a huge enhancement of the ratchet current . the ratio @xmath169 is governed by a factor @xmath170 and can reach about two orders of magnitude @xcite . it should be stressed that this enhancement occurs in moderate fields where the resistance exhibits weak sdh oscillations . heating damps the conductivity oscillations , therefore , depending on the sign of the correction , heating can either increase or decrease the conductivity because ( @xmath171 ) changes sign as a function of @xmath2 . as a result , the heating - induced ratchet current is an oscillating function of the field with a zero mean value . and @xmath151 calculated after eqs . , for ( cd , mn)te qw sample at @xmath126 k. the structure parameters are relevant to the experiment presented in fig . [ jxjy4k ] : @xmath172 mev , @xmath173 and @xmath174 . for other parameters we used the literature values for _ n_-(cd , mn)te : @xmath175 , @xmath176 mev , see ref . @xcite , and @xmath127 , see ref . for the calculations we also used @xmath177 and literature data of @xmath178 k and @xmath179 for @xmath180 , see ref . solid and dashed curves show the results for two asymmetries of equal magnitudes but opposite signs of @xmath83 . the oscillation amplitude is much higher than the ratchet current @xmath181 at zero magnetic field ( the latter is not seen at the chosen scale ) . ] the calculated magnetic field dependence of the ratchet currents @xmath150 and @xmath151 are shown in fig . [ theoryjxjy ] . the results are presented for two gate voltage combinations with equal magnitude but opposite sign of @xmath83 . equations -([jy ] ) reveal the most important features of the magnetic ratchet effect : the current is proportional to @xmath182 , oscillates around zero with the same period as the longitudinal magnetoresistance , the amplitude of the current oscillations is immensely larger than the ratchet current at zero magnetic field , and the sign of the photocurrent reverses by changing the sign of @xmath183 . all these features have been observed experimentally , see figs . [ jxjy4k]-[cdte ] , [ asymmetry ] , and [ temperatur ] . equations ( [ j])- also describe the more complicated oscillating ratchet current behavior observed in the dms structures at low temperatures , see figs . [ temperatur ] and [ theorytemp ] for experiment and theory , respectively . indeed , due to the temperature dependence of the brillouin function at low temperatures the zeeman splitting becomes comparable to the landau levels separation . calculated curves for three values of the lattice temperatures and , consequently , different contributions of the zeeman effect , are shown in fig . [ theorytemp ] . the calculations show , as in experiment , beats in the oscillations of @xmath184 . they stem from the interplay of zeeman and landau splittings and become more pronounced with temperature decreasing . comparison of figs . [ temperatur ] and [ theorytemp ] demonstrates that eq . ( [ j ] ) describes the qualitative behavior of the photocurrent quite well . we note that the beats in @xmath185 are substantially influenced by the mn spin system temperature , which , because of electron gas heating , can be higher than the lattice one @xcite . this increase due to the strong dependence of the brillouin function on @xmath163 , results in a reduced exchange part of @xmath158 in eq . ( [ deltaz ] ) . this shifts the onset of the exchange enhanced zeeman splitting to higher magnetic fields . allowance for this effect should further improve the agreement between experimental data and calculations . calculated for ( cd , mn)te qw sample at three different temperatures . curves are obtained for the same values of @xmath186 , @xmath11 , @xmath24 , @xmath23 , @xmath187 , @xmath137 , @xmath188 , @xmath165 and @xmath164 as used in fig . 14 . solid and dashed curves show , similarly to fig . [ theoryjxjy ] , the results for two opposite lateral asymmetries . ] to summarize , we have experimentally demonstrated and theoretically explained the magnetic quantum ratchet effect , i.e. , the ratchet effect in quantizing magnetic fields . to generate the ratchet current we use an asymmetric interdigitated top gate superlattice on top of the investigated quantum wells . electric currents driven by terahertz electric field exhibit sign - changing magneto - oscillations with an amplitude giantly enhanced as compared to the photocurrent at zero magnetic field . the current amplitude and direction can be controllably changed by the variation of voltages applied to individual gate sublattices . the effect is observed in structures with various qw compositions and superlattice parameters . the photocurrent generation mechanism can be well described in terms of semiclassical theory of magnetic ratchet effects . the observed effect is driven by the periodic modulation of the electron temperature caused by near field diffraction . the theory of the seebeck ratchet effect in the presence of quantizing magnetic field shows that the ratchet current follows the oscillations of the longitudinal resistance . in the dms structures at low temperatures the investigated effect is strongly influenced by the exchange enhanced zeeman splitting caused by the exchange interaction of electrons with mn@xmath21 ions . finally we note , that while here we have dealt with a homogeneous magnetic field , the observed ratchet photocurrent in superlattices made of hard magnetic material ( dy ) opens a possibility to study the ratchet effect driven by an inhomogeneous periodic magnetic field recently suggested in ref . this kind of experiments would apply remnant magnetization of dysprosium together with the enhanced magnetic properties of dms qws and is a future task . we thank i. a. dmitriev and d. r. yakovlev for helpful discussions . the support from the dfg priority program sfb 689 , the volkswagen stiftung program and rfbr ( projects 15 - 02 - 02989 , 16 - 02 - 01037 , 16 - 02 - 00375 , and 17 - 52 - 53063 ) is gratefully acknowledged . the work of dvf was supported by the grant from the president of the russian federation ( no . mk-5447.2016.2 ) . the research in poland was partially supported by the national science centre ( poland ) through grant no . dec-2012/06/a / st3/00247 and by the foundation for polish science through the ira programme financed by eu within sg op programme . f. jlicher , a. ajdari , and j. prost , rev . phys . * 69 * , 1269 ( 1997 ) . p. reimann , phys . rep . * 361 * , 57 ( 2002 ) . h. linke ( ed . ) , _ ratchets and brownian motors : basics , experiments and applications _ , special issue appl . phys . a : mater . sci . process . a * 75 * , 167 ( 2002 ) . p. hnggi and f. marchesoni , rev . phys . * 81 * , 387 ( 2009 ) . v. v. popov , d. v. fateev , e. l. ivchenko , and s. d. ganichev , phys . b * 91 * , 235436 ( 2015 ) . p. olbrich , j. karch , e. l. ivchenko , j. kamann , b. mrz , m. fehrenbacher , d. weiss , and s. d. ganichev , phys . b * 83 * , 165320 ( 2011 ) . p. olbrich , j. kamann , m. knig , j. munzert , l. tutsch , j. eroms , d. weiss , m .- h . liu , l. e. golub , e. l. ivchenko , v. v. popov , d. v. fateev , k. v. mashinsky , f. fromm , th . seyller , and s. d. ganichev , phys . b * 93 * , 075422 ( 2016 ) . t. watanabe , s. a. boubanga - tombet , y. tanimoto , d. fateev , v. popov , d. coquillat , w. knap , y. m. meziani , yuye wang , h. minamide , h. ito , and t. otsuji , ieee sensors j. * 3 * , 89 ( 2013 ) . y. kurita , g. ducournau , d. coquillat , a. satou1 , k. kobayashi , s. boubanga tombet , y. m. meziani , v. v. popov , w. knap , t. suemitsu , and t. otsuji , appl . 104 * , 251114 ( 2014 ) . s. a. boubanga - tombet , y. tanimoto , a. satou , t. suemitsu , y. wang , h. minamide , h. ito , d. v. fateev , v. v. popov , and t. otsuji , appl . . lett . * 104 * , 262104 ( 2014 ) . v. v. popov , d. v. fateev , t. otsuji , y. m. meziani , d. coquillat , and w. knap , appl . . lett . * 99 * , 243504 ( 2011 ) . p. faltermeier , p. olbrich , w. probst , l. schell , t. watanabe , s. a. boubanga - tombet , t. otsuji , and s. d. ganichev , j. appl . phys . * 118 * , 084301 ( 2015 ) . s. d. ganichev , w. weber , j. kiermaier , s. n. danilov , d. schuh , w. wegscheider , ch . gerl , d. bougeard , g. abstreiter and w. prettl , j. appl . physics * 103 * , 114504 ( 2008 ) . s. n. danilov , b. wittmann , p. olbrich , w. eder , w. prettl , l. e. golub , e. v.beregulin , z. d. kvon , n. n. mikhailov , s. a. dvoretsky , v. a. shalygin , n. q. vinh , a. f. g. van der meer , b. murdin , and s. d. ganichev , j. appl . phys . * 105 * , 013106 ( 2009 ) . p. olbrich , c. zoth , p. lutz , c. drexler , v. v. belkov , ya . v. terentev , s. a. tarasenko , a. n. semenov , s. v. ivanov , d. r. yakovlev , t. wojtowicz , u. wurstbauer , d. schuh , and s. d.ganichev , phys . b * 86 * , 085310 ( 2012 ) . j. a. gaj , r. planel , and g. fishman , solid state commun . * 29 * , 435 ( 1979 ) . j. k. furdyna , j. appl . phys . * 64 * , r29 ( 1988 ) . t. dietl , in _ handbook on semiconductors _ , vol.3b , ed . t.s . moss ( north - holland , amsterdam , 1994 ) . j. karch , p. olbrich , m. schmalzbauer , c. zoth , c. brinsteiner , m. fehrenbacher , u. wurstbauer , m.m . glazov , s. a. tarasenko , e. l. ivchenko , d. weiss , j. eroms , and s. d. ganichev , phys . lett . * 97 * , 227402 ( 2010 ) . j. karch , c. drexler , p. olbrich , m. fehrenbacher , m. hirmer , m. m. glazov , s. a. tarasenko , e. l. ivchenko , b. birkner , j. eroms , d. weiss , r. yakimova , s. lara - avila , s. kubatkin , m. ostler , t. seyller , and s. d. ganichev , phys . * 107 * , 276601 ( 2011 ) . m. m.glazov and s. d.ganichev , physics reports * 535 * , 101 ( 2014 ) . m.dantscher , d.a.kozlov , m.t.scherr , s.gebert , j.brenfnger , m.v.durnev , s.a.tarasenko , v.v.belkov , n.n.mikhailov , s.a.dvoretsky , z.d.kvon , d.weiss , and s.d.ganichev , arxiv cond - mat : 1612.08854 ( 2016 ) . k .- m.dantscher , d. a. kozlov , p. olbrich , c. zoth , p.faltermeier , m.lindner , g. v.budkin , s. a. tarasenko , v. v.belkov , z. d.kvon , n. n. mikhailov , s. a. dvoretsky , d.weiss , b. jenichen , and s. d.ganichev , phys . b * 92 * , 165314 ( 2015 ) .

the naked reissner - nordstrm ( r - n ) singularity is a classical general relativistic solution in electrovacuum . the solution is expected to have a very limited meaning , due to the fact that such singularities can not be created neither by a gravitational collapse , nor by dropping a charge into the black hole . ( according to the weak cosmic censorship conjecture general naked singularities should be prohibited in general theory of relativity , although there are indications that by including quantum effects the violations of the conjecture could be considered @xcite . ) moreover , a naked singularity created from some exotic initial data conditions should become quickly neutralized ( classically , or via quantum pair production ) . some results also indicate that if one considers electro - gravitational perturbations the r - n naked singularity becomes linearly unstable @xcite . however it was discovered that the scalar field scattering problem on such a singular background can be still well defined @xcite , since the waves remain regular at the origin . ( however , the back reaction of the given scalar field configuration might still excite some of the unstable electro - gravitational modes and this would eventually lead to a breakdown of the perturbation approach . ) despite the nice regularity property of the scattering problem , the spacetime is non - globally hyperbolic and the time evolution of the fields is not unique @xcite . this means one has to specify an additional boundary condition at the singularity to obtain a fully unique time evolution . another way of seeing the problem is through the language of operators : one can understand the spatial part of the wave operator as a positive symmetric operator acting on a @xmath4 hilbert space , and then obtain the scalar field dynamics through a suitable positive self - adjoint extension of such a symmetric operator @xcite . ( one `` preferred '' way in which such a self - adjoint extension can be always realized is through the so called friedrich s extension @xcite , which will also be the case of this paper . ) anyway , after uniquely specifying the dynamics , one should be able to characterize the scattering by a set of characteristic oscillations , the quasi - normal modes . low damped quasi - normal modes are in general used as a possible source of information about potential astrophysical objects ( such as neutron stars , black holes ) , and the highly damped modes are potentially interesting from the point of view of quantum gravity @xcite . since a lot of work was devoted to the problem of quasinormal modes of the reissner - nordstrm black hole , it might be interesting to observe what happens if one transits from the r - n black hole case to the r - n naked singularity case ( with a reflective boundary condition ) . information about `` what happens '' shows how many features of the quasinormal modes of the black hole spacetimes are specific to the black holes themselves and what features survive much more general conditions . let us also give one concrete example why asking what happens with the quasinormal modes of the naked r - n singularity might be interesting : in the black hole case the behavior of the asymptotically highly damped modes is widely suspected to be linked to the properties of the black hole spacetime horizon(s ) , ( or more specifically it is considered to carry information about _ quantum _ black holes ) . what does then happen with the asymptotically highly damped modes in case there are _ no _ horizons present ? ( if the behavior of the highly damped modes in case of r - n naked singularity would , for instance , resemble the behavior of the highly damped modes of the black hole , it would be a disturbing fact from the point of view of the popular conjectures linking the modes to the horizon s quantum area spacing . moreover , investigating the r - n naked singularity from this point of view is especially attractive , as the r - n naked singularity is obtained by a continuous transition in the @xmath5 parameters from the r - n black hole spacetime . ) thus , briefly , we hope that despite the fact that most likely the r - n naked singularity model does not correspond to a realistic physical situation , there are still many interesting things one can learn from such a model . the structure of this paper is as follows : in the second section we analyse the problem of the uniqueness of the time evolution of scalar fields on a r - n naked singularity background . in the sections three and four we analyse the properties of the effective potential for the scalar fields scattering and the geometrical optics ( eikonal ) limit of such a scattering problem . in the fifth section we define analytically solvable potentials that can give good approximations to the problem of the low damped qnm frequencies ( such that characterize the given scattering problem ) . in the sixth section we use those analytical approximations to derive semi - analytical results for the qnm frequencies in the eikonal limit . in the seventh section we use the numerical characteristic integration to obtain the low damped frequencies for the low values of @xmath3 . in the eighth section we analyse what happens with the asymptotically highly damped modes and we suggest that in case of naked r - n singularity such modes do _ not _ exist ( as one might expect considering some presently popular conjectures @xcite ) . we give the final conclusions in the section nine . we provide also appendices with more detailed results and some further suggestions for the analytical approximations of the problem . in this section we will follow the standard analysis of the scalar field evolution in a curved background . ( as an example of such an analysis see the treatment of schwarzschild black hole perturbations in @xcite . for a review that presents also such techniques see for example @xcite . ) take the klein - gordon equation for the complex ( charged ) scalar field : @xmath6 with the metric line element given as @xmath7 for the reissner - nordstrm ( r - n ) singularity the function @xmath8 is in planck units given as : @xmath9 take the decomposition of the field into the spherical harmonics @xmath10 after we separate the variables we obtain the following reduced equation @xmath11 \nonumber \\ & -&\frac{l(l+1)f(r)}{r^{2}}\psi_{l}(t , r ) . \end{aligned}\ ] ] we are interested only in compactly supported data initial value problem : first assume that @xmath12 is everywhere bounded and hence the laplace transform of @xmath12 exists : @xmath13 the laplace transformed equation gives the following equation : @xmath14 \nonumber \\ & -&\frac{l(l+1)f(r)}{r^{2}}\tilde\psi_{l}(s , r)+i_{l}(s , r),\end{aligned}\ ] ] where @xmath15_{|t=0}.\ ] ] the solution that corresponds to the initial data term @xmath16 is obtained by the inverse laplace transform of @xmath17 and the function @xmath17 is given as @xmath18 here @xmath19 is a green s function satisfying @xmath20 \\ & + & \left[s^{2}+\frac{l(l+1)f(r)}{r^{2}}\right ] g_{l}(s , r , r')=\delta(r - r ' ) . \nonumber\end{aligned}\ ] ] since @xmath12 was bounded , its laplace transform @xmath17 must be also bounded in @xmath21 . this translates to the boundedness of the green s function in @xmath21 . the unique solution of the given initial data problem is obtained if the condition of boundedness of the green s function leads to a _ unique _ way to construct green s function from the two linearly independent solutions @xmath22 of the homogeneous equation @xmath23 \\ & + & \left[s^{2}+\frac{l(l+1)f(r)}{r^{2}}\right ] u_{l1,2}(s , r)=0.\nonumber\end{aligned}\ ] ] if @xmath8 goes to 1 at spatial infinity ( the metric is asymptotically flat ) , then there is only one solution of that stays bounded as @xmath24 . if there exists only one solution of the equation , such that it is linearly independent from the solution bounded at infinity and in the same time it is bounded at @xmath25 then these two solutions uniquely define the green s function . if all the solutions of are singular at 0 , there is no solution ( bounded in @xmath21 ) of the given initial value problem such that it can be laplace transformed ( this can be taken as an indication that there is no solution at all ) . if both of the linearly independent solutions are regular at 0 , and ( at least ) two different green functions lead to a function in the domain of the inverse laplace transform , then there is no _ uniquely _ defined solution to the initial value problem . in the `` worst '' case there are infinitely many solutions , given by arbitrary linear combination of @xmath22 , that are linearly independent to the solution bounded at @xmath24 . in such a case the problem is underdetermined and one needs one more condition at @xmath26 that selects a unique green s function between the different green s functions marking different time evolutions . for each one of the choices of the green s function , one can reproduce the calculation from @xcite and see that the quasinormal modes defined by 1 . the choice of the green s function close to 0 , 2 . the outgoing radiation condition , characterize the time evolution of the field at a fixed point within some specific time interval . unfortunately for the case of r - n naked singularity ( @xmath27 , @xmath28 ) both of the linearly independent solutions @xmath22 are regular at 0 and the problem is underdetermined . it is easy to show that both of the solutions are regular at the origin . write as : @xmath29\frac{d u_{l1,2}(s , r)}{dr}\nonumber\\ & + & \left[s^{2}+\frac{l(l+1)f(r)}{r^{2}}\right ] u_{l1,2}(s , r)=0.\end{aligned}\ ] ] now taking the @xmath25 limit of the equation one obtains the following @xmath30\frac{d u_{l1,2}(s , r)}{dr}\nonumber \\&+&\frac{l(l+1)}{q^{2 } } u_{l1,2}(s , r)=0.\end{aligned}\ ] ] this means the solutions @xmath31 behave close to 0 as @xmath32 with @xmath33 hence both are regular . the fact that the problem is underdetermined is not surprising , since the space - time is _ not _ globally hyperbolic and anything can fall out at any time from the singularity . this means the singularity has `` hair '' ( carries some other information beyond the metric ) and the quasi - normal modes obviously depend on the hair " . let us add here that , as we already mentioned in the introduction , the `` hair '' of the singularity relates to the existence of many different self - adjoint extensions of the `` hamiltonian '' operator in the equation . ( in @xcite one can find a nice analysis of the uniqueness of the self - adjoint extensions of such operators for many different types of naked spacetime singularities including the r - n naked singularity . ) is there any intuitive physical condition that we can further impose on the fields , that will uniquely select the appropriate green s function ? at least to get the geometrical optics continuous extension of the black hole case one can impose the condition that nothing falls in or out of the singularity . this means there is neither absorption nor superradiation in the scattering and the s - matrix of the k - g field is a unitary operator . what does this condition mean ? the conserved current 4-vector for the complex klein - gordon field is given by @xmath34 let us integrate the 4-current along a cyllindrical hypersurface given by @xmath35 and @xmath36 $ ] . then since is a conserved current it holds that @xmath37 here @xmath38 is the integral along a hypersurface given by the interior of the cylinder at the constant time ( @xmath39 or @xmath40 ) , with a future orientated surface normal vector . @xmath41 is a cylindrical hypersurface given by @xmath35 and @xmath42 is an induced density given as @xmath43 . furthermore @xmath44 is a normal vector to @xmath41 given as @xmath45 . since we want @xmath46 to remain constant with respect to time as we take the limit @xmath47 ( nothing flows out or into the singularity ) , the equation reduces to the following @xmath48 = 0.\ ] ] let us analyse the equation in decomposition into spherical harmonics , hence @xmath49 then we can rewrite the equation as @xmath50_{|r = r_{0 } } = 0,\end{aligned}\ ] ] for arbitrary @xmath51 . since @xmath52 as @xmath47 , in order to fulfill the equation we impose for every @xmath3 and every @xmath53 @xmath54_{|r=0}=0.\ ] ] but this means that the functions @xmath12 should be always constrained either by the condition @xmath55 or @xmath56 . ( one might argue that it will be enough to claim that the fields and their first @xmath21 derivatives should be real at zero , but that does not put any general constraint on the normal modes . ) now we obtain the function @xmath12 from the normal modes as @xmath57,\ ] ] and take the condition @xmath55 for arbitrary time @xmath53 . this translates to @xmath58=0.\ ] ] but since both @xmath59 ( see ) and fourier - like transform given by should not map non - zero functions to zero , one obtains the condition @xmath60 . this means we compose the relevant wave packet only from the following modes : @xmath61 the same line of reasoning applies to the condition @xmath56 and the wave - packets that fulfill such condition must be composed entirely from modes given as ( @xmath62 ) : @xmath63 for @xmath64 , we have the coefficient @xmath65 and the solution @xmath66 behaves as a constant for @xmath67 ( with the value set to 1 ) , which means that the condition does not make strictly sense ; in such case the modes are given simply by the function @xmath66 . this shows that the condition of s - matrix being a unitary operator gives additional constraints on the green s function . the previous conditions mean that we shall consider only a linear space of wave - packets formed purely from modes that vanish at zero , or a linear space of wave - packets formed purely from modes whose first @xmath21 derivatives vanish at zero . ( of course one can not superpose wave - packets formed from modes having the vanishing radial derivative at 0 with wave packets formed from modes that are vanishing at 0 , since the 4-current does not linearly depend on a wave - function . ) so let us pick one of those two types of modes , fix the particular green s function and employ the following reasoning : since @xmath68 is conserved and the coefficients in the equation are everywhere outside 0 a smooth function of @xmath21 , the function @xmath69 , arising from compactly supported initial data could be unbounded in time only in the case the wave - packet becomes concentrated around the singularity and slowly growing asymptotically with time into delta function @xmath70 but this scenario is prevented by the boundary condition that `` nothing flows in or out of the singularity at any time '' . this means we expect the solution to be bounded with respect to both time and space and all the reasoning based on the assumption of the existence of laplace transform is justified . further in the text we will employ the field vanishing condition at 0 . ( this boundary condition at the singularity corresponds to what is known as friedrich s extension of a symmetric operator . ) thus the quasinormal modes will relate to the scattering problem following from the time evolution determined by the boundary condition @xmath71 . but everything that we will do in the following text can be repeated for any other meaningful normal modes boundary condition giving another time evolution of the fields . the general dependence of the quasi - normal modes ( characteristic oscillations ) on such a boundary condition ( given by some linear combination @xmath72 ) describes the way the quasi - normal modes depend on the hair " of the singularity . since the quasi - normal modes are the ones that carry astrophysical information about the astrophysical sources , their dependence on the hair " of the singularity ( in some simplifying sense given by the complex parameter @xmath73 ) has a potential astrophysical importance ( of course in case naked singularities have any astrophysical importance ) . using @xmath74 defined as @xmath75 and @xmath76 the tortoise coordinate given by the condition : @xmath77 one can rewrite the equation into the following form @xmath78 with @xmath79 f\big(r(x)\big).\ ] ] and , for the normal modes @xmath80 , we can write @xmath81\phi_{l}(x)=0.\ ] ] if @xmath82 , we can see that @xmath8 given by eq . ( [ tortoise ] ) has no zeros for real arguments , but eq . ( [ x ] ) can still be integrated to give @xmath83 where @xmath84 is an integration constant . we remark here @xmath85 as @xmath86 , but for @xmath25 we have @xmath87 as can be seen in fig . ( [ fig : rstar ] ) . further in the text we take the tortoise coordinate with the boundary condition @xmath88 . it means that with respect to the usual ( @xmath89 ) tortoise coordinate such tortoise coordinate is shifted to the origin by the transformation @xmath90 . ) for a spacetime with @xmath91 and different values of @xmath92 , taking @xmath93 . ] the potential has for the ratio @xmath94 less than approximately @xmath95 and the relevant @xmath76 ( in the naked singularity case the domain of @xmath76 is constrained ) 3 extrema , one smaller `` outer '' maximum , one dominant `` inner '' maximum and minimum in the potential valley between them . ( for @xmath25 the function @xmath96 . ) for @xmath94 more than approximately @xmath95 the potential has only one maximum ( thus only one peak ) . these features of the potential can be seen in the figure [ fig : pot ] . moreover , in the case @xmath94 less than approximately @xmath95 and @xmath97 , the inner maximum becomes completely dominant , making the outer peak negligible as compared to the size of the inner peak . for the outer peak and for @xmath97 the terms @xmath98 represent only a small correction ( as compared to the term proportional to @xmath99 ) and the second peak will vanish in this approximation at @xmath100 . ( this is because the outer peak lies always at @xmath101 . it is also quite obvious that the inner peak exists only due to the fact that the terms @xmath98 become dominant for @xmath21 close to 0 . ) the features described in this paragraph can be observed in the figure [ fig : pot_l ] . further in the text we will use the following notation related to the potential parameters : by @xmath102 we mean the heights of the two peaks . ( @xmath103 is the first larger peak , @xmath104 the second smaller peak . ) by @xmath105 we call the curvature of the first peak and by @xmath106 the curvature of the second peak . furthermore , by @xmath107 we mean the point of the location of the top of the first peak and by @xmath108 the point of the location of the second peak . many of these parameters can be exactly calculated in the @xmath97 limit . the results of these calculations are given in the appendix [ b ] . given by eq . ( [ potential ] ) with @xmath109 , @xmath91 for @xmath110 and 0.56 . ( the curves from left to right correspond to the increase of charge . ) note that the dashed part of the potential ( for @xmath111 and 0.5 ) is inside the black hole horizon . ] first let us have a look at the geometrical optics ( eikonal ) limit of the klein - gordon equation , which is valid for @xmath97 . in particular what does the geometrical optics limit tell us about the fundamental mode ? the effective potential for the massless particle is @xmath112 the effective potential goes to plus infinity as @xmath25 and its extremes are given as @xmath113 this means the potential has one minimum and one maximum for @xmath114 for @xmath115 there are no extrema . the maximum represents the unstable orbit of a massless particle , and for the black hole case it is only the maximum that is relevant ( because it is located above the horizon ) . for the naked singularity with @xmath116 there exists also a stable orbit of a massless particle . however the fundamental mode in the geometrical optics limit is related to the unstable orbit as @xmath117 ( for a very good paper that discusses this topic see @xcite . ) note that here @xmath118 gives the unstable orbit decay rate and furthermore holds the following @xmath119 ( again @xmath76 is the tortoise coordinate . ) particularly for the reissner - nordstrm black hole / naked singularity ( with @xmath120 ) equation can be expressed as : @xmath121 with @xmath122 , given as before . given by eq . ( [ potential ] ) with @xmath91 for @xmath123 ( @xmath124 ) and different values of @xmath3 ( same caption as on the right plot ) . note in this case the existence of the secondary peak , reminiscent of the black hole potential . below : the same , but this time for @xmath125 ( @xmath126 ) . there is no secondary peak in this case.,title="fig : " ] given by eq . ( [ potential ] ) with @xmath91 for @xmath123 ( @xmath124 ) and different values of @xmath3 ( same caption as on the right plot ) . note in this case the existence of the secondary peak , reminiscent of the black hole potential . below : the same , but this time for @xmath125 ( @xmath126 ) . there is no secondary peak in this case.,title="fig : " ] from the black hole qnm - s boundary conditions one can see that such massless particles decaying from their unstable orbit and eventually escaping either to the b - h horizon , or the infinity represent the fundamental qnm - s in the @xmath97 ( geometrical ) limit . one can observe ( see the discussion later ) that the same holds in the naked singularity case . in the @xmath97 case the fundamental mode of the massless perturbations is described by the picture of a particle decaying from its unstable `` photon '' orbit . ( note that in the naked singularity case the stable photon orbit makes the particle decaying inwards oscillate around the stable orbit . ) this should suggest that in the geometrical optics limit 1 . one shall expect continuity of the fundamental mode ( as a function of @xmath127 ) when turning from the black hole to the naked singularity case , 2 . for @xmath128 there is a clear indication that there do not exist low damped quasi - normal modes . the way how to understand the continuity ( for @xmath97 ) of the fundamental mode as one transfers from the black hole case to the naked singularity case could be the following : consider the partial s - matrix @xmath129 . within the s - matrix one can look for poles on the complex plane : obviously any of the poles must be necessarily non - real . for the relativistic time dependence convention chosen as @xmath130 the poles with @xmath131 must be purely imaginary ( as a result of the given hamiltonian being a symmetric operator ) and they represent bound states with the non - relativistic energies given as @xmath132 . the poles with @xmath133 represent quasinormal frequencies . some of the quasinormal frequencies correspond to the resonances in the phase factor related to the partial s - matrix , and the wave packets formed out of the resonance energies represent quantum particles tunnelling out of the potential valley ; for the non - relativistic quantum particles ( for example ) with the decay rate given as @xmath134 . now in the naked singularity case the fundamental mode is represented by a resonance that has a low enough energy ( at the level of the smaller peak ) and gets trapped in the potential valley in between the peaks for some time and then radiated away . the larger peak behind the smaller peak is hugely dominant ( large @xmath3 ) and effectively acts to the wave packet as an infinite barrier . there are probably much higher resonance energies related to higher qnm frequency overtones that are determined by the higher peak , but left unaffected by the details of the small peak . so there are two different effective regimes for the quasinormal modes , in the first regime they are sensitive only to the details of the smaller peak , in the second regime they are sensitive only to the details of the larger peak . this splitting disappears in the case of small @xmath3-s since the two peaks are of comparable heights and effectively interfere . let us finish this section with one more remark : in the black hole case one can relate the fundamental mode to the peak of the potential ( independently of @xmath97 ) following the way of thinking in @xcite . the ratio of the amplitudes of the reflected and transmitted waves @xmath135 is for quasinormal modes 1 . the same happens for real @xmath136-s , when the energies @xmath137 are close ( `` almost at '' ) to the peak of the potential . so for `` almost '' real @xmath137 ( fundamental mode ) the continuity of the ratio of reflected and transmitted amplitudes analytically continued to the complex plane suggests that the real part of the frequency should be near the square root of the peak of the potential . ( this intuition is then to some extend confirmed also semi - analytically , such as by the ferrari and mashhoon approach using the poeschl - teller potential @xcite , or by using the inverted harmonic oscillator potential . ) in the case of a naked singularity life is not so easy ( in the reduced 1d problem we have no transmitted amplitudes ) , so this simple logic fails . ( one can say luckily it fails as if it remained valid one would also for @xmath97 expect a discontinuity in the fundamental mode as one jumps from the black hole to the naked singularity . this is because the second , dominant peak suddenly appears in the domain of the @xmath76 coordinate , due to the fact that the domain of @xmath76 discontinuously jumps when passing from the black hole to the naked singularity case . ) certainly one can say that any of the quick fits , such as were done in the black hole case do not lead to anything close to the numerical data obtained in this paper . in this section we want to suggest some ways to model the naked singularity scattering ( with a general time evolution ) by solutions of the equation with @xmath138 replaced by analytically solvable potentials . the reasons are the following : first , this section serves as a basis for the analysis of the @xmath97 cases provided in the section [ largelsection ] . second , it demonstrates that the naked singularity scattering is _ in principle _ treatable via analytical approximations and thus shows the general power of analytical techniques . ( also the approximations obtained here might produce some future results for the quasinormal frequencies . for some further calculations see the appendix [ a ] . furthermore , for a very nice overview of the results for the quasi - normal frequencies and related transmission resonances of the analytically tractable potentials see @xcite . ) third , it might bring more insights into the physics obtained through the exact scalar field potential . the key point is to split the scalar field potential into different domains and approximate it on each of those domains ( or directly the solution of the equation ) by a different , analytically solvable potential . then one has to impose the standard procedure : the logarithmic derivatives of the solutions on different domains must be glued on the domain s boundary . we might add here that since the analytically solvable potentials are typically fitted by the parameters of the peaks one expects such approximation to work for the modes that are _ not _ too damped ( the _ low _ damped qnm modes ) . the approximation _ can not _ be taken too seriously for the _ highly _ damped modes . let us suggest modeling all the solutions of the relevant cases as follows : for this case it might be interesting to consider the morse potential . ( for the definition and origins of all the potentials used in this section see also @xcite . ) on the other hand the infinite valley behind the one and only peak may not be well modelled by only a finite valley that is behind the peak of the morse potential . thus we consider the solution in the region ` behind' morse potential region to be given as for @xmath139 and the logarithmic derivative gluing condition originates from : 1 . @xmath140 : use function given as ( for the details see equation ): @xmath141 2 . @xmath142 : use the appropriate solution of the equation with the morse potential : @xmath143 here @xmath144 is a point most conveniently chosen where the morse potential is 0 ( for any arbitrary @xmath3 there is such point for @xmath145 ) . this means @xmath144 can be analytically given as @xmath146 let us make here one remark : instead of approximating the original scalar field equation for @xmath67 one can consider taking the equation in the @xmath25 approximation . this leads to the following : @xmath147 the equation has solutions @xmath148 . the solution such that fulfills @xmath149 must have @xmath150 . this approximation of the solution we think to be less exact as the approximation given as , as it is in fact effectively only the first non - zero term of the power series expansion of taken at 0 . ( for such reasons we decided to use in our analytical approximations the approximate solution . ) let us consider now the limit of large @xmath3 . in this case , since the peak already lies close to 0 and the change in the highest power of inverted @xmath76 becomes more and more dominant ( with higher @xmath3-s ) , one might simplify the approximation given in case [ one ] , by using the following potential : 1 . @xmath151 : use function given as 2 . @xmath152 : use the appropriate solution of the equation with the potential ( poeschl - teller potential ) : @xmath153.\ ] ] here we also used the poeschl - teller potential instead of the morse potential , because the logarithmic derivative gluing condition is much easier to solve for the solutions of the poeschl - teller than for the solutions of the morse potential . the expense of the simplicity is that the approximation by such potential might be slightly less exact than by using the morse potential , but one still expects it to be accurate enough . if one does not want to neglect the second smaller peak ( so one is interested in resonances related to the valley between the two peaks ) , the analytically treatable potential describing all the resonances ( and also higher damped qnm frequencies ) could be : 1 . @xmath140 : use function given as , 2 . @xmath155 : solutions of the equation with the morse potential @xmath156 3 . @xmath157 : solution of the equation with the poeschl - teller potential : @xmath158.\ ] ] here @xmath144 is best chosen as in the case [ one ] and @xmath159 is chosen to be such that the resulting potential is continuous . generally @xmath159 has to be obtained numerically , but for larger @xmath3-s ( certainly @xmath160 is more than enough , as we checked ) , the curvature of the second peak is very small comparing to the curvature of the first peak and also the curvature of the first peak becomes ( for large @xmath3-s ) very large comparing to the scales of the potential , so that one can calculate ( for @xmath160 with a good approximation at least to the 6 decimal places ) @xmath159 just by 1 . taking the poeschl - teller potential to be constant and given by the hight of the outer peak 2 . simplifying the morse potential by the following approximation : @xmath161 the resulting formula then becomes @xmath162 then one can easily show that @xmath163 as @xmath164 . now consider what happens with the resonances related to the valley between the two peaks in the large @xmath3 limit . in such case the dominant peak grows to infinity as compared to the smaller peak , the infinite valley behind the dominant peak shrinks to 0 . also @xmath165 and the valley between the two peaks becomes flat as compared to the difference between the height of the first peak and the value of the potential at the bottom of the valley . if the resonances related to the smaller peak locate close to the top of the peak ( they should since the first peak grows with @xmath3 an makes the tunnelling harder ) , then one might effectively approximate the case [ three ] by the poeschl - teller potential : @xmath158,\ ] ] for @xmath166 . ( one can also derive this approximation straight from the logarithmic derivative gluing condition in the appendix [ section ] by taking the @xmath164 limit . ) in this section we want to analytically confirm the results obtained directly from the eikonal limit . the case [ four ] from the previous section is easily solvable . the solution in the region @xmath168 must fulfil the outgoing radiation condition . ( to be exact after analytically extending the solution to the complex plane it must give asymptotically the outgoing waves on the line @xmath169 . ) such solution is given as follows @xmath170 where @xmath171 is the standard hypergeometric function ( of the type 2 - 1 ) and @xmath172 the boundary condition at 0 gives @xmath173 , leading to @xmath174 unless @xmath94 is not too close to the upper limit given for @xmath3 large as 9/8 , one can reliably approximate @xmath175 and the condition turns to be : @xmath176 now can be rewritten through gamma functions and reduces to the simple problem of finding poles of the product of gamma functions @xmath177\gamma\left[-\frac{i\omega+i\sqrt{v_{2}}}{\alpha_{2}}+\frac{1}{2}\right]=0.\ ] ] the poles are located at @xmath178 for @xmath179 being a natural number and this gives @xmath180 this is precisely the formula for the lowest damped black hole qnm frequencies . this means in the @xmath97 limit we see ( at least for the lowest modes ) a continuous transition from the black hole case to the naked singularity case , as expected . ( at least in the case where it holds that @xmath175 , which becomes a less accurate approximation for @xmath94 close to @xmath95 . on the other hand for such ratios of @xmath94 the peak is almost vanished and one can assume that the decaying circular orbit eikonal picture has already quite limited sense . ) all this means that in the regime @xmath97 the lowest qnm frequencies should be located close to the first peak of the potential and are ( up to certain @xmath179 ) given by the formula . they represent resonances such that they are insensitive to the details of the larger peak as in their case the larger peak can be already seen to act effectively as an infinite potential barrier . hence we can conclude that in the @xmath97 case we are able to match the geometrical ( eikonal ) limit of the original problem considered . take the case [ two ] from the previous section . on the left hand side of the logarithmic derivative gluing condition we need a logarithmic derivative of the function : @xmath181 on the right hand side of the logarithmic derivative gluing condition we need the logarithmic derivative of the solution on the interval @xmath182 $ ] . it is given as in , only the parameters are related to the first and only peak in this case . then the logarithmic derivative gluing condition leads to the following : @xmath183.\ ] ] if one takes the @xmath164 limit then the equation becomes : @xmath184.\ ] ] here @xmath185 is defined as @xmath186 . for not highly damped frequencies in the @xmath97 limit it must hold that @xmath187 . but by plotting the right side of the equation minus the left side of the equation in mathematica it can be shown that the equation does not have any real solutions . for more than large intervals of values of @xmath185 the absolute value of the left hand side of the equation minus the right hand side of the equation always seems to grow ( almost ) linearly with respect to @xmath185 from the value approximately given as 0.35 at @xmath188 to infinity . it might not be hard to prove also analytically that does not have real solutions by using many relatively simple properties of gamma functions . this means there are _ no _ low damped modes in the case of @xmath189 and @xmath97 , as expected from the eikonal limit / particle picture . ( it means the damping of the fundamental mode grows to infinity as @xmath164 and grows no less rapidly than @xmath105 , hence no less rapidly than cubically with @xmath3 . ) in @xcite , the quasi - normal modes of a scalar field in an electrically charged vaidya background were studied using a similar numerical setup to what we are going to use here . an interesting investigation done in @xcite tried to determine what would happen with the quasi - normal modes as the time - dependent background approached a naked singularity , but the numerical code used was not suitable for following the field evolution after the extremal configuration ( @xmath190 ) was reached . in our present work , we have a static configuration that describes a naked a singularity in order to study the properties of the quasi - normal modes . our objective in this section is to solve eq . ( [ wave ] ) with potential ( [ potential ] ) numerically , in the case where @xmath191 as described in the last section . to do this , we rewrite eq . ( [ wave ] ) in terms of the light - cone variables @xmath192 and @xmath193 , where @xmath76 corresponds to the tortoise coordinate ( [ x ] ) , as @xmath194 that can be integrated with the boundary conditions @xmath195 where condition ( [ bc1 ] ) is a necessary condition on the field @xmath196 near the origin ( see the discussion on fig . [ fig : phi_r ] below ) and condition ( [ bc2 ] ) defines an `` arbitrary '' relevant initial signal to be propagated . we use the algorithm @xmath197 see where @xmath198 and @xmath199 are the integration steps in @xmath200 and @xmath201 , respectively , and the definitions of @xmath202 etc can be seen on fig . [ fig : grid ] . note that here @xmath203 is the potential ( [ potential ] ) evaluated at the same @xmath21 coordinate as @xmath204 ( and @xmath202 ) . ) with boundary conditions ( [ bc1 ] ) and ( [ bc2 ] ) . the points marked with `` x '' are out of our domain , values of @xmath196 at the positions marked with filled circles are given by the boundary conditions and the values at the empty circles are obtained with the algorithm ( [ algo ] ) . ] as we can see in fig . [ fig : phi_r ] , the boundary conditions ( [ bc1 ] ) and ( [ bc2 ] ) ensure the necessary conditions on the fields @xmath196 and @xmath205 near the center . as we discussed previously in section [ sec2 ] , the physically correct boundary condition for @xmath205 is @xmath206 . from this we must have for @xmath207 that @xmath208 and @xmath209 . in fact , the quantity @xmath210 dx\ ] ] is invariant along the @xmath53-evolution governed by ( [ waveuv ] ) with the boundary conditions @xmath211 . taking into account ( [ x ] ) , the integrand can be expressed near @xmath26 as @xmath212 with the prime denoting the derivative with respect to @xmath21 . any initial condition with finite @xmath213 , as those ones obeying ( [ bc1 ] ) and ( [ bc2 ] ) , is such that @xmath214 for @xmath25 , with @xmath215 , implying that @xmath216 for initial condition with finite e obeying @xmath217 . moreover , we have checked that our boundary conditions for @xmath196 also reproduce the @xmath209 condition , by testing the code with the condition @xmath218 , where @xmath219 is a very small value , and we found no differences in the late time evolution of the fields . with @xmath220 as a function of @xmath21 near the center @xmath26 for a late time @xmath221 , shown here in order to exemplify the effect of conditions ( [ bc1 ] ) and ( [ bc2 ] ) in the numerical integration , for a spacetime with @xmath222 and @xmath123 . below : the same as in the left plot , but this time for the function @xmath223.,title="fig : " ] with @xmath220 as a function of @xmath21 near the center @xmath26 for a late time @xmath221 , shown here in order to exemplify the effect of conditions ( [ bc1 ] ) and ( [ bc2 ] ) in the numerical integration , for a spacetime with @xmath222 and @xmath123 . below : the same as in the left plot , but this time for the function @xmath223.,title="fig : " ] in the left plot of fig . [ fig : field_qm ] we present some typical time evolutions of @xmath196 , for a @xmath220 and different @xmath224 ratios . in the right plot we present the obtained frequencies of the qnms ( fundamental mode ) in the @xmath225 plane . we can see a discontinuity in the frequencies as @xmath226 , as was expected from the discussion of the potential @xmath227 ( see fig . [ fig : pot ] ) . we also point here that we see no significant changes , but rather a smooth behavior as @xmath228 ( @xmath229 in the plot ) . but we see a point of inflection in @xmath230 at @xmath231 , for which we did not find an analytical explanation . we remark here some numerical issues that prevented us from extending the results shown in fig . [ fig : field_qm ] for lower and higher @xmath224 rations . as @xmath226 and @xmath196 becomes less damped , different modes subsist for longer times and and longer evolutions are needed in order to obtain the clear frequency of the fundamental mode . on the other hand , as we increase @xmath224 , @xmath196 is damped so quickly that we can not observe enough oscillation cycles to obtain the frequencies . these are the issues that have limited our results . with @xmath220 at @xmath232 for a spacetime with @xmath233 and different values of @xmath234 . below : frequencies of the fundamental mode with @xmath220 in the @xmath225 plane , parametrized by the @xmath224 ration.,title="fig : " ] with @xmath220 at @xmath232 for a spacetime with @xmath233 and different values of @xmath234 . below : frequencies of the fundamental mode with @xmath220 in the @xmath225 plane , parametrized by the @xmath224 ration.,title="fig : " ] finally , in fig . [ fig : field_l ] we explore how the frequencies of the qnms change with @xmath3 . as usual in black hole scattering problems , we see that the oscillation frequency @xmath230 increases with @xmath3 . but the qualitative behaviour of @xmath235 changes significantly with @xmath224 . in the upper plots ( @xmath236 ) , @xmath237 decreases with @xmath3 , that is , the damping time is longer . in the lower plots ( @xmath2 ) , we have the opposite tendency . this behaviour is connected to the potential @xmath227 shown in fig . [ fig : pot_l ] . it might be also interesting to mention that in case @xmath2 , there is a qualitative similarity in the behaviour of the imaginary part of the frequencies as a function of @xmath3 , between the case when @xmath3 is small and the @xmath3 large limit . for @xmath91 and @xmath123 ( @xmath124 ) . below : same as above , but this time for @xmath91 and @xmath125 ( @xmath126 ) . ] let us explore the following limit , @xmath3 will be fixed and @xmath238 , assuming that there exist an infinite number of qnms . assume further that @xmath239 diverge as @xmath240 . in such a limit , to have a good understanding of what the qnm boundary conditions are , it is convenient to take an analytic continuation of the solution into the complex @xmath76 plane and impose the outgoing wave boundary conditions on the ( stokes ) line @xmath241 . this is for almost imaginary asymptotic frequencies ( @xmath242 as @xmath240 ) the line @xmath243 . the key point is to realize that , in the asymptotic case , @xmath244 holds everywhere apart of a tiny region around @xmath245 , where the potential @xmath246 can be already approximated by @xmath247 ( there is one subtlety : the potential , and also later the solutions , are multi - valued functions in the complex variable , but after specifying a branch cut together with the branch , this does not cause any problems to our analysis . ) so let us make the following statement : the solutions of the equation @xmath248 are _ everywhere _ good approximations to the solutions of the scalar field equation for the asymptotically highly damped modes . ( the region where the approximations @xmath249 and @xmath250 cease to hold is the region where we can already neglect the potential as a whole with respect to the @xmath137 term . ) the equation has a general solution given as @xmath251~,\ ] ] where @xmath252 is the bessel function . the solution for which @xmath253 fulfils the vanishing boundary condition at 0 is given by @xmath254 . our qnm boundary condition says that this solution taken along the stokes line @xmath241 , should give purely outgoing radiation for @xmath255 . but in the @xmath256 limit we know that we can approximate the solution ( via bessel functions approximations in such a limit ) as @xmath257 this linear combination is independent on @xmath136 and does not give the purely outgoing radiation ( @xmath258 ) , which suggests that our qnm boundary conditions can not be fulfilled for asymptotically highly damped modes . this can be taken as an evidence that the asymptotically highly damped modes ( @xmath259 ) do _ not _ exist for the r - n naked singularity and this can be taken as a confirmation that the highly damped qnms could link to the black hole spacetime horizon s properties ( non - existence of the horizon leads to non - existence of the modes in the asymptotic limit ) . one can see also some link to the fact that for the extremal r - n black hole the spacing between the asymptotic frequencies goes to zero as proportional to the surface gravity ( this might make a whole infinite tower of arbitrarily highly damped modes eventually `` collapse '' to a single mode ) . moreover , we assume that the non - existence of the highly damped modes can be proven along the same lines also for another naked singularity spacetimes ( like negative mass schwarzschild spacetime etc @xcite ) . in this paper we analysed the problem of the scalar field scattering on a r - n naked singularity background from the point of view of quasi - normal modes . the evolution on the r - n naked singularity is non - unique unless one specifies an additional boundary condition representing a `` hair '' of the singularity . the quasi - normal modes then carry information about the `` hair '' . we applied a particular boundary condition , that nothing comes out , or in from the singularity and analysed analytically , as well as numerically , the characteristic oscillations of the scalar field perturbations ( low damped quasi - normal modes ) . we analysed the eikonal @xmath97 case via the analytical approach confirming the intuition obtained through the massless particle viewpoint , and showed that an approach based on analytical approximations can be useful also for the small @xmath3 wave mode numbers . for the small @xmath3-s we calculated the frequencies numerically via the characteristic integration method . we also suggested arguments showing that the asymptotically highly damped modes ( limit @xmath3 fixed and @xmath260 ) do _ not _ exist in case of r - n naked singularity . this might confirm the intuition one has about such modes from the black hole physics ( and from the presently popular conjectures @xcite ) . the basic results can be summarized as follows : for the low modes and the large @xmath3 there is a continuous transition in the low damped qnm modes between the r - n black hole and the r - n naked singularity . however , when the ratio @xmath94 becomes larger than approximately @xmath95 then the picture becomes significantly different and the low damped modes do _ not _ exist for large @xmath3-s . ( this is a very different picture from the bh based intuition . ) for the small @xmath3 numbers the modes face a discontinuous transition when transiting from the black hole to the naked singularity . furthermore , the @xmath3 dependence @xmath237 ( for small @xmath3 ) changes as @xmath94 becomes larger than approximately @xmath95 : @xmath237 decreases for @xmath261 and increases for @xmath262 . it might be interesting to notice that for @xmath262 the increase of @xmath263 as a function of @xmath3 ( for small @xmath3-s ) matches the behaviour of @xmath263 for large @xmath3-s . in the case of large @xmath3-s and @xmath262 we have shown that @xmath263 of the fundamental mode grows at least cubically with @xmath3 and thus , as we already mentioned , the low damped modes do not exist . for the asymptotically highly damped modes our results seem to suggest that they do not exist , which means that the imaginary parts of the frequencies are bounded . the authors are grateful to fapesp , cnpq , and the max planck society for the financial support . in this appendix section we provide a list of some of the quantities ( also some of their derivations ) that occur in the calculations relevant to the problem analysed in this paper . write @xmath227 in a convenient form : @xmath264 with @xmath265,\\ c&\equiv & q^{2}l(l+1)-(2m)^{2}-2q^{2},\\ d&\equiv & 6mq^{2},\\ e&\equiv & -2q^{4}.\end{aligned}\ ] ] now let us calculate the second derivative of the potential with respect to the tortoise coordinate : @xmath266 at @xmath267 the derivative @xmath268 is zero , then necessarily @xmath269_{|x_{max}}\ ] ] and since @xmath270 we obtain @xmath271 for the global extremum condition holds : @xmath272 this equation can have maximally 4 roots . from it can be easily seen that for @xmath24 becomes for any @xmath273 positive ( dominant term is @xmath274 ) and goes to 0 . in case of naked singularity for @xmath25 the potential goes always to @xmath275 ( dominant term is @xmath276 ) , in case of black hole for @xmath277 it always goes to 0 ( dominant term is @xmath8 ) . this means , together with , that for any @xmath278 ( black hole or not ) there are always either two local maxima and one local minimum , or one local maximum without local minima . for @xmath97 the following approximations hold : for large @xmath3 the maximum of the inner peak can be approximately found analytically and behaves as : @xmath279 also holds the following : for @xmath280 @xmath281 the parameters @xmath282 turn in the large @xmath3 limit to be : @xmath283^{3/2}}{|q|},\ ] ] @xmath284^{3/2}}{3^{3/2 } |q|}.\ ] ] this implies also the following results : for @xmath164 @xmath285 write the two linearly independent solutions on the domain @xmath286 in the form convenient for the logarithmic derivative gluing condition : @xmath287 . if the domain extends to the infinity and @xmath288 is taken to be the asymptotically incoming wave solution and @xmath289 the asymptotically outgoing wave solution , then @xmath290 is the s - matrix ( ratio of the coefficients of the outgoing and incoming waves ) . anyway , the logarithmic derivative gluing condition at the boundary of the regions @xmath286 and @xmath291 turns to be : this is the condition connecting the @xmath293 coefficient with @xmath294 coefficient through the values of the solutions and their derivatives at the boundary of the regions . on one side consider asymptotically free - wave region , whereas on the other side consider infinite potential barrier - like boundary condition for @xmath295 . in such case one can through equations of the type connect ( after finite number of steps ) the s - matrix of the external ( asymptotic ) region with the boundary condition at the origin ( where we can conveniently put @xmath296 to be zero ) . then one gets the s - matrix expressed through a function of the boundary values ( at all the boundaries ) of the solutions and their derivatives . this will be algebraically more difficult condition than the one coming just from posing the outgoing wave condition in the external region , as we did in the previous sections ( case of large @xmath3 and @xmath94 larger than approximately 9/8 ) . the reason for posing the more complicated condition of the type is that it might be easier ( and also brings more insight ) to look for a real energy resonances ( by plotting phase change and looking for rapid phase shifts , hence rapid change of the logarithm of the s - matrix ) , than to try to numerically solve the more simple outgoing wave condition ( with the help of appropriate computer software ) . such resonances should then 1 to 1 correspond to the low damped quasi - normal frequencies . let us apply these ideas and try to calculate the expressions for the partial s - matrix . this is the case [ one ] . we are gluing only two regions . the solution on the left side of the boundary @xmath298 , such that fulfils @xmath299 , is again given as . on the right side ( the morse potential side ) the solution is given as : by @xmath301 we mean asymptotically outgoing and incoming wave solutions in the sense that @xmath302 and @xmath303 exactly . furthermore @xmath304 is being the kummer ( confluent ) hypergeometric function and @xmath305 also @xmath306 is the s - matrix . the logarithmic derivatives gluing condition at @xmath307 is then @xmath308= \frac{\alpha_{1}z(a_{1})}{2}+i\omega+\frac{s'_{1}}{s'_{2}}\frac{m\left(s'_{1}+1 , s'_{2}+1,z(a_{1})\right)}{m\left(s'_{1 } , s'_{2},z(a_{1})\right)}.\ ] ] considering that the following holds : @xmath309 one can rewrite into : @xmath310= i2\sqrt{v_{1}}+i\omega+\frac{s'_{1}}{s'_{2}}\frac{m\left(s'_{1}+1 , s'_{2}+1,\frac{i4\sqrt{v_{1}}}{\alpha_{1}}\right)}{m\left(s'_{1 } , s'_{2},\frac{i4\sqrt{v_{1}}}{\alpha_{1}}\right)}.\ ] ] the partial s - matrix can be obtained from ( @xmath311 ) as : @xmath312 since for real @xmath136 holds that @xmath313 and @xmath298 is a real valued function , one can immediately observe that ( for @xmath314 ) @xmath315 , hence partial s - matrix is unitary . @xmath316-\left[i2\sqrt{v_{1}}-i\omega+\frac{s_{1}}{s_{2}}\frac{m\left(s_{1}+1 , s_{2}+1,\frac{i4\sqrt{v_{1}}}{\alpha_{1}}\right)}{m\left(s_{1 } , s_{2},\frac{i4\sqrt{v_{1}}}{\alpha_{1}}\right)}\right]}{f(a_{1})\left[\frac{1}{a_{1}}+\frac{\beta_{1}e^{\beta_{1}a_{1}}-\beta_{2}e^{\beta_{2}a_{1}}}{e^{\beta_{1}a_{1}}-e^{\beta_{2}a_{1}}}\right]-\left[i2\sqrt{v_{1}}+i\omega+\frac{s'_{1}}{s'_{2}}\frac{m\left(s'_{1}+1 , s'_{2}+1,\frac{i4\sqrt{v_{1}}}{\alpha_{1}}\right)}{m\left(s'_{1 } , s'_{2},\frac{i4\sqrt{v_{1}}}{\alpha_{1}}\right)}\right ] } \nonumber\end{aligned}\ ] ] the two peak case is a more complicated problem . anyway , one can write the logarithmic derivative gluing condition as @xmath317^{-1}\}}{f_{21}\{g_{1},g_{2},g_{3 } , [ 1+\exp(2\alpha_{2}(a_{2}-x_{2max}))]^{-1}\}}. \nonumber\end{aligned}\ ] ] here @xmath318 is the s - matrix from the case [ one ] and is given by the formula . the s - matrix related to the case [ three ] ( call it @xmath319 ) is related to @xmath318 through the condition . in the gluing formula @xmath319 stands for @xmath293 and @xmath318 for @xmath294 . also @xmath320 are the solutions of the poeschl - teller case @xmath321^{-1}\right],\ ] ] and @xmath322^{-1}\right].\ ] ] @xmath323 are the morse potential solutions given as . we see that if there is a regime in which the modes do not feel the smaller peak , then they are given by the poles of @xmath318 and must be in the same time poles of @xmath319 . the formula for @xmath319 in this case ( for frequencies that represent poles of @xmath318 ) simplifies to @xmath324_{|_{a_{i}}}.\ ] ] there is a possibility that for the case of the low modes in the two peak model one might be able to neglect the infinite depth of the valley between the infinite barrier and the morse peak . in such case one can take a simpler model where the morse potential directly follows after the infinite barrier . the @xmath318 term appearing in the formula will be in such case given by a much more simple expression : @xmath325

in recent years , it has become widely accepted that electron - phonon ( ep ) interaction plays a crucial role in systems with strong electronic correlations , such as high - temperature superconducting cuprates@xcite or colossal magnetoresistive manganites.@xcite a large amount of the available data on optical and transport properties has been interpreted in terms of polaronic states . in view of the complex underlying physics , however , the application of single - polaron theories seems to be problematic . on the other hand , a reliable theory of strongly coupled ep systems with finite carrier densities is not available yet . while limiting cases , such as weak coupling or large phonon frequencies compared to the electronic bandwidth can to some degree be described theoretically , especially the adiabatic regime of small phonon frequencies at intermediate - to - strong ep coupling constitutes a long - standing , open problem . since analytical results can only be obtained at the expense of uncontrolled approximations , progress toward an understanding of many - polaron systems requires unbiased numerical methods . owing to the availability of high performance computers , the latter can yield insight into the properties of basic microscopic models such as the spinless holstein model considered here . nevertheless , apart from work on the half - filled band case , where the physics is dominated by a cross over from a luttinger liquid to a peierls state , the many - polaron problem has received little attention in the past . motivated by this situation , in the present work , we address the important issue of the effect of carrier density on the character of the quasiparticles of the system . while for very strong couplings no significant changes are expected due to the existence of rather independent small ( self - trapped ) polarons with negligible residual interaction , a density - driven cross over from a state with large polarons to a metal with weakly dressed electrons may occur in the intermediate coupling regime . this problem has recently been investigated experimentally in terms of optical measurements on @xmath0 films.@xcite while small polarons have been identified as the charge carriers in the low - temperature metallic phase of the ca doped system , the weaker electron - lattice coupling in the sr doped system gives rise to large - polaron - like behavior of the optical conductivity . the effect of the finite carrier density on such extended polaronic quasiparticles has been analyzed@xcite using a recently developed weak - coupling theory.@xcite for the case of the holstein model , the abovementioned density - driven transition from small to large polarons and finally to weakly ep - dressed electrons is expected to be possible only in one dimension ( 1d ) , where large polarons extending over more than one lattice site exist at weak coupling . existing work indicates that in higher dimensions , the electrons remain quasifree at weak coupling , and become self - trapped above a critical value of the ep coupling strength.@xcite by contrast , approximate variational calculations suggest that large polarons may also exist in @xmath1.@xcite the situation is different for frhlich - type models@xcite with long - range ep interaction , in which large polaron states exist even for strong coupling and in higher dimensions . here , we make use of quantum monte carlo ( qmc ) and exact diagonalization ( ed ) methods , and calculate the photoemission spectra for finite clusters at zero and finite temperature . due to the limitations of existing algorithms , we develop a new grand - canonical qmc approach which is free of autocorrelations and therefore enables us to study small phonon frequencies and low temperatures . the organization of this paper is as follows . in sec . [ sec : model ] , we present the model and discuss previous work , while sec . [ sec : methods ] contains details about the methods we employed . section [ sec : results ] gives a discussion of our numerical data , and finally , we conclude in sec . [ sec : conclusions ] . the new qmc technique is described in the appendix . the spinless holstein model is defined by the hamiltonian @xmath2 where @xmath3 ( @xmath4 ) creates ( annihilates ) a spinless fermion at lattice site @xmath5 , @xmath6 ( @xmath7 ) creates ( annihilates ) a phonon of energy @xmath8 ( @xmath9 ) at site @xmath5 , and @xmath10 . while the first term describes hopping processes between neighboring lattice sites @xmath11 , the second term corresponds to the elastic and kinetic energy of the phonons . finally , the last term in eq . ( [ eq : model : h ] ) represents a local coupling of the lattice displacement to the electron occupation @xmath12 . the parameters of the model are the hopping integral @xmath13 , the phonon energy @xmath8 , and the coupling constant @xmath14 . the latter is related to the polaron binding energy @xmath15 via @xmath16 . by defining the dimensionless quantities @xmath17 , and @xmath18 , we are left with two independent parameters . in the sequel , we express all energies in units of @xmath13 . the lattice constant is taken to be unity , and periodic boundary conditions are applied . the holstein hamiltonian ( [ eq : model : h ] ) represents a generic model for a many - polaron system . due to the restriction to spinless fermions , on - site bipolaron formation is suppressed.@xcite quantum fluctuations of the lattice enable the electrons to hop even in the strong - coupling regime , while in models with classical phonons translation symmetry is broken , so that no itinerant polaron states exist . along this line , the half - filled spinless holstein model has been studied in one dimension using qmc,@xcite ed,@xcite dmrg@xcite and variational methods.@xcite similar to the spinfull model at half filling , the spinless model with one electron per two lattice sites displays a quantum phase transition from a metallic state to a peierls insulating state with charge - density - wave ( cdw ) order at @xmath19 . while the insulating cdw state of the spinless model is stable for any @xmath20 in the adiabatic limit @xmath21 , it can be destroyed by quantum phonon fluctuations for @xmath22.@xcite capone _ et al._@xcite used ed to study the cross over from a single polaron to a many - electron system away from half filling . although their work is actually for the holstein - hubbard model , results are very similar to the spinless holstein model in the limit of large @xmath23 considered in ref . . the authors found that the critical coupling for polaron formation is unaffected by the electron density in the antiadiabatic regime . finally , we would like to point out previous analytical efforts to take into account squeezing and correlation effects in many - polaron systems.@xcite in this work we use qmc , and ed in combination with the kernel polynomial method ( kpm ) , to calculate , in particular , single - particle spectra . these two approaches are characterized by different advantages and shortcomings . ed in the form used here yields exact spectra at @xmath19 with extremely high energy resolution . however , even for relatively small clusters of ten sites the size used here this method requires parallel supercomputers to handle the enormous hilbert space ( matrix dimension @xmath24@xmath25 ) . consequently , the momentum resolution , determined by the cluster size , is rather crude . in contrast , the grand - canonical qmc simulations can be performed on personal computers , and for larger clusters . the drawback is that ( a ) the method is limited to finite temperatures , and ( b ) the calculation of spectral properties requires the analytic continuation from imaginary time to real frequencies , which is an ill - conditioned problem . therefore , regularization techniques such as the maximum entropy method ( mem ) used here have to be applied , which offer only limited energy resolution and may introduce uncontrolled errors . for this reason , we will also present a direct comparison of ed and qmc results . in applying existing qmc methods for the holstein model to the parameter regime considered here , one faces extremely long autocorrelations times , as well as large statistical errors.@xcite this makes it very difficult to obtain accurate results for spectral properties , due to the use of the mem . to overcome these problems , the algorithm presented in the appendix is based on the canonical lang - firsov transformation.@xcite the latter explicitly contains the shift of the lattice equilibrium position in the presence of an electron , which is favorable for the qmc simulations . moreover , the absence of a direct coupling between the electron density and the lattice displacements in the transformed hamiltonian ( [ eq : app : hspinless ] ) permits exact , uncorrelated sampling of the phonon degrees of freedom ( momenta ) in terms of principal components . here we have mainly used a value of @xmath26 for the inverse temperature @xmath27 . for the critical parameters studied below , accurate simulations at even lower temperatures become very demanding due to a minus - sign problem ( sec . [ sec : app : sign_results ] ) . nevertheless , we shall present results for the density of states at @xmath28 and half filling but for a reduced cluster size . note that already for @xmath26 we have @xmath29 for the values @xmath30 used in the sequel , so that direct thermal excitation of phonons is expected to be rather unimportant . for the trotter decomposition , we chose @xmath31 ( see appendix ) for static observables . owing to the larger numerical effort associated with the calculation of dynamical properties , for the latter we use @xmath32the larger error being small compared to the uncertainties introduced by the mem inversion . the trotter error in the holstein model becomes important mainly for large phonon frequencies and very strong ep interaction.@xcite since we do not consider this regime here , the abovementioned values of @xmath33 yield satisfactorily small systematic errors . away from half filling , the electron density has to be adjusted to the values of interest by varying the chemical potential @xmath34 . due to the computational effort associated with this trial - and - error procedure , the actual band filling @xmath35 , say , is usually only very close to a desired value , e.g. , @xmath36 , but not exactly the same . furthermore , within the qmc simulations , the filling can only be determined up to a statistical error @xmath37 , similar to other observables . for the results presented here , the relative deviation @xmath38 of the actual value in the qmc simulations from the value @xmath39 reported is always smaller than @xmath40 . the same is true for the relative statistical error @xmath41 . to perform the continuation of the one - electron green function from imaginary time to real frequencies , we use the mem ( for a review see , e.g. , ref . ) . the algorithm employed here incorporates an improved evidence approximation,@xcite which is less susceptible to overfitting resulting in peaks whose position , weights and even existence are not clear than the conventional evidence approximation.@xcite on the other hand , groups of narrowly spaced peaks are represented by an envelope if the accuracy of the qmc data is not sufficient . the proposal of an alternative evidence approximation leading in general to stronger regularization and , consequently , lower energy resolution than the conventional approximation may not seem advantageous at first . however , considering the great variety of different spectra compatible with the qmc data within the error bars due to the extremely ill - conditioned analytical continuation problem , we argue that it rather corresponds to choosing the `` most uninformative '' solution in the spirit of the maximum entropy principle . in fact , we found that the alternative evidence approximation yields much more consistent results than the `` classic '' evidence approximation@xcite when we subjected artificial test data to different noise while keeping the signal - to - noise ratio fixed . a detailed comparison of different regularization schemes available will be presented elsewhere . in order to scrutinize the resulting spectra , the following test has been performed . following the idea of silver _ et al._,@xcite the results from the mem applied to the original qmc data @xmath42 have been taken as an `` exact '' spectral function @xmath43 , from which mock data @xmath44 for the imaginary - time green function has been created by using eq . ( [ eq : app : maxent ] ) and adding gaussian random noise of absolute size @xmath45@xmath46 , comparable to the statistical errors of the qmc results . again using the mem to invert these data , the thereby obtained spectral function @xmath47 allows to check the stability of the mem result @xmath43 with respect to statistical errors . we find that the features in our results for @xmath43 remain stable if @xmath48 varies within the statistical errors , so that the data presented below are satisfactorily reliable apart from the aforementioned broadening and the limited energy resolution . while our method is completely free of autocorrelations between subsequent phonon configurations , results for the imaginary - time green function at different times are usually not statistically independent.@xcite however , in the present case , owing to the exact sampling and the large number of measurements , these correlations are very small and have been neglected since they have negligible influence on the results . an additional , stringent test of the qmc results consists of a comparison to ed data which will be presented in sec . [ sec : results : intermediate ] . finally , we note that all qmc simulations have been run on personal computers ( e.g. , intel xeon 2600 mhz ) , with cpu times varying between several hours and a couple of days . the combination of ed and kpm has been described in detail in ref . . in this work , we are mainly interested in the single - particle spectral functions @xmath49 \ , , \end{split } \\ & & \begin{split } a^-(k,{\omega } ) = \sum_l | & { { \langle}\psi_{l , k}^{(n_\text{e}-1)}| } c^{{\phantom{\dag}}}_{q - k } { |\psi_{0,q}^{(n_\text{e})}{\rangle } } |^2\ , \\ & \times \delta[{\omega}+ ( e_{l , k}^{(n_\text{e}-1 ) } - e_{0,q}^{(n_\text{e } ) } ) ] \ , , \end{split}\end{aligned}\ ] ] where @xmath50 denotes the @xmath51th eigenstate with @xmath52 electrons , momentum @xmath53 and energy @xmath54 . the cluster size has been fixed at @xmath55 , with as many as 15 dynamical phonons and resulting truncation errors @xmath56 for the ground - state energy.@xcite additionally , the convergence with the number of phonons has been monitored by means of the phonon distribution function.@xcite it is important to point out that such a small number of dynamical phonons only yields accurate results for the parameters considered here if the symmetric phonon mode with @xmath57 is separated from the hamiltonian analytically before the diagonalization.@xcite in the present case , this phonon mode corresponding to a homogeneous shift of the oscillator equilibrium positions together with that for @xmath58 represents the dominant contribution to phonon excitations , so that the separation is very efficient at reducing the computational effort . the largest matrix dimension in the present calculations was about @xmath59 , running on six compute nodes of a hitachi sr8000-f1 . the energy resolution of the kpm spectral function with 512 moments has been enhanced by a factor of 8 by the mem,@xcite resulting in a significantly better resolution compared to , e.g. , the spectral decoding method . we would like to point out that in contrast to the case of qmc , the ed data to which the mem is applied does not have a statistical error . it is long known that within the holstein model ( [ eq : model : h ] ) , a single electron undergoes a cross over to a small polaron at a critical value of the ep coupling.@xcite the important question to be addressed here is what happens if the density of electrons ( or polarons ) is increased . we do not expect significant changes in the strong - coupling regime due to the rather localized " nature of the small polaron state . on the contrary , for intermediate ep interaction and @xmath60 , each electron is surrounded by a phonon cloud extending over several lattice sites . now , at larger densities , a substantial overlap of the single - particle wavefunctions occurs , leading to a dissociation of the individual polarons and finally to a restructuring of the whole many - particle ground state . furthermore , it has been shown that squeezing effects reduce the polaronic band - narrowing in the many - electron case,@xcite thereby favoring such a scenario . since within the holstein model , a large polaron state is expected to be stable in 1d , we will only consider this case in the sequel . in view of the situation in most polaronic materials , we mainly consider the adiabatic regime , taking @xmath61 . to illustrate the important differences between @xmath62 and @xmath63 , however , some results for the antiadiabatic value @xmath64 will also be shown . as we shall see below , the spectra in the intermediate coupling region turn out to have a fairly complex structure . it is therefore helpful to begin with the limiting cases of weak and strong ep coupling . for weak coupling @xmath65 , the sign problem is not severe ( sec . [ sec : app : sign_results ] ) and the qmc simulations can easily be performed for large lattices with @xmath66 , thereby making the dispersion of quasiparticle features well visible . as indicated in eq . ( [ eq : app : maxent ] ) , the mem inversion yields @xmath67 , where @xmath34 denotes the chemical potential . one - electron spectral function @xmath67 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 , and weak coupling @xmath65.,title="fig:",scaledwidth=23.85% ] one - electron spectral function @xmath67 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 , and weak coupling @xmath65.,title="fig:",scaledwidth=23.85% ] + one - electron spectral function @xmath67 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 , and weak coupling @xmath65.,title="fig:",scaledwidth=23.85% ] one - electron spectral function @xmath67 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 , and weak coupling @xmath65.,title="fig:",scaledwidth=23.85% ] figure [ fig : results : qmc_lambda0.1 ] shows the evolution of the one - electron spectral function @xmath67 with increasing electron density @xmath39 . at first sight , we see that the spectra bear a close resemblance to the free - electron case , i.e. , there is a strongly dispersive band running from @xmath68 to @xmath69 . the latter can be attributed to weakly dressed electrons with an effective mass similar to the noninteracting case . as expected , the height ( width ) of the peaks increases ( decreases ) significantly in the vicinity of the fermi momentum @xmath70 , which is determined by the crossing of the band with the chemical potential . one - electron density of states @xmath71 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 and @xmath65.,scaledwidth=47.5% ] since the peaks in fig . [ fig : results : qmc_lambda0.1 ] are generally very sharp , it is important to check the results from the inversion with the mem . by computing the approximate number of electrons corresponding to the fillings @xmath720.4 of the 32-site system , we can roughly determine @xmath70 . although the nonzero ep interaction will affect the band dispersion @xmath73 , the weak coupling considered here justifies the use of a rigid band approximation , as reflected in the free - particle character of the spectra in fig . [ fig : results : qmc_lambda0.1 ] . for @xmath72 , i.e. , about three electrons for @xmath66 , the fermi momentum is expected to lie between @xmath74 and @xmath75 , corresponding to the second and third curve from the top in fig . [ fig : results : qmc_lambda0.1](a ) . in fact , we see that the largest peaks are those for @xmath74 and @xmath76 , having almost the same height . above and below these momenta , the peaks are noticeably smaller . a similar analysis carried out for the other fillings in figs . [ fig : results : qmc_lambda0.1](b)(d ) also shows a very good agreement between the location of the highest peaks and the estimated values of @xmath70 . finally , we would like to point out that the apparent absence of any phonon signatures in fig . [ fig : results : qmc_lambda0.1 ] is not a defect of the mem , but results form the large scale of the @xmath77-axis chosen . as a consequence , the peaks running close to the bare band dominate the spectra and suppress any small phonon peaks present . if we enlarge the resolution , for all densities @xmath720.4 , we observe the band flattening@xcite at large wavevectors which originates from the intersection of the approximately free - electron dispersion with the optical phonon energy occurring at @xmath78 . to complete our discussion of the weak - coupling regime , we show in fig . [ fig : results : dos_wc ] the one - electron density of states ( dos ) @xmath71 given by eq . ( [ eq : app : dosgen ] ) . clearly , for small @xmath39 , there is a peak with large spectral weight at the fermi level . in contrast , for large @xmath39 , the tendency toward formation of a peierls ( band ) insulating state at @xmath79 suppresses the dos at the fermi level , although we are well below the critical value of @xmath80 at which the cross over to the insulating state takes place at @xmath19.@xcite the additional small features separated from @xmath34 by the bare phonon energy @xmath8 will be discussed in the intermediate coupling case below . one - electron spectral function @xmath67 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 , and strong coupling @xmath81.,title="fig:",scaledwidth=23.85% ] one - electron spectral function @xmath67 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 , and strong coupling @xmath81.,title="fig:",scaledwidth=23.85% ] + one - electron spectral function @xmath67 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 , and strong coupling @xmath81.,title="fig:",scaledwidth=23.85% ] one - electron spectral function @xmath67 from qmc for different band fillings @xmath39 , @xmath66 , @xmath26 , @xmath61 , and strong coupling @xmath81.,title="fig:",scaledwidth=23.85% ] we now turn our attention to the opposite limit of the spinless holstein model , namely the case of strong ep coupling @xmath82 . while calculations with many other numerical methods such as , e.g. , standard grand - canonical qmc or ed , become very difficult in this regime , the qmc approach used here yields quite accurate results even for large systems . the improved statistics in the strong - coupling regime is also reflected in the dependence of the average sign on @xmath80 ( cf . the discussion of fig . [ fig : app : sign_beta_omega ] in the appendix ) . as before , we study the dependence of the spectral function on the band filling , again for @xmath66 . the results are depicted in fig . [ fig : results : qmc_lambda2.0 ] . in the low - density regime @xmath72 [ fig . [ fig : results : qmc_lambda2.0](a ) ] , we expect the well - known almost flat polaron band having exponentially reduced spectral weight ( given by @xmath83 in the single - electron , strong - coupling limit ) which , nevertheless , can give rise to coherent transport at @xmath19 . unfortunately , the qmc / mem has difficulties resolving such extremely weak signatures , especially at larger @xmath39 , where the statistical errors are more noticeable , and for temperatures such that the thermal energy is comparable to the polaron band width . obtained by integration over the spectrum is not sensitive enough to yield the correct value of @xmath34 . ] therefore the polaron band can hardly be resolved in figs . [ fig : results : qmc_lambda2.0](a ) or ( b ) , and is not even detectable in figs . [ fig : results : qmc_lambda2.0](c ) and ( d ) . however , if we consider a density @xmath84 , corresponding to a single electron on a cluster with @xmath85 , at a lower temperature @xmath28 , the qmc results do indeed give a polaron band around @xmath34 ( fig . [ fig : results : qmc_lambda2.0_polaron ] ) . due to the sign problem , similar calculations for @xmath86 would be very demanding . besides , the spectrum consists of two incoherent features located above and below the chemical potential , which reflect the phonon - mediated transitions to high - energy electron states . here , the maximum of the photoemission spectra ( @xmath87 ) follows a tight - binding cosine dispersion . the incoherent part of the spectra is broadened according to the phonon distributions . we would like to mention that the spectrum of a single polaron at strong ep coupling has been studied analytically by alexandrov and ranninger.@xcite dispersion of the polaron band in @xmath67 for @xmath84 , @xmath85 and @xmath28 ( drawn to a larger scale as compared to fig . all other parameters as in fig . [ fig : results : qmc_lambda2.0].,scaledwidth=23.85% ] with increasing band filling [ see panels ( b ) and ( c ) ] , the chemical potential is shifted to lower energies @xmath88 , but is expected to be still located in a narrow band with little spectral weight . regardless of the aforementioned shortcomings , the qmc results provide a clear picture of the physics . there exists a finite gap to the photoemission ( inverse photoemission ) parts of the spectrum , so that the system typifies as a polaronic metal . we shall see below that a completely different behavior is observed at intermediate coupling . notice that the incoherent inverse photoemission ( photoemission ) signatures are more pronounced at small ( large ) wavevectors . finally , for @xmath36 [ fig . [ fig : results : qmc_lambda2.0](d ) ] , the incoherent features lie rather close to the fermi level , thus being accessible by low - energy excitations . now , the photoemission spectrum for @xmath89 becomes almost symmetric to the inverse photoemission spectrum for @xmath90 and already reveals the gapped structure which occurs at @xmath79 due to the cdw - formation accompanied by a peierls distortion . as in the weak - coupling case discussed above , the properties of the system also manifest itself in the dos , shown in fig . [ fig : results : dos_sc ] . owing to the strong ep interaction , the spectral weight at the chemical potential is exponentially small for all fillings @xmath39 . at half filling , the dos exhibits particle - hole symmetry , and the system can be described as a peierls insulator , consisting of a polaronic superlattice . in contrast to the weak - coupling case , the ground state is characterized as a polaronic insulator rather than a band insulator . moreover , here the gap at the fermi level persists even at the finite temperature used in the simulation , while for @xmath65 and @xmath79 the insulating state found at @xmath19 is destroyed by thermal fluctuations ( fig . [ fig : results : dos_wc ] ) . note that , owing to the sign problem , the result for @xmath79 in fig . [ fig : results : dos_sc ] is for a reduced cluster size @xmath85 . however , in the strong - coupling regime considered here , finite - size effects are expected to be very small . one - electron density of states @xmath71 from qmc for different band fillings @xmath39 and cluster sizes @xmath91 , @xmath26 , @xmath61 and @xmath81.,scaledwidth=47.5% ] + as discussed in the introduction , a cross over from a polaronic state to a system with weakly dressed electrons can be expected in the intermediate coupling regime . here we choose @xmath92 , which corresponds to the critical value for the small - polaron cross over in the one - electron problem ( cf . , e.g. , fig . 5 in ref . ) . we first discuss the qmc results . owing to the sign problem , which is particularly noticeable for @xmath92 ( see appendix ) , we have to decrease the system size as we increase the electron density @xmath39 . figure [ fig : results : qmc_lambda1.0 ] shows the spectral function for @xmath92 and increasing band filling . owing to the overlap of large polarons in the intermediate coupling regime , we start with the very low density case @xmath84 [ fig . [ fig : results : qmc_lambda1.0](a ) ] . compared to the behavior for @xmath82 [ fig . [ fig : results : qmc_lambda2.0](a ) ] , we notice that the polaron band now lies much closer to the incoherent features , and that there is a mixing of these two parts of the spectrum at small values of @xmath53 . nevertheless , the almost flat polaron band is well visible for large @xmath53 . with increasing density , the polaron band merges with the incoherent peaks at higher energies . this is the above - anticipated density - driven cross over from a polaronic to a ( diffusive ) metallic state , with the main band crossing the fermi level . at @xmath36 , we find a band with large effective width ranging from @xmath93 to @xmath94 . a more detailed discussion of the spectral function will be given below when we present ed results . ( color online ) one - electron density of states @xmath71 from qmc for different band fillings @xmath39 , cluster sizes @xmath91 and inverse temperatures @xmath95 . here @xmath61 and @xmath96.,scaledwidth=47.5% ] further information about the density dependence can be obtained from the one - electron dos . the latter is presented in fig . [ fig : results : dos_ic ] for different fillings @xmath840.5 . as in fig . [ fig : results : qmc_lambda1.0 ] , the cluster size is reduced with increasing @xmath39 in order to cope with the sign problem . to illustrate the rather small influence of finite - size effects , fig . [ fig : results : dos_ic ] also contains results for @xmath55 . for low density @xmath84 , the dos in fig . [ fig : results : dos_ic ] lies in between the results for weak and strong coupling discussed above . although the spectral weight at the chemical potential is strongly reduced compared to @xmath65 , @xmath97 is still significantly larger than for @xmath81 . when the density is increased to @xmath98 , the dos at the chemical potential increases , as a result of the dissociation of the polarons . as we increase @xmath39 further , a pseudogap begins to form at @xmath34 , which is a precursor of the cdw gap at half filling and zero temperature . finally , in the case of half filling @xmath79 , the dos has become symmetric with respect to @xmath34 . there are broad features located either side of the chemical potential , which take on maxima close to @xmath99 . however , apart from the strong - coupling case , where the single - polaron binding energy is still a relevant energy scale , the position of these peaks is rather determined by the energy of the upper and lower bands , split by the formation of a peierls state . the gap of size @xmath100 expected for the insulating charge - ordered state at @xmath19 is partially filled in due to the finite temperature considered here . furthermore , we find additional , much smaller features roughly separated from @xmath34 by the bare phonon frequency @xmath8 , whose height decreases with decreasing temperature , as revealed by the results for @xmath28 ( fig . [ fig : results : dos_ic ] ) . these peaks not present at @xmath19 ( see ref . ) arise from thermally activated transitions to states with additional phonons excited , and are also visible in figs . [ fig : results : dos_wc ] and [ fig : results : dos_sc ] . while for weak coupling [ @xmath101 , @xmath65 , fig . [ fig : results : dos_wc ] ] , the maximum of these features is almost exactly located at @xmath102 , it moves to @xmath103 for intermediate coupling [ @xmath104 , @xmath92 , fig . [ fig : results : dos_ic ] ] , and finally to @xmath105 for strong coupling [ @xmath106 , @xmath82 , fig . [ fig : results : dos_sc ] ] . although the exact positions of the peaks are subject to uncertainties due to the mem , this reflects the shift of the maximum in the phonon distribution function with increasing coupling . the mem yields an envelope of the multiple peaks separated by @xmath8 . + at this point , we would like to mention the exact relation@xcite @xmath107 for the first moment of the one - particle spectral function of the spinless holstein model , with @xmath108 . while the zeroth moment is identical to the normalization of @xmath43 for each @xmath53 , @xmath109 depends in a nontrivial way on the parameters of the system . in principle , eq . ( [ eq : results : moment ] ) may be used as a boundary condition in the mem.@xcite however , in the present case , we have found that while the normalization is virtually exact for all results shown , the first moment deviates from the exact values . moreover , the use of the exact values for @xmath110 as a condition in the inversion causes the mem not to converge for some parameters . additional calculations on small systems reveal that this disagreement results from the finite trotter error , despite the relatively small value @xmath32 used . away from parameters such that @xmath111 , @xmath110 as determined from the mem spectra deviates from the exact values by maximally 1020 percent . nevertheless , the peak positions and weights in the spectra shown are fairly accurate , as illustrated by the comparison with ed below . finally , we would like to point out that calculations of @xmath43 for @xmath112 become very time - consuming due to the increasing requirement in disk space ( see appendix ) and longer simulation times . despite the wealth of information contained in the qmc results presented so far , it is obvious that an interpretation of the various excitations is far from being easy . in addition to the finite temperature , the qmc method used here does not allow to separately calculate the photoemission and inverse photoemission parts @xmath113 and @xmath114 , respectively . moreover , the use of the mem limits the energy resolution and introduces uncertainties concerning the positions and weights of structures in the spectra . these circumstances make it difficult , e.g. , to distinguish between phononic and electronic contributions , especially if @xmath115 . to gain further insight , we therefore supplement the qmc data with ed results . in order to overcome the abovementioned limitations of qmc , here we use ed in combination with the kpm , as described in sec . [ sec : methods : ed ] , to calculate the one - particle spectral functions defined by eq . ( [ eq : methods : apm ] ) . to test the quality of the qmc spectra , we directly compare the two methods , keeping in mind that the qmc calculations have been performed at a finite temperature , whereas ed yields ground - state results . ( color online ) comparison of the integrated spectral weight obtained from the qmc and ed data of fig . [ fig : results : ed_qmc_lambda1.0](a ) for three different momenta @xmath53.,scaledwidth=45.0% ] figure [ fig : results : ed_qmc_lambda1.0 ] displays the ed results for the same parameters as in fig . [ fig : results : qmc_lambda1.0 ] , i.e. , @xmath61 , @xmath92 , and @xmath720.4 . additionally shown are the integrated spectral weights of @xmath113 and @xmath114 ( dashed lines ) , as well as qmc results for the same system size and @xmath116 ( solid thick lines ) , which have been shifted by @xmath34 . the fermi level lies between the highest - energy peak in @xmath113 and the lowest - energy peak in @xmath114 . we see from fig . [ fig : results : ed_qmc_lambda1.0 ] that the qmc results are compatible with the ed data . because of the finite temperature and the use of the mem , qmc , of course , can not reproduce the sharp ed peaks . to further test the agreement between the two methods , we compare in fig . [ fig : results : ed_qmc_weight ] the integrated spectral weight of the data shown in fig . [ fig : results : ed_qmc_lambda1.0](a ) , and find a very satisfactory accord . moreover , this agreement becomes even better for larger densities @xmath39 , for which the sharp peaks in the ed data evolve into bands [ see fig . [ fig : results : ed_qmc_lambda1.0](b)(d ) ] . in addition to the possibility of distinguishing between @xmath114 and @xmath113crucial for an identification of polaron bands at intermediate coupling the higher energy resolution of ed also allows to resolve signatures of phonon excitations . starting with the case @xmath72 [ fig . [ fig : results : ed_qmc_lambda1.0](a ) ] , we notice that there is a polaron band at the fermi level , excitations to / from which are given by the highest ( lowest ) peak of @xmath113 ( @xmath114 ) . the photoemission part @xmath113 reflects the poisson - like phonon distribution of the one - electron ground state ( @xmath72 and @xmath55 ) . the integrated spectral weight gives a measure for the electronic weight of the various poles in the one - electron spectrum . for example , it reveals that the phonon peaks in @xmath113 have very little weight . in fact , the integrated weight jumps to a finite value at the first peaks near the fermi level for @xmath117 and @xmath118 , while it changes very little as one moves further down in energy . for @xmath119 , the small spectral weight contained in @xmath113 is continuously distributed among the phonon peaks . we also observe the well - known flattening of the polaron band at large values of @xmath53 . similar to the one - electron case,@xcite the low - energy states have mostly electronic character at small @xmath53 , and become mostly phononic at large values of @xmath53 . while this effect is expected to occur in the low - density regime , we find that is persists even for @xmath36 [ fig . [ fig : results : ed_qmc_lambda1.0](d ) ] . moreover , fig . [ fig : results : ed_qmc_lambda1.0 ] reveals that the maximum in the incoherent contribution follows closely the free - electron dispersion for all densities @xmath720.4 . with increasing electron density , the equally spaced peaks in @xmath113 broaden significantly , until at @xmath36 they have merged to form a broad band . the polaron band is still visible at @xmath98 [ fig . [ fig : results : ed_qmc_lambda1.0](b ) ] , but becomes indistinguishable from the incoherent excitations at even larger @xmath39 . eventually , at @xmath36 [ fig . [ fig : results : ed_qmc_lambda1.0](d ) ] , we find an almost symmetric behavior of @xmath114 and @xmath113 with respect to the fermi level at @xmath120 . compared to the strong - coupling case [ fig . [ fig : results : qmc_lambda2.0](d ) ] , there exist incoherent excitations with almost zero energy . this indicates the metallic character of the system expected for large polaron densities and intermediate ep coupling . ( color online ) one - electron spectral functions @xmath121 and @xmath122 from ed for @xmath55 , @xmath123 , and @xmath81 . here @xmath72 ( @xmath113 : red dot - dashed lines , @xmath114 : black solid lines ) and @xmath124 ( @xmath113 : green dashed two - dotted lines , @xmath114 : blue dotted lines ) . dashed lines correspond to the integrated spectral weight of @xmath121 and @xmath122 , respectively.,scaledwidth=45.0% ] until here , we have only presented results for the adiabatic case @xmath61 . although the latter is most relevant for strongly correlated materials , it is important to compare the above findings to the antiadiabatic strong - coupling regime . to this end , we chose @xmath64 as well as @xmath82 . the condition for the formation of a single small polaron in this case , however , is @xmath125 , because the phonon frequency is the important energy scale . in contrast , the criterion @xmath126 is purely based on the balance of kinetic and potential energy.@xcite the spectral function for @xmath72 and @xmath124 is displayed in fig . [ fig : results : w4 ] for @xmath127 , i.e. , at the transition point where self - trapping occurs for a single particle . contrary to the case @xmath92 and @xmath61 considered above , the absorption features are now separated by the phonon energy @xmath8 , with their width being determined by electronic excitations . a comparison of the spectral functions for @xmath72 and @xmath124 in fig . [ fig : results : w4 ] indicates that there is no density - driven cross over of the system as observed in the adiabatic case . in particular , owing to the large phonon energy , there are no low - energy excitations close to the polaron band , so that the latter remains well separated from the incoherent features even for @xmath124 . furthermore , the spectral weight of the polaron band also remains almost unchanged as we increase the density from @xmath72 to @xmath124 . consequently , almost independent small polarons are formed also at finite electron densities , in accordance with previous findings of capone _ et al._@xcite for small systems . normalized kinetic energy as a function of ep coupling @xmath80 for different band fillings @xmath39 and adiabatic ratios @xmath128 . also shown are the results for a single polaron , obtained with the method of ref . . errorbars are smaller than the symbols used and @xmath31.,scaledwidth=45.0% ] previously , the electronic kinetic energy has been studied extensively to monitor the polaron and bipolaron cross over in the holstein model with one and two electrons ( see , e.g. , refs . and ) . here , at least in the adiabatic regime , polaron formation manifests itself in a drop of the kinetic energy around the critical coupling . in the following , however , we illustrate that the kinetic energy [ eq . ( [ eq : app : ek ] ) ] does not reflect the significant changes in the one - particle spectrum we found with increasing density in the intermediate coupling regime ( see sec . [ sec : results : intermediate ] ) . to compare different band fillings , we consider the normalized quantity @xmath129 , shown in fig . [ fig : results : ek ] as a function of ep coupling for different fillings and phonon frequencies , for a fixed system size @xmath130 and @xmath26 . in the adiabatic regime @xmath61 , owing to the overlap of the phonon clouds surrounding the individual quasiparticles , the carriers clearly become more mobile with increasing density , but the change in the kinetic energy is nevertheless surprisingly small . in addition to these two fillings , we have also included in fig . [ fig : results : ek ] the result for a single polaron with the same parameters , calculated with the qmc method of ref . , which is indistinguishable from the result for @xmath72 . as expected from the almost identical spectra for @xmath72 and 0.3 presented in sec . [ sec : results : nonadiabatic ] for the antiadiabatic regime , @xmath131 is almost independent of @xmath39 in this case . strongly correlated electron - phonon systems in dependence on the carrier density have been studied by means of unbiased numerical methods . using exact diagonalization together with the kernel polynomial method , as well as a novel quantum monte carlo approach , we were able to present accurate results for the one - dimensional spinless holstein model in the most difficult regime of small phonon frequencies and intermediate to strong electron - phonon interaction . we have calculated the one - particle spectral function , the density of states and the electronic kinetic energy at different band fillings . while exact diagonalization is restricted to relatively small clusters , thereby limiting the momentum resolution of the spectra , quantum monte carlo can be used to study larger systems , but yields a lower energy resolution for dynamical quantities . the reliability of the results from the maximum entropy inversion has been scrutinized , and we find a very good agreement between the two methods . in the adiabatic case , characteristic for , e.g. , the manganites , we observe a carrier density - driven cross over from a polaronic state to a metallic system at weak and intermediate electron - phonon coupling . note that the nature of the quasiparticles is changed in the course of this transition . this points toward an insufficiency of simple single - polaron theories to explain experimental results for polaronic materials . in fact , very similar effects due to carrier interaction have recently been observed experimentally for the manganites.@xcite on the contrary , for large phonon frequencies and strong coupling , the individual polarons remain virtually unaffected upon increasing the electron density . consequently , the system remains polaronic even at large band fillings . finally , the kinetic energy is shown to be too insensitive in order to reveal the above cross over . for weak electron - phonon coupling , the system becomes a peierls or band insulator at half filling , whereas for strong coupling a polaronic insulating state with an exponentially small polaron band at the fermi level , and a finite gap to the incoherent excitations , prevails for all fillings . finally , reliable calculations of the optical conductivity are highly desirable in order to explain the large amount of available experimental data for polaronic materials . work along this line is in progress . this work was supported by the austrian science fund ( fwf ) , project no . p15834 , the deutsche forschungsgemeinschaft through spp1073 , and the bavarian competence network for high performance computing ( konwihr ) . m. h. is grateful to doc ( doctoral scholarship program of the austrian academy of sciences ) , and to hpc - europa . h. f. acknowledges the hospitality at graz university of technology . we would like to thank a. alvermann , h. g. evertz , t. lang and a. weie for useful discussion . furthermore , we acknowledge generous computer time granted by the lrz munich . in this appendix we extend the one - electron qmc algorithm developed in ref . , henceforth also referred to as i , to the spinless holstein model with many electrons . we begin by applying the lang - firsov transformation@xcite @xmath132 with @xmath133 to the hamiltonian ( [ eq : model : h ] ) in first quantization and with dimensionless phonon variables @xmath134 and @xmath135 , which takes the form@xcite @xmath136 the coupling constant @xmath137 is related to that in eq . ( [ eq : model : h ] ) by @xmath138 . using the result @xmath139 ( see i ) , we define the grand - canonical hamiltonian @xmath140 where @xmath34 denotes the chemical potential and @xmath15 is the polaron binding energy . we would like to point out that in the spinfull case , the transformed hamiltonian would contain an attractive hubbard term . for the determinant qmc method to be applicable , the latter has to be decoupled using auxiliary fields.@xcite in contrast , for the spinless model considered here , the algorithm is almost identical to the one - electron case . the following derivation assumes a hypercubic lattice in d dimensions , consisting of @xmath142 sites . we use the suzuki - trotter decomposition@xcite @xmath143 with @xmath144 . the trace appearing in the partition function @xmath145 can be split up into a bosonic and a fermionic component leading to @xmath146 where @xmath147 , and with periodic boundary conditions @xmath148 . we have chosen the time ordering , which is arbitrary , in increasing order . the phonon coordinates @xmath134 in @xmath149 can be integrated out analytically in the same manner as in i. moreover , the momenta @xmath135 can be replaced by their eigenvalues on each time slice , and the partition function takes the form @xmath150 with @xmath151 , @xmath152^{nl/2}$ ] and @xmath153 the bosonic action @xmath154 may be expressed in terms of principal components ( see i ) . the fermion degrees of freedom can be integrated out exactly leading to@xcite @xmath155 where the @xmath156 matrix @xmath157 is given by @xmath158 with @xmath159 here @xmath160 is the usual tight - binding hopping matrix.@xcite to save computer time , we employ the checkerboard breakup@xcite @xmath161 using eq . ( [ eq : app : checker ] ) , the numerical effort scales as @xmath162 instead of @xmath163 , while the error due to this additional approximation is of the same order @xmath164 as the trotter error in eq . ( [ eq : app : suzuki - trotter ] ) . defining the bosonic and fermionic weights @xmath165 and @xmath166 respectively , the partition function finally becomes @xmath167 one of the advantages of blankenbecler _ et al._s@xcite formalism as well as the current approach is the close relation to the one - electron green function @xmath168 working in real space and imaginary time , we have@xcite @xmath169 and @xmath170 we would like to mention that despite the formal similarity to the method of blankenbecler _ et al._,@xcite the numerical realization of the present approach is quite different . while the grand - canonical method of ref . benefits enormously with respect to performance from a local updating scheme for the phonons , here we use a global updating in terms of the principal components together with the reweighting method described in i. although this requires us to recalculate the full matrix @xmath171 in each mc step , the resulting statistically independent configurations clearly outweigh the loss in performance , especially for small @xmath8 for which autocorrelation times can exceed @xmath172 sweeps when using the original grand - canonical qmc method.@xcite the approach proposed here allows one to perform uncorrelated measurements after each update . moreover , our calculations show that numerical stabilization by means of a time - consuming singular value decomposition is not necessary for all parameters considered in sec . [ sec : results ] . finally , owing to the phase factors in the transformed hopping term , all the matrices become complex - valued . an important practical test in simulations therefore is the reality of expectation values of observables , i.e. , the disappearance of the imaginary part within statistical errors . the phase factors in the hopping term [ eq . ( [ eq : app : hspinless ] ) ] give rise to a minus - sign problem , i.e. , @xmath173 can take on negative values , whereas @xmath174 is always positive . while for a fixed number of electrons the sign problem diminishes with increasing system size,@xcite the opposite is true if we fix the electron density . in that case , we find an exponential decrease of the average sign with increasing cluster size , and a strong dependence on the band filling ( sec . [ sec : app : sign_results ] ) . the calculation of observables within the formalism presented here is similar to the standard determinant qmc method.@xcite an important difference is that we have to use the lang - firsov - transformed operators , i.e. , @xmath175 as mentioned above , the mc sampling is based purely on the bosonic weight @xmath174 . this corresponds to a reweighting of the probability distribution so that the fermionic weight @xmath173 is treated as part of the observables.@xcite as for a single electron,@xcite the overlap of the probability distributions involved is large enough to permit this procedure . hence , we have @xmath176 where the expectation value @xmath177 for an equal - time observable @xmath178 is defined as @xmath179 we begin with the electron density @xmath180 the expectation value in eq . ( [ eq : app : n ] ) can be calculated from the diagonal elements of the green function @xmath181 [ eq . ( [ eq : app : ga ] ) ] , i.e. , @xmath182 . similarly , the absolute value of the electronic kinetic energy per site is given by @xmath183 equal - time two - particle correlation functions such as @xmath184 may be calculated with the help of wick s theorem.@xcite an observable of great interest is the time - dependent one - particle green function @xmath185 which is related to the spectral function @xmath186 through @xmath187 the inversion of the above relation is ill - conditioned and requires the use of the mem.@xcite fourier transformation leads to @xmath188 the allowed imaginary times are @xmath189 , with nonnegative integers @xmath190 . within the qmc approach , we have@xcite @xmath191 another interesting quantity is the one - electron density of states @xmath192 where @xmath193 . it may be obtained numerically via @xmath194 and subsequent use of the mem . average sign @xmath195 [ eq . ( [ eq : app : sign ] ) ] of the fermionic weight @xmath173 as a function of electron - phonon coupling @xmath80 in one dimension ( a ) for different inverse temperatures @xmath95 , and ( b ) for different values of the adiabatic ratio @xmath128 . here and in subsequent figures @xmath31 , lines are guides to the eye only , and errorbars are suppressed if smaller than the symbols shown.,title="fig:",scaledwidth=35.0% ] + average sign @xmath195 [ eq . ( [ eq : app : sign ] ) ] of the fermionic weight @xmath173 as a function of electron - phonon coupling @xmath80 in one dimension ( a ) for different inverse temperatures @xmath95 , and ( b ) for different values of the adiabatic ratio @xmath128 . here and in subsequent figures @xmath31 , lines are guides to the eye only , and errorbars are suppressed if smaller than the symbols shown.,title="fig:",scaledwidth=35.0% ] while hamiltonian ( [ eq : app : hspinless ] ) is symmetric with respect to a particle - hole transformation for @xmath141 ( half filling ) , this symmetry is broken if we use the checkerboard approximation ( [ eq : app : checker ] ) , so that @xmath196 for the above choice of the chemical potential . to simplify calculations , the results for the average sign at @xmath79 presented below have therefore been obtained using the exact hopping term . for general band fillings , we find that the checkerboard breakup requires different values of @xmath34 to perform simulations at the same electron densities , but results are identical within statistical errors if @xmath34 is adjusted accordingly . the derivation of the qmc algorithm in sec . [ sec : app : qmc_partition ] is valid for any dimension d of the lattice under consideration . nevertheless , here we only report results for the sign problem in @xmath197 , the case considered in sec . [ sec : results ] , and make some remarks about the influence of the dimensionality at the end . the average sign is defined as @xmath198 with the fermionic weight in the present case given by @xmath166 [ eq . ( [ eq : app : omega ] ) ] . we begin with the dependence of @xmath195 on the electron - phonon coupling strength . for simplicity , we show results for @xmath79 , while the effect of band filling will be discussed later . the choice @xmath79 is convenient since we know the chemical potential , and we shall see below that the sign problem is most pronounced for a half - filled band . moreover , all existing qmc results for the spinless holstein model are for half filling , so that it is interesting to see how the sign problem affects simulations within the current approach . average sign as a function of ( a ) band filling @xmath39 , and ( b ) system size @xmath91.,title="fig:",scaledwidth=35.0% ] + average sign as a function of ( a ) band filling @xmath39 , and ( b ) system size @xmath91.,title="fig:",scaledwidth=35.0% ] figure [ fig : app : sign_beta_omega](a ) reveals that the average sign takes on a minimum near the point where depending on the filling the cross over to small polarons or a cdw state occurs . in the adiabatic regime @xmath60 , the cross over condition is @xmath126 [ fig . [ fig : app : sign_beta_omega](a ) ] , whereas for @xmath199 we have @xmath200 [ fig . [ fig : app : sign_beta_omega](b ) ] . at weak and strong coupling , the average sign is close to 1 , so that accurate simulations can be carried out . as shown in fig . [ fig : app : sign_beta_omega](b ) , we find that @xmath195 becomes very small for @xmath60 and intermediate @xmath80 , while it increases noticeably in the nonadiabatic regime @xmath199 . owing to the absence of any autocorrelations in our approach , the present method therefore represents a significant improvement of existing algorithms , as the latter face severe autocorrelations even for @xmath201 , thereby limiting the temperature range and cluster size . as illustrated in fig . [ fig : app : sign_n_n](a ) , the average sign depends strongly on the band filling @xmath39 . while @xmath202 in the vicinity of @xmath203 or @xmath204 , a significant reduction is visible near half filling @xmath79 . the minimum occurs at @xmath79 , and the results display particle - hole symmetry as expected . here we have chosen @xmath26 , @xmath61 and @xmath92 , for which the sign problem is most noticeable according to fig . [ fig : app : sign_beta_omega ] . note that there is no sign problem in simulations for , e.g. , the hubbard model at half filling , where the ( fermionic ) weights of the @xmath205 and @xmath206 electrons are equal and their product entering the partition function is positive.@xcite this can also be expected within the current approach if we consider the half filled spinfull holstein or holstein - hubbard model , and work along this line is in progress . finally , in fig . [ fig : app : sign_n_n](b ) , we report the average sign as a function of system size , again for @xmath79 . the behavior is strikingly different from the one - electron case considered in i. while in the latter @xmath207 as @xmath208 , here the average sign decreases nearly exponentially with increasing system size . obviously , this limits the applicability of our method . however , as illustrated in sec . [ sec : results ] , we can nevertheless obtain accurate results at low temperatures , small phonon frequencies , and over a large range of the electron - phonon interaction . moreover , we would like to point out that for such parameters , existing methods suffer strongly from autocorrelations , rendering simulations extremely difficult . the dependence of the sign problem on the dimension of the system is similar to the single - electron case.@xcite for the same parameters @xmath91 , @xmath128 , @xmath209 and @xmath80 , the minimum in @xmath195 at intermediate @xmath80 becomes more pronounced as one increases the dimension of the cluster . to conclude with , we would like to point out that , in principle , the sign problem can be compensated by performing sufficiently long qmc runs , although we have to keep in mind that the statistical errors increase proportional to @xmath210 ( ref . ) . owing to the purely phononic updates , the present algorithm is very fast and we have made up to about @xmath211 single measurements to obtain the results presented in sec . [ sec : results ] . the use of the reweighting method requires us to store the measured values of observables in every mc step , and to perform the jackknife error analysis@xcite at the end of the run . for example , in the case of the green function ( [ eq : app : gb_tau ] ) , @xmath212 complex numbers have to be stored for each momentum @xmath53 in every step . depending on the observables of interest , this sets a practical limit for the maximal number of measurements due to restrictions in available disk space .

convection is a major process taking place in reg giant ( rg ) envelopes . in a previous letter ( @xcite , hereafter at97 ) , results of a study of this subject were reported briefly . in the present paper , more results of this study are presented . throughout the paper , the reference one - dimensional ( 1d ) model of convection is the mixing length theory ( mlt , @xcite ) , according to the prescription of @xcite . laboratory experiments , observations of the outer layers of the solar convective zone , and theoretical understanding have demonstrated that convection is much more complicated than mlt assumes . analysis of experiments of convection in a box , in which a temperature gradient was established by using helium at about 5k and controlling the boundary conditions , have shown that convective flow is non - local with defined narrow streams going from the top down ( and vice versa ) , i.e. the existence of a downstream at a given depth at a given time depends upon the existence of a downstream in a higher point at previous time and not upon local parameters only . this phenomenon is known as hard turbulence ( see @xcite ) . observation of the outer layers of the sun have shown the existence of two dimensional cells of upstreams , in a variety of sizes and shapes , with the boundaries of these cells being the downstreams , i.e. the flow is not symmetric ( see @xcite ) . during the years , many studies were made in order to overcome the simplifications of mlt . several non - local and non - instantaneous theories were proposed ( @xcite ) . a common feature in these theories is using turbulence theory techniques in which the hydrodynamical quantities at each point are constructed of two components : a constant value and a varying value . equations are formulated for each of the varying quantities which are then used to determine the convective velocity and flux . these equations include creation and diffusion terms and thus constitute a set of coupled partial difference equations . maybe because of this complexity , there is no wide usage of these theories . xiong ( 1998a , b ) have used one of these theories to compute linear pulsation times of rr lyrae and long period variables . @xcite are using a simplified 1d model of turbulent energy diffusion and convection with eight free parameters to describe cepheid and rr lyrae variables . @xcite used turbulence models to construct an alternative model for mlt which includes a full spectrum of eddy sizes ( see also @xcite ) . this model is still local and instantaneous , but is easy to include in astrophysical codes . it was found that this model yields higher values of convective flux than mlt at efficient convection regimes , and lower values of convective flux at inefficient convection regimes . this model was found to be preferable to mlt in many studies ( see @xcite ) . these studies included red giant evolution ( @xcite ) , heliosesmology ( @xcite ) and calculations of uvby colors of a f and g stars ( @xcite ) . the superiority of this model , for example in modeling red giant evolution , was the usage of a constant parameter for various masses , while when using mlt there was a need to change the mixing length parameter with mass . however , @xcite have shown that a varying calibration is needed for this model too . alternatives to the usual mixing length used in mlt ( which is a free parameter times the pressure scale height ) were suggested . these included using the density scale height ( @xcite ) and the distance from the boundary of the convection zone as in @xcite . for a correct modeling of red giant variables , the lack of time dependence and the locality of mlt are major disadvantages . the convection velocities , as predicted by mlt , and the radial velocities associated with pulsation are of the same order of magnitude ( of @xmath4 ) . thus , interaction of the two phenomena is expected . ad hoc prescriptions were made in order to take into account time dependence in dynamical calculations with mlt ( for example , pulsation calculation of miras , see @xcite ) . in these calculations , the convective flux at a given time is determined by using mlt and the flux at a previous time assuming a typical convective time scale estimated by convection velocity . in other calculations , like evolution of massive stars , a non locality is presented to mlt in particular when treating penetration ( or overshoot ) of convection to stable neighboring zones . a common assumption is that this overshoot region is proportional to the mixing length at the edge of the convection region . these methods are obviously not rigorous and not intended to describe these processes precisely ( for example see @xcite . being mainly a hydrodynamical phenomenon , an appealing method for modeling convection is by using numerical multidimensional simulations . in this approach , no explicit model of convection is needed , and thus no mixing length parameter is assumed . time dependence and non locality are naturally present . however , convection is multidimensional and the flow has many length scales and time scales , that might be smaller by far than other typical scales of the problem . this is why such calculations are not suitable for all problems with convection . numerical simulations of the hard turbulence experiments mentioned above where made ( werne 1993,1994 , @xcite ) . these calculations reproduced the typical narrow streams of the experiments , and revealed that the origin of the flow was the thin boundary layers at the bottom and at the top of the experimental box . many studies were made in order to understand general features of convective flow ( @xcite ) . typically , these studies assumed simplified physical assumptions , such as ideal gas equation of state , constant opacity and constant gravitation . the systems were confined in a box with an initial unstable polytropic profile . boundary conditions were periodic on the sides of the box , no flow was assumed in and out of the bottom and top of the box , a constant temperature gradient at the bottom and a constant temperature at the top . some of the simulations assumed planar symmetry using 2d codes , while other used three - dimensional ( 3d ) codes . summary of some the results of these studies is : there is an asymmetry in the flow with narrow rapid downstreams and wider upstreams and the 3d picture resembles the observations of the sun with cells of upstreams ; the flow in the wider upstreams is somewhat more turbulent ; the size of the cells increases with depth so downstreams merge . this picture is revealed when using high resolution , while low resolution simulations show laminar flow with big convection cells . the internal energy flux in the downstreams is relatively large , but a large kinetic energy flux with opposite sign almost cancels the net energy flux in the downstreams . near the outer boundary , convection might be supersonic . other `` simulation in a box '' studies were dealing with invoking of p - modes in the convection zone ( @xcite ) , producing g - modes below ( and above ) the convection region ( @xcite ) and penetration of convection to neighboring layers ( @xcite singh 1994,1995 ) . attempts have been made to simulate the outer layers of the solar convection region ( stein & nordlund 1989,1998 , @xcite ) . since the computation box in these simulations included only a fraction of the convection region , penetrative boundary conditions were applied at the bottom of the computation box . the outer layers which were included in the simulations and in which a superadiabatic gradient exist , are the most sensitive to the convection model . these simulations reproduced the granulation pattern of convective streams of solar convection , as well as the heat flux and spectrum ( @xcite ) . comparison of various measurements form the simulations ( such as maximal temperature gradient near the outer boundary ) enables calibration of the mlt parameter . @xcite and @xcite have published such calibration for various models of solar - type stars . they found that mlt parameter is 1.2 to 1.8 for starts with effective temperatures of 7100k to 4300k ( with dependency upon g - gravity and metallicity as well ) . they also investigated the @xcite model and found it needs varying calibration as well . simulations of convection in zz - ceti type white dwarfs were done by @xcite . they found that different calibration models yield different mlt parameter values : 1.5 according to one measurement , and 4 according to another . two - dimensional ( 2d ) calculations of rr - lyra were performed by deupree ( 1975,1977 , 1985 ) ; the ( very ) low resolution simulations of the narrow convection region in these variables stars have shown that convection has a damping effect upon pulsation , and vice versa . in these stars , there is little difference between a radiative model and a convective model , due to the narrow convection region . thus , the relaxation time in the 2d simulations is relatively small , and might be calculated . this is not true for rg in which the convection region is wide , and that is why simulating convection in rg s envelopes is much more difficult . the above discussion , as well as the topic of this paper , concerns convective envelopes , in which the convection is not efficient enough to enforce an almost adiabatic profile throughout the convective region . multidimensional studies of other astrophysical cases have been published . this includes core convection ( @xcite ) , shell burning ( @xcite , bazan & arnett 1994,1998 ) , nova bursts ( @xcite ) , core collapse ( see @xcite and ref . therein ) , and accretion disks ( for example @xcite ) . in these studies , some of the interesting questions are : the amount of mixing and penetration due to convective streams , nuclear energy sources and sinks in the convection region , the coupling of convection and rotation . an interesting question in the context of multidimensional simulations of convection , is the validity of 2d calculations . while it is obvious that 3d simulations are preferable , it is not clear what are the quantitative differences between 2d and 3d results . in many 2d studies of convection , there are large convective cells ( eddies ) that occupy the whole convective region ( see for example @xcite ) , this phenomena is not seen in 3d simulations . however , it does not mean that there are no small size eddies in 2d simulations , when increasing spatial resolution large eddies tend to brake up to small eddies ( see @xcite ) , and even low resolution 2d simulations have small eddies ( see @xcite and discussion in @xcite ) . the qualitative features of convective flow , revealed by 3d simulations , are present in 2d simulations as well ( see @xcite and ref . therein ) . there are few quantitative measurements of the differences between 2d and 3d simulations : in a study of convection in proto - neutron star , @xcite have shown that the importance of small scale features , and the kinetic energies are different when comparing 3d to 2d simulations ; @xcite have simulated novae in 3d and compared the results to their 2d simulations ( @xcite ) , they found the difference in the spectrum of eddy sizes to have an influence on the amount of overshoot and mixing , thus yielding a different answer to the question of the existence of thermonuclear runaway ; in a study of convection in solar type stars , which is more comparable to this paper , @xcite have results of first 3d calculations , these results indicate little difference between the calibrated mixing length parameter from 3d simulations , relative to the calibrated mixing length parameter from 2d simulations . modeling the convection region of rg envelopes is necessary when studying phenomena related to these stars . for example , consider the observational relation of the pulsation time to the luminosities of mirae : there is ambiguity concerning the pulsation mode ( is it the fundamental mode or the first mode ? ) . the period in the first mode is about half of the period in the fundamental mode , however changing the mixing length parameter from 1 to 1.5 is enough to change the calculated period by a factor of 2 ( @xcite ) . the value of the mixing length parameter affects the calculated age of globular clusters ( see @xcite and ref . therein ) . due to the importance of modeling convection in studying rg s , and the lack of previous hydrodynamical calculations of this phenomena , i focused on 2d simulation of rg s envelope convection . in 2 i describe the code used for the simulations . summary of the results published in at97 is presented in 3 and more results of the at97 simulations are summarized in 4 . a survey of the importance of the various parameters of the simulation techniques is presented in 5 and 6 . a short summary of this research is given in 7 . as was described in at97 , the hydrodynamical code vulcan ( @xcite ) was used for this study , after utilizing several modifications to the code . this version of the code solves the hydrodynamic equations and radiative diffusion equations by splitting each time step into hydrodynamic time step followed by a radiative diffusion time step . both steps are solved implicitly to have an accurate and stable solution . the code uses an arbitrary lagrangian eulerian ( ale ) scheme in which each lagrangian step is followed by a relaxation of the numerical grid . in this relaxation stage , the grid is modified to eliminate distortion and to allow calculation of flow with vorticity . a quasi 1d lagrangian relaxation was used , in which each radial row of cells kept its mass . this was done for each row of cells by calculating an average outer radius , so that the sum of the mass fluxes through it would be zero . using this method , one was able to follow the average radial expansion or shrinking of the envelope . the mass , momentum and energy fluxes in the relaxation stage were calculated by a second order donor scheme , in which the gradiants of these variables were used to estimate the fluxes at the boundaries of the cells ( according to @xcite ) . . like other numerical studies of convection ( @xcite ) , i implemented subscale viscosity and diffusion terms using the @xcite model . the inclusion of these terms in the equations solved by the code should compensate for the use of a finite number of cells which is much less than needed when simulating turbulent flow if one wishes to simulate the effects of all scales up to the scale of physical dissipation . in the simple smagorinsky model these terms are the usual viscosity terms with a coefficient which depends upon derivatives of the velocities and a typical length scale of the numerical grid . when there is a scalar field that is characteristic of the fluid , one should add a diffusion term of this scalar field with a coefficient which is proportional to the viscosity coefficient . in convective flow , this scalar is the entropy ( @xcite , ludwig 1994 , @xcite ) and thus entropy diffusion was added to the code . the two parameters in these large eddies simulation ( les ) terms were the smagorinsky coefficient ( which multiplies the viscosity coefficient ) and the prandtl number ( which is the ratio of the viscosity coefficient to the diffusion coefficient ) . both the viscosity and the entropy diffusion were added as explicit terms since they were relatively small . the inner boundary in the simulations was located deeper than the convection zone to minimize the effects of not including the entire envelope . however , in preliminary simulations , hydrodynamical waves were invoked below the convection zone , and there was a convective flux near the inner boundary . by comparing simulations in which the inner boundary was placed at different depths , it was found that this convective flux was not physical because it was always located near the inner boundary . realizing this was due to an improper boundary condition ( in which only the radial velocities were set to zero for points on the inner boundary ) , i changed the condition and set both components of the velocities at the inner boundary to zero and allowed only radial velocities at the lowest three rows of points . with this new boundary condition , the flux disappeared . the inner boundary condition for the radiative diffusion was of a constant entering flux equal to the luminosity of the static model ( i.e. equivalent to @xmath2 ) . for the hydrodynamic step a constant ( usually zero ) outer pressure was used , without limiting the direction of the velocities at the outer boundary . however , at the mesh relaxation stage , an average radius was calculated for the outer boundary so that the mass flux through the boundary was zero . the remapping fluxes ( i.e. the fluxes at the mesh relaxation stage ) , through this boundary were calculated using the values of the bundary cells ( i.e. , using an acceptor scheme for the incoming flow ) . for the radiative diffusion , an outgoing flux was calculated using @xmath5 where t is the temperature of the boundary cell and @xmath6 is the stefan constant ( i.e. , assuming it to be at optical depth of about two thirds ) . for the sides of the computational sector either reflective boundary conditions , or periodic boundary conditions were used . when using the reflective boundary conditions , all fluxes through these boundaries were set to be zero , and thus only downstreams or upstreams existed on the boundaries . this was different when using periodic boundary condition , for which each boundary cell has an artificial neighbor at the other side of the sector ( though , since i assumed cylindrical symmetry , there is no direct physical meaning to periodic boundary conditions ) . in order to analyze the results of the simulations i used printouts of time averaged energy fluxes for each radial layer . the fluxes included radiative flux as was calculated by the radiative diffusion step , remapping fluxes of internal energy and kinetic energy calculated by the grid relaxation stage , and hydrodynamical fluxes that can be calculated using @xmath7 where @xmath8 is the total energy density , @xmath9 is the pressure , @xmath10 is the velocity vector and @xmath11 is the mass density . the convective flux is the sum of the remapping and the hydrodynamical fluxes . both the radiative and the remapping fluxes were calculated explicitly in the simulations and had only to be time averaged . however , the hydrodynamical fluxes were intrinsic and thus had to be estimated . i used two methods : a. explicitly using the above formula , and b. by energy balance of each radial layer in the hydrodynamical step ( i.e. the flux from one layer to the next in each time step is equal to the change in total energy of the layer during the time step minus the flux from the previous layer ) . both methods had inconsistencies : the first uses a formula which is not one of the formulae that control the simulation and the second is correct only if full conservation of energy exists in the simulation . additional discussion of this issue appears in 4 , where a flux plot is presented . as a test case for the simulation techniques i reproduced the results of a `` convection in a box '' study ( published by @xcite ) . the configuration for these simulations was of two stable layers with an unstable layer in between . the velocity field and the averaged fluxes in the simulations were similar to the published existing results . in addition to the 2d hydrodynamic code , two 1d codes were used : a static code ( see @xcite for description ) , and a hydrodynamic code ( @xcite ) . the static code was used to integrate initial models of the envelope , and the dynamic code was used to follow the hydrodynamic evolution of the models for comparison with the 2d results ( as explained in the following sections ) . both codes use @xcite prescription of mlt to calculate convective flux . the equation of state and opacities used by these 1d codes are identical to those used by the 2d code . the equation of state is based on tabular form and includes sum of components for electrons , ions , molecules of hydrogen and helium and radiation ( see @xcite for details ) . opacities are calculated according to @xcite with water molecules contribution according to @xcite ( taken from @xcite ) . the study was focused on models of an envelope of a red giant with total mass of @xmath0 core mass of @xmath1 , luminosity of @xmath2 and metallicity of @xmath3 . the models were integrated using the mixing length theory with the mixing length proportional to the pressure scale height . five models were integrated using different values of the proportion coefficient ( -mixing length parameter ) @xmath12 : @xmath13 ( i.e. no convection ) , @xmath14 , @xmath15 , @xmath16 , and @xmath17 . these five models had different total energy as well as different energetic structure as an outcome of different energy transport efficiency , as was dictated by the value of the mixing length parameter . the aim of the study was to find which of these five models was more consistent with the results of 2d dynamical simulations . the final configuration reached by the end of 2d dynamical simulations of each of these five models should be the same , since they all correspond to the same physical envelope ( the same mass , luminosity and structure below the convection zone ) . the difference in the simulations is in the relaxation time needed for the different models to adjust their total energy and structure and to reach thermal equilibrium . due to the long thermal time scale of these envelopes , and to computational limitations , this relaxation phase could not be fully simulated , however , as was demonstrated in figure 1 , there was enough data in the beginning of the simulations to reach conclusions about the preferred model . in this figure the luminosity as a function of time for each of the five simulations was presented . the most obvious result in this figure is that in the simulations that started from models with @xmath18 the envelope lost energy while for simulations started from models with @xmath19 the envelope gained energy . that was the main reason for the conclusion that the 1d model integrated with @xmath20 was the most consistent with the 2d simulations . a consistency check of this result was done by using 1d dynamic simulation with @xmath20 mlt starting from the above different initial models . it was found out that the 1d simulations mimic the 2d simulations , i.e. the luminosity ( and radius ) as a function of time for the different initial models looked similar . a more precise analysis of the results revealed that the final configuration of the _ dynamical _ simulations ( both one and 2d ) had slightly less total energy than the _ static _ 1d model of @xmath20 . a typical velocity field in the simulations is presented in figure 2 . one can identify narrow downstreams and wider upstreams in the flow . the length scale of the streams increased with depth ( as well as the pressure scale height ) . the existence of hydrodynamical waves below the convection zone is apparent . as mentioned before , by comparison of the five simulations a conclusion could be made that the @xmath20 model was the most consistent with the final 2d configuration , however from figure 3 it is clear that this model was not in a full thermal equilibrium . in this figure , the ( time ) averaged fluxes in the 2d simulation of this model are presented , including the radiative flux and the total flux ( the sum of radiative and convective fluxes ) . for comparison , the 1d ( static ) radiative flux is presented , too . the radiative flux in the 2d simulation resembled the static model , though a modest widening of the convective zone is clear , and an increase in the radiative flux above the constant luminosity is apparent near the bottom of the convective zone ( see @xcite ) . the value of the total flux is indicative of the thermal equilibrium of the model . the total flux was slightly less than @xmath21 below the convective zone , and reached a maximum of about @xmath22 near the upper boundary and then it declined to the surface luminosity of about @xmath23 . while it is clear that this is a result of a yet non equilibrium configuration , one should nt forget that due to the lack of algebric energy conservation , this plot is influenced by the uncertainty in the value of the hydrodynamical flux ( as was described in 2 ) : the energy conservation in each time step was of the order of @xmath24 of the energy . this seems to be small , but when one divides it by the time step ( about 80 seconds ) , one gets @xmath25 . i think it is important to note that this has little effect upon the simulation itself , which i confirmed by adding artificial non - energy - conserving term to the _ 1d _ dynamic code and comparing the results of several simulations where i changed the value of this term . the different values of the luminosity in the simulations ( figure 1 ) were dominated by the values of of the outer radii ( compare this figure with figure 3 in at97 ) . in the initial configurations there was a relatively large difference in the effective temperatures : 3675k for @xmath26 , 4600k for @xmath27 and 5110k for @xmath28 . however , in the dynamical simulations , very quickly ( less than a month ) the difference shrank dramatically , and the effective temperature became : 4520k , 4600k , and 4650k ( for the same models as above ) with time fluctuations of about 100k . in other words , using the dynamical simulations , one gets almost immediately , an effective temperature which is very close to the effective temperature of the @xmath20 model . this gives more confidence in the above conclusion that the @xmath20 is the preferred model . the 1d consistency simulations featured the same phenomenon . the five ( 2d ) simulations used similar numerical parameters . the grid consisted of 36 angular cells in each row , with 53 to 72 rows ( for the simulations of models with @xmath19 , i included deeper regions of the envelope - down to eight @xmath29 - because of the deeper convection zone ) . the angular sector was @xmath30 near the equatorial plane . the fluxes in the relaxation stage were calculated using a second order donor scheme , and the energy flux consisted of both internal and kinetic energies . i used reflective boundary conditions on the sides of the computational domain . les viscosity was used , with smagorinsky coefficient of 0.5 and without entropy diffusion . i made many tests of the dynamical simulation of the model that started with the @xmath20 profile . i changed the numerical scheme , the number of cells and the angular width of the computational domain and the parameters of the les terms . the changes in numerical scheme included : calculating the remapping fluxes using full donor ( i.e. without taking the gradients of the hydrodynamical variables into account ) , using internal energy fluxes in the grid relaxation stage ( without adding the kinetic energy ) and the use of periodic boundary conditions at the sides . i found that the full donor simulation yielded a rapid expansion of the envelope - implying it is necessary to include gradients in the calculation of the fluxes . the use of the internal energy ( only ) for the remapping energy fluxes caused a small constant loss of kinetic energy and thus a modest shrinking of the envelope without changing the luminosity . the velocity field of the simulation with periodic boundary conditions is presented in figure 4 . the non radial velocities at the sides are expected , but no other significant change can be seen , compared to the velocity field of reflecting boundary conditions ( figure 2 ) , i.e. the size of the eddies and the changes in the velocities with depth look similar . the luminosity and outer radius as a function of time look the same as well , as can be seen in figure 5 . changes in the angular width or number of cells had small effects . increasing both the angular width to @xmath31 and the number of cells in each row to 54 yielded similar results as the nominal simulation , as presented in the velocity field - figure 6 , and in the luminosity and outer radius - figure 7 . when i decreased the angular width to @xmath32 i got a larger variability in the luminosity and outer radius ( see figure 8) and a small increase in the average luminosity , as well . a small increase in the average luminosity was seen in a simulation with larger number of cells ( = 54 ) in each row ( with the nominal angular width of @xmath30 ) , as presented in figure 9 . increasing the number of rows to 81 , had no significant effect . the variations in the les terms included changing the numerical coefficient and adding an entropy diffusion term . when i had zero or small ( smagorinsky coefficient of 0.01 ) les viscosity i got somewhat larger variability in the luminosity and outer radius , but without changes of the average values . however , including an entropy diffusion term ( with prandtl number of one ) increased the average luminosity by a small amount ( see figure 10 ) . when i had this term , a further increase in the luminosity was evident when the smagorinsky coefficient was increased to one a summary of these tests reveals that there was a tendency to increase the luminosity in two cases : when i increased the angular resolution , and when i added an entropy diffusion term in the les scheme . these two cases have one physical meaning : an increase in the efficiency of convection by adding smaller scale turbulent fluxes . since there was a very small luminosity increase in the above simulations , i performed yet another simulation where i increased the resolution by a ( relatively ) large factor . i started with a profile of the nominal simulation ( of @xmath33 cells ) and increased the number of cells almost by an order of magnitude ( @xmath34 cells ) , without the entropy diffusion term . this simulation , being much more cpu demanding , was followed for a small physical time but nevertheless a clear conclusion can be made , as can be seen in figure 11 . the luminosity increased quite significantly from about @xmath23 in the nominal simulation to about @xmath35 . an increase in luminosity means that the envelope loses more energy , and from the five simulation with different initial models , this means that the 1d model which is consistent with the 2d simulations should have a larger value of @xmath12 , i.e. @xmath19 . to test this conclusion i performed high resolution simulation starting with the initial model based upon @xmath28 . again , i started with a profile of the nominal simulation ( with @xmath36 cells ) and increased the resolution to @xmath37 cells . this yielded an increase in the luminosity from about @xmath38 to about @xmath39 , as expected . in order to check systematically the dependence of the luminosity in the spatial resolution and in the les terms , i performed the following simulations , which started with the same initial profile . the initial profile was a 2d profile with an already developed convective flow based on the initial 1d model with @xmath20 . from this profile i interpolated four equivalent profiles with different spatial resolution : @xmath40 cells , @xmath41 cells , @xmath42 cells and @xmath43 cells , thus having factor of about 64 between the number of cells in the finer grid and in the coarser grid . from the above discussion it is clear that the significant term in the les scheme is the entropy diffusion term , so i used the nominal 0.5 value of the smagorinsky coefficient , and changed the importance of the entropy diffusion term by changing the prandtl number , using the values : @xmath44 ( i.e. the last value corresponds to increasing the diffusion term by a factor of five compared to the first value ) . i used these values of the prandtl number for each of the four resolutions , and thus having sixteen simulations . the flow in four of these sixteen simulations is presented in figure 12 , for different resolutions . the flow is presented by contours of the vorticity and indeed increasing the resolution yielded more small scale eddies ( and higher values of vorticity ) . the luminosities of these simulations are summarized in figure 13 , and in table 1 i present the average luminosities . from these results one can conclude that the best estimate of the luminosity is @xmath45 , and that the better the resolution is , the less important the value of prandtl number is . in a perfect les scheme one should get the converged result for each resolution for given set of numerical parameters ( i.e. prandtl number in our case ) . however , the value of about @xmath46 was obtained in each resolution with a different value of prandtl number , this value seems to increase when decreasing the resolution . for two of these 16 simulations i checked the importance of the les entropy diffusion term comparing to the les viscosity terms . i performed simulations where i changed both smagorinsky coefficient and prandtl number so that the amplitude of the diffusion term would be the same , while changing the amplitude of the viscosity terms . i performed a @xmath42 cell simulation with smagorinsky coefficient of 1.5 and prandtl number of 1 and got an average luminosity of @xmath47 ( equivalent to a prandtl number @xmath48 simulation ) , and a @xmath49 cell simulation with smagorinsky coefficient of 1 . and prandtl number of 1 where i got an average luminosity of @xmath46 ( equivalent to a prandtl number @xmath14 simulation ) . the summary of these results is that the 2d simulations of an initial model calculated with @xmath20 mlt , yields an energy - losing envelope with an average luminosity of @xmath46 . to find what value of @xmath12 can reproduce this i performed several 1d dynamic simulations . in each of these simulations , i started with the same @xmath20 profile , and assuming different values for the @xmath12 in the dynamical model i got different luminosities . for the following values of @xmath12 : 1.2 , 1.3 , 1.4 and 1.5 i got average luminosities of about : 230 , 240 , 250 and 270 ( @xmath50 ) . from this i conclude that the consistent value of @xmath12 is 1.4 . as was presented in at97 the convective envelope of a red giant ( of mass @xmath0 , core mass of @xmath1 , luminosity of @xmath2 , and metallicity @xmath3 ) was studied by comparing 2d simulations of five different initial models . while a full convergence of these simulations to a balanced final configuration was not possible , due to the large thermal time scale of this envelope , it was found that an initial model calculated with mlt using @xmath20 , was almost energetically balanced , and thus was the best estimate of the final configuration . after a comprehensive survey of the importance of the various parameters used in the simulation technique , a sensitivity to the spatial resolution and to the les entropy diffusion term was found . this sensitivity was tested by using four resolutions and four values of the les term , and it was concluded that the preferred value of @xmath12 is 1.4 , for the envelope tested in this study . it was found that , at least for this code and for these kind of simulations , one needs to calibrate the les terms . the conclusion for the value of @xmath12 , to be used in modeling the ged giant envelope studied in this research , being larger than one is consistent with the results of @xcite where they calculated evolutionary sequences and compared them to observation . this value of 1.4 is also consistent with _ extrapolation _ of the results of @xcite and @xcite where numerical simulations were applied to the outer convective zone of solar like stars . this conclusion is important for studies in which an accurate modeling of red giants is necessary , like obtaining pulsation times of miras ( @xcite ) or calculating isochrones of globular clusters(@xcite ) . an obvious limitation of this study is the use of 2d simulations ( and not 3d simulations ) . the spectrum of eddy sizes is probably different in 3d , but the influence of this upon the calibrated mixing length parameter is not known . however , @xcite have first estimation of the change in the calibrated mixing length parameter for their study of solar type convection , due to the difference between 2d and 3d simulations . they conclude that the calibrated mixing length parameter for the sun from 3d simulation is higher by 0.07 than from 2d simulation . i am grateful to yitzchak tuchman for his encouraging cooperation and to eli livne for his code . i thank ami glasner , yossi stein and zalman barkat for many fruitful discussions , and dave arnett and the referee for useful suggestions to the manuscript . abbett , w.p . , beaner , m. , davids , b. , georgobiani , d. , rathbun , p. , & stein r.f . 1997 , , 480 , 395 arnett , w.d . 1994 , , 427 , 932 asida , s.m . , & tuchman , y. 1997 , , 491 , l47 basu , s. & antia , h. m. 1994 , journal of astrophysics and astronomy , 15 , 143 bazan , g. , & arnett , w.d . 1994 , , 433 , l41 bazan , g. , & arnett , w.d . 1998 , , 496 , 316 bogdan , t.j . , cattaneo , f. , & malagoni , a. , 1993 , , 407 , 316 bohm - vitense e. , 1958 , z. astrophys . 46 , 108 buchler , j.r . , yecko , p. , koll@xmath51th , z. , & goupil , m .- j . 1999 , astro - ph/9905025 canuto , v.m . , 1996 , , 467 , 385 canuto , v.m . , 1997 , , 482 , 827 canuto , v.m . , & christensen - dalsgaard , j. , 1998 , annu fluid mech . , 30 , 167 canuto , v.m . , & dubovikov , m. , 1998 , , 493 , 834 canuto , v.m . , goldman , i. , & mazzitelli , i. , 1996 , , 473 , 550 canuto , v.m . , & mazzitelli , i. , 1991 , , 370 , 295 castaing , b. , gunaratne , g. , & heslot , f. , 1989 , j. fluid mech . , 204 , 1 cattaneo , f. , brummell , n.h . , toomre , j. , malagoni , a. , & hurlburt , n.e . 1991 , , 370 , 282 cattaneo , f. , hurlburt , n.e . , & toomre , j. 1990 , 349 , l63 chan , k.l . , & gigas , d. 1992 , , 389 , l87 chan , k.l . , & sofia , s. 1986 , , 307 , 222 chan , k.l . , sofia , s. , & wolff , c.l . 1982 , , 263 , 935 cox , a.n . , & stewart , j.n . 1970 , , 19 , 243 cox , j.p . , & giuli , r.t . 1968 , principles of stellar structure ( new york : gordon & breach ) demarque . p. , guenther , d.b . , & kim , y .- c . 1997 , , 474 , 790 deupree , r.g . 1975 , 201 , 183 deupree , r.g . 1977 , 211 , 509 deupree , r.g . 1985 , 296 , 160 deupree , r.g . 1998 , 499 , 340 freytag , b. , ludwig , h .- g . , & steffen , m. 1999 , in proceedings contribution to the workshop on stellar structure : `` theory and tests of convective energy transport '' granada , spain , 1998 , eds . a. gimenez , e. guinan , b. montesinos freytag , b. & salaris , m. 1999 , , 513 , l49 gehmeyr , m. & winkler , k. -h . a. 1992 , , 253 , 92 glasner , s.a . , & livne , e. 1995 , , 445 , l149 glasner , s.a . , & livne , e. , & truran , j.w . 1997 , , 475 , 754 grossman , s. a. 1996 , , 279 , 305 grossman , s.a . , narayan , r. 1993 , , 89 , 361 grossman , s.a . , narayan , r. , & arnett , d. 1993 , , 407 , 284 helsot , f. , castaing , b. , & libchaber , a. 1987 , , 36 , 5870 hossain , m. , & mullan , d.j . 1991 , , 380 , 631 hurlburt , n. e. , toomre , j. , & massaguer , j. m. 1984 , , 282 , 557 hurlburt , n. e. , toomre , j. , & massaguer , j. m. 1986 , , 311 , 563 hurlburt , n. e. , toomre , j. , massaguer , j. m. , & zahn , j. -p . 1994 , , 421 , 245 kercek , a. , hillebrant , w. , & truran , j.w . 1998 , , 337 , 379 kercek , a. , hillebrant , w. , & truran , j.w . 1998 , , 345 , 831 kerr , r.m . 1996 , j. fluid mech . , 310 , 139 kim , y. -c . , fox , p. a. , sofia , s. , & demarque , p. 1995 , , 442 , 422 kim , y. -c . , fox , p. a. , demarque , p. , & sofia , s. 1996 , , 461 , 499 livne , e. 1993 , , 412 , 634 ludwig , h .- g . , freytag , b. , & steffen , m. 1999 , , 346 , 111 ludwig , h. -g . , jordan , s. & steffen , m. 1994 , , 284 , 105 maeder , a. , & conti , p.s . 1994 , , 32 , 227 malagoni , a. , cattaneo , f. , & brummell , n.h . 1990 , , 361 , l33 mezzacappa , a. , calder , a. c. , bruenn , s. w. , blondin , j. m. , guidry , m. w. , strayer , m. r. & umar , a. s. 1998 , , 495 , 911 m@xmath52ller , e. , & janka h .- 1997 , , 317 , 140 nordlund , a. , & stein , r.f . 1995 , stellar evolution : what can be done , 32@xmath53 liege int . nordlund , a. , stein , r.f . , & brandenburg , a. 1996 , bull astron . india , 24 paczinski , b. 1969 , acta astr . 19 , 1 porter , d.h . , & woodward , p.r . 1994 , , 93 , 309 rast , m.p . , & toomre , j. 1993 , , 419 , 224 schwarzchild , m. 1961 , , 134 , 1 singh , h. p. , & chan , k. l. 1993 , , 279 , 107 singh , h. p. , roxburgh , i. w. , & chan , k. l. 1994 , , 281 , l73 singh , h. p. , roxburgh , i. w. , & chan , k. l. 1995 , , 295 , 703 smagorinsky , j. 1963 , mon . weather rev . , 91(3 ) , 99 smalley , b. , & kupka , f. 1997 , , 328 , 349 spiegel , e.a . 1963 , , 138 , 216 spruit , h.c . , nordlund , a. , & title , a.m. 1990 , , 28 , 263 stein r.f . , & nordlund , a. 1989 , 342 , l95 stein r.f . , & nordlund , a. 1998 , 499 , 914 stellingwerf , r.f . 1982 , , 262 , 330 stone , j. m. , & balbus , s. a. 1996 , , 464 , 364 stothers , r. b. , & chin , c. -w . 1995 , , 440 , 297 stothers , r. b. , & chin , c. -w . 1997 , , 478 , l103 tligner , a. , belmonte , a. , & libchaber , a. 1993 , , 47 , r2253 tuchman , y. 1980 , `` dynamical phenomena in red giant envelopes '' , phd . thesis tuchman , y. , sack , n. , & barkat , z. 1978 , , 219 , 183 tuchman , y. , sack , n. , & barkat , z. 1979 , , 234 , 217 ulrich , r.k . 1970 , ap . space sci . , 7 , 71 ulrich , r.k . 1976 , , 207 , 564 van leer b. 1979 , j. comput . phys . , 32 , 101 werne , j. 1993 , , 48 , 1020 werne , j. 1994 , , 49 , 4072 wu , x .- z . , & libchaber , a. 1992 , , 45 , 842 xie , x. , & toomre , j. 1993 , , 405 , 747 xiong , d. r. , cheng , q. l. , & deng , l. 1997 , , 108 , 529 xiong , d. r. , deng , l. , & cheng , q. l. 1998 , , 499 , 355 xiong , d. r. , cheng , q. l. , & deng , l. 1998 , , 500 , 449 yaari , a. , & tuchman , y. 1998 , private communication zahn , j. -p . 1991 , , 252 , 179

particles with resonant short - range interactions display the efimov effect @xcite and related universal phenomena associated with a discrete scaling symmetry @xcite . in particular , efimov showed that three identical bosons form infinitely many trimer bound states with an accumulation point at the scattering threshold when the @xmath3-wave scattering length @xmath4 is tuned to the unitary limit @xmath5 : @xmath6 where @xmath0 is the mass of the particles , @xmath7 is a transcendental number , and @xmath8 is the binding wavenumber of the efimov state labeled by @xmath9 . the geometric spectrum in ( [ kappa - star ] ) is the signature of a discrete scaling symmetry with scaling factor @xmath10 . for a finite scattering length larger than the range of the interaction , the universal properties persist but there is only a finite number of efimov states . the full structure of the trimer spectrum as a function of the inverse scattering length @xmath11 is conveniently summarized in the so - called efimov plot illustrated in fig . [ fig:3body ] . this plot shows the scattering length dependence of the bound states in the @xmath12 plane for the 3-body problem . scattering thresholds are indicated by the cross - hatching while the universal window is given by the circle of radius @xmath13 . a few of the infinitely many branches of efimov states are shown by the solid lines.,width=302 ] @xmath14@xmath12 plane , where @xmath15 is a momentum variable that corresponds to the binding momentum for bound states with @xmath16 . the threshold for scattering states is indicated by the hatched area . the efimov trimers are represented by the solid lines below the threshold . there are infinitely many branches of efimov trimers , but only a few are shown . in the unitary limit @xmath17 , infinitely many trimer states accumulate at threshold . when the radial variable @xmath18 is of the order of the inverse interaction range @xmath13 or larger , universality breaks down and the states become sensitive to the details of the interaction at short distances . ultracold atoms are an ideal tool to study such phenomena since the scattering length can be tuned experimentally to be large compared to the range of the interaction using feshbach resonances . efimov trimers can be observed in ultracold atomic gases via their signature in three - body recombination rates @xcite . using this method , kraemer _ et al . _ provided the first evidence for efimov trimers in an ultracold gas of @xmath19cs atoms by observing the resonant enhancement of atom losses through three - body recombination caused by the trimers @xcite . efimov states have been observed in a variety of other atomic species since then @xcite . the scaling factor @xmath20 can be significantly reduced in heteronuclear systems with different mass particles which makes it easier to observe multiple states . in this paper , we focus on heteronuclear systems with two species of atoms where only the interspecies scattering length is large . for comparable masses the scaling factor is quite large in this case ( @xmath21 for equal masses ) as we now have only two resonant interactions out of three . however , in the case of two heavy bosonic atoms of mass @xmath1 ( denoted @xmath22 ) and one light atom of mass @xmath23 ( denoted @xmath24 ) , this factor can become significantly smaller than the value 22.7 for identical bosons @xcite . relaxation and recombination losses near an interspecies resonance have recently been investigated in mixtures of cesium and lithium atoms and three consecutive cs - cs - li efimov resonances were observed in agreement with the predicted scaling relations @xcite . the effect of the cs - cs scattering length on the efimov resonances was investigated in refs . @xcite . connected to the efimov states are universal bound states in the four- and higher - body sector . although there is no accumulation of bound states at threshold , there are universal four- and higher - body states attached to each efimov trimer . the binding energies of these higher - body states are uniquely determined by @xmath4 and @xmath8 and no higher - body parameters are required @xcite . in experiments with ultracold cesium atoms universal states up to @xmath25 have been observed @xcite . for identical bosons , exactly two tetramer states are attached to each trimer state @xcite . this pattern appears to hold also for the higher - body states and has been calculated explicitly up to @xmath26 @xcite . the ground states have been calculated up to @xmath27 @xcite . it would be very interesting to demonstrate the existence of the higher - body universal states in experiments with mixtures . focusing on four - body states , the @xmath28 states are most promising due to the favorable scaling factor . however , the relation of the higher - body energies to the trimer energies is not governed by the same universal numbers and even the number of higher - body states attached to the trimers will very likely depend on the mass ratio . the signatures of efimov physics in heteronuclear four - body systems were first investigated in ref . a detailed study of the four - body states in a mixture of cesium and lithium atoms in the vicinity of the unitary limit and the three - atom threshold was recently performed by blume and yan @xcite . for large enough mass ratios ( @xmath29 they found two universal four - body states in the unitary limit . they also found an excited four - body state for negative values of the scattering length for all mass ratios they investigated ( @xmath30 ) . for the mass ratio @xmath31 they showed that the excited tetramer becomes unbound at a certain positive value of the scattering length , leading them to the conjecture that the scattering length at which the excited tetramer becomes unbound depends on the mass ratio . in this paper , we build upon the investigation by blume and yan and extend it in several ways . we extend their calculation to positive scattering lengths further away from the unitary limit . in particular , we investigate the mass dependence of the spectrum and discuss the number of universal tetramer states as function of the mass ratio . one of our key findings is that the structure of the spectrum changes as the mass ratio @xmath2 decreases . in particular , the distance between the thresholds for trimer and tetramer states gets smaller until they almost coincide for @xmath32 . preliminary studies of this aspect were presented in ref . moreover , we summarize our insights in a multi - dimensional efimov plot which shows @xmath12 as a function of @xmath11 and the ratio of the heavy and light masses @xmath2 . a particularly interesting aspect of the @xmath28 system is the prospect of observing effective efimov states of a @xmath33 dimer and two heavy atoms @xmath22 which are denoted @xmath34 in the following . such an effective efimov effect must exist close to the crossing point of the @xmath35 trimer and the @xmath33 dimer threshold . this requires precise studies of the bound state spectrum near the @xmath33 dimer threshold . assuming that only the @xmath33 dimer and the two @xmath22 atoms interact , one would expect a scaling factor that is larger than 1986 , which is the limiting value for @xmath36 @xcite . for a cesium - lithium mixture , one would expect a scaling factor of about 2500 . using the born - oppenheimer approximation , wang and collaborators @xcite found evidence for the existence of two such states for the mass ratio @xmath37 with a scaling factor of approximately 20 . they explain the small scaling factor with an effective interaction between the @xmath22 particles that is induced by the exchange of the @xmath24 particle . in this paper , we will address this puzzle by investigating the effective @xmath34 efimov states for different mass ratios @xmath2 using explicit four - body calculations as well as using an effective field theory for heteronuclear three - body mixtures @xcite . to obtain the energies of the @xmath38 trimer and @xmath28 tetramer systems , we employ the gaussian expansion method ( gem ) @xcite . the method was successfully applied to various types of three- , four- and five - body hypernuclear systems @xcite and trimer and tetramer systems of @xmath39he @xcite . we use jacobian coordinate sets for trimer and tetramer systems as shown in fig . [ fig : three - atom ] . the hamiltonians for the three and four - body systems are described as @xmath40 respectively , where @xmath41 is the kinetic - energy operator , the @xmath42 and @xmath43 are two - body potentials and @xmath44 and @xmath45 are three - body potentials , which are the same as in ref . the details are given below . here it should be noted that we switch off the two - body interaction between the heavy bosonic atoms . the energies are obtained by solving the trimer and tetramer schrdinger equations given by @xmath46 at ( 0,0.99 ) and @xmath47 are heavy bosonic atoms and @xmath48 is a light atom . in ( b ) @xmath49 , @xmath47 , and @xmath48 are heavy bosonic atoms and @xmath50 is a light atom . the two bosonic atoms in the trimer and the three bosonic atoms in the tetramer are symmetrized.,title="fig:",scaledwidth=40.0% ] ; at ( 8,0 ) and @xmath47 are heavy bosonic atoms and @xmath48 is a light atom . in ( b ) @xmath49 , @xmath47 , and @xmath48 are heavy bosonic atoms and @xmath50 is a light atom . the two bosonic atoms in the trimer and the three bosonic atoms in the tetramer are symmetrized.,title="fig:",scaledwidth=50.0% ] ; at ( -2.55,0.5 ) @xmath51 ; at ( -0.63,0.85 ) @xmath52 ; at ( 5.5,2.2 ) @xmath51 ; at ( 6.7,2 . ) @xmath53 ; at ( 7.7,2.35 ) @xmath54 ; at ( 8.6,-1.88 ) @xmath55 ; at ( 6.7,-2.2 ) @xmath56 ; at ( 7.2,-1.3 ) @xmath57 ; the wave functions of three - body and four - body systems are described as a sum of amplitudes for the rearrangement channels ( @xmath58 for the trimer , @xmath59 for the tetramer ) illustrated in fig . [ fig : three - atom ] as follows : @xmath60_{jm } \\ \psi^{\rm tetramer}_{jm } & = \sum^4_{c=1 } \sum_{n_c , n_c,\nu_c}\sum_{\ell_c , l_c,\lambda_c } c^{(c)}_{n_c \ell_c n_c l_c \nu_c \lambda_c } \\ & \quad \times [ [ \phi^{(c)}_{n_c \ell_c}({\bf r}_c ) \psi^{(c)}_{n_c l_c}({\bf r}_c)]_i \xi_{\nu_c \lambda_c}^{(c ) } ( \mbox{\boldmath $ \rho$}_c)]_{jm}. \end{split}\ ] ] here , the two ( three ) bosonic atoms in the trimer ( tetramer ) must be symmetrized . for this purpose , we take the angular momenta @xmath61 to be even numbers in the corresponding rearrangement channels . we take the functional form of @xmath62 , @xmath63 , and @xmath64 as @xmath65 where @xmath66 , @xmath67 , @xmath68 denote the normalization constants . the gaussian range parameters are chosen according to geometrical progression : @xmath69 the gaussian ranges are very suitable for accurately describing both the short - range correlations and the long - range tail in the asymptotic region of few - body wave functions . the gaussian shape of basis functions makes the calculation of matrix elements simple even between basis functions of different channels . rrrrrrrrrrrrr + @xmath70 & @xmath71 & @xmath72 & @xmath73 & @xmath74 & @xmath75 & @xmath76 & @xmath77 & @xmath78 + 1 & 0 & 0 & 23 & 0.028 & 15 & 20 & 0.038 & 65 + 1 & 2 & 2 & 21 & 0.180 & 27 & 19 & 0.228 & 20 + 2 & 0 & 0 & 19 & 0.110 & 26 & 21 & 0.085 & 29 + + + the eigenenergies @xmath79 in eq . ( [ hamil ] ) and the coefficients @xmath80 in eq . ( [ wave ] ) are then determined by the raleigh - ritz variational method . in tables [ table : parameters ] and [ table : parameters2 ] , typical parameters of the basis functions for the trimer and tetramer are listed . rrrrrrrrrrrrrrrrr + @xmath70 & @xmath71 & @xmath72 & @xmath73 & @xmath74 & @xmath75 & @xmath76 & @xmath77 & @xmath78 & @xmath81 & @xmath82 & @xmath83 + 1 & 0 & 0 & 11 & 0.073 & 13 & 13 & 0.059 & 8 & 11 & 0.055 & 14 + 2 & 0 & 0 & 15 & 0.043 & 9 & 15 & 0.044 & 7 & 15 & 0.075 & 33 + 3 & 0 & 0 & 17 & 0.038 & 6 & 16 & 0.044 & 13 & 17 & 0.055 & 25 + 3 & 1 & 1 & 13 & 0.057 & 7 & 13 & 0.052 & 6 & 14 & 0.094 & 25 + 4 & 0 & 0 & 14 & 0.139 & 15 & 12 & 0.033 & 5 & 14 & 0.127 & 12 + + + the potential and kinetic energy terms are expanded in the truncated basis as shown in eq . ( [ wave ] ) . because the states are not orthogonal this transforms the schrdinger equation @xmath84_{jm } { \vphantom{\def{&}{}h - e c^{(c)}_{n_cl_cn_cl_c}[\phi^{(c)}_{n_c \ell_c}({\bf r}_c ) \psi^{(c)}_{n_c l_c}({\bf r}_c)]_{jm } } } \right| h - e \left| c^{(c)}_{n_cl_cn_cl_c}[\phi^{(c)}_{n_c \ell_c}({\bf r}_c ) \psi^{(c)}_{n_c l_c}({\bf r}_c)]_{jm } { \vphantom{\def{&}{}[\phi^{(c')}_{n_{c ' } \ell_{c'}}({\bf r}_{c ' } ) \psi^{(c')}_{n_{c ' } l_{c'}}({\bf r}_{c'})]_{jm}h - e } } \right > } = 0\ ] ] into a general eigenvalue problem @xmath85 where @xmath86 is a normalization matrix and @xmath87 is a shorthand notation for all the indices that are summed over , i.e. @xmath88 in the three - body case . the matrix elements can be calculated analytically for some potentials and the equation can be solved using standard linear algebra methods . since we want to study universal effects , i.e. effects that are independent of the exact shape of the potential and only depend on the long - range behavior of the potential , it is beneficial to use simple potentials . we have chosen gaussian potentials , because their matrix elements can be calculated analytically in the gaussian expansion method ( gem ) . there are many prior calculations with gaussian potentials @xcite , so we can easily validate our method . we use the same potentials as in ref . the two - body potential is given by @xmath89 and the three - body potential is @xmath90 where @xmath91 , and @xmath92 is the position of atom @xmath93 . the interaction takes place only between different atoms , where the @xmath94th atom is the special one ( type @xmath24 ) , and all other atoms are identical bosons ( type @xmath22 ) . the interaction between identical bosons is neglected , because it is assumed to be very weak compared to the resonant @xmath33 interaction . this is reflected in the summation in eq . ( [ eq : pot ] ) . gaussian potentials that are tuned to the universal regime can be understood as regularized contact terms in an eft expansion @xcite , so our results are valid only in the energy regime of this implicit eft expansion . this requires that all calculated energies are small compared to the natural energy scale set by the window of universality shown in fig . [ fig:3body ] . the natural energy scale connected to the potential ( [ eq:2bpot ] , [ eq:3bpot ] ) is @xmath95 the reduced mass of the interacting dimer ( @xmath33 ) . the two - body potential ( [ eq:2bpot ] ) is attractive , whereas the three - body potential ( [ eq:3bpot ] ) is chosen to be repulsive such that the calculated binding energies are in the window of universality , i.e. @xmath96 . the three - body force parameter , @xmath97 , is kept constant throughout the calculations presented here , except where noted otherwise . the units are chosen so that the calculations are numerically well - behaved , and the results are presented in terms of @xmath98 and @xmath4 , the @xmath33 s - wave scattering length . we vary @xmath99 to reproduce different scattering lengths @xmath4 as @xmath99 is roughly linearly proportional to @xmath11 . the length scale @xmath100 is then essentially a free parameter that sets the scale for @xmath4 . its numerical value is set to @xmath101 throughout our calculations . .comparison of our benchmarking results with @xcite for the mass ratio @xmath31 . the numbers are taken from fig . 2 of the supplement and were provided to us by d. blume @xcite . the first column gives the strength of the three - body force in units of the natural energy scale @xmath98 as defined in eq . . the following columns give the ratios of the trimer ground state and @xmath98 and the ratio of the lowest two trimer states , respectively . [ cols= " < , < , < , < , < " , ] in order to benchmark our method , we calculate the trimer and tetramer energies in the vicinity of the unitary limit and compare with the results for the mass ratio @xmath31 obtained by blume and yan @xcite . our calculations are carried out very close to the unitary limit at @xmath102 . the corresponding results for the trimer energies are given in table [ tab : trimere ] , while the corresponding tetramer energies are given in table [ tab : tetramere ] . for the different strengths of the three - body force @xmath97 , we find agreement with ref . @xcite to three digits or better , which gives us confidence in our method . however , the time spent optimizing the basis function parameters for the actual calculations discussed in sec . [ 4and3closetothresh ] was about @xmath103 times longer than for the benchmark calculations discussed above . moreover , the basis size used for the actual calculations was about twice as large as for the benchmark calculations . this ensures that our results in sec . [ 4and3closetothresh ] are converged to about 8 digits corresponding to a precision of @xmath104 . the energies are less sensitive to non - universal aspects of the finite - range model potentials for larger values of @xmath97 @xcite . as a consequence , we use @xmath105 in our calculations of the @xmath35 and @xmath28 systems below . this ensures absolute energies are small enough that range effects do not play a role and the results are close to the universal limit of zero - range interactions . in practice , we have ensured that all calculated energies satisfy @xmath106 . since the gaussian expansion method as employed here uses a truncated sum over rearrangement channels and relative angular momenta to represent the wave function ( cf . ( [ wave ] ) ) , contributions from higher relative angular momenta are neglected . we included as many configurations as possible without encountering numerical instabilities . in addition , the states are extremely shallow in the region where very high precision is required , such that higher relative angular momenta are strongly suppressed . the highest angular momentum quantum numbers that are included in our four - body calculations are @xmath107 . including the corresponding @xmath107 basis functions of the rearrangement channel @xmath108 ( cf . [ fig : three - atom ] ) , which contributes the most , leads to a correction of only @xmath109 near the threshold . this gives us confidence that including even higher relative angular momenta would not lead to significant contributions . in this work , we focus on three - body systems of one light particle @xmath24 of mass @xmath0 and two heavy bosons @xmath22 of mass @xmath1 and four - body systems of three heavy bosons @xmath22 and one light particle @xmath24 . in order to establish the qualitative behavior of trimer and tetramer states as a function of the mass ratio @xmath2 , we have calculated the tetramer and trimer energies for the cases @xmath110 , @xmath111 , and @xmath112 , which correspond to possible mixtures of ultracold atoms , as a function of @xmath113 where @xmath100 is the interaction range of the effective potentials considered ( cf . ( [ eq:2bpot ] , [ eq:3bpot ] ) ) . thus the scale @xmath98 defined in eq . is the natural energy scale of the problem , and all energies are quoted in units of @xmath98 . to embed our findings in a larger context , we show in fig . [ fig : overview ] an efimov plot for @xmath35 trimer and @xmath28 tetramer states for the mass ratios @xmath31 and @xmath114 . all states shown are well within the universal window indicated by the shaded area in fig . [ fig:3body ] . the results for @xmath115 are not plotted because they would overlap with the results for @xmath31 . to increase visibility of the features of the plot , the axes are rescaled by taking the fourth root . the variables @xmath22 and @xmath116 are introduced as @xmath117 and @xmath118 ( cf . ref . to @xmath119 small deviations from the linearity of the dimer threshold are strongly magnified . therefore we plot the energy differences to the @xmath33-@xmath22-@xmath22 dimer - atom - atom threshold on the right half plane and include an idealized dimer - atom - atom threshold . ] we only show the lowest trimer because we focus on the tetramer states attached to each trimer for two different mass ratios . we could confirm that , as already noted by blume at al . @xcite , the excited tetramer does exist in the unitary limit for @xmath31 and does not for @xmath114 . they reported that for mass ratios less than 16 , namely 12 and 8 , they could only find an excited tetramer away from the unitary limit on the negative scattering length side . for @xmath114 we did not find any excited tetramer , not even on the negative scattering length side . this indicates that the value of @xmath11 where the excited tetramer crosses into the trimer threshold recedes from a point on the positive scattering length side of the efimov plot for large mass ratios to points on the negative scattering length side for smaller mass ratios until it ceases to exist altogether . however , to prove this hypothesis , more data points are necessary . comparing the curves for the two mass ratios , one can see that the @xmath114 mixture has much shallower bound states , approaching the dimer threshold with an extremely small slope . this holds true even when the three - body potential strength is adjusted such that the trimer vanishes at the same value for @xmath113 for both mixtures . in addition , in both cases the tetramer and the trimer seem to vanish at the same point . in the next section , we show our results for this point with much higher resolution and discuss it in more detail . trimers and @xmath28 tetramers in a @xmath19cs-@xmath120li mixture ( lower group of lines ) and a @xmath121li-@xmath122li mixture ( upper group of lines ) . the blue dotted line marks the @xmath33 dimer + two @xmath22 atom threshold , the red solid lines the @xmath35 trimer + @xmath22 atom threshold , the dashed green lines the @xmath28 tetramer ground states , and the dot - dashed black line the excited state of the tetramer . to increase visibility of the features of the plot , the axes are rescaled by taking the fourth root . to this end , @xmath22 and @xmath116 are introduced as @xmath117 and @xmath118 ( cf . @xcite ) . , width=453 ] dimer threshold ( dotted line ) . @xmath123 is the @xmath33 dimer energy and @xmath98 is the natural energy scale defined in eq . . the inset shows a larger region of @xmath113 . we included our estimate ( solid black line ) of where the next excited efimov state should appear if the scaling factor were 20 ( see discussion in sec . [ resolutionestimate ] ) . we did not find evidence for such a state.,width=411 ] next we focus on the tetramer spectrum close to the atom - dimer threshold in the region where the trimer disappears . this region has not been studied in a full four - body approach for heteronuclear systems . the corresponding results are shown in figs . [ fig:133 - 6 ] and [ fig:87 - 7 ] , respectively . in all three cases ( @xmath124 ) , ( @xmath125 ) , and ( @xmath126 ) , the tetramer disappears through the dimer threshold after the trimer has already disappeared . however , the points in @xmath113 where the trimer and tetramer disappear approach each other as the mass ratio @xmath2 is decreased and the two states almost disappear at the same point in the ( @xmath126 ) case . this is a qualitatively new behavior that is not seen in systems with identical bosons . it might suggest that for equal masses @xmath127 , the two states disappear at the same value of @xmath113 . we have further investigated this below . if the mass ratio @xmath2 is even further decreased , we enter the regime of an inverted mass ratio where the scaling factor grows rapidly . in the current study we have not investigated this region . to quantify the behavior of the threshold crossings , we have extracted the difference @xmath128 between the points @xmath129 , denoting the value of @xmath113 where the @xmath28 tetramer vanishes into the @xmath33+@xmath22+@xmath22-threshold , and @xmath130 , the point where the @xmath35 trimer vanishes into the @xmath33+@xmath22-threshold , from figs . [ fig:133 - 6 ] and [ fig:87 - 7 ] . we show this coefficient as a function of the mass ratio @xmath2 in fig . [ fig : thresh ] . as a function of the mass ratio @xmath131 ( circle , square , triangle ) . the dashed ( solid ) line shows effective field theory calculations for an effective three - body system of an @xmath33 dimer and two heavy bosons @xmath22 of mass @xmath1 fitted to reproduce the point @xmath132 ( @xmath133 ) . , width=453 ] our results for the mass ratios @xmath131 are shown by the circle , square , and triangle , respectively . the errors in the extracted values are relatively large , but our calculations indicate that the relation is not linear . in order to understand this pattern , we have performed effective field theory calculations for effective three - body systems of an @xmath33 dimer of mass @xmath134 and two @xmath22 bosons of mass @xmath1 , where only the dimer and the bosons interact . using the effective field theory formalism for unequal mass particles with resonant interaction discussed in ref . @xcite , we can calculate the coefficient @xmath135 up to one free parameter , namely the effective three - body parameter of the @xmath33-@xmath22-@xmath22 system . ( see appendix [ app : eft ] for a brief discussion of this framework . ) therefore we can fit the effective three - body parameter either to the point @xmath132 ( dashed line ) or @xmath133 ( solid line ) . in each case the point @xmath136 is reproduced within error bands . a fit that reproduces all three points is not possible . this shows the limitations of the effective three - body calculation which is strictly valid only near @xmath137 . the effective three - body calculation shows that @xmath135 vanishes for @xmath138 , and is very small but finite at @xmath127 . we emphasize that this effective three - body approach is not exact and should only be valid near the point where the trimer vanishes through the @xmath33 dimer threshold . an estimate of the @xmath33-@xmath22 scattering length indicates that its value is at least two orders of magnitude larger for the case @xmath136 compared to the other two mass ratios . thus we expect the effective three - body picture to work best for this case . the tension between effective three - body and the gem calculations could be due to two potential problems . first , the gem calculations could be not as accurate as we estimate , possibly due to neglected higher partial wave contributions or hard to detect numerical pathologies . second , some essential aspect of the system might not be present in the effective three - body description , which might also explain why different approximation schemes yield different results as discussed in the next section . to discuss the effective efimov states that were already mentioned in the introduction , three aspects are important . to begin with , the region where they can occur has to be established . for discussions on whether or not they can be seen in numerical calculations or even in experiments , an important issue is the predicted scaling factor . we will briefly present two competing approximation schemes that lead to very different predictions . to conclude , we will comment on the current state of the evidence for the effective efimov states . near the dimer threshold , both trimer and tetramer become very weakly bound in comparison to the dimer . for this reason it is expected that at a certain point the system can be treated as an effective three - body problem . at the point were the trimer becomes unbound , the dimer - atom interaction is resonant , leading to the occurrence of the efimov effect in the effective three - body system . therefore , the effective efimov effect is expected to arise in a small region around the point where the trimer crosses the dimer threshold . this region is bounded by the tetramer ground state from below . as we have shown in the previous section , this region ( the size of which can be expressed by @xmath139 ) proved to be extremely small in our calculations . the effective efimov effect has already been studied in systems of four identical bosons by deltuva @xcite and in heteronuclear systems by wang et al . @xcite . for four identical bosons , deltuva found @xmath140 , which compares well to our value for @xmath141li/@xmath120li , @xmath142 . for equal masses , our effective three - body calculations yielded values between @xmath143 and @xmath144 , depending on the three - body parameter ( compare fig . [ fig : thresh ] and the discussion in the previous section ) . assuming that the system is away from the unitary limit and close to the dimer threshold the dimer - atom scattering length is much larger than the scattering length between identical bosons , which means that the system studied in @xcite is comparable to our calculations . deltuva also did effective three - body dimer - atom - atom calculations and showed that very close to the dimer - atom - atom threshold this approximation is valid . his estimated scaling factor for the effective efimov states is @xmath145 , which agrees with @xcite and our discussion in the next section . in a paper from 2012 , wang et al . @xcite used the born - oppenheimer approximation in heteronuclear systems with very high mass ratios ( @xmath146 and @xmath147 ) to calculate recombination features . they calculated trimer ( @xmath38 ) , tetramer ( @xmath148 ) and also @xmath34 states as effective two- and three - body problems . they also used correlated gaussian results to check their calculations where possible . moreover , they provide a qualitative sketch of the heteronuclear efimov spectrum . we note , however , that this sketch is somewhat misleading concerning the window where dimer - atom - atom efimov states could form , as its size is exaggerated in light of our results . comparing their results to ours , several differences can be noted . from fig . 2 in @xcite we extracted their @xmath139 for @xmath149 , which is roughly @xmath150 , an order of magnitude larger than for our calculations ( extrapolations via the effective three - body calculations to @xmath146 suggest something of order @xmath151 ) . @xcite , separable potentials of yamaguchi type between the @xmath33 pairs were used @xcite . so there are three obvious possible reasons for the discrepancy , a ) @xmath139 is not universal and in fact depends on the @xmath33 potential , b ) the bo approximation is problematic in this region , as may be suggested by the fact that the scaling factor between consecutive efimov states is 5.5 in their calculations , while the universal value is 3.96 @xcite , and c ) the effective three - body calculations we used are not valid for higher mass ratios because of the larger @xmath139 value , and therefore the extrapolation to @xmath149 is not sound . to rule out a ) , one would have to use a different potential and redo the calculations . moreover , it is unclear whether the calculations of wang et al . @xcite are conducted closer to the unitary limit than ours , but since they state that @xmath152 is not well satisfied in their calculations , we suppose @xmath153 is of comparable size for both their calculations and ours ( we have @xmath154 for the ( @xmath155 ) and ( @xmath156 ) systems , and @xmath157 for the ( @xmath158 ) system ) . in addition , because @xmath139 is very similar for the excited and ground state trimer in their calculation , and of roughly the same order of magnitude for different values of @xmath153 in our calculations , a strong dependence on @xmath153 is not likely . following the argument b ) , one might also add that our effective three - body calculation did not agree too well with our full four - body calculation , as shown in the previous section . ( note that the order of magnitude of @xmath139 is a fit parameter in our stm treatment , so it agrees by construction . ) this might indicate that approximations that reduce the problem to an effective three - body system are generally insufficient to accurately describe the behavior near the dimer threshold . aspect c ) should not play a huge role because the atom - dimer scattering length @xmath159 is estimated to be larger than @xmath160 for @xmath161 and does not change much between @xmath162 and @xmath163 , so extrapolating to @xmath146 should yield reasonable accuracy . the dimer - atom - atom efimov effect is conjectured to arise when the four - body system of @xmath148 can be treated as an effective three - body system that has an effective resonant interaction . this is expected to be the case in the vicinity where the @xmath38 three - body state vanishes into the dimer threshold . there , braaten and hammer @xcite argued , we can regard the @xmath33 dimer as much more strongly bound than the trimer , allowing us to treat the trimer as an effective two - body system @xmath164 with an effective @xmath165 scattering length @xmath159 which can be estimated by the inverse square root of the binding energy of the @xmath38 trimer . then the @xmath148 tetramer can be treated as an effective three - body system @xmath34 which is also governed by the dimer - atom scattering length @xmath159 . for large @xmath159 this leads to the emergence of the efimov effect . the scaling factor expected in this case is roughly 2000 , because there are only two interacting pairs @xmath165 , and there are two `` light '' particles @xmath22 and one that is slightly heavier @xmath166 . the scenario with the smallest scaling factor is @xmath167 , in which case it is 1986.1 @xcite . this is the picture we have been using in our stm calculations shown in fig . [ fig : thresh ] . because of the huge scaling factor , we did not expect to be able to find these efimov states with the accuracy we have reached . however , in the born - oppenheimer approximation picture that wang et al . @xcite used , the light atom @xmath24 is seen as mediating the interaction between the heavy bosons @xmath22 , leading to an effective three - body problem of three @xmath22 s , with an effective scattering length @xmath168 . this effective scattering length diverges for the values of @xmath169 where an efimov trimer becomes unbound , so it shows the same behavior as @xmath159 . but , since in this picture the @xmath24 is absorbed in the effective interaction , a symmetric system of three resonantly interacting @xmath22 s emerges . the efimov effect for this system , which is the well known three identical boson case , yields a much smaller scaling factor of @xmath170 @xcite . using the born - oppenheimer approximation , wang et al . @xcite find somewhat tenuous evidence for an excited tetramer separated from the lower tetramer by a scaling factor of @xmath171 on the positive @xmath168 side and @xmath172 on the negative side . this is in disagreement with the expectation from the eft approach outlined above . the main difference between the two approaches lies in the way the separation of scales is conducted . in the first approach , it is assumed that the @xmath33 dimer is much more deeply bound than the @xmath38 and @xmath148 states , allowing a separation of energy scales and associated length scales . this assumption is found to be valid , because @xmath173 for the whole region shown in figs . [ fig:133 - 6 ] and [ fig:87 - 7 ] . on the other hand , in the bo approximation it is assumed that the time scale of the motion of the light atom is much shorter than that of the heavy atoms , allowing us to solve the dynamics of the light atom for fixed heavy atoms and then solving the heavy atoms system independently . this assumes that the light atom interacts with all the heavy atoms while they are fixed at some points , thus generating an effective interaction . if however the light atom is very tightly bound to one heavy atom , as is the case here , its motion around the heavy atom might be on a shorter timescale than the motion of the heavy atom , but its motion towards the other heavy atoms is coupled to the motion of the heavy atom , which means it is on the same timescale and therefore not separable . by using the bo approximation one implies in the calculation that the dimer is not very deeply bound . this means that in our understanding , both effective treatments are mutually exclusive and have to be applied to different systems . if we regard the two approximations as two ends of a spectrum , the bo approximation would apply to the case where the dimer and the trimer are both very weakly bound , and the eft approximation would apply to the case where the dimer is infinitely strongly bound . between these two points , we conjecture there is a transition between the three identical boson case with its relatively small scaling factor which then becomes larger and larger until it reaches the limit of two light bosons with one heavy boson . for the case we are studying here , this could mean that the scaling factor is slightly smaller than we would expect from our stm calculation . this however can only be explored using a full four - body calculation that is accurate enough to resolve these effective efimov states . to estimate whether it would be possible to see the effective efimov states in our calculations , we made some further assumptions . first , we assumed that the universal tetramer can be interpreted at the same time as the lowest state of the effective efimov states . we do not expect this state to be fully universal , so the scaling factor between this state and the next may be smaller or larger , but we expect it to be of the same order of magnitude . this assumption was apparently also made by wang et al . @xcite , although they did not mention it explicitly . the second assumption is that the shape of the first excited effective efimov state would resemble the shape of the universal tetramer near the threshold . then we could scale the tetramer in figs . [ fig:133 - 6 ] and [ fig:87 - 7 ] using different scaling factors to assess whether this state could be found with our accuracy . an example for a scaling factor of 20 is shown in fig . [ fig:133 - 6 ] . our four - body calculations presented here are by far not accurate enough to find effective efimov states with a scaling factor of 2000 . but our accuracy might just be sufficient to detect effective efimov states if the scaling factor was indeed approximately 20 as suggested by wang et al . we did not find any states , but since they would be just at the limit of our resolution , this is no conclusive evidence of their non - existence . in this paper , we have presented a detailed study of the bound state spectrum of heteronuclear four - body systems of three heavy bosons @xmath22 and a light atom @xmath24 with resonant interspecies interaction in the region of positive scattering lengths . the interaction between the heavy bosons was assumed to be negligible . to obtain the energies of the @xmath38 trimer and @xmath28 tetramer systems , we employed the gaussian expansion method ( gem ) @xcite . the general structure of the spectrum for different mass ratios was summarized in the efimov plot of fig . [ fig : overview ] , a notable feature being that there is no excited efimov tetramer at all for the mass ratio @xmath114 . as the mass ratio increases , the excited state tetramer appears on the negative scattering length side of the efimov plot . for @xmath31 it exists also in the unitary limit and disappears through the trimer - atom threshold on the positive scattering length side of the efimov plot . our results confirm and extend the findings of blume and yan @xcite , who conducted calculations down to @xmath174 . next we focused on the dimer - atom(-atom ) threshold region . we found that the value of the scattering length where the trimer vanishes into the threshold and the point where the tetramer does the same are extremely close to each other . we have investigated the dependence of the distance between these points , @xmath135 , on the mass ratio @xmath2 and presented a detailed discussion of the sources of error in our calculation , such as a possible dependence of the results on the potential shape and missing higher partial wave contributions . we have estimated the size of errors from the latter to not be larger than @xmath104 near the dimer - atom - atom threshold . the correlation between @xmath135 and @xmath2 does not appear to be linear , and the results from the four - body calculation could be reconciled only partly with effective three - body results in the framework of a stm - like equation derived from effective field theory @xcite . we conclude that either an important aspect of the four - body system is missing in the effective three - body calculation close to the dimer - atom - atom threshold , or the errors of our four - body results are underestimated . the resolution of this question requires further work , such as repeating the calculations for different potential shapes , to test the influence on @xmath135 and the general shape of the tetramer near the threshold . in order to rule out finite range effects a study of higher excited trimers and the tetramer resonances belonging to them would be useful . in the last part , we discussed in depth the possibility of finding effective dimer - atom - atom efimov states near the dimer - atom - atom threshold . we did not observe such states and estimated it is unlikely to find them without a major improvement of the accuracy in our calculation . wang and collaborators @xcite have previously found such states , albeit only in effective three - body calculations within the born - oppenheimer approximation . we have compared our results to the results of @xcite and highlighted some important differences between the two approaches . the key question in reconciling the results involves the scaling factor for the effective efimov states . we expect a large scaling factor of order 2000 since only the @xmath33 dimer and the @xmath22 atoms of the effective three - body system are interacting resonantly . we estimated that for this scaling factor an observation of the effective efimov states in our calculation is very unlikely . in the born - oppenheimer picture , however , the light @xmath24 atom mediates an interaction between all three @xmath22 atoms such that there are effectively three resonant pair interactions and the scaling factor should be much smaller and close to 22.7 . as we discussed in detail , it is not obvious that the born - oppenheimer picture is still applicable near the dimer - atom - atom threshold . settling the question of the effective dimer - atom - atom efimov states in heteronuclear systems requires a significant increase in the accuracy of our four - body calculations and a coordinated study of the problem in both approaches . we acknowledge useful discussions with doerte blume on heteronuclear few - body systems and practical units , and with artem volosniev on the limitations of the born - oppenheimer approximation . we also thank doerte blume for providing data from her figures from @xcite . this research was supported in part by the deutsche forschungsgemeinschaft through sfb 1245 and by the german federal ministry of education and research under contract 05p15rdfn1 . thanks riken for support under the ipa program and the studienstiftung des deutschen volkes for support under the riken program . some of the calculations were carried out at supercomputing facilities of the yitp ( kyoto university ) and of nagoya university . an effective field theory for heteronuclear three - body systems consisting of an atom of mass @xmath175 and two identical bosonic atoms of mass @xmath176 was developed by helfrich et al . @xcite . this effective field theory leads to an analog of the skorniakov - ter - martirosian ( stm ) equation @xcite for the unequal mass case . similar equations also arise in the effective field theory description of two - neutron halo nuclei in nuclear physics @xcite . here we focus on the case of zero - range interactions between the unlike atoms characterized by a scattering length @xmath177 and no interaction between the like atoms ( @xmath178 ) . the trimer binding energies @xmath179 in this framework are given by the non - trivial solutions of the integral equation : @xmath180 where @xmath181 and @xmath182 are the reduced masses of the @xmath183 and @xmath184 systems . the cutoff @xmath185 plays the role of the three - body parameter and can be used to fix the energy @xmath179 of any three - body state in the system . ) , the log - periodicity of the three - body force term in the effective field theory was used to move the dependence on the three - body parameter to the upper limit of the integral ( see ref . @xcite for more details ) . ] all other three - body energies can then be predicted . alternatively , the energies can be predicted in units of @xmath186 . the wave function of the trimers can be obtained from the function @xmath187 but will not be required here . e. braaten and h .- w . hammer , phys . rept . * 428 * , 259 ( 2006 ) [ arxiv : cond - mat/0410417 ] . l. platter , few body syst . * 46 * , 139 ( 2009 ) [ arxiv:0904.2227 ] . p. naidon and s. endo , arxiv:1610.09805 [ quant - ph ] . e. nielsen and j.h . macek , phys . lett . * 83 * , 1566 ( 1999 ) . j. ulmanis , s. hfner , r. pires , e. d. kuhnle , y. wang , c. h. greene and m. weidemller , phys . * 117 * , 153201 ( 2016 ) [ arxiv:1608.01477 [ cond-mat.quant-gas ] ] . s. hfner , j. ulmanis , e. d. kuhnle , y. wang , c. h. greene , m. weidemller , arxiv:1701.08007 [ cond-mat.quant-gas ] . l. platter , h .- w . hammer and u. g. meiner , phys . a * 70 * , 052101 ( 2004 ) [ cond - mat/0404313 ] . a. zenesini , b. huang , m. berninger , s. besler , h .- c . ngerl , f. ferlaino , r. grimm , c. h. greene , and j. von stecher , new journal of physics * 15 * , 043040 ( 2013 ) [ arxiv:1205.1921 ] m. gattobigio , a. kievsky and m. viviani , phys . rev . a * 86 * , 042513 ( 2012 ) [ arxiv:1206.0854 [ physics.atm-clus ] ] . a. n. nicholson , phys . lett . * 109 * , 073003 ( 2012 ) [ arxiv:1202.4402 [ cond-mat.quant-gas ] ] . a. kievsky , n. k. timofeyuk and m. gattobigio , phys . a * 90 * , 032504 ( 2014 ) [ arxiv:1405.2371 [ cond-mat.quant-gas ] ] . d. blume and y. yan , phys . a * 92 * , 0333626 ( 2015 ) [ arxiv:1508.00081 [ cond-mat.quant-gas ] ] .

there are many routes to minkowski space . but the most physical one still seems to me via the law of inertia . and even along these lines alternative approaches exist . many papers were published in physics and mathematics journals over the last 100 years in which incremental progress was reported as regards the minimal set of hypotheses from which the structure of minkowski space could be deduced . one could imagine a hesse - diagram - like picture in which all these contributions ( being the nodes ) together with their logical dependencies ( being the directed links ) were depicted . it would look surprisingly complex . from a general - relativistic point of view , minkowski space just models an _ empty _ spacetime , that is , a spacetime devoid of any material content . it is worth keeping in mind , that this was not minkowski s view . close to the beginning of _ raum und zeit _ he stated : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in order to not leave a yawning void , we wish to imagine that at every place and at every time something perceivable exists . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ this already touches upon a critical point . our modern theoretical view of spacetime is much inspired by the typical hierarchical thinking of mathematics of the late 19th and first half of the 20th century , in which the _ set _ comes first , and then we add various structures on it . we first think of spacetime as a set and then structure it according to various physical inputs . but what are the elements of this set ? recall how georg cantor , in his first article on transfinite set - theory , defined a set : von bestimmten wohlunterschiedenen objecten @xmath0 unserer anschauung oder unseres denkens ( welche die ` elemente ' von @xmath1 genannt werden ) zu einem ganzen . '' ( @xcite , p.481 ) ] _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ by a ` set ' we understand any gathering - together @xmath1 of determined well - distinguished objects @xmath0 of our intuition or of our thinking ( which are called the ` elements ' of @xmath1 ) into a whole . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ do we think of spacetime points as `` determined well - distinguished objects of our intuition or of our thinking '' ? i think minkowski felt a need to do so , as his statement quoted above indicates , and also saw the problematic side of it : if we mentally individuate the points ( elements ) of spacetime , we as physicists have no other means to do so than to fill up spacetime with actual matter , hoping that this could be done in such a diluted fashion that this matter will not dynamically affect the processes that we are going to describe . in other words : the whole concept of a rigid background spacetime is , from its very beginning , based on an assumption of at best approximate validity . it is important to realise that this does not necessarily refer to general relativity : even if the need to incorporate gravity by a variable and matter - dependent spacetime geometry did not exist would the concept of a rigid background spacetime be of approximate nature , _ provided we think of spacetime points as individuated by actual physical events_. it is true that modern set theory regards cantor s original definition as too nave , and that for good reasons . it allows too many `` gatherings - together '' with self - contradictory properties , as exemplified by the infamous _ antinomies _ of classical set theory . also , modern set theory deliberately stands back from any characterisation of elements in order to not confuse the axioms themselves with their possible _ interpretations_. however , applications to physics require _ interpreted _ axioms , where it remains true that elements of sets are thought of as definite as in cantors original definition . modern textbooks on special relativity have little to say about this , though an increasing unease seems to raise its voice from certain directions in the philosophy - of - science community ; see , e.g. , @xcite@xcite . physicists sometimes tend to address points of spacetime as _ potential events _ , but that always seemed to me like poetry , begging the question how a mere potentiality is actually used for individuation . to me the right attitude seems to admit that the operational justification of the notion of spacetime events is only approximately possible , but nevertheless allow it as primitive element of theorising . the only thing to keep in mind is to not take mathematical rigour for ultimate physical validity . the purpose of mathematical rigour is rather to establish the tightest possible bonds between basic assumptions ( axioms ) and decidable consequences . only then can we in principle learn anything through falsification . the last remark opens another general issue , which is implicit in much of theoretical research , namely how to balance between attempted rigour in drawing consequences and attempted closeness to reality when formulating once starting platform ( at the expense of rigour when drawing consequences ) . as the mathematical physicists glance & wightman once formulated it in a different context ( that of superselection rules in quantum mechanics ) : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the theoretical results currently available fall into two categories : rigorous results on approximate models and approximate results in realistic models . _ ( @xcite , p.204 ) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ to me this seems to be the generic situation in theoretical physics . in that respect , minkowski space is certainly an approximate model , but to a very good approximation indeed : as global model of spacetime if gravity plays no dynamical rle , and as local model of spacetime in far more general situations . this justifies looking at some of its rich mathematical structures in detail . some mathematical background material is provided in the appendices . consider first the general situation where one is given a set @xmath2 . without any further structure being specified , the _ automorphisms _ group of @xmath2 would be the group of bijections of @xmath2 , i.e. maps @xmath3 which are injective ( into ) and surjective ( onto ) . it is called @xmath4 , where ` perm ' stands for ` permutations ' . now endow @xmath2 with some structure @xmath5 ; for example , it could be an equivalence relation on @xmath2 , that is , a partition of @xmath2 into an exhaustive set of mutually disjoint subsets ( cf . sect.[sec : groupactions ] ) . the automorphism group of @xmath6 is then the subgroup of @xmath7 that preserves @xmath5 . note that @xmath8 contains only those maps @xmath9 preserving @xmath5 whose inverse , @xmath10 , also preserve @xmath5 . now consider another structure , @xmath11 , and form @xmath12 . one way in which the two structures @xmath5 and @xmath11 may be compared is to compare their automorphism groups @xmath8 and @xmath12 . comparing the latter means , in particular , to see whether one is contained in the other . containedness clearly defines a partial order relation on the set of subgroups of @xmath4 , which we can use to define a partial order on the set of structures . one structure , @xmath5 , is said to be strictly stronger than ( or equally strong as ) another structure , @xmath11 , in symbols @xmath13 , iff the automorphism group of the former is properly contained in ( or is equal to ) the automorphism group of the latter . here . ] in symbols : @xmath14 . note that in this way of speaking a substructure ( i.e. one being defined by a subset of conditions , relations , objects , etc . ) of a given structure is said to be weaker than the latter . this way of thinking of structures in terms of their automorphism group is adopted from felix klein s _ erlanger programm _ @xcite in which this strategy is used in an attempt to classify and compare geometries . this general procedure can be applied to minkowski space , endowed with its usual structure ( see below ) . we can than ask whether the automorphism group of minkowski space , which we know is the inhomogeneous lorentz group @xmath15 , also called the poincar group , is already the automorphism group of a proper substructure . if this were the case we would say that the original structure is redundant . it would then be of interest to try and find a minimal set of structures that already imply the poincar group . this can be done by trial and error : one starts with some more or less obvious substructure , determine its automorphism group , and compare it to the poincar group . generically it will turn out larger , i.e. to properly contain @xmath15 . the obvious questions to ask then are : how much larger ? and : what would be a minimal extra condition that eliminates the difference ? these questions have been asked in connection with various substructures of minkowski space , whose definition is as follows : [ def : minkowskispace ] * minkowski space * of @xmath16 dimensions , denoted by @xmath17 , is a real @xmath18-dimensional affine space , whose associated real @xmath18-dimensional vector space @xmath19 is endowed with a non - degenerate symmetric bilinear form @xmath20 of signature @xmath21 ( i.e. there exists a basis @xmath22 of @xmath19 such that @xmath23 ) . @xmath17 is also endowed with the standard differentiable structure of @xmath24 . we refer to appendix[sec : affinespaces ] for the definition of affine spaces . note also that the last statement concerning differentiable structures is put in in view of the strange fact that just for the physically most interesting case , @xmath25 , there exist many inequivalent differentiable structures of @xmath26 . finally we stress that , at this point , we did not endow minkowski space with an orientation or time orientation . [ def : poincaregroup ] the * poincar group * in @xmath16 dimensions , which is the same as the * inhomogeneous lorentz group * in @xmath16 dimensions and therefore will be denoted by @xmath27 , is that subgroup of the general affine group of real @xmath18-dimensional affine space , for which the uniquely associated linear maps @xmath28 are elements of the lorentz group @xmath29 , that is , preserve @xmath30 in the sense that @xmath31 for all @xmath32 . see appendix[sec : affinemaps ] for the definition of affine maps and the general affine group . again we stress that since we did not endow minkowski space with any orientation , the poincar group as defined here would not respect any such structure . as explained in [ sec : affineframes ] , any choice of an affine frame allows us to identify the general affine group in @xmath18 dimensions with the semi - direct product @xmath33 . that identification clearly depends on the choice of the frame . if we restrict the bases to those where @xmath23 , then @xmath27 can be identified with @xmath34 . we can further endow minkowski space with an _ orientation _ and , independently , a _ time orientation_. an orientation of an affine space is equivalent to an orientation of its associated vector space @xmath19 . a time orientation is also defined trough a time orientation of @xmath19 , which is explained below . the subgroup of the poincar group preserving the overall orientation is denoted by @xmath35 ( proper poincar group ) , the one preserving time orientation by @xmath36 ( orthochronous poincar group ) , and @xmath37 denotes the subgroup preserving both ( proper orthochronous poincar group ) . upon the choice of a basis we may identify @xmath35 with @xmath38 and @xmath37 with @xmath39 , where @xmath40 is the component of the identity of @xmath41 . let us add a few more comments about the elementary geometry of minkowski space . we introduce the following notations : @xmath42 we shall also simply write @xmath43 for @xmath44 . a vector @xmath45 is called _ timelike , lightlike _ , or _ spacelike _ according to @xmath43 being @xmath46 , @xmath47 , or @xmath48 respectively . non - spacelike vectors are also called _ causal _ and their set , @xmath49 , is called the _ causal - doublecone_. its interior , @xmath50 , is called the _ chronological - doublecone _ and its boundary , @xmath51 , the _ light - doublecone _ : [ eq : vecdc ] @xmath52 a linear subspace @xmath53 is called timelike , lightlike , or spacelike according to @xmath54 being indefinite , negative semi - definite but not negative definite , or negative definite respectively . instead of the usual cauchy - schwarz - inequality we have [ eq : cauchyschwarz ] @xmath55 given a set @xmath56 ( not necessarily a subspace ) , its @xmath30-orthogonal complement is the subspace @xmath57 if @xmath45 is lightlike then @xmath58 . in fact , @xmath59 is the unique lightlike hyperplane ( cf . sect.[sec : affinespaces ] ) containing @xmath60 . in this case the hyperplane @xmath59 is called degenerate because the restriction of @xmath30 to @xmath59 is degenerate . on the other hand , if @xmath60 is timelike / spacelike @xmath59 is spacelike / timelike and @xmath61 . now the hyperplane @xmath59 is called non - degenerate because the restriction of @xmath30 to @xmath59 is non - degenerate . given any subset @xmath56 , we can attach it to a point @xmath62 in @xmath63 : @xmath64 in particular , the causal- , chronological- , and light - doublecones at @xmath65 are given by : [ eq : affdc ] @xmath66 if @xmath67 is a subspace of @xmath19 then @xmath68 is an affine subspace of @xmath63 over @xmath67 . if @xmath67 is time- , light- , or spacelike then @xmath68 is also called time- , light- , or spacelike . of particular interest are the hyperplanes @xmath69 which are timelike , lightlike , or spacelike according to @xmath60 being spacelike , lightlike , or timelike respectively . two points @xmath70 are said to be timelike- , lightlike- , or spacelike separated if the line joining them ( equivalently : the vector @xmath71 ) is timelike , lightlike , or spacelike respectively . non - spacelike separated points are also called causally separated and the line though them is called a causal line . it is easy to show that the relation @xmath72 defines an equivalence relation ( cf . [ sec : groupactions ] ) on the set of timelike vectors . ( only transitivity is non - trivial , i.e. if @xmath73 and @xmath74 then @xmath75 . to show this , decompose @xmath76 and @xmath77 into their components parallel and perpendicular to @xmath60 . ) each of the two equivalence classes is a _ cone _ in @xmath19 , that is , a subset closed under addition and multiplication with positive numbers . vectors in the same class are said to have the same time orientation . in the same fashion , the relation @xmath78 defines an equivalence relation on the set of causal vectors , with both equivalence classes being again cones . the existence of these equivalence relations is expressed by saying that @xmath63 is _ time orientable_. picking one of the two possible time orientations is then equivalent to specifying a single timelike reference vector , @xmath79 , whose equivalence class of directions may be called the _ future_. this being done we can speak of the future ( or forward , indicated by a superscript @xmath80 ) and past ( or backward , indicated by a superscript @xmath81 ) cones : [ eq : vecc ] @xmath82 note that @xmath83 and @xmath84 . usually @xmath85 is called the future and @xmath86 the past lightcone . mathematically speaking this is an abuse of language since , in contrast to @xmath87 and @xmath88 , they are not cones : they are each invariant ( as sets ) under multiplication with positive real numbers , but adding to vectors in @xmath89 will result in a vector in @xmath88 unless the vectors were parallel . as before , these cones can be attached to the points in @xmath63 . we write in a straightforward manner : [ eq : affc ] @xmath90 the cauchy - schwarz inequalities ( [ eq : cauchyschwarz ] ) result in various generalised triangle - inequalities . clearly , for spacelike vectors , one just has the ordinary triangle inequality . but for causal or timelike vectors one has to distinguish the cases according to the relative time orientations . for example , for timelike vectors of equal time orientation , one obtains the reversed triangle inequality : @xmath91 with equality iff @xmath60 and @xmath77 are parallel . it expresses the geometry behind the ` twin paradox ' . sometimes a minkowski ` distance function ' @xmath92 is introduced through @xmath93 clearly this is not a distance function in the ordinary sense , since it is neither true that @xmath94 nor that @xmath95 for all @xmath96 . in this section we consider general isometries of minkowski space . by this we mean general bijections @xmath97 ( no requirement like continuity or even linearity is made ) which preserve the minkowski distance ( [ eq : def : minkowskidistance ] ) as well as the time or spacelike character ; hence @xmath98 poincar transformations form a special class of such isometries , namely those which are affine . are there non - affine isometries ? one might expect a whole pandora s box full of wild ( discontinuous ) ones . but , fortunately , they do not exist : any map @xmath28 satisfying @xmath99 for all @xmath60 must be linear . as a warm up , we show [ thm : nowildisometries ] let @xmath28 be a surjection ( no further conditions ) so that @xmath100 for all @xmath32 , then @xmath9 is linear . consider @xmath101 . surjectivity allows to write @xmath102 , so that @xmath103 , which vanishes for all @xmath104 . hence @xmath105 for all @xmath106 , which by non - degeneracy of @xmath30 implies the linearity of @xmath9 . this shows in particular that any bijection @xmath97 of minkowski space whose associated map @xmath28 , defined by @xmath107 for some chosen basepoint @xmath108 , preserves the minkowski metric must be a poincar transformation . as already indicated , this result can be considerably strengthened . but before going into this , we mention a special and important class of linear isometries of @xmath109 , namely reflections at non - degenerate hyperplanes . the reflection at @xmath59 is defined by @xmath110 their significance is due to the following let the dimension of @xmath19 be @xmath18 . any isometry of @xmath109 is the composition of at most @xmath18 reflections . comprehensive proofs may be found in @xcite or @xcite . the easier proof for at most @xmath111 reflections is as follows : let @xmath112 be a linear isometry and @xmath45 so that @xmath113 ( which certainly exists ) . let @xmath114 , then @xmath115 so that @xmath116 and @xmath117 can not simultaneously have zero squares . so let @xmath118 ( understood as alternatives ) , then @xmath119 and @xmath120 . hence @xmath60 is eigenvector with eigenvalue @xmath121 of the linear isometry given by @xmath122 consider now the linear isometry @xmath123 on @xmath59 with induced bilinear form @xmath124 , which is non - degenerated due to @xmath113 . we conclude by induction : at each dimension we need at most two reflections to reduce the problem by one dimension . after @xmath125 steps we have reduced the problem to one dimension , where we need at most one more reflection . hence we need at most @xmath126 reflections which upon composition with @xmath112 produce the identity . here we use that any linear isometry in @xmath59 can be canonically extended to @xmath127 by just letting it act trivially on @xmath128 . note that this proof does not make use of the signature of @xmath30 . in fact , the theorem is true for any signatures ; it only depends on @xmath30 being symmetric and non degenerate . as already mentioned , theorem[thm : nowildisometries ] can be improved upon , in the sense that the hypothesis for the map being an isometry is replaced by the hypothesis that it merely preserve some relation that derives form the metric structure , but is not equivalent to it . in fact , there are various such relations which we first have to introduce . the family of cones @xmath129 defines a partial - order relation ( cf . sect.[sec : groupactions ] ) , denoted by @xmath130 , on spacetime as follows : @xmath131 iff @xmath132 , i.e. iff @xmath71 is causal and future pointing . similarly , the family @xmath133 defines a strict partial order , denoted by @xmath134 , as follows : @xmath135 iff @xmath136 , i.e. if @xmath71 is timelike and future pointing . there is a third relation , called @xmath137 , defined as follows : @xmath138 iff @xmath139 , i.e. @xmath62 is on the future lightcone at @xmath140 . it is not a partial order due to the lack of transitivity , which , in turn , is due to the lack of the lightcone being a cone ( in the proper mathematical sense explained above ) . replacing the future ( @xmath80 ) with the past ( @xmath81 ) cones gives the relations @xmath141 , @xmath142 , and @xmath143 . it is obvious that the action of @xmath144 ( spatial reflections are permitted ) on @xmath63 maps each of the six families of cones ( [ eq : affc ] ) into itself and therefore leave each of the six relations invariant . for example : let @xmath135 and @xmath145 , then @xmath146 and @xmath71 future pointing , but also @xmath147 and @xmath148 future pointing , hence @xmath149 . another set of ` obvious ' transformations of @xmath63 leaving these relations invariant is given by all dilations : @xmath150 where @xmath151 is the constant dilation - factor and @xmath152 the centre . this follows from @xmath153 , @xmath154 , and the positivity of @xmath155 . since translations are already contained in @xmath144 , the group generated by @xmath144 and all @xmath156 is the same as the group generated by @xmath144 and all @xmath156 for fixed @xmath0 . a seemingly difficult question is this : what are the most general transformations of @xmath63 that preserve those relations ? here we understand ` transformation ' synonymously with ` bijective map ' , so that each transformation @xmath9 has in inverse @xmath10 . ` preserving the relation ' is taken to mean that @xmath9 _ and _ @xmath10 preserve the relation . then the somewhat surprising answer to the question just posed is that , in three or more spacetime dimensions , there are no other such transformations besides those already listed : let @xmath157 stand for any of the relations @xmath158 and let @xmath159 be a bijection of @xmath63 with @xmath160 , such that @xmath161 implies @xmath162 and @xmath163 . then @xmath159 is the composition of an lorentz transformation in @xmath144 with a dilation . these results were proven by a.d.alexandrov and independently by e.c . a good review of alexandrov s results is @xcite ; zeeman s paper is @xcite . the restriction to @xmath160 is indeed necessary , as for @xmath164 the following possibility exists : identify @xmath165 with @xmath166 and the bilinear form @xmath167 , where @xmath168 . set @xmath169 and @xmath170 and define @xmath171 by @xmath172 , where @xmath173 is any smooth function with @xmath174 . this defines an orientation preserving diffeomorphism of @xmath166 which transforms the set of lines @xmath175 const . and @xmath176 const . respectively into each other . hence it preserves the families of cones ( [ eq : def : affcausalc ] ) . since these transformations need not be affine linear they are not generated by dilations and lorentz transformations . these results may appear surprising since without a continuity requirement one might expect all sorts of wild behaviour to allow for more possibilities . however , a little closer inspection reveals a fairly obvious reason for why continuity is implied here . consider the case in which a transformation @xmath159 preserves the families @xmath133 and @xmath177 . the open diamond - shaped sets ( usually just called ` open diamonds ' ) , @xmath178 are obviously open in the standard topology of @xmath63 ( which is that of @xmath179 ) . note that at least one of the intersections in ( [ eq : opendiamonds ] ) is always empty . conversely , is is also easy to see that each open set of @xmath63 contains an open diamond . hence the topology that is defined by taking the @xmath180 as sub - base ( the basis being given by their finite intersections ) is equivalent to the standard topology of @xmath63 . but , by hypothesis , @xmath159 and @xmath181 preserves the cones @xmath182 and therefore open sets , so that @xmath159 must , in fact , be a homeomorphism . there is no such _ obvious _ continuity input if one makes the strictly weaker requirement that instead of the cones ( [ eq : affc ] ) one only preserves the doublecones ( [ eq : affdc ] ) . does that allow for more transformations , except for the obvious time reflection ? the answer is again in the negative . the following result was shown by alexandrov ( see his review @xcite ) and later , in a different fashion , by borchers and hegerfeld @xcite : [ thm : borchershegerfeld ] let @xmath183 denote any of the relations : @xmath184 iff @xmath185 , @xmath184 iff @xmath186 , or @xmath184 iff @xmath187 . let @xmath159 be a bijection of @xmath63 with @xmath160 , such that @xmath184 implies @xmath188 and @xmath189 . then @xmath159 is the composition of an lorentz transformation in @xmath15 with a dilation . all this shows that , up to dilations , lorentz transformations can be characterised by the causal structure of minkowski space . let us focus on a particular sub - case of theorem[thm : borchershegerfeld ] , which says that any bijection @xmath159 of @xmath63 with @xmath160 , which satisfies @xmath190 must be the composition of a dilation and a transformation in @xmath15 . this is sometimes referred to as _ alexandrov s theorem_. it gives a precise answer to the following physical question : to what extent does the principle of the constancy of a finite speed of light _ alone _ determine the relativity group ? the answer is , that it determines it to be a subgroup of the 11-parameter group of poincar transformations and constant rescalings , which is as close to the poincar group as possibly imaginable . alexandrov s theorem is , to my knowledge , the closest analog in minkowskian geometry to the famous theorem of beckman and quarles @xcite , which refers to euclidean geometry and reads as follows under slightly weaker hypotheses : they allowed the map @xmath9 to be ` many - valued ' , that is , to be a map @xmath191 , where @xmath192 is the set of non - empty subsets of @xmath179 , such that @xmath193 for any @xmath194 and any @xmath195 . however , given the statement of theorem[thm : beckmanquarles ] , it is immediate that such ` many - valued maps ' must necessarily be single - valued . to see this , assume that @xmath196 has the two image points @xmath197 and define @xmath198 for @xmath199 such that @xmath200 for all @xmath201 and @xmath202 . then , according to theorem[thm : beckmanquarles ] , @xmath203 must both be euclidean motions . since they are continuous and coincide for all @xmath201 , they must also coincide at @xmath204 . ] : [ thm : beckmanquarles ] let @xmath179 for @xmath16 be endowed with the standard euclidean inner product @xmath205 . the associated norm is given by @xmath206 . let @xmath207 be any fixed positive real number and @xmath208 any map such that @xmath209 ; then @xmath9 is a euclidean motion , i.e. @xmath210 . note that there are three obvious points which let the result of beckman and quarles in euclidean space appear somewhat stronger than the theorem of alexandrov in minkowski space : * the conclusion of theorem[thm : beckmanquarles ] holds for _ any _ @xmath211 , whereas alexandrov s theorem singles out lightlike distances . * in theorem[thm : beckmanquarles ] , @xmath164 is not excluded . * in theorem[thm : beckmanquarles ] , @xmath9 is not required to be a bijection , so that we did not assume the existence of an inverse map @xmath10 . correspondingly , there is no assumption that @xmath10 also preserves the distance @xmath207 . in this subsection we wish to discuss the extent to which the law of inertia already determines the automorphism group of spacetime . the law of inertia privileges a subset of paths in spacetime form among all paths ; it defines a so - called _ path structure _ @xcite@xcite . these privileged paths correspond to the motions of privileged objects called _ free particles_. the existence of such privileged objects is by no means obvious and must be taken as a contingent and particularly kind property of nature . it has been known for long @xcite@xcite@xcite how to operationally construct timescales and spatial reference frames relative to which free particles will move uniformly and on straight lines respectively all of them ! ( a summary of these papers is given in @xcite . ) these special timescales and spatial reference frames were termed _ inertial _ by ludwig lange @xcite . their existence must again be taken as a very particular and very kind feature of nature . note that ` uniform in time ' and ` spatially straight ' together translate to ` straight in spacetime ' . we also emphasise that ` straightness ' of ensembles of paths can be characterised intrinsically , e.g. , by the desargues property @xcite . all this is true if free particles are given . we do not discuss at this point whether and how one should characterise them independently ( cf . @xcite ) . the spacetime structure so defined is usually referred to as _ projective_. it it not quite that of an affine space , since the latter provides in addition each straight line with a distinguished two - parameter family of parametrisations , corresponding to a notion of _ uniformity _ with which the line is traced through . such a privileged parametrisation of spacetime paths is not provided by the law of inertia , which only provides privileged parametrisations of spatial paths , which we already took into account in the projective structure of spacetime . instead , an affine structure of spacetime may once more be motivated by another contingent property of nature , shown by the existence of elementary clocks ( atomic frequencies ) which do define the same uniformity structure on inertial world lines all of them ! once more this is a highly non - trivial and very kind feature of nature . in this way we would indeed arrive at the statement that spacetime is an affine space . however , as we shall discuss in this subsection , the affine group already emerges as automorphism group of inertial structures without the introduction of elementary clocks . first we recall the main theorem of affine geometry . for that we make the following [ def : collineation ] three points in an affine space are called * collinear * iff they are contained in a single line . a map between affine spaces is called a * collineation * iff it maps each triple of collinear points to collinear points . note that in this definition no other condition is required of the map , like , e.g. , injectivity . the main theorem now reads as follows : [ thm : collareaffine ] a bijective collineation of a real affine space of dimension @xmath16 is necessarily an affine map . a proof may be found in @xcite . that the theorem is non - trivial can , e.g. , be seen from the fact that it is not true for complex affine spaces . the crucial property of the real number field is that it does not allow for a non - trivial automorphisms ( as field ) . a particular consequence of theorem[thm : collareaffine ] is that bijective collineation are necessarily continuous ( in the natural topology of affine space ) . this is of interest for the applications we have in mind for the following reason : consider the set @xmath212 of all lines in some affine space @xmath2 . @xmath212 has a natural topology induced from @xmath2 . theorem[thm : collareaffine ] now implies that bijective collineations of @xmath2 act as homeomorphism of @xmath212 . consider an open subset @xmath213 and the subset of all collineations that fix @xmath214 ( as set , not necessarily its points ) . then these collineations also fix the boundary @xmath215 of @xmath214 in @xmath212 . for example , if @xmath214 is the set of all timelike lines in minkowski space , i.e. , with a slope less than some chosen value relative to some fixed direction , then it follows that the bijective collineations which together with their inverse map timelike lines to timelike lines also maps the lightcone to the lightcone . it immediately follows that it must be the composition of a poincar transformation as a constant dilation . note that this argument also works in two spacetime dimensions , where the alexandrov - zeeman result does not hold . the application we have in mind is to inertial motions , which are given by lines in affine space . in that respect theorem[thm : collareaffine ] is not quite appropriate . its hypotheses are weaker than needed , insofar as it would suffice to require straight lines to be mapped to straight lines . but , more importantly , the hypotheses are also stronger than what seems physically justifiable , insofar as not every line is realisable by an inertial motion . in particular , one would like to know whether theorem[thm : collareaffine ] can still be derived by restricting to _ slow collineations _ , which one may define by the property that the corresponding lines should have a slope less than some non - zero angle ( in whatever measure , as long as the set of slow lines is open in the set of all lines ) from a given ( time-)direction . this is indeed the case , as one may show from going through the proof of theorem[thm : collareaffine ] . slightly easier to prove is the following : [ thm : timelikecollareaffine ] let @xmath159 be a bijection of real @xmath18-dimensional affine space that maps slow lines to slow lines , then @xmath159 is an affine map . a proof may be found in @xcite . if ` slowness ' is defined via the lightcone of a minkowski metric @xmath30 , one immediately obtains the result that the affine maps must be composed from poincar transformations and dilations . the reason is [ thm : samekernelpropto ] let @xmath19 be a finite dimensional real vector space of dimension @xmath16 and @xmath30 be a non - degenerate symmetric bilinear form on @xmath19 of signature @xmath21 . let @xmath216 be any other symmetric bilinear form on @xmath19 . the ` light cones ' for both forms are defined by @xmath217 and @xmath218 . suppose @xmath219 , then @xmath220 for some @xmath221 . let @xmath222 be a basis of @xmath19 such that @xmath223 . then @xmath224 for @xmath225 implies ( we write @xmath226 ) : @xmath227 and @xmath228 . further , @xmath229 for @xmath230 then implies @xmath231 for @xmath232 . hence @xmath220 with @xmath233 . this can be applied as follows : if @xmath234 is affine and maps lightlike lines to lightlike lines , then the associated linear map @xmath28 maps lightlike vectors to lightlike vectors . hence @xmath235 vanishes if @xmath236 vanishes and therefore @xmath237 by lemma[thm : samekernelpropto ] . since @xmath238 is timelike if @xmath60 is timelike , @xmath239 is positive . hence we may define @xmath240 and have @xmath241 for all @xmath242 , saying that @xmath243 is a lorentz transformation . @xmath9 is the composition of a lorentz transformation and a dilation by @xmath244 . as is well known , the two main ingredients in special relativity are the principle of relativity ( henceforth abbreviated by pr ) and the principle of the constancy of light . we have seen above that , due to alexandrov s theorem , the latter almost suffices to arrive at the poincar group . in this section we wish to address the complementary question : under what conditions and to what extent can the rp _ alone _ justify the poincar group ? this question was first addressed by ignatowsky @xcite , who showed that under a certain set of technical assumptions ( not consistently spelled out by him ) the rp alone suffices to arrive at a spacetime symmetry group which is either the inhomogeneous galilei or the inhomogeneous lorentz group , the latter for some yet undetermined limiting velocity @xmath245 . more precisely , what is actually shown in this fashion is , as we will see , that the relativity group must contain either the proper orthochronous galilei or lorentz group , if the group is required to comprise at least spacetime translations , spatial rotations , and boosts ( velocity transformations ) . what we hence gain is the group - theoretic insight of how these transformations must combine into a common group , given that they form a group at all . we do not learn anything about other transformations , like spacetime reflections or dilations , whose existence we neither required nor ruled out at this level . the work of ignatowsky was put into a logically more coherent form by franck & rothe @xcite@xcite , who showed that some of the technical assumptions could be dropped . further formal simplifications were achieved by berzi & gorini @xcite . below we shall basically follow their line of reasoning , except that we do not impose the continuity of the transformations as a requirement , but conclude it from their preservation of the inertial structure plus bijectivity . see also @xcite for an alternative discussion on the level of lie algebras . for further determination of the automorphism group of spacetime we invoke the following principles : * homogeneity of spacetime . * isotropy of space . * galilean principle of relativity . we take st1 to mean that the sought - for group should include all translations and hence be a subgroup of the general affine group . with respect to some chosen basis , it must be of the form @xmath246 , where @xmath247 is a subgroup of @xmath248 . st2 is interpreted as saying that @xmath249 should include the set of all spatial rotations . if , with respect to some frame , we write the general element @xmath250 in a @xmath251 split form ( thinking of the first coordinate as time , the other three as space ) , we want @xmath247 to include all @xmath252 finally , st3 says that velocity transformations , henceforth called ` boosts ' , are also contained in @xmath247 . however , at this stage we do not know how boosts are to be represented mathematically . let us make the following assumptions : * boosts @xmath253 are labelled by a vector @xmath254 , where @xmath255 is the open ball in @xmath256 of radius @xmath245 . the physical interpretation of @xmath257 shall be that of the boost velocity , as measured in the system from which the transformation is carried out . we allow @xmath245 to be finite or infinite ( @xmath258 ) . @xmath259 corresponds to the identity transformation , i.e. @xmath260 . we also assume that @xmath257 , considered as coordinate function on the group , is continuous . * as part of st2 we require equivariance of boosts under rotations : @xmath261 the latter assumption allows us to restrict attention to boost in a fixed direction , say that of the positive @xmath262-axis . once their analytical form is determined as function of @xmath60 , where @xmath263 , we deduce the general expression for boosts using ( [ eq : equivboostsrot ] ) and ( [ eq : so3imbedding ] ) . we make no assumptions involving space reflections . ) explicitly , though they all use it ( implicitly ) , usually in statements to the effect that it is sufficient to consider boosts in one fixed direction . once this restriction is effected , a one - dimensional spatial reflection transformation is considered to relate a boost transformation to that with opposite velocity . this then gives the impression that reflection equivariance is also invoked , though this is not necessary in spacetime dimensions greater than two , for ( [ eq : equivboostsrot ] ) allows to invert one axis through a 180-degree rotation about a perpendicular one . ] we now restrict attention to @xmath263 . we wish to determine the most general form of @xmath253 compatible with all requirements put so far . we proceed in several steps : 1 . using an arbitrary rotation @xmath264 around the @xmath262-axis , so that @xmath265 , equation ( [ eq : equivboostsrot ] ) allows to prove that @xmath266 where here we wrote the @xmath267 matrix in a @xmath268 decomposed form . ( i.e. @xmath269 is a @xmath270 matrix and @xmath271 is the @xmath270 unit - matrix ) . applying ( [ eq : equivboostsrot ] ) once more , this time using a @xmath272-rotation about the @xmath273-axis , we learn that @xmath239 is an even function , i.e. @xmath274 below we will see that @xmath275 . 2 . let us now focus on @xmath276 , which defines the action of the boost in the @xmath277 plane . we write @xmath278 we refer to the system with coordinates @xmath279 as @xmath280 and that with coordinates @xmath281 as @xmath282 . from ( [ eq : formboost3 ] ) and the inverse ( which is elementary to compute ) one infers that the velocity @xmath60 of @xmath282 with respect to @xmath280 and the velocity @xmath283 of @xmath280 with respect to @xmath282 are given by + [ eq : formboost4 ] @xmath284 + since the transformation @xmath285 is the inverse of @xmath286 , the function @xmath287 obeys @xmath288 hence @xmath289 is a bijection of the open interval @xmath290 onto itself and obeys @xmath291 3 . next we determine @xmath289 . once more using ( [ eq : equivboostsrot ] ) , where @xmath264 is a @xmath272-rotation about the @xmath273-axis , shows that the functions @xmath292 and @xmath293 in ( [ eq : formboost1 ] ) are even and the functions @xmath294 and @xmath245 are odd . the definition ( [ eq : formboost4b ] ) of @xmath289 then implies that @xmath289 is odd . since we assumed @xmath257 to be a continuous coordinatisation of a topological group , the map @xmath289 must also be continuous ( since the inversion map , @xmath295 , is continuous in a topological group ) . a standard theorem now states that a continuous bijection of an interval of @xmath296 onto itself must be strictly monotonic . together with ( [ eq : formboost6 ] ) this implies that @xmath289 is either the identity or minus the identity map . to save notation , so that ( [ eq : formboost6 ] ) now reads @xmath297 . first assume that @xmath289 is strictly monotonically increasing , then @xmath298 implies @xmath299 , a contradiction , and @xmath300 implies @xmath301 , likewise a contradiction . hence @xmath302 in this case . next assume @xmath289 is strictly monotonically decreasing . then @xmath303 is a strictly monotonically increasing map of the interval @xmath290 to itself that obeys ( [ eq : formboost6 ] ) . hence , as just seen , @xmath304 , i.e. @xmath305 . ] if it is the identity map , evaluation of ( [ eq : formboost5 ] ) shows that either the determinant of @xmath269 must equals @xmath306 , or that @xmath269 is the identity for all @xmath257 . we exclude the second possibility straightaway and the first one on the grounds that we required @xmath269 be the identity for @xmath307 . also , in that case , ( [ eq : formboost5 ] ) implies @xmath308 for all @xmath309 . we conclude that @xmath305 , which implies that the relative velocity of @xmath280 with respect to @xmath282 is minus the relative velocity of @xmath282 with respect to @xmath280 . plausible as it might seem , there is no a priori reason why this should be so . and @xmath283 are measured with different sets of rods and clocks . ] . on the face of it , the rp only implies ( [ eq : formboost6 ] ) , not the stronger relation @xmath310 . this was first pointed out in @xcite . 4 . we briefly revisit ( [ eq : formboosts2 ] ) . since we have seen that @xmath311 is the inverse of @xmath312 , we must have @xmath313 , so that ( [ eq : formboosts2 ] ) implies @xmath314 . but only @xmath315 is compatible with our requirement that @xmath316 be the identity . now we return to the determination of @xmath269 . using ( [ eq : formboost4 ] ) and @xmath305 , we write @xmath317 and @xmath318\,.\ ] ] equation @xmath319 is now equivalent to + [ eq : formboost8 ] @xmath320 + since , as already seen , @xmath292 is an even and @xmath294 is an odd function , ( [ eq : formboost8 ] ) is equivalent to @xmath321 , i.e. the unimodularity of @xmath253 . equation ( [ eq : formboost7 ] ) then allows to express @xmath294 in terms of @xmath292 : @xmath322\,.\ ] ] 6 . our problem is now reduced to the determination of the single function @xmath292 . this we achieve by employing the requirement that the composition of two boosts in the same direction results again in a boost in that direction , i.e. @xmath323 according to ( [ eq : formboost6b ] ) each matrix @xmath269 has equal diagonal entries . applied to the product matrix on the left hand side of ( [ eq : formboost10 ] ) this implies that @xmath324 is independent of @xmath60 , i.e. equal to some constant @xmath325 whose physical dimension is that of an inverse velocity squared . hence we have @xmath326 where we have chosen the positive square root since we require @xmath327 . the other implications of ( [ eq : formboost10 ] ) are + [ eq : formboost12 ] @xmath328 + from which we deduce @xmath329 conversely , ( [ eq : formboost11 ] ) and ( [ eq : formboost13 ] ) imply ( [ eq : formboost12 ] ) . we conclude that ( [ eq : formboost10 ] ) is equivalent to ( [ eq : formboost11 ] ) and ( [ eq : formboost13 ] ) . so far a boost in @xmath262 direction has been shown to act non - trivially only in the @xmath277 plane , where its action is given by the matrix that results from inserting ( [ eq : formboost9 ] ) and ( [ eq : formboost11 ] ) into ( [ eq : formboost6b ] ) : @xmath330 * if @xmath331 we rescale @xmath332 and set @xmath333 . then ( [ eq : formboost14 ] ) is seen to be a euclidean rotation with angle @xmath239 in the @xmath334 plane . the velocity spectrum is the whole real line plus infinity , i.e. a circle , corresponding to @xmath335 $ ] , where @xmath336 and @xmath337 are identified . accordingly , the composition law ( [ eq : formboost13 ] ) is just ordinary addition for the angle @xmath239 . this causes several paradoxa when @xmath60 is interpreted as velocity . for example , composing two finite velocities @xmath338 which satisfy @xmath339 results in @xmath340 , and composing two finite and positive velocities , each of which is greater than @xmath341 , results in a finite but negative velocity . in this way the successive composition of finite positive velocities could also result in zero velocity . the group @xmath342 obtained in this fashion is , in fact , @xmath343 . this group may be uniquely characterised as the largest connected group of bijections of @xmath344 that preserves the euclidean distance measure . in particular , it treats time symmetrically with all space directions , so that no invariant notion of time - orientability can be given in this case . * for @xmath345 the transformations are just the ordinary boosts of the galilei group . the velocity spectrum is the whole real line ( i.e. @xmath60 is unbounded but finite ) and @xmath247 is the galilei group . the law for composing velocities is just ordinary vector addition . * finally , for @xmath346 , one infers from ( [ eq : formboost13 ] ) that @xmath347 is an upper bound for all velocities , in the sense that composing two velocities taken from the interval @xmath290 always results in a velocity from within that interval . writing @xmath348 , @xmath349 , and @xmath350 , the matrix ( [ eq : formboost14 ] ) is seen to be a _ lorentz boost _ or _ hyperbolic motion _ in the @xmath334 plane : @xmath351 the quantity @xmath352 is called _ _ rapidity__. if rewritten in terms of the corresponding rapidities the composition law ( [ eq : formboost13 ] ) reduces to ordinary addition : @xmath353 . this shows that only the galilei and the lorentz group survive as candidates for any symmetry group implementing the rp . once the lorentz group for velocity parameter @xmath245 is chosen , one may fully characterise it by its property to leave a certain symmetric bilinear form invariant . in this sense we geometric structure of minkowski space can be deduced . this closes the circle to where we started from in section[sec : frommetrictoaffine ] . in the previous sections we always understood an automorphisms of a structured set ( spacetime ) as a bijection . mathematically this seems an obvious requirement , but from a physical point of view this is less clear . the physical law of inertia provides us with distinguished motions _ locally _ in space and time . hence one may attempt to relax the condition for structure preserving maps , so as to only preserve inertial motions _ hence we ask the following question : what are the most general maps that _ locally _ map segments of straight lines to segments of straight lines ? this local approach has been pursued by @xcite . to answer this question completely , let us ( locally ) identify spacetime with @xmath179 where @xmath16 and assume the map to be @xmath354 , that is , three times continuously differentiable . so let @xmath355 be an open subset and determine all @xmath354 maps @xmath356 that map straight segments in @xmath357 into straight segments in @xmath179 . in coordinates we write @xmath358 and @xmath359 , so that @xmath360 . a straight segment in @xmath357 is a curve @xmath361 ( the open interval @xmath362 is usually taken to contain zero ) whose acceleration is pointwise proportional to its velocity . this is equivalent to saying that it can be parametrised so as to have zero acceleration , i.e. , @xmath363 for some @xmath364 . for the image path @xmath365 to be again straight its acceleration , @xmath366 , must be proportional to its velocity , @xmath367 , where the factor of proportionality , @xmath368 , depends on the point of the path and separately on @xmath292 . hence , in coordinates , we have @xmath369 for each @xmath294 this must be valid for all @xmath370 in a neighbourhood of zero in @xmath371 . taking the second derivatives with respect to @xmath292 , evaluation at @xmath372 , @xmath373 leads to [ eq : projmap2 ] @xmath374 where @xmath375 here we suppressed the remaining argument @xmath294 . equation ( [ eq : projmap2 ] ) is valid at each point in @xmath357 . integrability of ( [ eq : projmap2-a ] ) requires that its further differentiation is totally symmetric with respect to all lower indices ( here we use that the map @xmath9 is @xmath354 ) . this leads to @xmath376 inserting ( [ eq : projmap2-b ] ) one can show ( upon taking traces over @xmath377 and @xmath378 ) that the resulting equation is equivalent to @xmath379 in particular @xmath380 so that there is a local function @xmath381 ( if @xmath357 is simply connected , as we shall assume ) for which @xmath382 . equation ( [ eq : projmap4 ] ) is then equivalent to @xmath383 so that @xmath384 for some @xmath385 and @xmath386 . using @xmath387 and ( [ eq : projmap4 ] ) , equation ( [ eq : projmap2-a ] ) is equivalent to @xmath388=0 $ ] , which finally leads to the result that the most general solution for @xmath9 is given by @xmath389 here @xmath390 is a @xmath391 matrix , @xmath292 and @xmath140 vectors in @xmath179 , and @xmath386 . @xmath62 and @xmath140 must be such that @xmath357 does not intersect the hyperplane @xmath392 where @xmath9 becomes singular , but otherwise they are arbitrary . , i.e. iff @xmath394 , the transformations ( [ eq : projmap5 ] ) are not affine . in this case they are called proper projective . are there physical reasons to rule out such proper projective transformations ? a structural argument is that they do not leave any subset of @xmath179 invariant and that they hence can not be considered as automorphism group of any subdomain . a physical argument is that two separate points that move with the same velocity cease to do so if their worldlines are transformed by by a proper projective transformation . in particular , a rigid motion of an extended body ( undergoing inertial motion ) ceases to be rigid if so transformed ( cf.@xcite , p.16 ) . an illustrative example is the following : consider the one - parameter ( @xmath395 ) family of parallel lines @xmath396 ( where @xmath397 is the parameter along each line ) , and the proper projective map @xmath398 which becomes singular on the hyperplane @xmath399 . the one - parameter family of image lines @xmath400 have velocities @xmath401 whose directions are independent of @xmath397 , showing that they are indeed straight . however , the velocity directions now depend on @xmath395 , showing that they are not parallel anymore . let us , regardless of this , for the moment take seriously the transformations ( [ eq : projmap5 ] ) . one may reduce them to the following form of generalised boosts , discarding translations and rotations and using equivariance with respect to the latter ( we restrict to four spacetime dimensions from now on ) : [ eq : focklorentztrans1 ] @xmath402 where @xmath403 represents the boost velocity , @xmath404 its modulus , and all functions of @xmath60 are even . the subscripts @xmath405 and @xmath406 refer to the components parallel and perpendicular to @xmath257 . now one imposes the following conditions which allow to determine the eight functions @xmath407 , of which only seven are considered independent since common factors of the numerator and denominator cancel ( we essentially follow @xcite ) : 1 . the origin @xmath408 has velocity @xmath257 in the unprimed coordinates , leading to @xmath409 and thereby eliminating @xmath410 as independent function . the origin @xmath411 has velocity @xmath412 in the primed coordinates , leading to @xmath413 and thereby eliminating @xmath293 as independent function . reciprocity : the transformation parametrised by @xmath412 is the inverse of that parametrised by @xmath257 , leading to relations @xmath414 , @xmath415 , and @xmath416 , thereby eliminating @xmath417 as independent functions . of the remaining three functions @xmath418 an overall factor in the numerator and denominator can be split off so that two free functions remain . transitivity : the composition of two transformations of the type ( [ eq : focklorentztrans1 ] ) with parameters @xmath257 and @xmath419 must be again of this form with some parameter @xmath420 , which turns out to be the same function of the velocities @xmath257 and @xmath419 as in special relativity ( einstein s addition law ) , for reasons to become clear soon . this allows to determine the last two functions in terms of two constants @xmath245 and @xmath421 whose physical dimensions are that of a velocity and of a length respectively . writing , as usual , @xmath422 the final form is given by [ eq : focklorentztrans2 ] @xmath423 in the limit as @xmath424 this approaches an ordinary lorentz boost : @xmath425 moreover , for finite @xmath421 the map ( [ eq : focklorentztrans2 ] ) is conjugate to ( [ eq : lorentzfock3 ] ) with respect to a time dependent deformation . to see this , observe that the common denominator in ( [ eq : focklorentztrans2 ] ) is just @xmath426 , whereas the numerators correspond to ( [ eq : lorentzfock3 ] ) . hence , introducing the deformation map @xmath427 and denoting the map @xmath428 in ( [ eq : focklorentztrans2 ] ) by @xmath9 , we have @xmath429 note that @xmath112 is singular at the hyperplane @xmath430 and has no point of the hyperplane @xmath431 in its image . the latter hyperplane is the singularity set of @xmath432 . outside the hyperplanes @xmath433 the map @xmath112 relates the following time slabs in a diffeomorphic fashion : [ eq : difftimeslabs ] @xmath434 since boosts leave the upper - half spacetime , @xmath435 , invariant ( as set ) , ( [ eq : difftimeslabs - a ] ) shows that @xmath9 just squashes the linear action of boosts in @xmath436 into a non - linear action within @xmath437 , where @xmath421 now corresponds to an invariant scale . interestingly , this is the same deformation of boosts that have been recently considered in what is sometimes called _ doubly special relativity _ ( because there are now two , rather than just one , invariant scales , @xmath421 and @xmath245 ) , albeit there the deformation of boosts take place in momentum space where @xmath421 then corresponds to an invariant energy scale ; see @xcite and also @xcite . in this section we wish to discuss in more detail some of the non - trivial structures in minkowski . i have chosen them so as to emphasise the difference to the corresponding structures in galilean spacetime , and also because they do not seem to be much discussed in other standard sources . let us start right away by characterising those vectors for which we have an inverted cauchy - schwarz inequality : [ lemma : stricticsi ] let @xmath19 be of dimension @xmath438 and @xmath45 be some non - zero vector . the strict inverted cauchy - schwarz inequality , @xmath439 holds for all @xmath106 linearly independent of @xmath60 iff @xmath60 is timelike . obviously @xmath60 can not be spacelike , for then we would violate ( [ eq : strictinvcsi ] ) with any spacelike @xmath77 . if @xmath60 is lightlike then @xmath77 violates ( [ eq : strictinvcsi ] ) iff it is in the set @xmath440 , which is non - empty iff @xmath438 . hence @xmath60 can not be lightlike if @xmath438 . if @xmath60 is timelike we decompose @xmath441 with @xmath442 so that @xmath443 , with equality iff @xmath60 and @xmath77 are linearly dependent . hence @xmath444 with equality iff @xmath60 and @xmath77 are linearly dependent . the next lemma deals with the intersection of a causal line with a light cone , a situation depicted in fig.[fig : robb ] . [ lemma : intlinecone ] let @xmath445 be the light - doublecone with vertex @xmath62 and @xmath446 be a non - spacelike line , i.e. @xmath447 , through @xmath448 . if @xmath60 is timelike @xmath449 consists of two points . if @xmath60 is lightlike this intersection consists of one point if @xmath450 and is empty if @xmath451 . note that the latter two statements are independent of the choice of @xmath452as they must be , i.e. are invariant under @xmath453 , where @xmath454 . we have @xmath455 iff @xmath456 for @xmath60 timelike we have @xmath457 and ( [ eq : lemmaintlinecone1 ] ) has two solutions @xmath458 indeed , since @xmath448 , the vectors @xmath60 and @xmath459 can not be linearly dependent so that lemma[lemma : stricticsi ] implies the positivity of the expression under the square root . if @xmath60 is lightlike ( [ eq : lemmaintlinecone1 ] ) becomes a linear equation which is has one solution if @xmath460 and no solution if @xmath461 [ note that @xmath462 since @xmath463 by hypothesis ] . ( -12,155)@xmath464 ( -22,114)@xmath140 ( -35,56)@xmath465 ( -72,98)@xmath62 ( -60,0)@xmath466 ( -42,15)@xmath60 ( -75,146)@xmath467 ( -75,62)@xmath468 ( -13,195)@xmath469 [ prop : einsteinsynch ] let @xmath469 and @xmath445 as in lemma[lemma : intlinecone ] with @xmath60 timelike . let @xmath464 and @xmath465 be the two intersection points of @xmath469 with @xmath445 and @xmath470 a point between them . then @xmath471 moreover , @xmath472 iff @xmath71 is perpendicular to @xmath60 . the vectors @xmath473 and @xmath474 are lightlike , which gives ( note that @xmath475 is spacelike ) : [ eq : prop : einsteinsynch23 ] @xmath476 since @xmath477 and @xmath478 are parallel we have @xmath479 with @xmath480 so that @xmath481 and @xmath482 . now , multiplying ( [ eq : prop : einsteinsynch3 ] ) with @xmath155 and adding this to ( [ eq : prop : einsteinsynch2 ] ) immediately yields @xmath483 since @xmath484 this implies ( [ eq : prop : einsteinsynch1 ] ) . finally , since @xmath477 and @xmath485 are antiparallel , @xmath486 iff @xmath487 . equations ( [ eq : prop : einsteinsynch23 ] ) now show that this is the case iff @xmath488 , i.e. iff @xmath489 . hence we have shown @xmath490 in other words , @xmath140 is the midpoint of the segment @xmath491 iff the line through @xmath62 and @xmath140 is perpendicular ( wrt . @xmath30 ) to @xmath469 . the somewhat surprising feature of the first statement of this proposition is that ( [ eq : prop : einsteinsynch1 ] ) holds for _ any _ point of the segment @xmath491 , not just the midpoint , as it would have to be the case for the corresponding statement in euclidean geometry . the second statement of proposition[prop : einsteinsynch ] gives a convenient geometric characterisation of einstein - simultaneity . recall that an event @xmath140 on a timelike line @xmath469 ( representing an inertial observer ) is defined to be einstein - simultaneous with an event @xmath62 in spacetime iff @xmath140 bisects the segment @xmath491 between the intersection points @xmath492 of @xmath469 with the double - lightcone at @xmath62 . hence proposition[prop : einsteinsynch ] implies [ corol : einsteinsim ] einstein simultaneity with respect to a timelike line @xmath469 is an equivalence relation on spacetime , the equivalence classes of which are the spacelike hyperplanes orthogonal ( wrt . @xmath30 ) to @xmath469 . the first statement simply follows from the fact that the family of parallel hyperplanes orthogonal to @xmath469 form a partition ( cf . sect.[sec : groupactions ] ) of spacetime . from now on we shall use the terms ` timelike line ' and ` inertial observer ' synonymously . note that einstein simultaneity is only defined relative to an inertial observer . given two inertial observers , [ eq : twoobservers ] @xmath493 we call the corresponding einstein - simultaneity relations _ @xmath469-simultaneity _ and _ @xmath494-simultaneity_. obviously they coincide iff @xmath469 and @xmath494 are parallel ( @xmath60 and @xmath283 are linearly dependent ) . in this case @xmath495 is @xmath469-simultaneous to @xmath470 iff @xmath470 is @xmath494-simultaneous to @xmath495 . if @xmath469 and @xmath494are not parallel ( skew or intersecting in one point ) it is generally not true that if @xmath495 is @xmath469-simultaneous to @xmath470 then @xmath470 is also @xmath494-simultaneous to @xmath495 . in fact , we have [ prop : skewsim ] let @xmath469 and @xmath494 two non - parallel timelike likes . there exists a unique pair @xmath496 so that @xmath497 is @xmath469-simultaneous to @xmath140 and @xmath140 is @xmath494 simultaneous to @xmath497 . we parameterise @xmath469 and @xmath494 as in ( [ eq : twoobservers ] ) . the two conditions for @xmath497 being @xmath469-simultaneous to @xmath140 and @xmath140 being @xmath494-simultaneous to @xmath497 are @xmath498 . writing @xmath499 and @xmath500 this takes the form of the following matrix equation for the two unknowns @xmath155 and @xmath501 : @xmath502 this has a unique solution pair @xmath503 , since for linearly independent timelike vectors @xmath60 and @xmath283 lemma[lemma : stricticsi ] implies @xmath504 . note that if @xmath469 and @xmath494 intersect @xmath505 . clearly , einstein - simultaneity is conventional and physics proper should not depend on it . for example , the fringe - shift in the michelson - morley experiment is independent of how we choose to synchronise clocks . in fact , it does not even make use of any clock . so what is the general definition of a ` simultaneity structure ' ? it seems obvious that it should be a relation on spacetime that is at least symmetric ( each event should be simultaneous to itself ) . going from one - way simultaneity to the mutual synchronisation of two clocks , one might like to also require reflexivity ( if @xmath62 is simultaneous to @xmath140 then @xmath140 is simultaneous to @xmath62 ) , though this is not strictly required in order to one - way synchronise each clock in a set of clocks with one preferred ` master clock ' , which is sufficient for many applications . moreover , if we like to speak of the mutual simultaneity of sets of more than two events we need an equivalence relation on spacetime . the equivalence relation should be such that each inertial observer intersect each equivalence class precisely once . let us call such a simultaneity structure ` admissible ' . clearly there are zillions of such structures : just partition spacetime into any set of appropriate spacelike hypersurfaces ( there are more possibilities at this point , like families of forward or backward lightcones ) . an _ absolute _ admissible simultaneity structure would be one which is invariant ( cf . sect.[sec : groupactions ] ) under the automorphism group of spacetime . we have [ prop : abssim ] there exits precisely one admissible simultaneity structure which is invariant under the inhomogeneous proper orthochronous galilei group and none that is invariant under the inhomogeneous proper orthochronous lorentz group . a proof is given in @xcite . there is a group - theoretic reason that highlights this existential difference : [ prop : abssimmath ] let @xmath249 be a group with transitive action on a set @xmath2 . let @xmath506 be the stabiliser subgroup for @xmath507 ( due to transitivity all stabiliser subgroups are conjugate ) . then @xmath2 admits a @xmath249-invariant equivalence relation @xmath508 iff @xmath509 is _ not _ maximal , that is , iff @xmath509 is properly contained in a proper subgroup @xmath510 of @xmath249 : @xmath511 . a proof of this may be found in @xcite ( theorem1.12 ) . regarding the action of the inhomogeneous galilei and lorentz groups on spacetime , their stabilisers are the corresponding homogeneous groups . now , the homogeneous lorentz group is maximal in the inhomogeneous one , whereas the homogeneous galilei group is not maximal in the inhomogeneous one , since it can still be supplemented by time translations without the need to also invoke space translations . this , according to proposition[prop : abssimmath ] , is the group theoretic origin of the absence of any invariant simultaneity structure in the lorentzian case . however , one may ask whether there are simultaneity structures _ relative _ to some _ additional _ structure @xmath512 . as additional structure , @xmath512 , one could , for example , take an inertial reference frame , which is characterised by a foliation of spacetime by parallel timelike lines . the stabiliser subgroup of that structure within the proper orthochronous poincar group is given by the semidirect product of spacetime translations with all rotations in the hypersurfaces perpendicular to the lines in @xmath512 : @xmath513 here the @xmath514 only acts on the spatial translations , so that the group is also isomorphic to @xmath515 , where @xmath516 is the group of euclidean motions in 3-dimensions ( the hyperplanes perpendicular to the lines in @xmath512 ) . we can now ask : how many admissible @xmath517 invariant equivalence relations are there . the answer is [ prop : realsim ] there exits precisely one admissible simultaneity structure which is invariant under @xmath517 , where @xmath512 represents am inertial reference frame ( a foliation of spacetime by parallel timelike lines ) . it is given by einstein simultaneity , that is , the equivalence classes are the hyperplanes perpendicular to the lines in @xmath512 . the proof is given in @xcite . note again the connection to quoted group - theoretic result : the stabiliser subgroup of a point in @xmath517 is @xmath518 , which is clearly not maximal in @xmath517 since it is a proper subgroup of @xmath516 which , in turn , is a proper subgroup of @xmath517 . here we wish to briefly discuss another important structure associated with causality relations in minkowski space , which plays a fundamental rle in modern quantum field theory ( see e.g. @xcite ) . let @xmath519 and @xmath520 be subsets of @xmath63 . we say that @xmath519 and @xmath520 are _ causally disjoint _ or _ spacelike separated _ iff @xmath521 is spacelike , i.e. @xmath522 , for any @xmath523 and @xmath524 . note that because a point is not spacelike separated from itself , causally disjoint sets are necessarily disjoint in the ordinary set - theoretic sense the converse being of course not true . for any subset @xmath525 we denote by @xmath526 the largest subset of @xmath63 which is causally disjoint to @xmath2 . the set @xmath526 is called the _ causal complement _ of @xmath2 . the procedure of taking the causal complement can be iterated and we set @xmath527 etc . @xmath528 is called the _ causal completion _ of @xmath2 . it also follows straight from the definition that @xmath529 implies @xmath530 and also @xmath531 . if @xmath532 we call @xmath2 _ causally complete_. we note that the causal complement @xmath526 of any given @xmath2 is automatically causally complete . indeed , from @xmath531 we obtain @xmath533 , but the first inclusion applied to @xmath526 instead of @xmath2 leads to @xmath534 , showing @xmath535 . note also that for any subset @xmath2 its causal completion , @xmath528 , is the smallest causally complete subset containing @xmath2 , for if @xmath536 with @xmath537 , we derive from the first inclusion by taking @xmath538 that @xmath539 , so that the second inclusion yields @xmath540 . trivial examples of causally complete subsets of @xmath63 are the empty set , single points , and the total set @xmath63 . others are the open diamond - shaped regions ( [ eq : opendiamonds ] ) as well as their closed counterparts : @xmath541 we now focus attention to the set @xmath542 of causally complete subsets of @xmath63 , including the empty set , @xmath543 , and the total set , @xmath63 , which are mutually causally complementary . it is partially ordered by ordinary set - theoretic inclusion @xmath544 ( cf . sect.[sec : groupactions ] ) and carries the ` dashing operation ' @xmath545 of taking the causal complement . moreover , on @xmath542 we can define the operations of ` meet ' and ` join ' , denoted by @xmath546 and @xmath547 respectively , as follows : let @xmath548 where @xmath199 , then @xmath549 is the largest causally complete subset in the intersection @xmath550 and @xmath551 is the smallest causally complete set containing the union @xmath552 . the operations of @xmath546 and @xmath547 can be characterised in terms of the ordinary set - theoretic intersection @xmath553 together with the dashing - operation . to see this , consider two causally complete sets , @xmath554 where @xmath199 , and note that the set of points that are spacelike separated from @xmath519 and @xmath520 are obviously given by @xmath555 , but also by @xmath556 , so that [ eg : capcup ] @xmath557 here ( [ eq : capcup1 ] ) and ( [ eq : capcup2 ] ) are equivalent since any @xmath548 can be written as @xmath558 , namely @xmath559 . if @xmath554 runs through all sets in @xmath542 so does @xmath560 . hence any equation that holds generally for all @xmath548 remains valid if the @xmath554 are replaced by @xmath561 . equation ( [ eq : capcup2 ] ) immediately shows that @xmath550 is causally complete ( since it is the @xmath562 of something ) . taking the causal complement of ( [ eq : capcup1 ] ) we obtain the desired relation for @xmath563 . together we have [ eg : meetjoin ] @xmath564 from these we immediately derive [ eg : negating ] @xmath565 all what we have said so far for the set @xmath542 could be repeated verbatim for the set @xmath566 of _ chronologically complete _ subsets . we say that @xmath519 and @xmath520 are _ chronologically disjoint _ or _ non - timelike separated _ , iff @xmath567 and @xmath568 for any @xmath523 and @xmath524 . @xmath526 , the _ chronological complement _ of @xmath2 , is now the largest subset of @xmath63 which is chronologically disjoint to @xmath2 . the only difference between the causal and the chronological complement of @xmath2 is that the latter now contains lightlike separated points outside @xmath2 . a set @xmath2 is _ chronologically complete _ iff @xmath569 , where the dashing now denotes the operation of taking the chronological complement . again , for any set @xmath2 the set @xmath526 is automatically chronologically complete and @xmath528 is the smallest chronologically complete subset containing @xmath2 . single points are chronologically complete subsets . all the formal properties regarding @xmath570 , @xmath546 , and @xmath547 stated hitherto for @xmath542 are the same for @xmath566 . one major difference between @xmath542 and @xmath566 is that the types of diamond - shaped sets they contain are different . for example , the closed ones , ( [ eq : closeddiamonds ] ) , are members of both . the open ones , ( [ eq : opendiamonds ] ) , are contained in @xmath542 but _ not _ in @xmath566 . instead , @xmath566 contains the closed diamonds whose ` equator'sphere in which the forward and backward light - cones in ( [ eq : closeddiamonds ] ) intersect . in the two - dimensional drawings the ` equator ' is represented by just two points marking the right and left corners of the diamond - shaped set . ] have been removed . an essential structural difference between @xmath542 and @xmath566 will be stated below , after we have introduced the notion of a lattice to which we now turn . to put all these formal properties into the right frame we recall the definition of a lattice . let @xmath571 be a partially ordered set and @xmath572 any two elements in @xmath573 . synonymously with @xmath574 we also write @xmath575 and say that @xmath292 is smaller than @xmath294 , @xmath294 is bigger than @xmath292 , or @xmath294 majorises @xmath292 . we also write @xmath576 if @xmath574 and @xmath232 . if , with respect to @xmath141 , their greatest lower and least upper bound exist , they are denoted by @xmath577called the ` meet of @xmath292 and @xmath294'and @xmath578called the ` join of @xmath292 and @xmath294'respectively . a partially ordered set for which the greatest lower and least upper bound exist for any pair @xmath572 of elements from @xmath573 is called a _ lattice_. we now list some of the most relevant additional structural elements lattices can have : a lattice is called _ complete _ if greatest lower and least upper bound exist for any subset @xmath579 . if @xmath580 they are called @xmath336 ( the smallest element in the lattice ) and @xmath121 ( the biggest element in the lattice ) respectively . an _ atom _ in a lattice is an element @xmath292 which majorises only @xmath336 , i.e. @xmath581 and if @xmath582 then @xmath583 or @xmath584 . the lattice is called _ atomic _ if each of its elements different from @xmath336 majorises an atom . an atomic lattice is called _ atomistic _ if every element is the join of the atoms it majorises . an element @xmath245 is said to _ cover _ @xmath292 if @xmath585 and if @xmath586 either @xmath587 or @xmath588 . an atomic lattice is said to have the _ covering property _ if , for every element @xmath294 and every atom @xmath292 for which @xmath589 , the join @xmath578 covers @xmath294 . the subset @xmath590 is called a _ distributive triple _ if [ eq : disttriple ] @xmath591 a lattice is called _ distributive _ or _ boolean _ if every triple @xmath592 is distributive . it is called _ modular _ if every triple @xmath592 with @xmath574 is distributive . it is straightforward to check from ( [ eq : disttriple ] ) that modularity is equivalent to the following single condition : @xmath593 if in a lattice with smallest element @xmath336 and greatest element @xmath121 a map @xmath594 , @xmath595 , exist such that [ eq : orthocomplement ] @xmath596 the lattice is called _ orthocomplemented_. it follows that whenever the meet and join of a subset @xmath597 ( @xmath598 is some index set ) exist one has demorgan s laws ) is redundant , as each of its two statements follows from the other , due to @xmath599 . ] : [ eq : gendemorgan ] @xmath600 for orthocomplemented lattices there is a still weaker version of distributivity than modularity , which turns out to be physically relevant in various contexts : [ def : orthomodular ] an orthocomplemented lattice is called _ orthomodular _ if every triple @xmath592 with @xmath574 and @xmath601 is distributive . from ( [ eq : modularity ] ) and using that @xmath602 for @xmath603 one sees that this is equivalent to the single condition ( renaming @xmath245 to @xmath604 ) : [ eq : orthomodularity ] @xmath605 where the second line follows from the first by taking its orthocomplement and renaming @xmath606 to @xmath607 . it turns out that these conditions can still be simplified by making them independent of @xmath245 . in fact , ( [ eq : orthomodularity ] ) are equivalent to [ eq : altorthomodularity ] @xmath608 it is obvious that ( [ eq : orthomodularity ] ) implies ( [ eq : altorthomodularity ] ) ( set @xmath609 ) . but the converse is also true . to see this , take e.g. ( [ eq : altorthomodularity2 ] ) and choose any @xmath610 . then @xmath611 , @xmath612 ( by hypothesis ) , and @xmath613 ( trivially ) , so that @xmath614 . hence @xmath615 , which proves ( [ eq : orthomodularity2 ] ) . complete orthomodular atomic lattices are automatically atomistic . indeed , let @xmath294 be the join of all atoms majorised by @xmath616 . assume @xmath232 so that necessarily @xmath617 , then ( [ eq : altorthomodularity2 ] ) implies @xmath618 . then there exists an atom @xmath245 majorised by @xmath619 . this implies @xmath620 and @xmath601 , hence also @xmath621 . but this is a contradiction , since @xmath294 is by definition the join of all atoms majorised by @xmath292 . finally we mention the notion of _ compatibility _ or _ commutativity _ , which is a symmetric , reflexive , but generally not transitive relation @xmath421 on an orthomodular lattice ( cf . sec.[sec : groupactions ] ) . we write @xmath622 for @xmath623 and define : [ eq : defcompatibility ] @xmath624 the equivalence of these two lines , which shows that the relation of being compatible is indeed symmetric , can be demonstrated using orthomodularity as follows : suppose ( [ eq : defcompatibility1 ] ) holds ; then @xmath625 , where we used the orthocomplement of ( [ eq : defcompatibility1 ] ) to replace @xmath626 in the first expression and the trivial identity @xmath627 in the second step . now , applying ( [ eq : altorthomodularity2 ] ) to @xmath628 we get @xmath629=(b\wedge a)\vee ( b\wedge a')$ ] , i.e. ( [ eq : defcompatibility2 ] ) . the converse , @xmath630 , is of course entirely analogous . from ( [ eq : defcompatibility ] ) a few things are immediate : @xmath622 is equivalent to @xmath631 , @xmath622 is implied by @xmath574 or @xmath632 , and the elements @xmath336 and @xmath121 are compatible with all elements in the lattice . the _ centre _ of a lattice is the set of elements which are compatible with all elements in the lattice . in fact , the centre is a boolean sublattice . if the centre contains no other elements than @xmath336 and @xmath121 the lattice is said to be _ the other extreme is a boolean lattice , which is identical to its own centre . indeed , if @xmath633 is a distributive triple , one has @xmath634 . after these digression into elementary notions of lattice theory we come back to our examples of the sets @xmath542 @xmath566 . our statements above amount to saying that they are complete , atomic , and orthocomplemented lattices . the partial order relation @xmath141 is given by @xmath635 and the extreme elements @xmath336 and @xmath121 correspond to the empty set @xmath543 and the total set @xmath63 , the points of which are the atoms . neither the covering property nor modularity is shared by any of the two lattices , as can be checked by way of elementary counterexamples . and @xmath140 . then @xmath636 whereas @xmath637 is given by the closed diamond ( [ eq : closeddiamonds ] ) . note that this is true in @xmath542 _ and _ @xmath566 . but , clearly , @xmath637 does not cover either @xmath638 or @xmath639 . ] in particular , neither of them is boolean . however , in @xcite it was shown that @xmath566 is orthomodular ; see also @xcite which deals with more general spacetimes . note that by the argument given above this implies that @xmath566 is atomistic . in contrast , @xmath542 is definitely _ not _ orthomodular , as is e.g. seen by the counterexample given in fig.[fig : countex ] . regarding this point , there are some conflicting statements in the literature . the first edition of @xcite states orthomodularity of @xmath566 in proposition4.1.3 , which is removed in the second edition without further comment . the proof offered in the first edition uses ( [ eq : altorthomodularity1 ] ) as definition of orthomodularity , writing @xmath640 for @xmath292 and @xmath641 for b. the crucial step is the claim that any spacetime event in the set @xmath642 lies in @xmath641 and that any causal line through it must intersect either @xmath640 or @xmath643 . the last statement is , however , not correct since the join of two sets ( here @xmath640 and @xmath643 ) is generally larger than the domain of dependence of their ordinary set - theoretic union ; compare fig.[fig : countex ] . : ( generally , the domain of dependence of a subset @xmath2 of spacetime @xmath1 is the largest subset @xmath644 such that any inextensible causal curve that intersects @xmath645 also intersects @xmath2 . ) ] it is also not difficult to prove that @xmath566 is irreducible . , the failure of irreducibility of @xmath646 is directly related to the existence of closed timelike curves ; see @xcite . ] ( -365,104)@xmath469 ( -344.5,72)@xmath292 ( -280,50)@xmath647 ( -322,85)@xmath648 ( -152,85)@xmath648 ( -180,50)@xmath294 ( -30,50)@xmath294 ( -105,50)@xmath647 ( -158,54)@xmath292 it is well known that the lattices of propositions for classical systems are boolean , whereas those for quantum systems are merely orthomodular . in classical physics the elements of the lattice are measurable subsets of phase space , with @xmath141 being ordinary set - theoretic inclusion @xmath635 , and @xmath546 and @xmath547 being ordinary set - theoretic intersection @xmath553 and union @xmath649 respectively . the orthocomplement is the ordinary set - theoretic complement . in quantum mechanics the elements of the lattice are the closed subspaces of hilbert space , with @xmath141 being again ordinary inclusion , @xmath546 ordinary intersection , and @xmath547 is given by @xmath650 . the orthocomplement of a closed subset is the orthogonal complement in hilbert space . for comprehensive discussions see @xcite and @xcite . one of the main questions in the foundations of quantum mechanics is whether one could understand ( derive ) the usage of hilbert spaces and complex numbers from somehow more fundamental principles . even though it is not a priori clear what ones measure of fundamentality should be at this point , an interesting line of attack consists in deriving the mentioned structures from the properties of the lattice of propositions ( quantum logic ) . it can be shown that a lattice that is complete , atomic , irreducible , orthomodular , and that satisfies the covering property is isomorphic to the lattice of closed subspaces of a linear space with hermitean inner product . the complex numbers are selected if additional technical assumptions are added . for the precise statements of these reconstruction theorems see @xcite . it is now interesting to note that , on a formal level , there is a similar transition in going from galilei invariant to lorentz invariant causality relations . in fact , in galilean spacetime one can also define a chronological complement : two points are chronologically related if they are connected by a worldline of finite speed and , accordingly , two subsets in spacetime are chronologically disjoint if no point in one set is chronologically related to a point of the other . for example , the chronological complement of a point @xmath62 are all points simultaneous to , but different from , @xmath62 . more general , it is not hard to see that the chronologically complete sets are just the subsets of some @xmath651 hypersurface . the lattice of chronologically complete sets is then the continuous disjoint union of sublattices , each of which is isomorphic to the boolean lattice of subsets in @xmath256 . for details see @xcite . as we have seen above , @xmath652 is complete , atomic , irreducible , and orthomodular ( hence atomistic ) . the main difference to the lattice of propositions in quantum mechanics , as regards the formal aspects discussed here , is that @xmath652 does _ not _ satisfy the covering property . otherwise the formal similarities are intriguing and it is tempting to ask whether there is a deeper meaning to this . in this respect it would be interesting to know whether one could give a lattice - theoretic characterisation for @xmath653 ( @xmath1 some fixed spacetime ) , comparable to the characterisation of the lattices of closed subspaces in hilbert space alluded to above . even for @xmath654 such a characterisation seems , as far as i am aware , not to be known . as is well known , the notion of a rigid body , which proves so useful in newtonian mechanics , is incompatible with the existence of a universal finite upper bound for all signal velocities @xcite . as a result , the notion of a perfectly rigid body does not exist within the framework of sr . however , the notion of a _ rigid motion _ does exist . intuitively speaking , a body moves rigidly if , locally , the relative spatial distances of its material constituents are unchanging . the motion of an extended body is described by a normalised timelike vector field @xmath655 , where @xmath214 is an open subset of minkowski space , consisting of the events where the material body in question ` exists ' . we write @xmath656 for the minkowskian scalar product . being normalised now means that @xmath657 ( we do _ not _ choose units such that @xmath658 ) . the lie derivative with respect to @xmath76 is denoted by @xmath659 . for each material part of the body in motion its local rest space at the event @xmath660 can be identified with the hyperplane through @xmath62 orthogonal to @xmath661 : @xmath662 @xmath663 carries a euclidean inner product , @xmath664 , given by the restriction of @xmath665 to @xmath663 . generally we can write @xmath666 where @xmath667 is the one - form associated to @xmath76 . following @xcite the precise definition of ` rigid motion ' can now be given as follows : [ def : rigidmotion ] let @xmath76 be a normalised timelike vector field @xmath76 . the motion described by its flow is _ rigid _ if @xmath668 note that , in contrast to the killing equations @xmath669 , these equations are non linear due to the dependence of @xmath216 upon @xmath76 . we write @xmath670 for the tensor field over spacetime that pointwise projects vectors perpendicular to @xmath76 . it acts on one forms @xmath239 via @xmath671 and accordingly on all tensors . the so extended projection map will still be denoted by @xmath672 . then we e.g. have @xmath673 it is not difficult to derive the following two equations : equation ( [ eq : rigmo2 ] ) simply follows from @xmath674 , so that @xmath675 for all @xmath676 . in fact , @xmath677 , where @xmath678 is the spacetime - acceleration . this follows from @xmath679)= g(a , x)-g(u,\nabla_xu)=g(a , x)$ ] , where @xmath680 was used in the last step . ] @xmath681 where @xmath9 is any differentiable real - valued function on @xmath214 . equation ( [ eq : rigmo1 ] ) shows that the normalised vector field @xmath76 satisfies ( [ eq : def : rigmo ] ) iff any rescaling @xmath682 with a nowhere vanishing function @xmath9 does . hence the normalization condition for @xmath76 in ( [ eq : def : rigmo ] ) is really irrelevant . it is the geometry in spacetime of the flow lines and not their parameterisation which decide on whether motions ( all , i.e. for any parameterisation , or none ) along them are rigid . this has be the case because , generally speaking , there is no distinguished family of sections ( hypersurfaces ) across the bundle of flow lines that would represent ` the body in space ' , i.e. mutually simultaneous locations of the body s points . distinguished cases are those exceptional ones in which @xmath76 is hypersurface orthogonal . then the intersection of @xmath76 s flow lines with the orthogonal hypersurfaces consist of mutually _ einstein synchronous _ locations of the points of the body . an example is discussed below . equation ( [ eq : rigmo2 ] ) shows that the rigidity condition is equivalent to the ` spatially ' projected killing equation . we call the flow of the timelike normalised vector field @xmath76 a _ killing motion _ ( i.e. a spacetime isometry ) if there is a killing field @xmath280 such that @xmath683 . equation ( [ eq : rigmo2 ] ) immediately implies that killing motions are rigid . what about the converse ? are there rigid motions that are not killing ? this turns out to be a difficult question . its answer in minkowski space is : ` yes , many , but not as many as navely expected . ' before we explain this , let us give an illustrative example for a killing motion , namely that generated by the boost killing - field in minkowski space . we suppress all but one spatial directions and consider boosts in @xmath262 direction in two - dimensional minkowski space ( coordinates @xmath684 and @xmath262 ; metric @xmath685 ) . the killing field is denote the vector fields naturally defined by the coordinates @xmath686 . pointwise the dual basis to @xmath687 is @xmath688 . ] @xmath689 which is timelike in the region @xmath690 . we focus on the ` right wedge ' @xmath691 , which is now our region @xmath214 . consider a rod of length @xmath469 which at @xmath692 is represented by the interval @xmath693 , where @xmath694 . the flow of the normalised field @xmath683 is [ eq : boostkill2 ] @xmath695 where @xmath696 labels the elements of the rod at @xmath697 . we have @xmath698 , showing that the individual elements of the rod move on hyperbolae ( ` hyperbolic motion ' ) . @xmath699 is the proper time along each orbit , normalised so that the rod lies on the @xmath262 axis at @xmath697 . the combination @xmath700 is just the flow parameter for @xmath280 ( [ eq : boostkill1 ] ) , sometimes referred to as ` killing time ' ( though it is dimensionless ) . from ( [ eq : boostkill2 ] ) we can solve for @xmath155 and @xmath699 as functions of @xmath684 and @xmath262 : [ eq : lambdatau ] @xmath701 from which we infer that the hypersurfaces of constant @xmath155 are hyperplanes which all intersect at the origin . moreover , we also have @xmath702 ( @xmath293 is just the ordinary exterior differential ) so that the hyperplanes of constant @xmath155 intersect all orbits of @xmath76 ( and @xmath280 ) orthogonally . hence the hyperplanes of constant @xmath155 qualify as the equivalence classes of mutually einstein - simultaneous events in the region @xmath703 for a family of observers moving along the killing orbits . this does not hold for the hypersurfaces of constant @xmath699 , which are curved . the modulus of the spacetime - acceleration ( which is the same as the modulus of the spatial acceleration measured in the local rest frame ) of the material part of the rod labelled by @xmath704 is @xmath705 as an aside we generally infer from this that , given a timelike curve of local acceleration ( modulus ) @xmath239 , infinitesimally nearby orthogonal hyperplanes intersect at a spatial distance @xmath706 . this remark will become relevant in the discussion of part2 of the noether - herglotz theorem given below . in order to accelerate the rod to the uniform velocity @xmath60 without deforming it , its material point labelled by @xmath704 has to accelerate for the eigentime ( this follows from ( [ eq : boostkill2 ] ) ) @xmath707 which depends on @xmath704 . in contrast , the killing time is the same for all material points and just given by the final rapidity . in particular , judged from the local observers moving with the rod , a rigid acceleration requires accelerating the rod s trailing end harder but shorter than pulling its leading end . in terms of the coordinates @xmath708 , which are co - moving with the flow of @xmath280 , and @xmath709 , which are co - moving with the flow of @xmath76 , we just have @xmath710 and @xmath711 respectively . the spacetime metric @xmath30 and the projected metric @xmath216 in terms of these coordinates are : [ eq : metricscomovcoord ] @xmath712 note the simple form @xmath30 takes in terms of @xmath704 and @xmath155 , which are also called the ` rindler coordinates ' for the region @xmath713 of minkowski space . they are the analogs in lorentzian geometry to polar coordinates ( radius @xmath704 , angle @xmath155 ) in euclidean geometry . let us now return to the general case . we decompose the derivative of the velocity one - form @xmath714 as follows : @xmath715 where @xmath716 and @xmath717 are the projected symmetrised and antisymmetrised derivatives respectively ) by @xmath547 and @xmath546 respectively and the symmetrised and antisymmetrised ( covariant- ) derivative by @xmath718 and @xmath719 . for example , @xmath720 and @xmath721 . note that @xmath722 is the same as the ordinary exterior differential @xmath723 . everything we say in the sequel applies to curved spacetimes if @xmath724 is read as covariant derivative with respect to the levi - civita connection . ] [ eq : thetaomega ] @xmath725 the symmetric part , @xmath716 , is usually further decomposed into its traceless and pure trace part , called the _ shear _ and _ expansion _ of @xmath76 respectively . the antisymmetric part @xmath717 is called the _ vorticity _ of @xmath76 . now recall that the lie derivative of @xmath30 is just twice the symmetrised derivative , which in our notation reads : @xmath726 this implies in view of ( [ eq : def : rigmo ] ) , ( [ eq : rigmo2 ] ) , and ( [ eq : thetaomega1 ] ) let @xmath76 be a normalised timelike vector field @xmath76 . the motion described by its flow is rigid iff @xmath76 is of vanishing shear and expansion , i.e. iff @xmath727 . vector fields generating rigid motions are now classified according to whether or not they have a vanishing vorticity @xmath717 : if @xmath728 the flow is called _ irrotational _ , otherwise _ rotational_. the following theorem is due to herglotz @xcite and noether @xcite : [ thm : noetherherglotz1 ] a rotational rigid motion in minkowski space must be a killing motion . an example of such a rotational motion is given by the killing field we now use standard cylindrical coordinates @xmath729 , in terms of which @xmath730 . ] @xmath731 inside the region @xmath732 where @xmath280 is timelike . this motion corresponds to a rigid rotation with constant angular velocity @xmath733 which , without loss of generality , we take to be positive . using the co - moving angular coordinate @xmath734 , the split ( [ eq : hormetric ] ) is now furnished by [ eq : rotsplitmetric ] @xmath735 the metric @xmath216 is curved ( cf . lemma[lemma : noetherherglotz1 - 1 ] ) . but the rigidity condition ( [ eq : def : rigmo ] ) means that @xmath216 , and hence its curvature , can not change along the motion . therefore , even though we can keep a body in uniform rigid rotational motion , we can not put it into this state from rest by purely rigid motions , since this would imply a transition from a flat to a curved geometry of the body . this was first pointed out by ehrenfest @xcite . below we will give a concise analytical expression of this fact ( cf . equation ( [ eq : omegaliezero ] ) ) . all this is in contrast to the translational motion , as we will also see below . the proof of theorem[thm : noetherherglotz1 ] relies on arguments from differential geometry proper and is somewhat tricky . here we present the essential steps , basically following @xcite and @xcite in a slightly modernised notation . some straightforward calculational details will be skipped . the argument itself is best broken down into several lemmas . at the heart of the proof lies the following general construction : let @xmath1 be the spacetime manifold with metric @xmath30 and @xmath736 the open region in which the normalised vector field @xmath76 is defined . we take @xmath214 to be simply connected . the orbits of @xmath76 foliate @xmath214 and hence define an equivalence relation on @xmath214 given by @xmath184 iff @xmath62 and @xmath140 lie on the same orbit . the quotient space @xmath737 is itself a manifold . tensor fields on @xmath738 can be represented by ( i.e. are in bijective correspondence to ) tensor fields @xmath739 on @xmath214 which obey the two conditions : [ eq : projectibility ] @xmath740 tensor fields satisfying ( [ eq : projectibility1 ] ) are called _ horizontal _ , those satisfying both conditions ( [ eq : projectibility ] ) are called _ projectable_. the @xmath741-dimensional metric tensor @xmath216 , defined in ( [ eq : hormetric ] ) , is an example of a projectable tensor if @xmath76 generates a rigid motion , as assumed here . it turns @xmath742 into a @xmath741-dimensional riemannian manifold . the covariant derivative @xmath743 with respect to the levi - civita connection of @xmath216 is given by the following operation on projectable tensor fields : @xmath744 i.e. by first taking the covariant derivative @xmath724 ( levi - civita connection in @xmath745 ) in spacetime and then projecting the result horizontally . this results again in a projectable tensor , as a straightforward calculation shows . the horizontal projection of the spacetime curvature tensor can now be related to the curvature tensor of @xmath738 ( which is a projectable tensor field ) . without proof we state [ lemma : noetherherglotz1 - 1 ] let @xmath76 generate a rigid motion in spacetime . then the horizontal projection of the totally covariant ( i.e. all indices down ) curvature tensor @xmath421 of @xmath746 is related to the totally covariant curvature tensor @xmath747 of @xmath742 by the following equation appears with a minus sign on the right hand side of ( [ eq : projectedcurvaturerelation ] ) because the first index on the hatted curvature tensor is lowered with @xmath216 rather than @xmath30 . this induces a minus sign due to ( [ eq : hormetric ] ) , i.e. as a result of our ` mostly - minus'-convention for the signature of the spacetime metric . ] : @xmath748 where @xmath749 is the total antisymmetriser , which here projects tensors of rank four onto their totally antisymmetric part . formula ( [ eq : projectedcurvaturerelation ] ) is true in any spacetime dimension @xmath18 . note that the projector @xmath750 guarantees consistency with the first bianchi identities for @xmath421 and @xmath747 , which state that the total antisymmetrisation in their last three slots vanish identically . this is consistent with ( [ eq : projectedcurvaturerelation ] ) since for tensors of rank four with the symmetries of @xmath751 the total antisymmetrisation on tree slots is identical to @xmath749 , the symmetrisation on all four slots . the claim now simply follows from @xmath752 . we now restrict to spacetime dimensions of four or less , i.e. @xmath753 . in this case @xmath754 since @xmath672 makes the tensor effectively live over @xmath125 dimensions , and any totally antisymmetric four - tensor in three or less dimensions must vanish . applied to ( [ eq : projectedcurvaturerelation ] ) this means that @xmath755 , for horizontality of @xmath717 implies @xmath756 . hence the right hand side of ( [ eq : projectedcurvaturerelation ] ) just contains the pure tensor product @xmath757 . now , in our case @xmath758 since @xmath745 is flat minkowski space . this has two interesting consequences : first , @xmath742 is curved iff the motion is rotational , as exemplified above . second , since @xmath747 is projectable , its lie derivative with respect to @xmath76 vanishes . hence ( [ eq : projectedcurvaturerelation ] ) implies @xmath759 , which is equivalent to . ] @xmath760 this says that the vorticity can not change along a rigid motion in flat space . it is the precise expression for the remark above that you can not rigidly _ set _ a disk into rotation . note that it also provides the justification for the global classification of rigid motions into rotational and irrotational ones . a sharp and useful criterion for whether a rigid motion is killing or not is given by the following [ lemma : noetherherglotz1 - 2 ] let @xmath76 be a normalised timelike vector field on a region @xmath761 . the motion generated by @xmath76 is killing iff it is rigid and @xmath762 is exact on @xmath214 . that the motion generated by @xmath76 be killing is equivalent to the existence of a positive function @xmath763 such that @xmath764 , i.e. @xmath765 . in view of ( [ eq : thetaomega1 ] ) this is equivalent to @xmath766 which , in turn , is equivalent to @xmath727 and @xmath767 . this is true since @xmath716 is horizontal , @xmath768 , whereas the first term in ( [ eq : matteo1 ] ) vanishes upon applying @xmath672 . the result now follows from reading this equivalence both ways : 1 ) the killing condition for @xmath769 implies rigidity for @xmath76 and exactness of @xmath762 . 2 ) rigidity of @xmath76 and @xmath770 imply that @xmath769 is killing , where @xmath771 . we now return to the condition ( [ eq : omegaliezero ] ) and express @xmath772 in terms of @xmath723 . for this we recall that @xmath677 ( cf . footnote[foot : lievel ] ) and that lie derivatives on forms commute with exterior derivatives , where @xmath773 is the map of inserting @xmath76 in the first slot . ] . hence we have @xmath774 here we used the fact that the additional terms that result from the lie derivative of the projection tensor @xmath672 vanish , as a short calculation shows , and also that on forms the projection tensor @xmath672 can be written as @xmath775 , where @xmath773 denotes the map of insertion of @xmath76 in the first slot . now we prove [ lemma : noetherherglotz1 - 3 ] let @xmath76 generate a rigid motion in flat space such that @xmath776 , then @xmath777 equation ( [ eq : omegaliezero ] ) says that @xmath717 is projectable ( it is horizontal by definition ) . hence @xmath778 is projectable , which implies @xmath779 using ( [ eq : decveloneform ] ) with @xmath727 one has @xmath780 antisymmetrisation in the first two tensor slots makes the first term on the right vanish due to the flatness on @xmath724 . the antisymmetrised right hand side is hence equal to @xmath781 . taking the lie derivative of both sides makes the left hand side vanish due to ( [ eq : hatnablaomega1 ] ) , so that @xmath782 where we also used ( [ eq : omegaliezero ] ) . so we see that @xmath783 if @xmath776 . ) is generally not true if @xmath728 ; see equation ( [ eq : proof : noeterherglotz2i ] ) . ] the last three lemmas now constitute a proof for theorem[thm : noetherherglotz1 ] . indeed , using ( [ eq : lieacczero ] ) in ( [ eq : lieomega1 ] ) together with ( [ eq : omegaliezero ] ) shows @xmath784 , which , according to lemma[lemma : noetherherglotz1 - 2 ] , implies that the motion is killing . next we turn to the second part of the theorem of noether and herglotz , which reads as follows : [ thm : noetherherglotzp2 ] all irrotational rigid motions in minkowski space are given by the following construction : take a twice continuously differentiable curve @xmath785 in minkowski space , where w.l.o.g @xmath699 is the eigentime , so that @xmath786 . let @xmath787 be the hyperplane through @xmath788 intersecting the curve @xmath789 perpendicularly . let @xmath214 be a the tubular neighbourhood of @xmath789 in which no two hyperplanes @xmath790 intersect for any pair @xmath791 of points on the curve . in @xmath214 define @xmath76 as the unique ( once differentiable ) normalised timelike vector field perpendicular to all @xmath792 . the flow of @xmath76 is the sought - for rigid motion . we first show that the flow so defined is indeed rigid , even though this is more or less obvious from its very definition , since we just defined it by ` rigidly ' moving a hyperplane through spacetime . in any case , analytically we have , @xmath793 in @xmath214 any @xmath262 lies on exactly one such hyperplane , @xmath794 , which means that there is a function @xmath795 so that @xmath796 and hence @xmath797 . this implies @xmath798 . using the expression for @xmath9 from ( [ eq : proof : noeterherglotz2a ] ) this is equivalent to @xmath799\,,\ ] ] where ` @xmath800 ' denotes the ` identity vector - field ' , @xmath801 , in minkowski space . note that in @xmath214 we certainly have @xmath802 and hence @xmath803 . in @xmath214 we now define the normalised timelike vector field , @xmath804 . ] @xmath805 using ( [ eq : proof : noeterherglotz2b ] ) , its derivative is given by @xmath806/(n^2c^2)\,,\ ] ] where @xmath807 this immediately shows that @xmath808 ( since @xmath809 ) and therefore that @xmath810 . hence @xmath76 , as defined in ( [ eq : proof : noeterherglotz2c ] ) , generates an irrotational rigid motion . for the converse we need to prove that any irrotational rigid motion is obtained by such a construction . so suppose @xmath76 is a normalised timelike vector field such that @xmath810 . vanishing @xmath717 means @xmath811 . this is equivalent to @xmath812 , which according to the frobenius theorem in differential geometry is equivalent to the integrability of the distribution it denotes an assignment of a linear subspace @xmath813 in the tangent space @xmath814 to each point @xmath62 of @xmath1 . the distribution @xmath815 is defined by @xmath816 . a distribution is called ( locally ) integrable if ( in the neighbourhood of each point ) there is a submanifold @xmath817 of @xmath1 whose tangent space at any @xmath818 is just @xmath813 . ] @xmath815 , i.e. the hypersurface orthogonality of @xmath76 . we wish to show that the hypersurfaces orthogonal to @xmath76 are hyperplanes . to this end consider a spacelike curve @xmath819 , where @xmath397 is the proper length , running within one hypersurface perpendicular to @xmath76 . the component of its second @xmath397-derivative parallel to the hypersurface is given by ( to save notation we now simply write @xmath76 and @xmath820 instead of @xmath821 and @xmath822 ) @xmath823 where we made a partial differentiation in the second step and then used @xmath727 . geodesics in the hypersurface are curves whose second derivative with respect to proper length have vanishing components parallel to the hypersurface . now , ( [ eq : proof : noeterherglotz2f ] ) implies that geodesics in the hypersurface are geodesics in minkowski space ( the hypersurface is ` totally geodesic ' ) , i.e. given by straight lines . hence the hypersurfaces are hyperplanes . theorem[thm : noetherherglotzp2 ] precisely corresponds to the newtonian counterpart : the irrotational motion of a rigid body is determined by the worldline of any of its points , and any timelike worldline determines such a motion . we can rigidly put an extended body into any state of translational motion , as long as the size of the body is limited by @xmath706 , where @xmath239 is the modulus of its acceleration . this also shows that ( [ eq : lieacczero ] ) is generally not valid for irrotational rigid motions . in fact , the acceleration one - form field for ( [ eq : proof : noeterherglotz2c ] ) is @xmath824 from which one easily computes @xmath825 from this one sees , for example , that for _ constant acceleration _ , defined by @xmath826 ( constant acceleration in time as measured in the instantaneous rest frame ) , we have @xmath784 and hence a killing motion . clearly , this is just the motion ( [ eq : boostkill2 ] ) for the boost killing field ( [ eq : boostkill1 ] ) . the lie derivative of @xmath762 is now easily obtained : @xmath827 showing explicitly that it is not zero except for motions of constant acceleration , which were just seen to be killing motions . in contrast to the irrotational case just discussed , we have seen that we can not put a body rigidly into rotational motion . in the old days this was sometimes expressed by saying that the rigid body in sr has only three instead of six degrees of freedom . this was clearly thought to be paradoxical as long as one assumed that the notion of a perfectly rigid body should also make sense in the framework of sr . however , this hope was soon realized to be physically untenable @xcite . in this appendix we spell out in detail some of the mathematical notions that were used in the main text . given a set @xmath2 , recall that an _ equivalence relation _ is a subset @xmath508 such that for all @xmath828 the following conditions hold : 1 ) @xmath829 ( called ` reflexivity ' ) , 2 ) if @xmath830 then @xmath831 ( called ` symmetry ' ) , and 3 ) if @xmath830 and @xmath832 then @xmath833 ( called ` transitivity ' ) . once @xmath421 is given , one often conveniently writes @xmath184 instead of @xmath830 . given @xmath507 , its _ equivalence class _ , @xmath834\subseteq s$ ] , is given by all points @xmath421-related to @xmath62 , i.e. @xmath834:=\{q\in s\mid ( p , q)\in r\}$ ] . one easily shows that equivalence classes are either identical or disjoint . hence they form a _ partition _ of @xmath2 , that is , a covering by mutually disjoint subsets . conversely , given a partition of a set @xmath2 , it defines an equivalence relation by declaring two points as related iff they are members of the same cover set . hence there is a bijective correspondence between partitions of and equivalence relations on a set @xmath2 . the set of equivalence classes is denoted by @xmath835 or @xmath836 . there is a natural surjection @xmath837 , @xmath838 $ ] . if in the definition of equivalence relation we exchange symmetry for antisymmetry , i.e. @xmath830 and @xmath831 implies @xmath839 , the relation is called a _ partial order _ , usually written as @xmath131 for @xmath830 . if , instead , reflexivity is dropped and symmetry is replaced by asymmetry , i.e. @xmath830 implies @xmath840 , one obtains a relation called a _ strict partial order _ , usually denoted by @xmath841 for @xmath830 . an _ left action _ of a group @xmath249 on a set @xmath2 is a map @xmath842 , such that @xmath843 ( @xmath844 group identity ) and @xmath845 . if instead of the latter equation we have @xmath846 one speaks of a _ right action_. for left actions one sometimes conveniently writes @xmath847 , for right actions @xmath848 . an action is called _ transitive _ if for every pair @xmath849 there is a @xmath850 such that @xmath851 , and _ simply transitive _ if , in addition , @xmath852 determine @xmath30 uniquely , that is , @xmath853 for some @xmath397 implies @xmath854 . the action is called _ effective _ if @xmath855 for all @xmath397 implies @xmath856 ( ` every @xmath857 moves something ' ) and _ free _ if @xmath855 for some @xmath397 implies @xmath856 ( ` no @xmath857 has a fixed point ' ) . it is obvious that simple transitivity implies freeness and that , conversely , freeness and transitivity implies simple transitivity . moreover , for abelian groups , effectivity and transitivity suffice to imply simple transitivity . indeed , suppose @xmath858 holds for some @xmath859 , then we also have @xmath860 for all @xmath861 and hence @xmath862 by commutativity . this implies that @xmath858 holds , in fact , for all @xmath397 . for any @xmath859 we can consider the _ stabilizer subgroup _ @xmath863 if @xmath112 is transitive , any two stabilizer subgroups are conjugate : @xmath864 . by definition , if @xmath112 is free all stabilizer subgroups are trivial ( consist of the identity element only ) . in general , the intersection @xmath865 is the normal subgroup of elements acting trivially on @xmath2 . if @xmath112 is an action of @xmath249 on @xmath2 , then there is an effective action @xmath866 of @xmath867 on @xmath2 , defined by @xmath868,s):=\phi(g , s)$ ] , where @xmath869 $ ] denotes the @xmath870-coset of @xmath870 in @xmath249 . the _ orbit _ of @xmath397 in @xmath2 under the action @xmath112 of @xmath249 is the subset @xmath871 it is easy to see that group orbits are either disjoint or identical . hence they define a partition of @xmath2 , that is , an equivalence relation . a relation @xmath421 on @xmath2 is said to be invariant under the self map @xmath3 if @xmath872 . it is said to be invariant under the action @xmath112 of @xmath249 on @xmath2 if @xmath873 for all @xmath850 . if @xmath421 is such a @xmath249-invariant equivalence relation , there is an action @xmath874 of @xmath249 on the set @xmath835 of equivalence classes , defined by @xmath875):=[\phi(g , p)]$ ] . a general theorem states that invariant equivalence relations exist for transitive group actions , iff the stabilizer subgroups ( which in the transitive case are all conjugate ) are maximal ( e.g. theorem1.12 in @xcite ) . [ def : affinespace ] an @xmath18-dimensional * affine space * over the field @xmath876 ( usually @xmath296 or @xmath877 ) is a triple @xmath878 , where @xmath2 is a non - empty set , @xmath19 an @xmath18-dimensional vector space over @xmath876 , and @xmath879 an effective and transitive action @xmath880 of @xmath19 ( considered as abelian group with respect to addition of vectors ) on @xmath2 . we remark that an effective and transitive action of an abelian group is necessarily simply transitive . hence , without loss of generality , we could have required a simply transitive action in definition[def : affinespace ] straightaway . we also note that even though the action @xmath879 only refers to the abelian group structure of @xmath19 , it is nevertheless important for the definition of an affine space that @xmath19 is , in fact , a vector space ( see below ) . any ordered pair of points @xmath881 uniquely defines a vector @xmath60 , namely that for which @xmath882 . it can be thought of as the difference vector pointing from @xmath140 to @xmath62 . we write @xmath883 , where @xmath884 is a map which satisfies the conditions [ thm : altdefaffspace ] let @xmath2 be a non - empty set , @xmath19 an @xmath18-dimensional vector space over @xmath876 and @xmath884 a map satisfying conditions ( [ eq : affdiffmap ] ) . then @xmath2 is an @xmath18-dimensional affine space over @xmath876 with action @xmath886 . one usually writes @xmath887 , which defines what is meant by ` @xmath80 ' between an element of an affine space and an element of @xmath19 . note that addition of two points in affine space is not defined . the property of being an action now states @xmath888 and @xmath889 , so that in the latter case we may just write @xmath890 . similarly we write @xmath891 , defining what is meant by ` @xmath81 ' between two elements of affine space . the minus sign also makes sense between an element of affine space and an element of vector space if one defines @xmath892 . we may now write equations like @xmath893 the formal proof of which is again left to the reader . it implies that considered as abelian group , any linear subspace @xmath56 defines a subgroup . the orbit of that subgroup in @xmath2 through @xmath507 is an affine subspace , denoted by @xmath68 , i.e. @xmath894 which is an affine space over @xmath67 in its own right of dimension @xmath895 . one - dimensional affine subspaces are called _ ( straight ) lines _ , two - dimensional ones _ planes _ , and those of co - dimension one are called _ affine morphisms , or simply affine maps , are structure preserving maps between affine spaces . to define them in view of definition[def : affinespace ] we recall once more the significance of @xmath19 being a vector space and not just an abelian group . this enters the following definition in an essential way , since there are considerably more automorphisms of @xmath19 as abelian group , i.e. maps @xmath28 that satisfy @xmath896 for all @xmath32 , than automorphisms of @xmath19 as linear space which , in addition , need to satisfy @xmath897 for all @xmath45 and all @xmath898 ) . in fact , the difference is precisely that the latter are all continuous automorphisms of @xmath19 ( considered as topological abelian group ) , whereas there are plenty ( uncountably many ) discontinuous ones , see @xcite . , then it is easy to see that @xmath896 for all @xmath32 implies @xmath897 for all @xmath45 and all @xmath899 ( rational numbers ) . for continuous @xmath9 this implies the same for all @xmath900 . all discontinuous @xmath9 are obtained as follows : let @xmath901 be a ( necessarily uncountable ) basis of @xmath296 as vector space over @xmath902 ( ` hamel basis ' ) , prescribe any values @xmath903 , and extend @xmath9 linearly to all of @xmath296 . any value - prescription for which @xmath904 is not constant gives rise to a non @xmath296-linear and discontinuous @xmath9 . such @xmath9 are ` wildly ' discontinuous in the following sense : for _ any _ interval @xmath905 , @xmath906 is dense @xcite . ] in the convenient way of writing introduced above , this is equivalent to @xmath911 for all @xmath912 and all @xmath45 . ( note that the @xmath80 sign on the left refers to the action @xmath879 of @xmath19 on @xmath2 , whereas that on the right refers to the action @xmath913 of @xmath914 on @xmath526 . ) this shows that an affine map @xmath159 is determined once the linear map @xmath9 between the underlying vector spaces is given and the image @xmath497 of an arbitrary point @xmath140 is specified . equation ( [ eq : affinemap2 ] ) can be rephrased as follows : setting @xmath915 equation ( [ eq : affinemap2 ] ) is equivalent to @xmath916 for all @xmath917 . in view of the alternative definition of affine spaces suggested by proposition[thm : altdefaffspace ] , this shows that we could have defined affine maps alternatively to ( [ eq : affinemap1 ] ) by ( @xmath918 is the difference map in @xmath526 ) @xmath919 affine bijections of an affine space @xmath878 onto itself form a group , the _ affine group _ , denoted by @xmath920 . group multiplication is just given by composition of maps , that is @xmath921 . it is immediate that the composed maps again satisfy ( [ eq : affinemap1 ] ) . for any @xmath45 , the map @xmath922 is an affine bijection for which @xmath923 . note that in this case ( [ eq : affinemap1 ] ) simply turns into the requirement @xmath924 for all @xmath106 , which is clearly satisfied due to @xmath19 being a commutative group . hence there is a natural embedding @xmath925 , the image @xmath926of which is called the subgroup of _ translations_. the map @xmath927 defines a group homomorphism @xmath928 , since @xmath929 . we have just seen that the translations are in the kernel of this map . in fact , the kernel is equal to the subgroup @xmath926 of translations , as one easily infers from ( [ eq : affinemap3 ] ) with @xmath923 , which is equivalent to @xmath930 for all @xmath917 . hence there exists a @xmath45 such that for all @xmath507 we have @xmath931 . the quotient group @xmath932 is then clearly isomorphic to @xmath933 . there are also embeddings @xmath934 , but no canonical one : each one depends on the choice of a reference point @xmath935 , and is given by @xmath936 , where @xmath937 for all @xmath507 . this shows that @xmath920 is isomorphic to the semi - direct product @xmath938 , though the isomorphism depends on the choice of @xmath935 . the action of @xmath939 on @xmath507 is then defined by @xmath940 which is easily checked to define indeed an ( @xmath108 dependent ) action of @xmath938 on @xmath2 . [ def : linearframe ] a * linear frame * of the @xmath18-dimensional vector space @xmath19 over @xmath876 is a basis @xmath941 of @xmath19 , regarded as a linear isomorphism @xmath942 , given by @xmath943 . the set of linear frames of @xmath19 is denoted by @xmath944 . there is a natural right action of @xmath946 on @xmath944 , given by @xmath947 . it is immediate that this action is simply transitive . it is sometimes called the _ passive interpretation _ of the transformation group @xmath946 , presumably because it moves the frames associated to the observer and not the points of @xmath19 . on the other hand , any frame @xmath9 induces an isomorphism of algebras @xmath948 , given by @xmath949 . if @xmath950 , then @xmath951 , where @xmath941 . restricted to @xmath952 , this induces a group isomorphism @xmath953 and hence an @xmath9-dependent action of @xmath946 on @xmath19 by linear transformations , defined by @xmath954 , where @xmath955 . this is sometimes called the _ active interpretation _ of the transformation group @xmath946 , presumably because it really moves the points of @xmath19 . [ def : affineframe ] an * affine frame * of the @xmath18-dimensional affine space @xmath878 over @xmath876 is a tuple @xmath956 , where @xmath108 is a base point in @xmath2 and @xmath942 is a linear frame of @xmath19 . @xmath159 is regarded as a map @xmath957 , given by @xmath958 . we denote the set of affine frames by @xmath959 . if we regard @xmath945 as an affine space @xmath961 , it comes with a distinguished base point @xmath108 , the zero vector . the group @xmath962 is therefore naturally isomorphic to @xmath963 . the latter naturally acts on @xmath945 in the standard way , @xmath964 , where group multiplication is given by @xmath965 the group @xmath963 has a natural right action on @xmath959 , where @xmath966 . explicitly , for @xmath967 and @xmath968 , this action reads : @xmath969 it is easy to verify directly that this is an action which , moreover , is again simply transitive . it is referred to as the _ passive interpretation _ of the affine group @xmath963 . conversely , depending on the choice of an affine frame @xmath970 , there is a group isomorphism @xmath971 , given by @xmath972 , and hence an @xmath159 dependent action of @xmath963 by affine maps on @xmath878 . if @xmath968 and @xmath973 , the action reads @xmath974 this is called the _ active interpretation _ of the affine group @xmath963 . an affine frame @xmath975 with @xmath941 defines @xmath976 points @xmath977 , where @xmath978 and @xmath979 for @xmath980 . conversely , any @xmath976 points @xmath977 in affine space , for which @xmath981 are linearly independent , define an affine frame . note that this linear independence does not depend on the choice of @xmath982 as our base point , as one easily sees from the identity @xmath983 which holds for any set @xmath984 of @xmath985 points in affine space . to prove it one just needs ( [ eq : affdistributivity ] ) . hence we say that these points are _ affinely independent _ iff , e.g. , the set of @xmath0 vectors @xmath986 is linearly independent . therefore , an affine frame of @xmath18-dimensional affine space is equivalent to @xmath976 affinely independent points . such a set of points is also called an _ affine basis_. this motivates writing the sums on the right hand sides of ( [ eq : affbasisexp ] ) in a perfectly symmetric way without preference of any point @xmath995 : @xmath996 where the right hand side is defined by any of the expressions ( [ eq : affbasisexp ] ) . this defines certain _ linear combinations _ of affine points , namely those whose coefficients add up to one . accordingly , the _ affine span _ of points @xmath997 in affine space is defined by @xmath998 harvey brown and oliver pooley . minkowski space - time : a glorious non - entity . in dennis dieks , editor , _ the ontology of spacetime _ , volume 1 of _ philosophy and foundations of physics _ , pages 6792 , 2006 . andrzej trautman . foundations and current problems of general relativity . in andrzej trautman , felix pirani , and hermann bondi , editors , _ lectures on general relativity _ , volume 1 of _ brandeis summer institute in theoretical physics _ , pages 1248 . prentice - hall , inc . , englewood cliffs , new jersey , 1964 .

it is now well - established that early - type galaxies possess substantial ism , with neutral hydrogen masses on the order of @xmath0 @xmath5 @xmath6 @xcite . accompanying this neutral gas are emission - line regions extending out to several kiloparsecs , with line ratios characteristic of gas in a relatively low state of ionization . they appear to be powered by the diffuse galactic background , rather than being a superposition of individual h ii regions around bright compact sources . in the past , much effort has been devoted to understanding the sources of ionizing radiation powering these nebulae . in passively - evolving ( non - star - forming ) galaxies , it is the old stellar population itself which appears likely to provide the dominant contribution to the local extreme - uv background ( e.g * ? ? ? * ) , at least outside of their innermost nuclei . there , low - luminosity agn had until recently been thought to provide at least a significant contribution , at least in many cases ( e.g. * ? ? ? * ) . however , increasing evidence suggests that the old stellar population may still provide a significant , or even dominant role in powering liner emission ( e.g. * ? ? ? * ) . within the stellar population , post - asymptotic giant branch stars ( pagbs ) remain the favoured candidate for powering the observed emission lines . in particular , pagb stars are expected to provide sufficient ionizing photons in order to account for the observed h@xmath7 flux @xcite , and the radial surface brightness profiles of the observed emission lines are consistent with ionization by a diffuse population , rather than a central source ( e.g. * ? ? ? * ; * ? ? ? * ) . at the same time , similar arguments have proven capable of ruling out other possible ionizing sources . recently , @xcite proposed that understanding the mechanism powering the diffuse line - emission in passively - evolving galaxies may shed light on the origin of type ia supernovae ( sne ia ) . there exist two families of models for how these tremendous explosions come about : in the double - degenerate scenario @xcite , these events result from the merger of a binary pair of white dwarfs ( wds ) . in the standard model of the single - degenerate channel @xcite , a carbon - oxygen wd accretes matter from a main sequence or red giant companion until triggering a thermonuclear explosion . if the latter is correct , then accreting , nuclear - burning wds should be an extremely luminous component of any stellar population . with temperatures on the order of @xmath8 @xmath4 , they should provide the dominant ionizing background in relatively young ( @xmath3 4gyr ) passively - evolving stellar populations . if this is the case , it should be readily apparent from spectroscopic observations of those ellipticals which host low - ionization emission - line regions @xcite . this provides a unique opportunity to constrain plausible progenitor models for sne ia . in particular , the observed line flux in recombination lines of ionized helium can provide an excellent test for the presence of relatively high temperature sources ( @xmath9 2 6 @xmath10 @xmath2 ) , and their total ionizing luminosity . this covers much of the temperature range we expect from accreting , nuclear - burning white dwarfs , such as supersoft sources ( ssss ) , or in particular , the so - called ` ultra - soft sources ' ( those accreting white dwarfs with greatly inflated photospheres , see * ? ? ? * ) . however , we expect higher temperatures ( up to @xmath0 @xmath4 ) for the most massive ( @xmath11 1.0 @xmath12 ) nuclear - burning white dwarfs . in the past , any possible population of such objects has been severely constrained based on the observed paucity of soft x - ray emission in early type galaxies , and the lack of individual soft x - ray sources ( e.g. * ? ? ? * ; * ? ? ? however , independent diagnostic tools remain extremely important . here we extend the work of @xcite to include a number of emission - line diagnostic tests for the presence of a population of very hot sources ( @xmath13 @xmath11 @xmath14 ) , as well as those too cool to strongly ionize he ii ( @xmath13 @xmath9 @xmath15 ) . in @xmath162 , we review our model assumptions for the ism in early - type galaxies , as well as the possible sources of ionizing radiation in passively - evolving stellar populations . in @xmath163 , we discuss those lines which are most sensitive to ionization by high temperature sources , in particular those of [ o i ] , [ n i ] , and [ c ii ] . together with observations of recombination lines of he ii , this extends the range over which emission lines can meaningfully constrain the hardness of the ionizing background in passively - evolving galaxies to cover all temperatures expected from steadily nuclear - burning white dwarfs , including those with inflated photospheres ( @xmath17 @xmath18 @xmath13 @xmath18 @xmath4 ) . in @xmath164 , we outline the even more pronounced effect of any putative sd progenitor population ( with @xmath19 @xmath11 @xmath2 ) in young ( t @xmath3 1gyr ) , post - starburst populations ( such as e+a galaxies ) . in this case the ionizing luminosity provided by sd progenitors would exceed that available from any other stellar source by up to two orders of magnitude , indicating the specific luminosity in any of the classical strong optical lines ( h@xmath20 , [ o ii ] 3727 , etc . ) should be similarly enhanced . we follow this in @xmath16 5 with estimates of the observational prospects for producing limits using a number of instruments and surveys , as well as how the presence of a significant population of accreting wds may compromise measures of the sfr in galaxies . finally , we discuss implications for the evolution of binary stellar populations and the progenitors of type ia supernovae . we model the emission - line regions in early - type galaxies using the 1-d photoionization code mappings iii ( see e.g. * ? ? ? our procedure follows closely that of @xcite ; here for convenience we summarize some of the essential ingredients . unless otherwise stated , we assume a solar @xcite metallicity for the gas phase , and 2.5 solar for the stellar population . the former is consistent with the mean oxygen abundance measured in the warm ism of nearby retired galaxies @xcite . for the ionizing continuum from the ` normally - evolving ' ( single star ) stellar population , we use the population synthesis calculations of @xcite . the most common morphology of the cold ism in such galaxies is a disky distribution extending from the nucleus out to several tens of kpc @xcite , with emission - line nebulae typically ionized by the diffuse background of the stellar population rather than nearby individual sources . therefore , we assume plane parallel geometry throughout . for our standard case , we assume a constant hydrogen number density of 100@xmath21 , consistent with the observed [ s ii ] 6717/6731 ratio observed in stacked passively - evolving galaxies from the sdss @xcite , although this may only be an approximate upper bound @xcite . in principle , the majority of the optical emission lines considered in this work are relatively insensitive to variations in the density . in our photoionization calculations , we assume that the modeled nebulae are ionization - bounded ( see * ? ? ? * and references therein ) , and terminate our calculations when the ionized hydrogen fraction falls below 1% . the ionization parameter ( u = @xmath22 ) is similarly well constrained by the observed ratio of [ o iii ] 5007 / h@xmath23 , generally lying in the range -4 @xmath3 log(u ) @xmath3 -3.5 ( e.g. * ? ? ? * ; * ? ? ? * ) . where appropriate , we normalize the line flux per unit stellar mass of the host population , and assume a covering fraction of unity . this can of course be scaled as needed , since the line luminosity is directly proportional to the stellar ionizing luminosity absorbed by the ism in ellipticals . in practice , it is often more convenient to provide line predictions normalized to the nearest recombination line of hydrogen , in order to minimize the importance of reddening in comparison with observed values . this has the added benefit of removing any dependence on the covering fraction of the gas , therefore we also provide predictions for each optical line luminosity relative to the nearest hydrogen recombination line . ] in early - type galaxies , pagbs remain favoured to be the predominant ionizing source ( at least within the stellar population ) . empirically , this has been strongly supported by energetic considerations ( e.g. * ? ? ? * ) , as well as more detailed emission - line studies @xcite . the spectral emission from these sources has been studied in detail by @xcite , whose models are included in the spectral synthesis calculations of @xcite used here . despite mounting evidence ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , some doubt remains about the longevity of post - agb stars , and therefore their expected numbers and contribution to the ionizing background @xcite . therefore , where appropriate , we will consider in our calculations the ionizing background with and without the contribution normally expected from pagbs . non - stellar sources , such as low - luminosity agn , likely also play a role in many galaxies @xcite , at least in their central regions . fast shocks have been thought to play a role in some cases , as well as ionization by the hot , x - ray emitting diffuse ism @xcite . however , such contaminating cases can be excluded with the careful use of line diagnostics ( e.g * ? ? ? * ) , allowing one to confine their attention to so - called retired galaxies where the old stellar population is the principle source of ionizing emission . other stellar sources , such as low - mass x - ray binaries , have been suggested , but do not appear to be significant in this regard @xcite . accreting , nuclear - burning wds have , however , been found to likely provide a significant contribution , and would be expected to dominate the ionizing contribution for relatively young , passively - evolving stellar populations , at least if they exist in sufficient numbers to account for the observed sn ia rate @xcite . even if they are not the primary evolutionary channel for sne ia , accreting wds may still account for a substantial contribution to the ionizing background ( see chen et al . in prep ) . the number of nuclear - burning wds expected in any stellar population , assuming the sd channel accounts for all sne ia , can be estimated from the sn ia rate , the mass accretion rate , and the characteristic mass accreted per sn ia ( as in * ? ? ? * ; * ? ? ? * ) : @xmath24 where @xmath25 is sn ia rate of the galaxy , and @xmath26 is the total mass accreted prior to explosion ( @xmath27 @xmath28 ) . this depends sensitively on our assumptions regarding the typical accretion rate , which in turn depends on as - yet poorly understood details of binary evolution . fortunately , we can remove any dependence on the accretion rate if we consider instead the total bolometric luminosity of all accreting , nuclear - burning wds , @xmath29 . the luminosity of any individual source is simply @xmath30 , with @xmath31erg / s the energy release from nuclear - burning of hydrogen , and @xmath32 the hydrogen fraction of solar metallicity gas . therefore , the total luminosity of any putative sd progenitor population can easily be estimated as : @xmath33 independent of the assumed accretion rate @xcite . for the snia rate we will use the delay time distribution derived for passively evolving galaxies by @xcite : @xmath34snia / century/@xmath35 ( assuming a chabrier imf and z = 0.05 ) . the spectral shape of the ionizing background in ellipticals can then be found from the sum of two terms : @xmath36 .predicted specific line luminosities ( in units of erg / s/@xmath12 ) for those lines most strongly effected by the introduction of a high temperature source population . we assume ionization by a 3 gyr - old ` normal ' stellar population only ( column 3 ) , and with the addition of a @xmath13 = @xmath4 sd sn ia progenitor population ( column 4 ) . in both cases , we take log(u ) = -3.5 . [ line_list ] [ cols="^,^,^,^,^ " , ] [ fits ] in figure [ abs_lums ] , we plot the total luminosity in ly@xmath20 , h@xmath20 , [ n ii ] 6583 , [ o iii ] 5007 and the [ o ii ] 3727 doublet , as a function of source temperature for ionization by sd progenitors alone . note that the predicted ly@xmath20 line luminosity should be only be considered an indicative estimate , as discussed in @xmath16 3 . we take a 500 myr - old ssp as our example of a post - starburst stellar population and assume our standard characteristic accreted mass @xmath37 . from fig.5 one can see that the presence of a high - temperature sd progenitor population leads to an enhancement for all lines by well over an order of magnitude ( except for the [ o iii ] 5007 line for high source temperatures ) . the [ n ii ] 6583 line luminosity exceeds that expected in the case of photoionization by the `` normally - evolving '' stellar population ( @xmath9 3.6 @xmath10 @xmath38 @xmath39 for @xmath40 = @xmath41 ) by a factor of @xmath42 across the considered temperature range . the most luminous of the considered lines is the [ o ii ] 3727 doublet , which lies in the near - uv on the boundary of the wavelength range typically considered the `` optical '' . however , a number of ground - based surveys have already been carried out specifically investigating this line ( typically as a proxy for the sfr , see @xmath16 [ prospects ] ) . the total luminosity in any line is determined by the incident ionizing flux , which in the sd scenario is proportional to the accreted ( and nuclear - processed ) mass ( recall eq . [ eq : ltot ] ) . therefore any measurement ( or upper limit ) on their luminosities associated with a given stellar population will constrain the total ionizing luminosity of any putative sd progenitors . this is particularly useful in the lower range of temperatures predicted for nuclear - burning wds ( @xmath43 ) , where both the he ii diagnostic and tracers of the str " omgren boundary fail . the significant enhancement in the line luminosity expected given photoionization by sd progenitors should allow the total mass processed by nuclear - burning accreting wds to be constrained to @xmath3 0.01 0.001 @xmath44/sn ia ( for example , using the [ o ii ] 3727 line , see fig . [ oii_ev_lums ] ) , should line fluxes be found to be consistent with ionization by pagbs alone . 6300 line and the [ o ii ] 3727 doublet as a function of stellar population age . we assume @xmath45 and an ionization parameter of log(u ) = -3.5 , with line fluxes given ionization by the `` normally - evolving '' stellar population only ( red , solid ) , and with the inclusion of our example sd progenitor population ( blue , dashed ) . ] a number of optical spectroscopic surveys of extended low - ionization emission line regions in early - type galaxies have already been conducted ( e.g. * ? ? ? * ; * ? ? ? * ) , or are presently ongoing @xcite . therefore , a large sample of such galaxies already exists which can be compared with our predictions . in estimating the feasibility of constraining the sd channel using optical forbidden lines , the most useful quantity is not the absolute luminosity per se , but rather the ew , as it characterizes the observability of the line . in fig . [ ew_hight ] , we show for a 3 gyr - old ssp the ew of [ n i ] 5200 , [ o i ] 6300 , and [ c ii ] 1335 ) , taking the continuum emission as predicted from @xcite . for significantly older stellar populations ( @xmath46 5gyr ) our diagnostics become signigicantly less discriminating , at least for the presence of sd sn ia progenitors , given their much lower expected luminosity at late times ( recall figs . [ 0.1__1gyr ] , [ oii_ev_lums ] ) . with the ew of the line and its broadening @xmath47 , we can estimate a minimum s / n of the spectrum needed to detect any line at a desired amplitude - to - noise ratio a / n using the formula given by @xcite : @xmath48{2\pi}\sigma}{\rm{ew}}\label{s_n}\ ] ] where we will take their characteristic broadening and minimum a / n ( @xmath47 = 120km / s , a / n = 4 ) . for sd source temperatures above @xmath49 , we predict for the [ o i ] line an ew of @xmath0 1 for the realistic case @xmath40 = 0.5 ( see fig . [ ew_hight ] ) . therefore , in order to barely detect [ o i ] 6300 emission , we would require a spectrum with s / n @xmath9 25 . such a s / n is only just out of reach of typical sdss spectra ( s / n @xmath0 10 15 ) , and a very robust detection should esily be possible with stacking . note that the [ o i ] line is also frequently detected in individual galaxies in the sample of @xcite . given the similar continuum level at 6563 and the much greater intrinsic strength of the h@xmath20 line ( see fig . [ abs_lums ] ) , the latter should similarly be quite easily detectable . as discussed in @xmath16 3 , normalizing our predictions for the flux in any line to that of the nearest hydrogen recombination line has the advantage of removing any dependence on the covering fraction or total mass of the stellar population . in fig . [ oi_dm ] , we plot the predicted [ o i ] 6300 / h@xmath20 ratio as a function of the total mass accreted by sd progenitors at an effective temperature of @xmath4 . note that , whereas we would expect a linear relation for the total line flux as a function of @xmath26 , this is not the case for diagnostic line ratios . this is primarily due to the contribution of the ionizing luminosity from sd progenitors to the ionization of hydrogen . as an illustrative example , we can compare our models with the observed [ o i ] 6300 / h@xmath20 ratio in the nucleus of the post - starburst galaxy ngc 3489 @xcite . as a conservative estimate ( see e.g. * ? ? ? * ) , we assume a 2 gyr - old ssp , with the available photoionizing continuum originating from both pagbs and sd progenitors . we see that , for a wide range of gas phase metallicities ( 0.5@xmath50 @xmath18 @xmath51 @xmath18 1.5@xmath50 ) , the [ o i ] 6300 / h@xmath20 diagnostic can in practice distinguish ( at @xmath9 1@xmath47 confidence ) between ionization by pagbs alone and the inclusion of a sd progenitor population which accretes more than @xmath90.04@xmath12 per sn ia at @xmath19 = @xmath4 . [ ni_dm ] which may be accreted by the typical wd in the sd scenario at a given photospheric temperature . upper limits are obtained assuming an upper limit of 10% on the he ii 4686 / h@xmath7 ratio ( red line ) , and for the measured [ o i ] 6300 / h@xmath20 ratio in the nucleus of ngc 3489 @xcite ( black , solid ) . for the latter , we also show 1@xmath47 upper limit ( black , dashed line ) . note that any upper limit on the accreted mass is equivalent to an upper limit on the total luminosity of the population of high temperature sources ( right y - axis ) . we assume 2 gyr - old stellar population , log(u ) = -3.5 , and solar metallicity for the gas phase . ] a similar case can be made for the [ n i ] 5200 / h@xmath7 line diagnostic ( see fig . [ ni_dm ] ) . though the [ n i ] 5200 line strength is always weaker than [ o i ] 6300 ( and , unfortunately , not detected in the sample of * ? ? ? * ) , it is still observed with confidence in a number of liers ( e.g. * ? ? ? [ ew_hight ] , we see that for sd source temperatures @xmath52 , the [ n i ] 5200 ew is roughly @xmath0 0.5 ( for @xmath53 and log(u)=-3.5 ) . using eq . [ s_n ] above , we find that this line could then be detected in a 3 gyr - old stellar population in spectra with s / n @xmath9 40 . this is well within what is typically achieved in individual sauron spectra ; with stacking , very robust detections ( or upper limits ) should be feasible . stacking of sdss spectra should allow for a confident detection with only @xmath0 20 galaxies . from fig [ ni_dm ] , we see that similar to the [ o i ] line , a detection or upper limit on the [ n i ] 5200 / h@xmath7 ratio of @xmath9 10% would constrain the average mass proccessed by nuclear - burning wds in the sd channel to a few @xmath54 per supernova . turning again to the [ o i ] 6300 / h@xmath20 ratio in ngc 3489 , we can extend the exercise in fig . [ ni_dm ] to investigate its dependence on wd temperature . in fig . [ dm_teff ] , we plot the maximum mass which can be accreted by nuclear - burning wds in the sd scenario for effective temperatures @xmath8 @xmath3 @xmath13 @xmath3 @xmath4 , should our models , we include the expected contribution from pagbs to the ionizing background . ] remain consistent with the observed ratio [ o i ] 6300 / h@xmath20 . note that the upper limit for any temperature in fig , [ dm_teff ] excludes any matter accreted at any other temperature . for example , accretion of @xmath55 of matter at @xmath56 k will produce an [ o i ] line luminosity equal to the observed value , therefore no more material at any other temperature could additionally be accreted without conflicting with observations . in practice , any realistic sd scenario would imply a population of wds in varying accretion states , with a range in temperatures . therefore any detailed sd scenario may be constrained considering the integrated luminosity and the spectrum of the ionizing radiation produced by the population of sd progenitors . for ngc 3489 , given the observed value of the [ o i ] 6300 / h@xmath20 ratio , no more than @xmath9 @xmath57 per supernova can be accreted and nuclear - processed by wds with photospheric temperatures @xmath13 @xmath11 @xmath58 . taking into account the [ o i ] line flux error , this translates to an upper limit of @xmath59 at @xmath60 . note that the flux errors in @xcite are not dominated by systematics , and in principle dedicated observations should be able to reduce the uncertainty considerably . if we compare this figure with the sensitivity of the he ii diagnostic in @xcite ( see their fig . 8 , or the red curve in our fig . [ dm_teff ] ) , we see that the two complement each other remarkably well ; upper limits on the he ii line luminosity ( here we assume he ii 4686 / h@xmath7 @xmath61 10% ) reach their peak effectiveness at @xmath9 2.5@xmath62 , just as the [ o i ] 6300 / h@xmath20 diagnostic becomes ineffective . an important point to consider is that , while we typically quote limits here in terms of the average mass accreted by nuclear - burning wds ( @xmath26 ) , in principle it is rather the total luminosity of any high - temperature sources which we constrain ( see right y - axis in fig . [ dm_teff ] ) . even if sd progenitors are not primarily responsible for the sn ia rate , accreting wds may still provide a significant contribution to the ionizing background in passively - evolving galaxies . therefore , their consideration would be of vital importance in any exercise in understanding the role of pagbs in low - ionization emission line regions , or their relative importance with respect to llagn ( as in , e.g. * ? ? ? therefore , careful modelling of the evolution of wd binaries and any phases of stable mass transfer are essential ( see chen et al , in prep ) . 5007 / h@xmath7 and [ o i ] 6300 / h@xmath20 for varying ionization parameter , assuming ionization by a 3 gyr - old ssp alone ( red ) , and with the include of a sd progenitor population which accretes @xmath26 = @xmath63 ( green ) and @xmath26 = @xmath64 ( blue ) with an effective temperature @xmath13 = @xmath4 . the ionization parameter increases along these lines from log(u ) @xmath9 -4.5 to -3 , in the direction indicated by the black arrow . shown in black are a subset of the results from @xcite . ] 3869/[o ii ] 3727and the ew of [ o ii ] 3727 / h for varying ionization parameter , assuming ionization by a 3 gyr - old ssp alone ( black ) , and with the inclusion of a sd progenitor population which accretes varying @xmath26 with photospheric temperature @xmath13 = @xmath14 and log(u ) = -3.5 ( red , solid ) and @xmath13 = @xmath4 with log(u ) = -3.5 ( blue , solid ) and log(u ) = -3.0 ( blue , dashed ) . for the tracks which include sd progenitors , @xmath26 increases from 0 to 0.3@xmath12 from left to right in the figure . ] finally , we note that in the presence of a hot sd progenitor population , the higher values of [ o i ] 6300 / h@xmath20 and [ n i ] 5200 / h@xmath7 would place most passively - evolving galaxies in the region of the bpt diagram associated strictly with shock ionization ( -0.5@xmath65([o iii ] 5007 / h@xmath7 ) @xmath66 , and @xmath67([o i ] 6300 / h@xmath20 ) @xmath46 -0.5 ) . however , this has not been observed ( e.g. * ? ? ? * ; * ? ? ? * ) . in fig . [ oiii_oi ] , we plot the predicted values of the [ o iii ] 5007 / h@xmath7 and [ o i ] 6300 / h@xmath20 ratios with varying ionization parameter assuming ionization by a 3gyr - old ssp alone , and with the inclusion of our standard sd progenitor population . we also include a subsample of the results from @xcite ; namely , those galaxies with stellar population ages @xmath61 4 gyr ( from * ? ? ? * ) , with ew(h@xmath20 ) < 1.5 @xcite . this restricts us to relatively young early - type galaxies for which the `` normally - evolving '' stellar population may plausibly power the observed nebular emission . the observed sample in fig . [ oiii_oi ] appears to be largely reproduceable with ionization by the ssp alone , although any interpretation is complicated by the considerable uncertainty in the majority of the emission line measurements . for a number of reasons , emission - line diagnostics in the near - uv provides an attractive alternative . as discussed in the previous section , the [ o ii ] 3727 doublet provides a much higher luminosity than h@xmath20 , in a waveband where the continuum is instrisically lower . the very nearby [ ne iii ] 3869 line is also available as a gauge of the ionization parameter ( through the ratio , e.g. * ? ? ? * ) . this allows one to construct a bpt - style diagnostic diagram using a remarkably narrow wavelength range ( as shown in fig . [ neiii_oii ] ) , using the ew of the [ o ii ] doublet to measure the specific ionizing luminosity of the observed stellar population , and [ ne iii ] 3869/[o ii ] 3727 to fix the ionization parameter . in fig , [ neiii_oii ] , we see that the ssp - only and sd progenitor ionizing populations occupy clearly distict regions at t = 3gyr . for a young , post - starburst galaxy ( t @xmath9 500myr ) , the ew of [ o ii ] 3727 would be enhanced by a further order of magnitude for the same ionization parameter ( see previous section ) . the nearby 5@xmath683 transition of he ii ( 3203 ) 3 times weaker than the 4 @xmath683 transtion at 4686 , the continuum is predicted to be @xmath04 times weaker at @xmath93200 . therefore the observability and utility of this transition is comparable ( see eq . [ s_n ] ) ] . would also allow one to constrain the hardness of the ionizing source population . together , this presents an excellent opportunity for near - uv spectroscopy to constrain the progenitors of sne ia . the prospect of calorimetry of the ionizing background in post - starburst galaxies offers a highly promising avenue for constraining the sd channel at the earliest delay - times . in fig . [ ew_stronglines ] , we plot the ew of the strongest optical lines for a 500 myr - old starburst . as discussed in @xmath16 4 , the most intuitive ( and least model - dependent ) line for this purpose may be the h@xmath20 recombination line . in fig . [ ew_stronglines ] , we see that in our standard sd case , we predict an h@xmath20 ew ranging from @xmath9 3.5 at @xmath69 to @xmath9 12 at @xmath70 , assuming @xmath40 = 0.5 . this may be compared with an h@xmath20 ew of @xmath9 0.2 expected from the ssp alone , giving a factor of @xmath71 enhancement . thus , if observations of post - starburst galaxies gave h@xmath20 line intensities consistent with ssp - only case , the contribution of the sd channel could be strongly constrained to a few per cent level , or , equivalently , to @xmath72 per supernova . the [ n ii ] 6583 line can provide a similar constraint ( see fig . [ ew_stronglines ] ) . however , the nitrogen abundance in the ism of external galaxies is known to be rather uncertain @xcite . therefore , there is little benefit in focussing on this line for calorimetry , unless it is unresolved from the h@xmath20 line . the [ o iii ] 5007 specific line luminosity drops off dramatically at the highest sd temperatures ; this is because with harder spectra , oxygen is increasingly ionized to @xmath73 near the illuminated face of the gas , reducing the emitting volume for [ o iii ] . at the same time , high - energy photons penetrate deeper into the nebula , preferentially ionizing neutral oxygen . note , however , that this itself may pose a problem for any high - temperature sd channel , as the ism in post - starburst ellipticals are observed to have relatively high values of the [ o iii ] 5007 / h@xmath7 ratio @xcite . the [ o ii ] 3727 cooling line , originating in the more low - ionized regions of nebulae ( the interior , in the geometry discussed here ) , is expected to have an ew @xmath11 15 ( for @xmath2 @xmath3 @xmath13 @xmath3 @xmath4 ) . this is a @xmath74-fold increase over that expected from the ssp alone ( @xmath75 , detectable with a minimum s / n @xmath9 40 ) . the [ o ii ] 3727 line ( as well as h@xmath20 ) has been invoked as a proxy for the star - formation rate ( more on this in the following section ) , and already the volumetric [ o ii ] 3727 luminosity function has been mapped at least to z @xmath18 5.4 ( see * ? ? ? * and references therein ) . individual nearby galaxies which have undergone recent , possibly quenched starbursts would make ideal targets for follow - up studies . as an example , ngc 3156 is a nearby ( @xmath9 22mpc ) early - type galaxy which the sauron survey found to host a post - starburst subpopulation ( with an estimated mass of @xmath9 @xmath6 ) . the proposed cubes nuv spectrograph would be able to detect the [ o ii ] 3727 line ( given our standard sd case from above , and assuming @xmath40 = 0.5 ) with a s / n @xmath9 10 @xmath76 in just 1800 seconds . although its specific luminosity is greatly enhanced with the presence of a high temperature source population ( see fig . [ ew_hight ] ) , the [ c ii ] 1335 line remains intrinsically weak relative to the strong uv recombination lines ( e.g. ly@xmath20 , he ii 1640 ) . the difficulty in detecting this line can best be illustrated with an example . at a distance of @xmath0 21 mpc , ngc 3607 is a nearby , relatively young ( mean stellar age @xmath0 3gyr , e.g. * ? ? ? * ) elliptical galaxy which hosts an extended low - ionization emission - line region . assuming our standard sd population from above and a covering fraction of @xmath77 , we can expect a total integrated [ c ii ] 1335 line flux of @xmath0 4.5 @xmath10 @xmath78 erg / s/@xmath79 , using the k - band luminosity to estimate the mass as in @xcite . the galex fuv spectrograph ( now no longer operational ) would have required a prohibitive @xmath0 51 ksec in order to reach s / n @xmath0 5 , with 1335 lying outside the optimal response of the instrument . however , this remains a promising target for future uv missions , such as the planned wso . until such time , the aforementioned he ii 1640 line remains an extremely promising target for observations ( see * ? ? ? * ) , as does the ly@xmath20 transition of hydrogen at 1216 , the latter being easily detected in any emission - line elliptical galaxy . in the limit where sd progenitors provide the only contribution to the ionizing continuum , the total luminosity expected in ly@xmath20 exceeds that of the he ii 1640 line by 1 2 orders of magnitude for sd source temperatures @xmath13 @xmath9 @xmath2 @xmath4 . in young stellar populations , the presence of strong emission lines indicates ongoing star formation ( in particular , for low values of n ii 6583 / h@xmath20 , e.g. * ? ? ? * ) , where interstellar matter is ionized by highly luminous o stars . in this case , the total luminosity in any given line may be used as a tracer of the total ionizing photon flux , and therefore a measure of the number of o stars , indirectly providing the sfr ( for an assumed imf ) . in practice , the h@xmath20 and [ o ii ] 3727 fluxes are the most practical in the optical , with the standard calibrations being @xcite : @xmath80\ ] ] @xmath81}[erg / s]\ ] ] however , this does not account for the possibility of an additional ionizing source within the stellar population , such as possible sd progenitors of sne ia . this would produce an offset in the calibration of any such sfr indicator , which would be dependent on age and mass of the stellar population . for example , for @xmath8 @xmath18 @xmath13 @xmath18 @xmath4 sd progenitor population , a 3 gyr - old @xmath82 ssp would still appear to have an ongoing star - formation rate of @xmath83 @xmath10 4@xmath12/yr from the [ o ii ] 3727 indicator . similarly , for a @xmath13 = @xmath2 sd progenitor population , a 3 gyr - old @xmath82 ssp would still appear to have an ongoing star - formation rate of @xmath84 @xmath10 1@xmath12/yr from the h@xmath20 indicator . because of their insensitivity to interstellar reddening , ir lines and the ir continuum itself are extremely useful sfr indicators . as an important coolant in the neutral ism , the [ c ii ] 157@xmath85 m line has been proposed as a useful sfr indicator , with the present best calibration given by @xcite @xmath86}[erg / s])^{0.983}}{1.028 \times 10^{40}}\ ] ] yet there remains considerable scatter in the observed relation . if there is indeed a significant contribution from populations of accreting , nuclear - burning wds , this should also provide an age - dependent offset . for example , for our standard case sd scenario , a 3 gyr - old @xmath82 ssp would still appear to have an ongoing star - formation rate of @xmath84 @xmath10 1@xmath12/yr . the [ o i ] 63@xmath85 m and 146@xmath85 m ir lines would provide only a small contribution to any sfr - indicator based on the continuum flux . however , as individual lines , both may in principle be a useful measure of the population of accreting wds in galaxies , in particular the 63@xmath85 m line , which is @xmath0 5 times stronger than the [ c ii ] 157@xmath85 m line for sd progenitor effective temperatures of @xmath8 @xmath18 @xmath13 @xmath18 @xmath4 . with the end of the herschel space observatory s operational lifespan , there is at present no instrument which could provide new observations in this waveband , however ample archival data is available . the field - imaging far - infrared line spectrometer ( fifi - ls ) , will also soon be available , observing in the 42 - 210 micron band . as a final word of caution , we recall that our results for the ir and fir forbidden lines discussed above are sensitive to the assumed density of the gas , which is observed to vary radially ( e.g. * ? ? ? * ) . therefore , any quantitative assessment of the contribution of a possible sd channel to the emission of these lines should consider carefully the environment under study . furthermore , these lines do not originate solely from photoionized regions , but arise also in the neutral medium @xcite . so far , we have considered only single - age stellar populations , and assumed the same metallicity throughout . although we have previously demonstrated that varying the metallicity of the stellar population will not significantly alter the ionizing continuum expected from sd progenitors @xcite , we have not tested the effect of varying star - formation histories . in a real galaxy , a much more complex superposition of populations of differing ages should be expected . this is particularly concerning , should the presence of a significantly older subpopulation be able to strongly dilute the observed ew in a galaxy with an otherwise young mean stellar age . this could in turn significantly reduce the detectability of our diagnostic lines . in order to evaluate this possibility , we perform the following experiment . we mix spectra of varying fractions of 1 gyr - old , z = 2.5@xmath87 and 10 gyr - old , z = @xmath87 stellar populations from @xcite . using the method of @xcite based on the lick indices @xcite , we find the luminosity - weighted age depending on the relative fractions of young and old populations ( johansson , private communication , see also johansson et al in prep ) . at the same time , we compute the expected flux in any of emission lines discussed above for the composite stellar population , using the combined spectra in our photoionization calculations . here we assume the same gas density @xmath88 with a gas phase metallicity of z = @xmath50 , as above , and with ionization parameter log(u ) = -3.5 . 6300 emission line as a function of time for single - age ssps ( blue ) , and for composite populations ( 1gyr and 10gyr , in red ) as a function of their measured age ( johansson et al , in prep ) . ] the result of this experiment is shown in fig . [ ew_mixed_pop ] , where we plot the ew of the [ o i ] 6300 emission line for such composite stellar populations , shown as a function of the derived `` mean age . '' for comparison , we also plot the [ o i ] ew for a single - aged ssp with z = 0.05 . note that the two tracks do not agree at 10 gyrs because of the different assumptions for the metallicity of the stellar population . as can be clearly seen , the ew is not significantly affected in the case of composite stellar populations . we have demonstrated how observations of forbidden lines of the most abundant metals , as well as recombination lines of hydrogen and ionized helium , can be used to constrain any population of sources with effective temperatures @xmath13 @xmath11 @xmath2 in early - type galaxies . these sources may not be directly detectable due to local interstellar absorption , as well as along the line of sight in the milky way , but may be revealed from the dramatic effect on the emission line fluxes predicted for their host galaxies . a confident detection , or upper limits on , any recombination line of he ii in the extended , low - ionization emission line regions of passively - evolving galaxies can constrain the total luminosity of sd progenitors with effective temperatures in the range @xmath89 @xmath3 @xmath13 @xmath3 @xmath90 . while this comfortably covers the range expected for symbiotics , many supersoft sourcess , and the so - called ultrasoft sources , other diagnostics must be invoked in order to constrain all possible sd progenitors . we have demonstrated that , in relatively young early - type galaxies which host significant neutral and weakly ionized ism , forbidden lines of [ o i ] , [ n i ] , and [ cii ] can be enhanced by up to @xmath91 times in the presence of a significant ( @xmath92(sd ) @xmath9 @xmath92(total ) , very hot ( @xmath13 @xmath11 @xmath14 ) population of accreting , nuclear - burning wds consistent with the sd progenitor scenario . in particular , past and presently ongoing optical spectroscopic surveys should already be able to either detect evidence of the sd channel or constrain its contribution to the sn ia rate down to a few per cent of the observed total for inferred delay - times 1 gyr @xmath3 t @xmath3 4 gyr . future observations with the planned wso ( the world space observatory * ? ? ? * ) should also be capable of making similar progress with observations of the [ c ii ] 1335 ( and he ii 1640 ) lines . the possible contribution of the sd channel at the earliest delay - times ( @xmath93 gyr ) can also be remarkably well - constrained with measurements of the specific ( per unit stellar mass ) luminosities of the `` classical '' strong emission lines , such as h@xmath94 , [ o ii ] 3727 , [ n ii ] 6583 . for a 500 myr - old stellar population , the h@xmath20 recombination line luminosity is predicted to be enhanced over the ssp - only case by up to @xmath0 two orders of magnitude for sd progenitor temperatures t @xmath3 2@xmath62 . a similar enhancement is expected for the [ o ii ] 3727 line luminosity for sd temperatures spanning the @xmath95 k range , suggesting calorimetry of post - starburst galaxies may be able to constrain the sd contribution to less than @xmath0 1% . alternatively , observations of emission - line nebulae consistent with ionization by the old ssp alone may be interpreted as a constraint on the total mass accreted in the steady nuclear - burning regime . assuming a hypothetical sd scenrio accounted for all sne ia , this would restrict the mass accreted at the @xmath95 k range of photospheric temperatures to @xmath3 @xmath96/yr . finally , if the sd scenario did indeed account for a significant fraction of the sn ia rate , than one might expect the resulting enhancement in the luminosities of emission - lines h@xmath20 , [ o ii ] 3727 , and [ c ii ] 157@xmath85 m to introduce an age - dependent offset in their calibration as indicators of the sfr in galaxies . at the same time , one would expect nuclear - burning wds to provide a vital ingredient in calculating the ionization budget in low - ionization emission line regions , necessitating a careful understanding of the evolution of accreting wd populations . we thank the referee and jonas johansson for their helpful comments and discussion , and brent groves for making his current version of mappings iii available at http://www.mpia-hd.mpg.de/@xmath0brent/mapiii.html .

networks of semiflexible filaments and fibers are common in biological systems , where dynamic structures of tunable strength , elasticity , and adjustable response are required @xcite . from the microscopic scale of cytoskeleton to the macroscopic scale of fibrous tissue , such networks utilize stiff or semiflexible filaments , allowing for rich mechanical response @xcite , dynamic remodelling @xcite , and controlled structural failure @xcite , while remaining an open structure allowing easy through access for solvents and solutes @xcite . inspired by these observations , filament networks are produced in industry , and now stand as a promising area in manufacturing functional materials @xcite . the physics of a single semiflexible chain or filament is well understood with the aid of worm - like chain model , in various implementations and approximations @xcite . the full theory is capable of accurately describing the force - extension relationship , as well as the magnitude of transverse fluctuations , for a range of filament stiffness spanning from very high ( a rigid elastic rod ) to very flexible ( a classical polymer coil ) and a range of end - to - end extensions from zero up to the full stretch where the entropic force has a characteristic divergence . when such filaments form a macroscopic crosslinked network , the main scaling features of its nonlinear mechanical response are still determined by the single filament properties , as reviewed in @xcite . there have been many key contributions to the elastic theory of stiff and semiflexible networks , including ones highlighting the issues of strain non - affinity @xcite . in order to obtain the constitutive relationship of a semiflexible network , a widely used approach follows the following procedure ( explained in detail in chapters 3,4 of ref . @xcite ) : the shear stress on @xmath1 plane is calculated by summing the contributions of the tension along the @xmath2 axis of all the chains crossing the @xmath3 plane . this approach has been very successful in deriving viscoelastic properties of polymer solutions and melts under shear flow @xcite ; storm et al . @xcite have applied the same stress - strain relation to a crosslinked network by assuming the affine elastic strain acting on each filament and the probability distribution of semiflexible filament length of wilhelm and frey @xcite . this model qualitatively describes how the network behaves when being deformed ; however , it can only be used numerically ( since no closed expressions are possible ) and it omits the pre - stress acting on the chains , as in this model the tension is assumed as zero when there is no deformation . wilhelm and frey also produced a numerical simulation of filament network elasticity on their own @xcite , using the mikado model of connectivity and athermal rigid rods as elastic elements . although important issues of percolation rigidity threshold are exposed , this work can not be used to describe most experimentally relevant ( and most biological ) filaments . more recently , palmer and boyce @xcite proposed a closed analytical form of elastic free energy , with the corresponding constitutive relation , using the same force - extension filament relationship as storm et al . they applied the ` 8-chain model ' originally introduced in the context of ordinary rubber elasticity @xcite with an approximate expression for individual filament elasticity applied for each strand . this might be the best attempt in formulating the nonlinear elasticity of semiflexible network to date . on the other hand , unterberger et.al . @xcite developed a ` 1-chain ' model for a network , by assuming filaments have a homogenous orientational distribution in the equilibrium system , and the average ( imposed ) stretch of the network @xmath4 is a @xmath5-root average of deformations of individual filaments : @xmath6^{1/p } / a^{1/p}$ ] , where @xmath7 is the deformation of an individual filament , @xmath5 is the averaging parameter and @xmath8 the area of the unit sphere . this procedure allows the local stretch to fluctuate around its average value , and may partly account for the nonaffinity . this model could only be applied numerically to reproduce several key mechanical features ( shear and normal stress ) . most good models discussed here make a successful fitting of shear stress data of different filament networks , which captures the generic stress - stiffening effect originating from the characteristic divergence of the force - extension curve of an individual filament . usually , the authors employ the versions of celebrated marco - siggia interpolation formula @xcite , which gives the correct response for filaments near full - extension ( but much less so for more flexible filaments ) . in fact , the well - known mackintosh scaling of differential shear modulus with stress , @xmath9 , in the stress - stiffening regime is entirely based on the mentioned divergence of an individual filament near full extension @xcite ( it holds even for an athermal network of undulating filaments , as long as the tensile force is proportional to the inverse - square of the compression , as is the case near full extension @xcite ) . therefore , a mere agreement ( good fitting ) of shear stress - strain curves is not a sufficient test of different theories . in particular , they must simultaneously descibe the effect of negative normal stress in the network of stiffer filaments , with a positive normal stress ( also known as the weissenberg effect ) for networks made of more flexible chains @xcite . it turns out that none of the mentioned theories , including the 8-chain model of palmer and boyce , are able to produce the correct normal stress , or address the network stability increasing with filament pre - tension@xcite . in this paper we develop and discuss a continuum theory of semiflexible network ( in closed analytical form ) that addresses these issues , while retaining accurate fitting of a wide range of shear stress data . before introducing free energy of a semiflexible network , we need to discuss the properties of a single semiflexible filament . a filament connecting two neighboring crosslinks in a network is sketched in fig . [ fig_1](a ) , with coordinates @xmath10 , where @xmath11 is the arc - length coordinate along the chain . different approaches to the chain ( in)extensibility have been tested over the years @xcite , from the strict constraint to the requirement that the length of filament remains constant on average ( while small local fluctuations are allowed ) @xcite , to the models that explicitly include filament stretching @xcite . it turns out that in the regime of high extension , when there are no foldbacks ( hairpins ) on the chain , all length - constraining models give the same divergence of the force - extension , @xmath12 , with the end - to - end ratio @xmath13 , where @xmath14 is the end - to - end length of the filament , fig . [ fig_1](a ) . this was used by the famous interpolation formula of marko and siggia @xcite . more recently , a complete theory of filament entropic elasticity has been developed , which spans the full range of extensions and the full range of bending modulus @xcite . in the worm - like chain model @xcite , the bending energy can be expressed by @xmath15 , where @xmath16 is the bending rigidity . the key physical quantity for describing the stiffness of a polymer chain is the persistence length @xmath17 . when the contour length of the filament , @xmath18 , is comparable with its persistence length @xmath19 , the chain is regarded as semiflexible " . combining the effects of enthalpy arising from bending and entropy of conformation fluctuations , the closed form of the single chain free energy @xcite can be expressed as a function of its end - to - end factor , @xmath13 : @xmath20 where @xmath21 is a dimensionless stiffness parameter reflecting the competition between bending and thermal energy ; in our notation @xmath22 . in ref.@xcite one can find the comparison with several notable models of semiflexible filament ( marko - siggia and ha - thirumalai ) and where they deviate from the accurate expression ( [ achain ] ) . , and end - to - end length @xmath14 being the mesh size . ( b ) the relationship @xcite between the tensile force , @xmath23 , and end - to - end factor , @xmath24 , plotted for semiflexible chains with different stiffness @xmath25 , and marking the position @xmath26 for a chain to stay at the force - free state . ] as shown in fig [ fig_1](b ) , if the value of @xmath25 is smaller than a critical value @xmath27 , the minimum of the free energy ( or the force - free natural length ) will be at @xmath28 ; such a chain can be regarded as flexible . when @xmath29 , one recovers the gaussian entropic spring form : @xmath30 . the marko - siggia ( in fact , fixman - kovac ) limit commonly used for quick fitting of force - extension curves is reached for flexible chains at high extension @xmath31 : @xmath32 . when @xmath33 , the chain can be regarded as a stiff rod with a natural length @xmath34 , and its energy is dominated by elastic bending : @xmath35 $ ] . under compression @xmath36 such a filament undergoes euler buckling instability . note that on forming a crosslinked network , @xmath37 does not need to be equal to @xmath26 for each strand in equilibrium , meaning there can be pre - tension in the network . we now construct the continuum elastic free energy of a network of such filaments using the methodology that was successfully developed in rubber elasticity @xcite . if the sample shape is changed with a deformation tensor @xmath38 , then the corresponding cauchy - green tensor @xcite is @xmath39 , with eigenvalues : @xmath40 , @xmath41 and @xmath42 . the values @xmath43 , @xmath44 , @xmath45 can be interpreted as stretching ratios along the principal directions of deformation . a class of simple , yet powerful theories of rubber elasticity is formulated in terms of invariants of the c - g tensor , @xmath46 , @xmath47 and @xmath48 . famous examples of this class are neo - hookean ( gaussian ) , mooney - rivlin , and gent models ( see @xcite for review ) . note that polymer networks are frequently treated as incompressible , so @xmath49 . for a polymer network , the chains are distributed and crosslinked randomly throughout the material , which is the main difficulty in obtaining an analytical expression of the total free energy of the network ( unless the individual chain is gaussian ) . to incorporate a more complicated chain , such as eq . ( [ achain ] ) , into a constitutive framework , it is necessary to have a model that relates the chain deformation to the applied affine strain . this can be accomplished by representing an element of volume in the average network as a `` mesh cell '' , which is then symmetrically multiplied to fill the volume . here we need to distinguish affinity " and non - affinity " in a filament network constructed by the unit - cell method . a mesh cell " is deformed differently from the material , as these cells are constructed in orientation aligned with the principal stretching direction of the material , which is referred to as non - affinity " by palmer and boyce @xcite . all mesh cells deform in the same way , with repeated deformed structure . however , the real non - affinity can arise from the different responses among different mesh cells " , due to local force relaxation . in this work we refer to the concepts of affinity " or non - affinity " in the local context of individual filament junctions and mesh cells , rather than the global one between the orientation of a mesh cell and the macroscopic deformation of the material . in assuming a symmetry of repeated mesh cells , we automatically discard the effects of the local non - affine deformations @xcite , i.e. take all mesh cells deforming uniformly , which we know must be the case at least for stiff filaments @xcite . since there is no quantitative way of assessing the degree of the error introduced by the affine approximation ( in our interpretation ) , we will have to look at the fits to the experimental data for validation . within a mesh cell , chains or crosslinks have several possible arrangements , reflecting what one assumes about the topology of the network mesh , see fig . [ models ] . most acceptable structures include the homogeneous sphere ( hs ) in 1-chain network model , the primitive cubic ( pc ) in 3-chain model , tetrahedral ( th ) in 4-chain model , and body - centered cubic ( bcc ) in 8-chain model ( see @xcite and a review @xcite for detail ) . in 1-chain model @xcite , one end of a polymer chain is fixed at the center of a sphere , while the other end is on the sphere surface at an arbitrary orientation @xmath50 , distributed isotropically . when deformed by @xmath38 , with stretching ratios along principal directions @xmath43 , @xmath44 , and @xmath51 , the lengths of the three semi - axes will change from @xmath14 , to @xmath52 , @xmath53 and @xmath54 , respectively . the elastic energy density of the network can be expressed as an orientational average of a deformed filament : @xmath55,\end{aligned}\ ] ] where @xmath56 is the density of crosslinked chains and @xmath57 . if one takes the ` entropic spring ' limit of the gaussian chain , the average @xmath58 reduces to the classical neo - hookean rubber - elastic expression @xmath59 $ ] . however , in the general case the free energy in 1-chain model can not be expressed in analytical form as a function of strain invariants , which renders it less convenient . in the following , we compare the 3- and 8-chain models , and decide on the 3-chain model preference , in particular due to the failure of 8-chain model in reproducing the normal stress . in the 3-chain model , a primitive cubic is constructed with lattice points representing the crosslinking sites , and the edges are aligned along the principle directions of deformation tensor @xmath38 . three chains are linked with their end - to - end vectors along the edges and the equilibrium mesh size @xmath14 . on deformation , the lengths of three perpendicular edges at one lattice point become @xmath60 , @xmath61 and @xmath62 , respectively . then the free energy density of a semiflexible network can be expressed as @xmath63 equation ( [ 3chain1 ] ) can be rearranged as a function of the strain invariants : @xmath64\end{aligned}\ ] ] where @xmath22 and @xmath13 as used in eq . ( [ achain ] ) . the body - centered cubic cell of 8-chain model is constructed with eight filaments connected from the center point to all eight lattice points @xcite . the edges of the cell are @xmath14 as we define the mesh size , while the chains are shorter by factor @xmath65 . the important feature of the high - symmetry 8-chain model is that on deformation all chains change their distance by exactly the same amount , from @xmath66 to simply @xmath67 , since the mesh cell is aligned along the principal axes of deformation tensor @xmath38 . the free energy density of the network in this case is given by the single - chain expression directly : @xmath68 when the elastic energy is expressed as a function of strain invariants , the stress tensor of an incompressible material ( with @xmath49 ) can be obtained as @xcite : @xmath69 - p\delta_{ij } , \label{strain}\end{aligned}\ ] ] where @xmath70 is the cauchy strain , and @xmath71 the lagrangian multiplier for incompressibility , the value of which determined by the boundary conditions . let us consider the simple shear deformation such that in cartesian coordinates a point @xmath72 in an original material will change to @xmath73 after being sheared ; @xmath74 is the shear strain . the incompressibility is satisfied automatically , and the remaining strain invariants are : @xmath75 . the shear stress in 3-chain model can be obtained from the general constitutive relation ( [ strain ] ) as : @xmath76_{\mathrm{3c } } = \frac{2}{3}n k_\mathrm{b } t\gamma x^2 \left [ \frac{\left(1-x^4\right)}{c \pi \left[1-\left(2+\gamma^2\right ) x^2+x^4\right]^2}- c \pi ^2\right],\end{aligned}\ ] ] while the corresponding expression in 8-chain model is : @xmath77_{\mathrm{8c}}= \frac{2}{3}nk_\mathrm{b}t\gamma x^2\left [ \frac{9}{c \pi \left[3-\left(3+\gamma^2\right ) x^2\right]^2}- c \pi ^2\right].\end{aligned}\ ] ] one must distinguish the nominal shear modulus @xmath78 , from the differential shear modulus @xmath79 @xcite , while the linear shear modulus of the network @xmath80 is equal to the differential modulus @xmath81 at @xmath82 . , which has a limit of 1 at @xmath82 ; the fitting parameters for 3-chain model are given in table [ thenetwork ] . ] take actin and fibrin network under simple shear deformation as an example . both 3-chain and 8-chain models fit the shear - experiment data from refs . @xcite equally well , in fact perfectly , as shown in fig . [ constitutiverelation](a ) . the stiffness , @xmath25 , and the initial end - to - end factor , @xmath83 , obtained by fitting with 3-chain model are a bit smaller than those in 8-chain model , but both in the semiflexible regime . this is because the eight chains in a bcc are stretched equally ( in principal axes ) and share the deformation , while the three chains in a pc are stretched differently and the most stretched one contributes the most to the nonlinear elastic energy . we believe the intrinsic heterogeneity in the 3-chain model , and the fact that the chain lying along the maximum principle stretch direction dominates the response of the whole cubic in semiflexible networks , is closer to the realistic case of filament network . in addition , we shall see below that the 8-chain model can not explain negative normal stress . hence we apply the 3-chain model in the following . figure [ constitutiverelation](b ) shows fits to experimental data for a wide variety of semiflexible filaments . in this figure we plot the shear modulus @xmath84 instead of stress , because the fitting convergence is much better when the initial data section is close to 1 . fitted parameters @xmath85 are listed in table [ thenetwork ] , along with other parameters for each material that we list from the literature . .fitting parameters @xmath85 for collagen , actin , vimentin , fibrin and neurofilament data obtained from @xcite . also shown are the linear modulus @xmath0 extracted from the original data and used for scaling in fig . [ constitutiverelation ] , the literature values of @xmath86 , and the calculated mesh size @xmath87 . [ cols="^,^,^,^,^,^",options="header " , ] first of all , one may be surprised that the persistence length of collagen fibers is quoted as @xmath88 m , when there is a large body of literature that would claim that collagen fibers are stiff athermal rods with persistence length of centimeters . this is all to do with the way a sample is prepared , and we use / quote the data @xcite where the collagen was apparently less aggregated than in a typical extra - cellular matrix . the same ambiguity will apply to actin networks as well , below . another point to note about the fitted values in table [ thenetwork ] is about neurofilament , which has @xmath89 , that is , rather flexible chains . taking the fitted value for @xmath90 , this leans that @xmath91 in this network was @xmath92 m , i.e. about 3.6 times longer than @xmath86 . by calculating @xmath34 for different filaments listed in table [ thenetwork ] , we see that the fitted @xmath37 is smaller than @xmath26 , indicating all of the filaments in fig . [ constitutiverelation](b ) are pre - compressed , rather than pre - stretched in the equilibrium state of the network . as table [ thenetwork ] shows , no matter what the effective stiffness of the examined biofilaments , the ratio of fitted parameters @xmath93 remains close to 1 . later , when we discuss the network stability ( fig . 6 ) , it will become clear that all the networks we examined here lie very near the stability boundary . it is not clear to us whether the fact that the mesh size is close to the filament persistence length is an unintended result of different crosslinking density in experiments @xcite , or is a relevant and universal biological feature . it is clear that networks of biological filaments have a great variety even within the same substance . figure [ actin32 ] shows the published data for in - vitro crosslinked actin networks reported by different groups , all performing the simple - shear experiment ( the data is digitized from a reference given in the plot ) . the stress - strain plotted in log - log format allows a clear identification of the linear regime @xmath94 ( the modulus varying between 95pa and 1pa for different sets ) , and the subsequent stiffening at higher shear . all curves in fig . [ actin32](a ) are fitted by the same eq . ( [ const3 ] ) with the linear modulus @xmath0 and the two parameters @xmath85 taking values : @xmath95 pa , @xmath96 ; @xmath97 pa , @xmath98 ; @xmath99 pa , @xmath100 ; @xmath101 pa , @xmath102 ; @xmath103 pa , @xmath104 and @xmath105 pa , @xmath106 for the six sets from top to bottom . the difference between two data sets from @xcite is the density of heavy meromyosin ( hmm ) crosslinker ( labelled on the plot ) ; the difference between two data sets from @xcite is the degree of actin filament bundling : an initial network of f - actin filaments turned into a more sparse collection of bundles under repeated shear cycles @xcite . as a result the network at ` cycle 7 ' has lower @xmath0 but higher stiffening . , before the stiffening sets in at higher shear . ( b ) the stress - stiffening of actin networks represented by the @xmath107 scaling relation . all data sets are the same as labelled in the plot ( a ) , and solid lines are theoretical curves plotted with eq . ( [ const3 ] ) using the same fitted parameters . ] to find the onset of non - linearity in our theory , the differential shear modulus @xmath108 can be expanded in powers of shear strain @xmath74 : @xmath109 \label{ktaylor } \\ & & + 4 nk_\mathrm{b}t x^4 \frac{1 + x^2}{c \pi ( 1 - x^2)^5 } \ , \gamma^2 . \nonumber\end{aligned}\ ] ] the crossover point when these two terms are comparable with each other , in a more stiff network with @xmath110 , can be approximated as @xmath111 . the stress at this point is @xmath112 $ ] . this crossover stress @xmath113 increases with the temperature as @xmath114 , closely matching the observations @xcite . at the end of range , when the shear strain approaches the point of divergence in stress @xmath115 , ( @xmath116 in eq . [ const3 ] ) , the stress and the differential modulus can be approximated as : @xmath117 ^ 2 } , \label{limit } \\ \frac{k(\gamma)}{n k_\mathrm{b } t } & \simeq & \frac{2 ( 1 - x^4 ) \left[x^2 - ( 2 - 3 \gamma^2 ) x^4 + x^6\right ] } { 3 \pi c \left[1 - ( 2 + \gamma^2 ) x^2 + x^4\right]^3}. \label{limit2}\end{aligned}\ ] ] both expression diverge due to the same vanishing denominator , while maintaining the obvious scaling relation @xmath107 , which has been reported by different theories and experiments @xcite . we see this limit exposed clearly in fig . [ actin32](b ) for very different actin networks . in this plot , the data points are the same as in fig . [ actin32](a ) , and we also plot the predicted curves of @xmath118 obtained by differentiating eq . ( [ const3 ] ) , using parameters @xmath119 and @xmath37 from the fitting in fig . [ actin32](a ) , for each data set . one has to make a comment here , in the context of @xmath107 scaling and different experiments . in many cases , such as in unterberger et.al . @xcite , the actin network was crosslinked by hmm and apparently retains some transient activity , producing the stress - softening and plasticity at higher stress ( this is also the case in @xcite before the actin filaments formed bundles , or with f - actin crosslinked by filamin ) . of course , our theory is not intended to deal with network plasicity ( we assumed all crosslinkes permanent ) and therefore we only retained the experimental data points in the early stress - stiffening regime to see the 3/2 scaling . on the other hand , just by examining the actual data for the actin network of storm et.al . @xcite , one might conclude that it strongly and systematically deviates from the 3/2 scaling . in fact , the authors of @xcite develop their own theory invoking various additional factors ( e.g. filament extensibility ) to account for this data . however , our basic theory , assuming permanent crosslinks , bulk incompressibility and inextensible filaments , evidently fits both @xmath120 and @xmath118 data very well . the fact that the data ( and the predicted curve ) do not appear to follow the 3/2 scaling is be due to the fact that , for these values of @xmath25 and @xmath37 , the final crossover to this characteristic scaling regime would occur at an even higher stress ( at which point the network would probably not survive in practice ) . though simple shear deformation was helpful in this analysis , the important issue of normal stress remains controversial . this is mainly because of the uncertain boundary conditions , see ref.@xcite for detail . since the experiments reporting normal stress measurements are most commonly conducted in rotating cylindrical geometry of a standard rheometer @xcite , we will consider this geometry and realistic boundary conditions to describe the response of the material in its normal direction when a shear is applied , as shown in fig . [ poynting](a ) . suppose the height and the radius of the undeformed cylinder are @xmath121 and @xmath122 , respectively ; on deformation they may become @xmath123 and @xmath124 . incompressibility maintains @xmath125 . in the @xmath126 coordinate system , after rotating the top plate by the angle @xmath127 , the coordinates change as : @xmath128 : @xmath129 where @xmath130 denotes the shear strain , which is a function of the radial position in this parallel - plate geometry . the strain invariants become : @xmath131/\lambda_h$ ] and @xmath132/\lambda_h^{2}$ ] . given the shear strain at the outermost surface , @xmath133 , the total free energy then becomes a function of the stretching ratio @xmath134 along the @xmath135 axis , after integration over radius : @xmath136 the equilibrium @xmath134 can be obtained by minimizing this free energy . when the cylinder radius becomes smaller ( or larger ) upon shear , with its height becoming correspondingly larger ( or smaller ) this phenomenon is called the positive ( or negative ) poynting effect @xcite . this geometric effect with stress - free top plate corresponds to the positive ( negative ) normal stress required to maintain the fixed plate separation in a more common rheometry experiment @xcite . the 8-chain model fails in obtaining the correct normal stress in a sheared network : due to its core assumption that eight chains in a cell are identically deformed ( stretched ) upon shear , the elongation ratio of the network along the stretching direction is always larger than 1 , making the normal stress always positive . at the outer surface of the cylinder ( black curve ) and the normal strain @xmath137 for a flexible network ( red curve ) with @xmath138 , and a more stiff network ( blue curve ) with @xmath139 . ] figure [ poynting](b ) shows the results for equilibrium @xmath134 , presented as a response to an oscillating imposed shear @xmath140 , for two model materials with different filament stiffness and pre - tension . a flexible network ( @xmath138 ) has a positive poynting effect , or positive normal stress is required to counter the expansion along the height ( in other setting , this is called the weissenberg effect ) . in a flexible network ( @xmath141 ) , when the mesh size is much smaller than the contour length of the subchains connecting the neighboring crosslinks , _ i.e. _ , @xmath142 , the entropic energy plays a main role . though chains are stretched along the principal extension direction in the shear geometry illustrated in fig . [ poynting](a ) , the chains in other two principle directions are more likely being compressed , leading to the material contraction along the radial and circumferential directions . however , if the mesh size is comparable with the contour length of the subchains , @xmath143 , the material can behave in a negative poynting effect manner : the height of the cylindrical sample contracts , or negative normal stress is required to counter that and maintain the fixed height . this is because the force acting on the chains directed along the principal extension in fig . [ poynting](a ) is close to a divergence if @xmath144 , and it causes less energy when the material is compressed in the longitudinal direction , rather than stretched . similar reason works also for a network of more stiff filaments ( see red curve for @xmath145 ) . we did not specifically calculate the normal stress here ( this would be a cumbersome process involving the full tensor form of eq . ( [ strain ] ) , first fixing the pressure @xmath71 from the condition of zero radial stress on the free outer surface ) . however , the magnitude of normal stress can be easily estimated from the linear relationship @xmath146 , which uses the fact that the normal strain is quite small ( i.e. the linear regime is justified ) and the young modulus is @xmath147 . taking the fibrin values in table [ thenetwork ] as an example , the normal stress is about @xmath148pa under a shear strain of @xmath149 . these are very close to the observations in experiments @xcite . the magnitude of @xmath150 pa for the same shear also accurately matches the actin results obtained in @xcite . figure [ map ] gives the full ` phase diagram ' of the stiffness - tension parameter space ( @xmath151 ) with phase boundaries separating positive / negative normal stress regions . in order to generate this diagram , for each parameter set ( @xmath152 ) we have calculated the value of @xmath153 by minimizing @xmath154 , under the given shear . in this way the boundary separating the positive ( @xmath155 , @xmath156 ) and the negative ( @xmath157 , @xmath158 ) normal stress regions in fig . [ map ] is calculated . there is also a weak dependence of this boundary on the magnitude of the applied shear strain , which is due to the inherent non - linearity of stress - strain response ; however , we are not showing this in the figure to avoid clutter . figure [ map ] shows that a loose flexible network usually has a positive poynting effect , while a network of more stiff filaments has a negative poynting effect , especially when the filaments are crosslinked with increasing pre - tension . , and filament pre - tension , @xmath37 , showing the boundaries of positive / negative poynting effect and the boundary of network stability ( the dashed line of the neutral filament , @xmath26 , was defined in fig . [ fig_1 ] ) . three ` softer ' filament networks from table [ thenetwork ] are shown in this map : @xmath159 - vimentin , @xmath160 - fibrin , and @xmath161 - neurofilament networks . three actin networks from fig . [ actin32 ] also fit on this map : @xmath162 - @xcite , @xmath163 - @xcite , and @xmath164 - @xcite . ] one can see from figs . [ constitutiverelation](b ) and [ actin32](b ) that the linear regime when @xmath94 persists for a different range of strain in different filaments ; the onset of hyperelastic regime is determined by how much pre - tension is in the filaments , i.e. how @xmath37 compares with @xmath165 from eq . ( [ achain ] ) . the linear shear modulus @xmath80 can be easily obtained from our theory : in 3-chain model it is given by the first term in eq . ( [ ktaylor ] ) . the marginal rigidity condition @xmath166 determines the strand pre - tension that is required for achieving a stable network . the stability criterion here is : @xmath167 note that the expression in the right hand side is always greater than @xmath168 as defined in section 2 . flexible chains with @xmath169 are therefore always stable in the rubbery network , that is , @xmath170 always . on the other hand , stiff filaments have to be crosslinked with @xmath171 exceeding the pre - tension threshold given by eq . ( [ stabil ] ) , which turns out to be slightly lower than @xmath165 defined in fig . [ fig_1](b ) . in other words , there have to be tensile forces acting on the crosslinked filaments in the network in order for it to be mechanically stable with a non - zero shear modulus . this notion is familiar from the `` tensegrity '' concept in biology and engineering @xcite . for stiff athermal filament network , the window of pre - tension between the linear modulus @xmath172 at @xmath173 , and @xmath174 at @xmath110 ( we assume inextensible chains ) is very narrow . the stability boundary of the network is plotted in fig . the condition for @xmath172 gives the equilibrium case labelled as @xmath175 in the phase diagram . however , the full analysis shows that the magnitude of the shear strain modifies the stability condition , as represented by the coloured curves for @xmath176 and @xmath177 . this shift means that the region of mechanical stability of filament network expands on increasing deformation , which matches exactly what the recent paper by sharma _ states @xcite . to summarize , in this work we develop a continuum elastic theory of a network of semiflexible filaments , by implementing the general free energy of one semiflexible chain into that of a disordered network . on reflection , we choose the 3-chain model as most closely matching the realistic system and achieve various quite stringent fits a number of different experimental data sets over the full range of nonlinear stress - strain range . we demonstrate that the general theory produces the conditions for positive / negative poynting effect ( normal stress ) in a network under imposed shear . the greatest weakness of this model is omitting the effects of local strain non - affinity , which might play an important role when crosslink density of the network is small and filaments stiff . the effect of tensegrity , or linking of filament pre - tension to the network stability and the magnitude of the shear modulus @xmath178 is an unexpected result of this theory . the stringent match of experimental data is reassuring , and we believe that the presented analytical continuum model can provide as an efficient and portable tool for studying the mechanical properties of semiflexible networks . + * acknowledgments : * this work has been funded by the tcm critical mass grant from epsrc ( ep / j017639 ) . we are grateful for informative discussions with masao doi and zhongcan ouyang , and thank cornelius storm for providing their raw experimental data @xcite . b. wickstead and k. gull , _ j. cell . biol . _ , 2011 , * 194 * , 513525 g. faury , _ pathol . _ , 2001 , * 49 * , 310325 s. zbek , p. g. balasubramanian , r. chiquet - ehrismann , r. p. tucker and j. c. adams , _ mol . biol . cell _ , 2010 , * 21 * , 43004305 k. e. kasza , a. c. rowat , j. liu , t. e. angelini , c. p. brangwynne , g. h. koenderink and d. a. weitz , _ curr . cell biol . _ , 2007 , * 19 * , 101107 p. bursac , g. lenormand , b. fabry , m. oliver , d. a. weitz , v. viasnoff , j. p. butler and j. j. fredberg , _ nature mater . _ , 2005 , * 4 * , 557561 d. a. fletcher and r. d. mullins , _ nature _ , 2010 , * 463 * , 485492 x. trepat , g. lenormand and j. j. fredberg , _ soft matter _ , 2008 , * 4 * , 17501759 t. lecuit , p .- f . lenne and e. munro , _ annu . biol . _ , 2011 , * 27 * , 157184 e. w. wong , p. e. sheehan and c. m. lieber , _ science _ , 1997 , * 277 * , 19711975 m. r. falvo , g. j. clary , r. m. taylor , v. chi , f. p. brooks , s. washburn and r. superfine , _ nature _ , 1997 , * 389 * , 582584 m. fixman and j. kovac , _ j. chem . phys . _ , 1973 , * 58 * , 15641568 j. f. marko and e. d. siggia , _ macromolecules _ , 1995 , * 28 * , 87598770 b. y. ha and d. thirumalai , _ j. chem . phys . _ , 1996 , * 106 * , 14 c. bouchiat , m. d. wang , j .- f . allemand , t. strick , s. m. block and v. croquette , _ biophys . j. _ , 1999 , * 76 * , 409413 j. r. blundell and e. m. terentjev , _ macromolecules _ , 2009 , * 42 * , 53885394 f. c. mackintosh , _ cytoskeletal mechanics : models and measurements _ , cambridge university press , cambridge , 2006 c. p. broedersz and f. c. mackintosh , _ rev . _ , 2014 , * 86 * , 9951036 d. a. head , a. j. levine and f. c. mackintosh , _ phys . rev . lett . _ , 2003 , * 91 * , 108102 c. heussinger , b. schaefer and e. frey , _ phys . e _ , 2007 , * 76 * , 031906 m. doi and s. f. edwards , _ the theory of polymer dynamics _ , oxford university press , oxford , 1986 d. c. morse , _ macromolecules _ , 1999 , * 32 * , 59345943 c. storm , j. j. pastore , f. mackintosh , t. lubensky and p. a. jamney , _ nature _ , 2005 , * 435 * , 191194 j. wilhelm and e. frey , _ phys . _ , 1996 , * 77 * , 25812584 j. wilhelm and e. frey , _ phys . rev _ , 2003 , * 91 * , 108103 j. s. palmer and m. c. boyce , _ acta biomater . _ , 2008 , * 4 * , 597612 m. j. unterberger , k. m. schmoller , a. r. bausch and g. a. holzapfel , _ j. mech _ , 2013 , * 22 * , 95114 m. l. gardel , j. h. shin , f. c. mackintosh , l. mahadevan , p. matsudaira and d. a. weitz , _ science _ , 2004 , * 304 * , 13011305 t. van dillen , p. onck and e. van der giessen , _ j. mech solids _ , 2008 , * 56 * , 22402264 g. agar , p. r. onck and e. van der giessen , _ macromolecules _ , 2011 , * 44 * , 70267033 p. a. janmey , m. e. mccormick , s. rammensee , j. l. leight , p. c. georges and f. c. mackintosh , _ nature mater . _ , 2007 , * 6 * , 4851 h. kang , q. wen , p. a. janmey , j. x. tang , e. conti and f. c. mackintosh , _ j. phys . chem . b _ , 2009 , * 113 * , 37993805 a. sharma , a. j. licup , k. a. jansen , r. rens , m. sheinman , g. h. koenderink and f. c. mackintosh , _ nature phys . _ , 2016 , * 12 * , 584587 m. otto , j. eckert and t. a. vilgis , _ macromol . theory sim . _ , 1994 , * 3 * , 543555 d. c. morse , _ macromolecules _ , 1998 , * 31 * , 70307043 o. kratky and g. porod , _ recueil des travaux chimiques des pays - bas _ , 1949 , * 68 * , 11061122 m. c. boyce and e. m. arruda , _ rubber chem . _ , 2000 , * 73 * , 504523 l. r. g. treloar , _ the physics of rubber elasticity _ , oxford university press , oxford , 1975 a. f. bower , _ applied mechanics of solids _ , crc press , boca raton , 2009 c. o. horgan , _ intl . j. non - lin . mech . _ , 2015 , * 68 * , 916 a. zaccone , j. r. blundell and e. m. terentjev , _ phys . rev . b _ , 2011 , * 84 * , 174119 c. p. goodrich , a. j. liu and s. r. nagel , _ nature phys . _ , 2014 , * 10 * , 578581 a. m. stein , d. a. vader , l. m. jawerth , d. a. weitz and l. m. sander , _ j. microscopy _ , 2008 , * 232 * , 463475 f. gittes , b. mickey , j. nettleton and j. howard , _ j. cell biol . _ , 1993 , * 120 * , 923934 n. mcke , l. kreplak , r. kirmse , t. wedig , h. herrmann , u. aebi and j. langowski , _ j. mol . biol . _ , 2004 , * 335 * , 12411250 o. i. wagner , s. rammensee , n. korde , q. wen , j. f. leterrier and p. a. janmey , _ exp . _ , 2007 , * 313 * , 22282235 k. schmoller , p. fernandez , r. arevalo , d. blair and a. bausch , _ nature comm . _ , 2010 , * 1 * , 134 p. h. j. kouwer , m. koepf , v. le sage , m. jaspers , a. m. van buul , z. h. eksteen - akeroyd , t. woltinge , e. schwartz , h. j. kitto , r. hoogenboom , s. j. picken , r. j. m. nolte , e. mendes and a. e. rowan , _ nature _ , 2013 , * 493 * , 651655 m. l. gardel , j. h. shin , f. c. mackintosh , l. mahadevan , p. a. matsudaira and d. a. weitz , _ phys lett . _ , 2004 , * 93 * , 188102 y. c. lin , n. y. yao , c. p. broedersz , h. herrmann , f. c. mackintosh and d. a. weitz , _ phys . lett . _ , 2010 , * 104 * , 058101 y. c. lin , c. p. broedersz , a. c. rowat , t. wedig , h. herrmann , f. c. mackintosh and d. a. weitz , _ j. mol . biol . _ , 2010 , * 399 * , 637644 c. o. horgan and j. g. murphy , _ j. elast . _ , 2010 , * 98 * , 205221 l. a. mihai and a. goriely , _ proc . a _ , 2011 , * 467 * , 36333646 j. h. poynting , _ proc . a _ , 1909 , * 82 * , 546559 d. e. ingber , _ annu . physiol . _ , 1997 , * 59 * , 579599

the radiation mechanisms of many high energy astrophysical objects are still an active issue since they often contain features which are not easily explained by the conventional synchrotron and inverse compton emissions . recently , much attention has been paid to the radiation signatures from the turbulent electromagnetic fields ( e.g. , medvedev 2000 , fleishman 2006 , kelner , aharonian , & khangulyan 2013 , mao & wang 2013 , teraki & takahara 2013 ) . the main scene of the emission regions of high energy astrophysical objects is collisionless shocks , and the turbulent electromagnetic fields would be generated in the shock region . therefore , the electromagnetic turbulence should be taken into account when we consider the radiation . however , for the major emission mechanisms of synchrotron radiation and inverse compton scattering , effects of small scale turbulence are not taken into account . there is a room for novel emission signatures in the consideration of turbulence , which may be of relevance to observations . by reproducing the observed spectra , we can extract physical parameters of astrophysical objects . thus , researches of radiation signatures from the turbulent field would play a key role for the understanding of physical mechanisms of the high energy astrophysical objects . radiation spectra from a small scale turbulent magnetic field have been well studied as the `` jitter radiation '' or `` diffusive synchrotron radiation '' ( medvedev et al . 2011 , fleishman & urtiev 2010 and references therein ) . differences from the synchrotron radiation become significant when the typical spatial scale of an eddy @xmath20 is smaller than the photon formation length ( pfl ) of the synchrotron photons @xmath21 , where @xmath22 is the larmor radius and @xmath23 is the lorentz factor of the electron . such a small scale magnetic field is thought to be generated by weibel instability around the shock front . when the strength of this small scale turbulent magnetic field is dominant , the radiation spectra are determined by the turbulence and reveal various signatures of the turbulent field . for example , when @xmath24 the peak frequency in @xmath25 spectrum becomes @xmath26 , and the spectrum in the highest frequency region shows a power law @xmath27 when the turbulence exists up to the maximum wave number @xmath28 . the power law index @xmath29 reflects that of the turbulent magnetic field @xmath30 . the spectra show more complex signatures when @xmath31 ( medvedev 2011 , reville & kirk 2010 , teraki & takahara 2011 ) . we note that the anisotropy of turbulence also affects the radiation spectra ( e.g. kelner et al . 2013 , reynolds & medvedev 2012 , medvedev 2006 ) . the radiation from an electron which moves in non - uniform magnetic field in a laboratory is well studied using the insertion device of synchrotron orbital radiation factory , where a series of magnets are line - upped to make the particle deflect periodically . it is called `` wiggler '' or `` undulator '' ( jackson 1999 ) . for undulator , the strength of magnets @xmath32 and gaps between them @xmath0 are chosen to satisfy the condition that the observer is always in the beaming cone . on the other hand , for wiggler , the observer is periodically in and off the beaming cone . we estimate the critical distance @xmath33 which divides wiggler and undulator . the deflection angle in one deflection is @xmath34 , where @xmath35 is the typical curvature radius of the orbit . the radiation from a relativistic particle is concentrated into small cone with opening angle @xmath19 . therefore , the critical condition dividing wiggler and undulator is @xmath36 , which is rewritten as @xmath37 . thus , the device is called undulator when @xmath38 , while it is called wiggler when @xmath39 . the radiation spectrum of undulator shows a sharp peak at @xmath40 , while wiggler shows a broad spectrum with peak frequency @xmath41 . the relation between typical frequencies and deflection angle is a key point for understanding of the radiation spectra . perturbative jitter radiation or perturbative dsr is recognized as extensions of the undulator radiation , since the spatial scale of turbulence @xmath0 is assumed to be much smaller than @xmath42 . the fact that @xmath43 determines the radiation feature was also noted in astrophysics in early seventies by , e.g. , rees , ( 1971 ) , and getmantsev & tokarev ( 1972 ) . rees assumed that the strong electromagnetic wave is emitted from the crab pulsar with frequency @xmath44 ( 30hz ) according to the `` oblique rotator '' model by ostriker & gunn ( 1969 ) , and argued that the radiation from an electron moving in the strong wave is not inverse compton scattering ( with frequency @xmath45 ) but synchrotron - like ( synchro - compton ) radiation , because the deflection angle @xmath43 is estimated as @xmath46 , where the factor @xmath47 comes from the ratio of cyclotron frequency to the wave frequency @xmath48 . the magnetic field strength @xmath49 is estimated by equating the spindown luminosity and power of magnetic dipole radiation of the crab pulsar and by considering the distance from the pulsar . he argued that the radiation signature of synchro - compton radiation resembles that of the synchrotron radiation . the typical frequency is determined by the field strength as @xmath50 , and the low frequency spectral index is roughly @xmath18 . getmantsev & tokarev ( 1972 ) studied the radiation spectra under general electromagnetic fields and stated that the radiation signature from a single charged particle is determined by the frequency or wavelength of the field for @xmath51 , while it is determined by the field strength for @xmath52 in the same way as synchrotron radiation . note that they did not discussed explicit expressions of the radiation spectra from a single electron for @xmath53 , they presented a general expression of radiation spectrum from an ensemble of electrons with a power spectrum . thirty years after these pioneering works , radiation spectra from a relativistic particle interacting with turbulent fields are gathering a renewed attention . for example , the radiation mechanism from langmuir turbulence treated in the present paper has become an interesting topic . the langmuir turbulence has been pointed out to be generated around the shock front of the relativistic shocks ( silva 2006 , dieckmann 2005 , bret , dieckmann , & deutch 2006 ) , so it should be as important as the radiation from turbulent magnetic field . fleishman & toptygin ( 2007a , b ) have made a systematic treatment of diffusive radiation in langmuir turbulence ( see references cited in fleishman & toptyigin 2007b for other earlier relevant works ) . their method is the most sophisticated treatment for langmuir turbulence , which is based on toptygin & fleishman ( 1987 ) . for later discussions , we shortly review their treatment . they treat the electron motion by a statistical approach and use a perturbative treatment for calculation of the radiation . the calculation formula for the radiation spectra is based on the one written in landau & lifshitz ( 1971 ) . for @xmath54 , a rectilinear trajectory with constant velocity is assumed , but non - zero acceleration from the external field is taken into account . the wavenumber of the langmuir waves is assumed to satisfy the condition @xmath55 , where @xmath4 is the plasma frequency . they calculate the correlation between the acceleration and the langmuir waves . the peak frequency is @xmath11 . the spectrum shows an abrupt cutoff above the peak , and becomes @xmath56 in higher frequencies when the turbulence exists up to the maximum wavenumber @xmath57 for a power law turbulent spectrum @xmath58 . the spectrum just below the peak is @xmath59 , and becomes @xmath60 in lower frequencies and @xmath61 in even lower frequency region . spectrum is predicted in the lowest frequency region , which comes from the effect of wave dispersion in the plasma . we do not discuss such effects in this paper . ] @xmath61 spectrum comes from the effect of multiple scattering . the angle between the velocity and observer direction becomes larger than the beaming cone after many deflections even when the deflection angle in one deflection @xmath43 is much smaller than @xmath62 , and the approximation of rectilinear trajectory is broken . therefore , this treatment is beyond the perturbative treatment , and they call it `` non - perturbative treatment '' . they treat the changing of direction of motion in many deflections by diffusion approximation . in consequence , a spectral break emerges in the low frequency region with a suppression of low frequency photons . the spectrum becomes @xmath61 from this effect and the index of @xmath63 comes from the diffusivity . they claimed that the break frequency approaches the peak frequency as @xmath43 becomes large and the break and peak merge for @xmath53 . they stated that even for @xmath52the spectrum in frequency region just below the peak of @xmath64 is @xmath61 . this is inconsistent with the statement by getmantsev & tokarev ( 1972 ) . this is one of the objectives of investigations in the present paper . the effect of large angle deflection would come into play in forming the radiation spectra for langmuir turbulence as in the wiggler radiation for @xmath65 . this point has been discussed for magnetic turbulence in kelner et al . ( 2013 ) , medevev et al . ( 2011 ) , and teraki & takahara ( 2011 ) . in this paper , we investigate general properties of the radiation spectra from relativistic electrons in a langmuir turbulence for various cases including this regime . before proceeding to the formulation of the calculation employed in the present paper , we point out the differences between the magnetic field generated by weibel instability ( entropy modes ) and the langmuir waves . the first is time variability . the entropy mode does not oscillate , so the typical timescale is determined by the turnover time of an eddy . it is longer than the crossing time of a relativistic electron if we assume that the background plasma is sub - relativistic . therefore , we can treat the magnetic field as a static field when we calculate the radiation spectra for the zeroth - order approximation . on the contrary , we should not treat the langmuir turbulence as a static field even for the zeroth - order approximation . the crossing time can be comparable to or longer than the period of the langmuir waves , because the typical spatial scale is about inertial length @xmath66 , governed by the plasma frequency @xmath4 ( diekmann 2005 ) . therefore , the time variability of the electric field can not be ignored . this effect is well studied by fleishman & toptygin ( 2007a , b ) . the second is the energy change of the radiating electrons . for the case of turbulent magnetic field , the energy of electrons is conserved if we ignore the radiation back reaction . however , the energy change can not be ignored for the langmuir turbulence , because the electric field can accelerate the electrons parallel to their velocity . the lorentz factor of the electron can change in a short time by strong langmuir waves . this effect may play a role for calculation of the radiation spectra . the calculation of electron trajectory for the strong turbulence by analytical approach is hard to perform . thus , we calculate the radiation spectra by numerical approach from first principle . we calculate radiation spectra for a wide range of the field parameters . we study about three factors , the scale length , time variability , and strength . by sweeping the parameter plane , we generally investigate the radiation spectra from a relativistic electron moving in langmuir turbulence . in section 2 , we describe calculation method , and we show the results in section 3 . in section 4 , we give the physical interpretations of the discovered spectral features using radiation for a spatially uniform plasma oscillation . in section 5 , we make a summary and some discussions . we use lienard - wiechert potential directly to calculate the radiation spectra . it is the same approach as employed in teraki & takahara 2011 for magnetic turbulence . this method is computationally expensive because the time step is restricted by observed frequency and the integration time has to be longer than the photon formation time ( pft ) . pft is defined as the coherence time for forming a photon with a given frequency ( reville & kirk 2010 , akhiezer & shulga 1987 ) . therefore , pft is a function of the frequency . since pft is @xmath67 times larger than the inverse of the observed frequency and since the time step should be shorter than pft , the integration time is at least @xmath67 times larger than the inverse of the observed frequency . as a result , large number of time steps are needed to calculate radiation spectra for highly relativistic case . although less heavy approach is proposed by reville & kirk 2010 to overcome this problem by using appropriate approximations , we use the first principle method we show below , because it does not include any approximations in the calculation formula of the radiation spectra from prescribed trajectory . therefore , this method is suitable for studying the cases for which any approximations are difficult to adopt . by using moderately relativistic particles , we reduce the computational expenses . we neglect radiation back reaction of electrons , because the strength of electric field is sufficiently weaker than @xmath68 ( cf . jackson 1999 ) . we use 3d isotropic langmuir turbulence in this paper . if the langmuir waves are generated near the shock front , they may be highly anisotropic on the spot . however , we assume the isotropic turbulence for two reasons . first , the langmuir turbulence in large part of the emission region would be isotropic , since langmuir waves interact with background ions and form the kolmogorov type isotropic turbulence ( treumann & baumjohan 1997 ) . second , the radiation spectra for 3d and 1d turbulence are not so different except for a particular case that the direction of radiating electrons and the wave vector are parallel . for this case , the radiation is linear acceleration emission . however , emission from the perpendicular acceleration dominates as long as the direction differs from the wave vector direction even slightly ( fleishman & toptygin 2007a ) . if the shock is ultra - relativistic , the downstream plasma may be relativistically hot . for simplicity , we assume a mildly relativistic shock , so that the downstream plasma is sub - relativistic . we ignore the thermal velocity of background plasma in the dispersion relation of langmuir waves @xmath69 . thus , we use the propagating langmuir waves with the same frequency @xmath4 . we generate 3d isotropic langmuir turbulence by using fourier transform description , which is slightly modified from the description for magnetic turbulence developed by giacalone & jokipii ( 1999 ) . it is described by superpositions of @xmath70 fourier modes , each with random phase , and direction @xmath71 here , @xmath72 , @xmath73 , @xmath74 , and @xmath4 are the amplitude , phase , wave vector , frequency of the @xmath75th mode , respectively . the amplitude @xmath72 of each mode is defined as @xmath76^{-1},\ ] ] where the variable @xmath1 represents the amplitude of turbulent field . we use the following form for the power spectrum @xmath77 where @xmath78 is the correlation length of the field . here , @xmath79 is chosen such that there is an equal spacing in logarithmic @xmath80-space , over the finite interval @xmath81 and @xmath82 , where @xmath83 , @xmath84 and @xmath85 is chosen to be @xmath86 or @xmath87 . the spectrum has a peak at @xmath88 and the spectral index is for 3-dimensional isotropic langmuir turbulence . then we define two parameters which characterize radiation spectra . the first one is @xmath89 where @xmath90 and @xmath91 . we call @xmath5 `` strength omega '' , and @xmath7 `` spatial omega '' . the strength omega @xmath5 accounts for the effect of the field strength . the meaning of strength omega can be understood by considering the curvature of orbit and beaming effects for the relativistic particles as follows . for @xmath92 , the local curvature radius of the orbit suffering from perpendicular acceleration by the electric field is @xmath93 . the radiation is concentrated in the beaming cone with an angle @xmath19 , so the searchlight sweeps the observer in the timescale of @xmath94 . it is an analogy of the cyclotron frequency in the mechanism of synchrotron radiation . therefore , @xmath95 represents the effect of the field strength on the radiation spectra . as for the spatial omega @xmath7 , since the electron moves nearly at the light speed , the rate of change in the force direction for the electron is @xmath96 . the ratio @xmath97 parametrizes the field feature , which represents the ratio of the deflection angle @xmath43 to @xmath62 . therefore , the spectra from the turbulent magnetic field changes for @xmath98 to @xmath99 drastically ( kelner et al . 2013 , medvedev et al . 2011 , and teraki & takahara 2011 ) . note that @xmath14 is well known as the `` strength parameter '' . it means the change of the lorentz factor of an electron which is accelerated along the electric field on the spatial scale @xmath100 . we can understand it clearly by using the work by an electric field on an electron as @xmath101 where we assume the electric field is parallel to the velocity . we should note that it is different from the one we used in teraki & takahara 2011 by a factor of @xmath102 , and it agrees with the definition of fleishman & toptygin ( 2007a , b ) . the second one is the ratio of the frequency of the langmuir waves to @xmath7 , @xmath103 for @xmath104 , the force direction changes with frequency @xmath4 for all electrons . for @xmath105 , a change of the force direction is mainly from the spatial fluctuation . summarizing above , although there are three parameters of the turbulent field ( @xmath7 , @xmath5 , @xmath4 ) , we can reduce them to two parameters of ( @xmath106 ) when our interest is in the spectral signature . we investigate the spectral features in the parameter plane of ( @xmath106 ) . we inject monoenergetic electrons with lorentz factor @xmath107 into the generated turbulent field . the number of electrons used for each calculation is more than @xmath108 and is written in the caption of each figure . the initial velocity is randomly chosen to achieve an isotropic distribution . next we solve the equation of motion @xmath109 by using the method which has second order accuracy for each time step . the electrons get energy in the langmuir turbulence from the parallel ( to the velocity ) component . if we pursue electrons much longer time than the pft of a given frequency , the radiation spectrum in high frequency region corresponds to an `` integrated spectrum '' . this spectrum would be understood by a superposition of `` instantaneous spectra '' . here , the term of `` instantaneous spectrum '' means a spectrum which does not contain the effect of secular energy change of the radiating electrons . if we do not obtain the instantaneous spectra , we can not discern either the intrinsic feature of instantaneous spectra or energy change of radiating electrons determines the spectral features . therefore , in this paper we concentrate on `` instantaneous spectra '' . on the other hand , spectral broadening due to finite integration time of particle orbit is inevitable . to compromise the accuracy and the instantaneousness , we choose the integration time as @xmath110 times pft of the peak or break frequency . let us take an example of the spectrum for well known jitter radiation , for @xmath111 . the break frequency of the jitter radiation is @xmath112 , and corresponding pft is @xmath113 . we can understand the origin of this frequency by using `` the method of virtual quanta '' , ( cf . jackson 1999 ) . the spatial fluctuation of the electric field with a spatial scale @xmath114 . this fluctuation can be regarded as a photon which has the frequency with @xmath115 in the electron rest frame . the electron scatters the photon , and transformed back to the observer frame , the frequency of the scattered photon is written as @xmath116 . in other words , we can understand this process as an analogy of the inverse compton scattering for the photon with frequency @xmath7 . we note that this analogy is based on the assumption that the observer is in almost the center of the beaming cone over the pft . it is valid when @xmath117 . the deflection angle in the time scale of @xmath118 is @xmath119 , and when we require that it is smaller than @xmath62 , we get the condition @xmath120 . it can be transformed to @xmath121 . an alternative explanation can be done by using `` the photon chasing effect '' ( rybiki & lightmani 1979 ) . since the radiation from an ultra - relativistic particle is concentrated into the beaming cone , the particle chases the photon . therefore , the observer sees the emitted radiation in a time span @xmath122 , so the frequency is @xmath123 . as a consequence , the typical frequency is @xmath112 . as we see in the next section , pft for the peak frequencies of the spectra is the one of the typical timescales of @xmath118 , @xmath124 , and @xmath125 . we choose a suitable one for each case . the integration time @xmath110 times the pft of the peak frequency is sufficiently long to resolve the spectral shape in the low frequency regions . we calculate the radiation spectra using the acceleration , velocity , and position of electrons . the energy @xmath126 emitted per unit solid angle @xmath127 ( around the direction @xmath128 ) and per unit frequency @xmath129 to the direction @xmath128 is computed as @xmath130 } { ( 1 - \bm{\beta } \cdot \bm{n } ) ^2 } \exp\bigl\{{i\omega ( t^{\prime } - \frac{\bm{n } \cdot \bm{r}(t^{\prime})}{c})}\bigr\ } \bigr|^2,\ ] ] where @xmath131 is the electron trajectory , @xmath132 is retarded time ( jackson 1999 ) . since we have not any specific constraints for @xmath14 and @xmath133 which are realized in the high energy astrophysical object , we sweep wide parameter range of @xmath14 and @xmath133 . first , we show the radiation spectra for @xmath134 ( figure [ jit ] ) , i.e. , for the situation where typical spatial scale is shorter than the inertial length . we set @xmath135 to @xmath136 , and fix @xmath137 ; specifically we set @xmath138 and @xmath139 , and change @xmath5 from @xmath140 to @xmath136 , and take @xmath141 . the inequality @xmath105 can be transformed to @xmath142 , which means that the fluctuation scale is much smaller than the inertial length . these fluctuations would be damped by landau damping and may not be realized in high energy astrophysical objects ( treumann & baumjohan 1997 ) . however , we can not reject the possibility that @xmath143 is realized for some time span in relativistic plasmas , so that we study the spectra for @xmath144 . to explore the regime for @xmath144 clearly , we set an extreme value @xmath145 . an example of the temporal variation of the lorentz factor of an electron is depicted in figure [ jit_ene ] . for @xmath146 , the low frequency spectrum is as flat as @xmath60 , and there is a break at @xmath147 , and the spectrum declines with power law @xmath148 in the high frequency region . for @xmath149 , the break frequency becomes higher than that for @xmath146 , and the high frequency spectrum deviates from a power law . for @xmath150 , the spectrum in low frequency region becomes hard with an index @xmath151 , and the spectrum in higher frequencies reveals a cutoff feature . for @xmath152 , we see further different features . the spectrum in low frequency side of the peak becomes softer , with the spectral index of about @xmath18 , and we see a small deviation from a power law in the lowest frequency region . the features for these spectra can be understood by using the analogy to the radiation theory from the stochastic magnetic field ( medvedev et al . 2011 , teraki & takahara 2011 ) . since the wavelength of langmuir waves for @xmath105 is very short , the oscillation of the electric field can be neglected in the particle crossing time of the wavelength . in fact , @xmath105 is also written by @xmath153 . the crossing time @xmath154 corresponds to the pft of the typical frequency . as explained above , we can use the straight analogy to the radiation theory of the stochastic magnetic fields for the radiation spectra from langmuir turbulence for @xmath144 by substituting the electric field strength @xmath155 with magnetic field strength @xmath32 . for @xmath156 and @xmath144 , we can use the radiation theories of the dsr and jitter radiation theory ( fleishman 2006 , medvedev 2006 ) . on the contrary , for @xmath157 and @xmath144 jitter radiation of strong deflection regime can be applicable ( medvedev et al . 2011 , teraki & takahara 2011 ) . we call @xmath156 and @xmath144 regime as `` jitter radiation '' regime because the jitter radiation is basically perturbative theory for @xmath156 . we call the radiation for @xmath157 and @xmath144 regime as `` wiggler radiation in langmuir turbulence '' , or `` wrl '' in short . although the wiggler radiation is not the radiation mechanism from the stochastic field but that for a fixed field , it has a common picture that the observer is in and off the beaming cone in the course of time . first , we describe the signature of the radiation spectra of jitter radiation or diffusive synchrotron radiation . the radiation signatures are determined by acceleration perpendicular to the motion , and the observer is always in the beaming cone , which is the same situation as the langmuir turbulence for @xmath98 and @xmath144 . for @xmath117 , the spectrum is written by a broken power law @xmath60 in the low frequency region and @xmath158 in the high frequency region . the break frequency is @xmath159 . for @xmath160 , while the break frequency remains the same as @xmath161 , the multiple deflection effect comes into play in the spectral features near the break frequency . the spectrum in the low frequency region becomes @xmath61 . since the multiple deflection makes the angle between observer direction and velocity larger than the beaming cone angle @xmath62 , the observer sees the radiation over the timescale which is determined by the deflection condition . fleishman supposed that the angle changes diffusively . the break frequency of @xmath162 is calculated from the angle diffusion ( fleishman 2006 ) . the spectral index of @xmath63 comes from the diffusivity , too . next , we describe the signatures for strong deflection regime of @xmath163 . for magnetic turbulence , the spectral shape resembles synchrotron radiation in the middle frequency region and deviations from it would be seen in low frequency region and highest frequency region ( teraki & takahra 2011 ) . the peak frequency of @xmath164 is from the sweeping of the beaming cone on the observer . the picture is the same as the synchrotron radiation . for short wavelength regime @xmath165 of langmuir turbulence , the physical picture is the same as this , because the spatial fluctuation dominates the changing rate of the deflection angle . a single deflection angle is @xmath166 , which is larger than @xmath62 for @xmath167 . as a result , the beaming cone sweeps out of the observer within one deflection . the intensity of the radiation off the beaming cone is weak . therefore , the timescale which sweeps the observer @xmath168 corresponds to the pft of the typical frequency . considering the photon chasing effect , we get the peak frequency @xmath169 . the spectral break in the low frequency region would correspond to the deviation from local circular orbit , but the numerical error from finite integration time is also becoming large in the lower frequency region . we do not discuss this point further here , since it is not the main point of contents in this paper . the power law component in the highest frequency region comes from the smaller scale part of turbulence . it is the same as the spectra of jitter regime . we note here that the power law component in high frequency region can not be seen for @xmath150 and @xmath152 . the reason may be as follows . in contrast to the magnetic turbulence , the energy of radiating electrons changes for @xmath99 . therefore , the high frequency region is determined by only the later part of integration time , because the peak frequency is @xmath170 for @xmath157 , and the electrons get energy and radiate higher energy photon in later time . the power law component for the highest frequency region in our calculation is hidden by the component that small numbers of electrons with larger energy radiate by wiggler mechanism . finally , we consider effects of the energy change in wrl regime , which is the one of the different points from the magnetic case . based on an example of the change of lorentz factor is depicted in figure [ jit_ene ] , we estimate the energy change in pft @xmath171 for peak frequency . the pft is @xmath124 . the change of lorentz factor is estimated as @xmath172 therefore , the change of the lorentz factor in this timescale is smaller than @xmath17 . thus , we do not need to consider the energy change in pft for peak frequency . however , for sufficiently large @xmath14 , we have to take it into account for lower frequency spectrum . we calculate the spectra for modest @xmath14 in this paper , so we omit this problem . it will be studied in future works . next we show the radiation spectra for @xmath156 and @xmath13 ( figure [ drl ] ) , i.e. , a situation where long wavelength turbulence with weak amplitudes dominates ; the spectra for @xmath173 and @xmath17 are also depicted for comparison . an example of the temporal variation of the lorentz factor of an electron is depicted in figure [ drl_ene ] . firstly , we consider the meaning of the parameters of @xmath156 and @xmath13 , which correspond to @xmath174 . in this regime , the changes of the direction of acceleration are mainly due to wave oscillation , rather than the spatial fluctuations because the crossing time is longer than the oscillation time . moreover , @xmath175 is derived from @xmath176 . therefore , we can regard the orbit as straight in the time scale of plasma oscillation @xmath177 . the condition @xmath13 can be transformed to @xmath178 , which means that the inertial length is shorter than the wavelength . therefore , this regime is likely to occur in the astrophysical objects . we set @xmath138 , @xmath179 and change @xmath4 from @xmath140 to @xmath47 , so that @xmath180 and @xmath173 to @xmath47 , and take @xmath141 . as was discussed in the previous subsection for @xmath173 , the spectrum shows jitter radiation signature . the peak frequency is @xmath161 and @xmath181 in the low frequency region , and @xmath148 above the peak frequency reflecting the spectrum of the turbulent electric field @xmath182 . as @xmath133 becomes larger , the peak shifts to higher frequency . for @xmath183 , the peak frequency is @xmath184 , which is identified with @xmath185 . the spectral index of low frequency side is @xmath186 . this feature coincides with the result of drl theory ( fleishman 2006 ) . we call this regime `` drl regime '' . the drl theory predicts the spectral shape for @xmath156 and @xmath13 as follows . the peak frequency is @xmath185 with an abrupt cutoff in the higher frequency side and a power law component emerges in the highest frequency region for @xmath187 . in the low frequency side , the spectrum becomes @xmath12 . the peak frequency is determined by the timescale of plasma oscillation . the shortest timescale of the electron motion is @xmath177 , and the observer located along the initial velocity direction can see this radiation , because the orbit is regarded as straight in this time scale as we showed above . we consider doppler boosting , and we get the peak frequency @xmath188 in the observer frame . the origin of the power law component in the highest frequency region is the same as jitter radiation . the hard spectral index @xmath17 in the frequency region lower than the peak is from the photon chasing effect . the electrons oscillate by the same oscillating frequency @xmath4 and move different angles to the observer in the observer frame . the photon chasing effect shifts the photon frequency from each electron and makes the @xmath12 spectrum ( see fleishman 2007a , b ) . it is regarded as the emission spectrum from an electron integrated over the solid angle , which can be understood in an analogy to the undulator theory ( jackson 1999 ) , although the force direction changes in this case not spatially but temporally . when the particle mean velocity and the wavenumber are fixed , that makes no difference for the radiation spectra . the peaky shape of the spectra is the most remarkable feature of the spectra for drl case . a difference is that the wave number is a vector , while the frequency is a scalar . for dsr , the electron feels spatial fluctuation with wavenumber along the velocity @xmath189 , therefore the low frequency spectrum becomes flat . on the other hand , for drl , all electrons feel the same frequency of @xmath4 for the langmuir turbulence ( c.f . fleishman & toptygin 2007a , b and medvedev 2006 ) . the remaining interesting regime of @xmath157 and @xmath13 is an open issue . as we explained in section [ int ] , the radiation spectrum for langmuir turbulence with @xmath15 predicted by fleishman & toptygin ( 2007b ) is @xmath61 below the peak of @xmath190 , which is inconsistent with the statements by getamantsev and tokarev ( 1972 ) . here we investigate this regime and clarify the radiation spectra . the inequations of @xmath157 and @xmath13 mean that the wavelength is longer than the inertial length and that the typical scale of pft is @xmath191 . we set @xmath192 , @xmath193 to @xmath194 , so that @xmath195 to @xmath196 , and @xmath138 . we set @xmath197 , and we change @xmath4 from @xmath136 to @xmath194 , and @xmath85 is chosen to be @xmath87 here . we show the interesting results for @xmath157 and @xmath13 ( figure [ wig ] ) . an example of energy change for @xmath198 and @xmath193 is depicted in figure [ wig_ene ] . for clear discussion of the radiation spectra in this regime , we discuss the radiation spectra for @xmath199 and for @xmath200 separately . for @xmath200 , the peak frequency is @xmath201 . the spectral index of low frequency region is @xmath17 , and cutoff feature can be seen above the peak . this region is regarded as the drl regime from these signatures . for @xmath200 , i.e. , @xmath202 , the particle is not deflected by large angle because the direction of the electric field changes in a time shorter than the time for which the beaming cone sweeps the observer . we set @xmath203 , therefore , a power law component in high frequency region is not seen at all . on the contrary , different features emerge for @xmath199 . the peak frequency becomes larger than @xmath11 . moreover , the spectra below the peak frequency become softer , the index changes from @xmath17 to @xmath18 . the energy change of electrons may cause the change of the peak frequency , but it can not explain the soft spectrum . rather , it would be naturally understood that the peak frequency is @xmath164 and @xmath204 by using the analogy of the wiggler radiation . we consider that we should use wrl theory not only for @xmath205 , but also for @xmath15 , because the deflection angle in one deflection is also larger than @xmath62 for this case . this is in contrast to the drl theory , which predicts the same spectral signatures for the parameter range of @xmath15 as for @xmath206 . thus , our numerical calculations reveal new features which have not been known previously for langmuir turbulence . we ascribe that the parameter regime @xmath15 is in the wrl regime in @xmath10 plane . this radiation feature seems to be consistent with the statements by getmantsev & tokarev ( 1972 ) . to clarify the spectral features for @xmath15 in more detail , and to confirm their statement and our consideration , we examine the radiation from a relativistic electron moving in pure plasma oscillation in the next section . in this section , we investigate the emission of a relativistic electron suffering from pure plasma oscillation in order to discuss the interpretation of the features of the radiation spectra for @xmath15 . to clarify the origin of the peak frequency @xmath207 , we set a simple configuration of the electric field , where electron motion is deterministic compared to stochastic character in turbulent fields . we calculate the electron velocity analytically and the radiation spectra numerically . by comparing the motion and spectra , we interpret the mechanism which determines the peak frequency . lastly we consider the radiation spectra from the turbulent field by using these results . we use a single langmuir wave which has infinitely large wavelength @xmath208 , in other words , @xmath209 . therefore , it is a pure plasma oscillation . we set @xmath210 with @xmath211 we characterize the field by using a single parameter of @xmath212 we inject an electron along the @xmath213-axis at @xmath214 with the initial lorentz factor @xmath215 and solve the equation of motion ( [ eom ] ) . therefore , the orbit is determined by @xmath215 and @xmath216 . solving the equation of motion , we get the momentum @xmath217 since the velocity is a periodic function , we can define the mean velocity by @xmath218 . the mean velocity @xmath219 can not be represented elementarily in a general form . then , we take the parameter @xmath220 and approximate the motion hereafter . we expand the lorentz factor and the velocity , and get the mean velocity and the mean lorentz factor in the lowest order of @xmath221 . @xmath222 for @xmath223 , @xmath224 , while for @xmath225 , @xmath226 . we note that @xmath227 can be much smaller than @xmath215 for @xmath225 . using this approximated velocity , we calculate the maximum lorentz factor in the mean velocity frame , to clarify the fact that the radiation signatures depend on @xmath216 as @xmath228.\ ] ] for @xmath223 , @xmath229 the motion in this frame is non - relativistic , therefore , the radiation in this frame is dipole radiation . on the other hand , for @xmath225 , the maximum lorentz factor is @xmath230 therefore , the motion is relativistic even in this frame and the radiation spectrum consists of higher harmonics , because @xmath231 approaches @xmath17 . it should be noted that for @xmath232 , the motion in the mean velocity frame is mildly relativistic with lorentz factor @xmath233 , and @xmath234 . we can see that the transition from non - relativistic harmonic motion to relativistic motion occurs around @xmath235 a few from this fact . next we show numerically calculated radiation spectra from the electron and their features are interpreted in terms of the properties of the orbit . we fix @xmath236 , and change @xmath4 to change the parameter @xmath216 . the observer is on the @xmath213-direction . we calculate radiation spectra using much longer integration time than the pft , because the electron moves perfectly periodically . as a consequence , the spectra show very sharp features , which makes it easier to understand the relation between spectral features and orbit . first , we show the spectrum for @xmath237 ( @xmath238 , figure [ peaks](a ) ) . we see a sharp peak like a delta function at the frequency @xmath239 . this is understood in terms of the motion of the electron in the mean velocity frame . for @xmath223 , in the mean velocity frame , electron motion is a non - relativistic simple harmonic motion with the frequency @xmath240 . therefore , the radiation is the dipole radiation with the frequency of @xmath241 . since @xmath242 for @xmath223 , the radiation frequency in the observer frame is @xmath243 thus , we ascribe the frequency @xmath243 in the radiation spectra of perturbative regime ( @xmath202 ) to the doppler shifted dipole radiation . next we show the spectrum for @xmath244 ( @xmath245 , figure [ peaks](b ) ) . we can see the higher harmonics of @xmath246 . it is from effects of retarded time , as is clearly seen in the mean velocity frame . however , the effect is very weak , as the power of the second harmonics is about @xmath247 times smaller than the fundamental mode in the frequency resolution of this calculation . the ratio of the power of the second harmonics to the fundamental mode is proportional to @xmath248 , so the second harmonics is much smaller than the fundamental mode in this case . for @xmath249 , the spectral shape changes significantly . first , the higher harmonics stand more strongly , because @xmath231 approaches @xmath17 . many harmonics are as strong as the fundamental mode for @xmath250 ( figure [ peaks](c ) ) , and the envelope of the peaks of the harmonics shows an exponential cutoff . we note that the spectrum in the frequency region higher than @xmath251 comes from numerical error . second , the frequency of the fundamental mode deviates from @xmath243 , because the difference between @xmath227 and @xmath215 becomes larger . the mean lorentz factor @xmath227 is @xmath252 for @xmath232 , thus the frequency of the fundamental mode in the original frame is @xmath253 . the difference between @xmath254 and @xmath255 is small , but we can discern it in figure [ peaks](c ) . next we discuss the peak frequency ( cutoff frequency ) for @xmath256 . we show the spectra for @xmath257 , and @xmath196 in figure [ nc_peak ] . the fundamental frequency is @xmath254 for @xmath232 , @xmath258 for @xmath259 , and @xmath260 for @xmath261 , but the cutoff frequency around a few hundreds does not change . we see the cutoff frequency is not of the fundamental mode , but it is determined by the higher harmonics for @xmath256 . the radiation spectra in the observer frame also can be derived by regarding it as a doppler boosted emission . however , since the mechanism which determines the peak frequency is the same as the wiggler radiation we can understand the peak frequency more easily by considering the pft in the observer frame . the condition @xmath262 is equivalent to that pft of the typical frequency in wiggler mechanism , where @xmath124 is shorter than the oscillating time @xmath125 . on the other hand , the lorentz factor relevant for the peak frequency is not @xmath227 , but @xmath215 , because the beaming cone sweeps the observer in the phase around @xmath263 , where @xmath75 is a natural number . we note that the change of the lorentz factor in the time scale of @xmath124 is @xmath17 at most , as seen in equation ( [ lfchange ] ) . as a result , the cutoff frequency is @xmath264 . in this way , we get a clear understanding of the mechanism of the peak frequency shift around @xmath265 . here , we compare the results in this section with the radiation spectra obtained in the preceding section . the case of langmuir turbulence with @xmath15 ( @xmath266 ) corresponds to the case of pure plasma oscillation with @xmath256 , since @xmath267 and @xmath209 for pure plasma oscillation . moreover , the approximation we used in pure plasma oscillation of @xmath268 is also applicable for the case of langmuir turbulence , since @xmath269 and @xmath107 for the spectra shown in figure [ wig ] . for pure plasma oscillation with @xmath256 , we have shown that the peak frequency is @xmath270 , and it consists of the higher harmonics of @xmath271 . therefore , the peak frequency for the langmuir turbulence with @xmath272 in figure [ wig ] is interpreted as @xmath164 , and it naturally explains the fact that the peak frequency is larger than @xmath11 . lastly we consider the spectral index for the langmuir turbulence with @xmath15 . as we see above , the spectral index for @xmath15 is neither @xmath17 predicted by perturbative drl nor @xmath63 predicted by the angle diffusion effect . we regard that the spectral index is around @xmath18 , because the radiation mechanism is identified as wiggler mechanism , and the angle integrated spectral index is @xmath18 in wiggler mechanism . the energy change in pft for the frequency lower than peak can be larger than @xmath17 for @xmath15 . this is a different point from the wiggler , and it can affect forming the spectral index . however , we choose modest value of @xmath273 for langmuir turbulence in the present paper , so that this effect does not stand out . summarizing above , we confirm that the peak frequency is @xmath274 and the spectral index in low frequency side of the peak @xmath275 for the langmuir turbulence with @xmath15 and @xmath276 . the mechanism which determine these radiation features is wiggler mechanism , which is consistent with the statements by getamantsev and tokarev ( 1972 ) . lastly we show an example of the extreme case of @xmath277 . for @xmath278 , the motion becomes strongly nonlinear and can not be treated analytically . thus , we show numerically calculated electron orbit . we show the radiation spectra and the orbit for @xmath279 ( figure [ peaks](d ) , figure [ tr500 ] ) . as we expected , the peak frequency is @xmath264 , because the observer sees this electron mainly in the phase that the electric field and the electron velocity is nearly perpendicular . we note that the spectral index of @xmath280 is the same as the wiggler radiation when the observers located in a particular direction , i.e. , the spectrum is not the angle integrated spectrum . this spectral index is an collateral evidence for the spectral index of @xmath18 for the case of turbulence with @xmath15 for which angle integrated spectrum is calculated . the lorentz factor changes from @xmath47 to @xmath281 , but @xmath215 is relevant to the observer situated at @xmath213-axis . in other words , the beaming cone sweeps the observer when @xmath282 in the present geometry . the lorentz factor relevant to an observer oriented in different angle is significantly different to each other . moreover , in general the energy change in the pft of peak frequency becomes larger than @xmath283 , because the electric force in some phase of oscillation accelerates the electron linearly , and the curvature radius becomes larger . we have to consider the linear acceleration emission in this case . thus , we draw the line on @xmath284 and @xmath285 , and divide the @xmath15 region . we call the radiation for this parameter range as `` non - linear trajectory '' radiation . this part of the spectra from an electron moving in the 3d turbulent electric field is determined by the chaotic trajectory . the generalization of the features of this regime is a future work . summarizing this section , we have considered the motion and radiation in a single mode plasma oscillation . we clarify that the cutoff frequency for @xmath256 ( @xmath286 ) is @xmath270 , which consists of higher harmonics of the fundamental frequency of @xmath271 . it is from the effect that the beaming cone sweeps the observer , in the same way as the wiggler radiation . using this result , we interpret that the peak frequency for 3d langmuir turbulence for @xmath15 ( @xmath266 ) is @xmath164 , where @xmath23 is determined by the acceleration . the shallower spectrum for @xmath15 can be explained by wrl mechanism . lastly , we show numerically calculated radiation spectra for the extreme case of @xmath277 . it shows wiggler like spectra in the middle and high frequency region , while the flattening can be seen in the low frequency region . it is from the effect of elongated trajectory to the electric field direction . the radiation signatures are summarized as a chart in the @xmath10 plane in figure [ a - b_plane ] . we have calculated the radiation spectra from relativistic electrons moving in a langmuir turbulence by using first principle numerical calculation . we characterize the radiation spectra by two parameters . the one is @xmath287 , where @xmath3 is the strength omega , and @xmath288 is the spatial omega . the strength omega accounts for the effect of the field strength to the radiation spectra , and the spatial omega accounts for the effect of spatial fluctuation with a typical scale of @xmath0 . the other is @xmath289 , where @xmath4 is the plasma frequency , which accounts for effects of the time variability of the waves . we investigate the spectral signatures in the @xmath10 plane ( figure [ a - b_plane ] ) . for @xmath156 and @xmath144 , the spectral features are the same as those of jitter radiation or diffusive synchrotron radiation . for @xmath206 and @xmath290 , the theory of the diffusive radiation in langmuir turbulence is confirmed , where time variability plays a primary role . for @xmath15 , the spectra show novel features which are not predicted by drl theory . in this regime , the peak frequency is @xmath291 , which is higher than the predicted frequency @xmath11 from the drl theory . the spectral index of the frequency region lower than the peak is @xmath275 . these features are explained by the wiggler mechanism . to clarify the radiation features in this regime , we calculate the radiation spectra from an electron moving in an oscillating electric field , i.e. , for vanishing spatial omega . we analytically calculate the motion of the electron , and numerically calculate the radiation spectra form this electron . we show that for @xmath292 , the spectrum around the peak frequency consists of the higher harmonics of the fundamental mode , by considering the radiation in the mean velocity frame . the electron motion becomes relativistic for @xmath256 even in this frame , so that strong higher harmonics photons are emitted because of the retarded time effect . as a result , the spectra in the observer frame consists of the higher harmonics of @xmath64 . the peak frequency is characterized by @xmath207 , which is understood by the analogy of the wiggler radiation . the feature that the radiation spectra from langmuir turbulence have a wide range of spectral indices can be important for high energy astrophysical objects , in particular gamma ray bursts . the emission mechanism of grb is not settled for now . the spectral indices of low frequency side of the band function are distributed as a gaussian with the central value of @xmath293 . non - negligible number of grbs have spectral index harder than the theoretical limit for synchrotron radiation @xmath18 . the photospheric emission model can overcome this difficulty , but it also has a difficulty that intrinsic photosheric emission is too hard with a low energy spectral index of @xmath294 and it is difficult to make it softer . on the other hand , the radiation mechanism from langmuir turbulence in this paper has some advantages . not only the spectral index is harder than the synchrotron radiation and it can reproduce very hard spectra of observed grb ( fleishman 2007b ) , but also it may explain a wide range of spectral indices . because the parameters of @xmath14 and @xmath133 are likely to have a value around 1 near the shock front ( silva 2006 , dieckmann 2005 ) , the radiation spectra change drastically around these parameters . we are grateful to the referee for his / her constructive and helpful comments and for informing us of the interesting paper by getmantsev & tokarev ( 1972 ) . we also thank k. toma and s. kimura for useful discussions . this work is supported by the jsps research fellowships for young scientists ( y.t . , 24593 ) . 99 akhiezer , a. i. , & shulga , n. f. 1987 , sov . phys.usp . , 30 , 197 bret , a. , dieckmann , m. e. , & deutch , c. 2006 , phpl , 13 , 082109 dieckmann , m. e. 2005 , phys . 94 , 155001 fleishman , g. d. 2006 , apj , 638 , 348 fleishman , g. d. , & toptygin , i. n. 2007a , phys . e. , 76 , 017401 fleishman , g. d. , & toptygin , i. n. 2007b , mnras , 381 , 1473 fleishman , g.d . , & urtiev , f. a. 2010 , mnras , 406 , 644 giacalone , j. , & jokipii , j. r. 1999 , apj , 520 , 204 getmantsev , g. g. , & tokarev , yu . v. 1972 , ap&ss , 18 , 135 jackson , j. d. 1999 , classical electrodynamics ( 3rd ed . ; new york : wiley ) kelner , s. r. , aharonian , f. a. , & khangulyan , d. 2013 , apj , 774 , 61 landau , l. d. , & lifshitz , e. m. 1980 , the classical theory of fields ( 4th ed.;butterworth - heinemann ) mao , j. , & wang , j. 2013 , apj , 776 , 17 medvedev , m. v. 2000 , apj , 540 , 714 medvedev , m. v. 2006 , apj , 637 , 869 medvedev , m.v . , frederiksen , j. t. , haugbolle , t. , & nordlund , a. 2011 , apj , 737 , 55 ostriker , j. p. & gunn , j. e. 1969 , apj , 157 , 1395 rees , m. j. 1971 , iaus , 46 , 407 reville , b. , & kirk , j. g. 2010 , apj , 724 , 1283 reynolds , s. j. , & medvedev m. v. 2012 , phpl , 19 , 023106 rybicki , r. d. , & lightman , j. g. 1979 , radiative processes in astrophysics ( new york : wiley ) silva , l. o. aipc , 856 , 109 treumann , r. a. , baumjohann , w. 1997 , advanced space plasma physics ( london : imperial college press ) teraki , y. , & takahara , f. 2011 , apj , 735 , l44 teraki , y. , & takahara , f. 2013 , apj , 763 , 131 toptygin , i. n. & fleishman , g. d. 1987 , ap&ss , 132 , 213 and @xmath296 , and @xmath47 from top to down in low frequency range . the number of electrons used for these calculations is @xmath297 . the straight lines show a power law spectrum with index @xmath17 and @xmath298 . we see the transition from jitter radiation to drl as @xmath133 increases . , width=453 ] with @xmath193 , @xmath300 , @xmath301 , and @xmath194 from top to down . the number of electrons used in these calculations is @xmath194 . the straight lines show the power law spectra with index @xmath18 and @xmath17 . we see the transition from the drl to wrl as @xmath133 decreases . , width=453 ] , and @xmath196 ( @xmath303 , and @xmath304 ) from top to down . a factor of @xmath305 is multiplied to the spectrum for @xmath232 , and @xmath306 is multiplied for @xmath261 . the fundamental frequency is @xmath307 . it is @xmath254 for @xmath232 , @xmath258 for @xmath259 and @xmath260 for @xmath261 . the cutoff frequency is a few times of @xmath308 , which does not depend on @xmath309 explicitly . , width=453 ] , and vertical axis is @xmath311 . for the jitter regime with @xmath156 and @xmath144 , the radiation spectra are determined by the spacial fluctuations , because @xmath7 is the largest of the three . the typical frequency for this case is @xmath161 . for @xmath206 , i.e. , @xmath312 , the radiation spectra are represented by drl theory , and typical frequency is @xmath313 . the line @xmath314 divides the drl region and wrl region , and the spectral features for the wrl regime @xmath15 are newly clarified in this paper . the typical frequency is @xmath164 and the spectral index of frequency region lower than the peak is @xmath275 , in the same way as synchrotron radiation . for @xmath315 and @xmath316 , the orbit of a radiating electron depicts non - linear trajectory , and its signature appears at the low frequency region of the spectrum . , width=340 ]

any conceivable theory of structure formation in the universe predicts statistical properties of observable quantities . therefore , the analysis of the present inhomogeneity of the universe is inevitably statistical . however , what kind of statistics is useful is not obvious , since we can invent infinite number of statistics to be analyzed . whether we can adopt better statistical descriptions is of great importance in this sense . each statistic has both advantages and disadvantages . the power spectrum , for example , can fully characterize the random gaussian fields , while it does not contain any information on the non - gaussianity , which contains significant information in the gravitational instability theory . among various statistical quantities , there is a promising class of statistics which utilizes smoothed cosmic fields . the smoothed field has less noisy property than the actual ( unsmoothed , or raw ) cosmic fields , such as galaxy distributions , temperature fluctuations of cosmic microwave background ( cmb ) , shear fields of the gravitational lensing , and so forth . perhaps the simplest example of such statistics is the variance @xmath0 of smoothed density contrast @xmath1 , which is a function of smoothing length @xmath2 . similarly , higher - order cumulants @xmath3 ( @xmath4 ) are also simple statistics . the density probability distribution function ( pdf ) @xmath5 , which in principle can be constructed from the hierarchy of cumulants ( e.g. , * ? ? ? * ) , is an another example of popular statistics of smoothed cosmic fields . rather recently , more complex statistics of smoothed cosmic fields have become popular in cosmology , such as the genus statistic @xcite , density peak statistics @xcite , area , length , level - crossing statistics @xcite , minkowski functionals @xcite , etc . these statistics provide assuring characterizations of the clustering pattern that can not be perceived only by the hierarchy of cumulants or by the pdf . the genus statistic is a powerful measure of the morphology in the 3-dimensional @xcite , and 2-dimensional @xcite clustering of galaxies and clusters , and also of the pattern of temperature fluctuations of cmb radiation @xcite . the peak statistics of the 3-dimensional density field are frequently used in connection with the statistics of the collapsing object @xcite , while the peak statistics of the cmb fluctuations @xcite and of the weak lensing fields @xcite are suitable for constraining cosmological models . the area , length , and level - crossing statistics directly quantify the amount of contour surfaces @xcite . the minkowski functionals @xcite , which are closely related to the above statistics , are also applied to smoothed cosmic fields @xcite . these statistics of smoothed density fields are considered as powerful descriptors of the statistical information of the universe . it is therefore essential to establish theoretical predictions of the behavior of such statistics so that we can effectively and ideally analyze the data of our universe . recent developments of the perturbation theory ( for review , see * ? ? ? * ) in calculating the variance , the cumulants , and the pdf are remarkable . the perturbation theory becomes more and more useful because the recent developments of observations enable us to have widely covered sample volume of the universe , which can minimize undesirable strongly nonlinear effects which we do not understand well . the direct comparison of the theoretical predictions of the perturbation theory with the actual data is a promising field of research in this sense . in the case of top - hat smoothing function , @xcite and @xcite used the tree - level ( i.e. , lowest - order ) perturbation theory to obtain the third- and forth - order cumulants , i.e. , the skewness and the kurtosis . the same quantities with gaussian smoothing are calculated by @xcite , @xcite and @xcite . quite cleverly , @xcite took advantage of special properties of the top - hat smoothing function , and succeeded to obtain full hierarchy of higher order moments in the tree - level perturbation theory . he also obtained the pdf from this hierarchy of cumulants , which remarkably describes the nonlinear behavior of the gravitational instability in numerical simulations . beyond the tree - level calculation , the perturbation theory with loop - corrections has been also developed @xcite . because of the simplicity of the statistics , calculating the variance , the cumulants and the pdf were primarily the playground of perturbation - theorists . the evaluation of other statistics by the perturbation theory is less trivial . a quite useful technique is the edgeworth expansion , which was first applied to cosmology in calculating the pdf from seeds perturbations by @xcite , and from non - linear perturbation theory by @xcite , and @xcite . the analytic expression of the genus statistics in the perturbation theory is derived by @xcite , whose technique corresponds to multivariate version of the edgeworth expansion . in some literatures the edgeworth expansion is used in connecting the statistics and dynamics of the universe @xcite . the purpose of this paper is to give a comprehensive description of the formalism by which the perturbative evaluation is possible for wide range of non - trivial statistics of smoothed cosmic fields in general . since the number of spatial dimensions of our universe is three , the statistics of large - scale cosmic fields are defined in either one- , two- , or three - dimensional space . for example , the space of the density field in a redshift map of galaxies or quasars is three - dimensional . a projected galaxy map on the sky , a shear field of gravitational lensing , and temperature fluctuations of cmb on the sky are fields in two - dimensional space . the absorption lines of quasar spectra and pencil - beam surveys of galaxies define fields in one - dimensional space . thus it is useful to develop our statistical method in @xmath6-dimensions for generality . as illustrative examples of applications of our method , we calculate level - crossing statistic ( or equivalently length and area statistics ) , genus statistics , density - extrema statistics and minkowski functionals . as cosmic fields , we consider density and velocity fields in three dimensions , the projected density field of galaxies in two dimensions . the basic formalism and results of the second order theory are presented in this paper . we will give results of the third order theory in a future paper , which are technically more involved . the primary purpose of the present paper is to fully describe the basic formalism for the statistics of smoothed cosmic fields in perturbation theory . applications of the second order theory to popular statistics and cosmic fields are systematically presented . thus , this paper is in some way a mixture of the new results and comprehensive review of the old results . the major new results in this paper are : * explicit formulas of the lowest non - gaussian correction in various smoothed cosmic fields in @xmath6-dimensions , including all the minkowski functionals . they are presented as functions of both density threshold and rescaled threshold by volume - fraction . * relations among several statistics of the smoothed field are clarified . * derivatives of skewness parameter of the velocity field and projected 2d field are derived in second order perturbation theory of gravitational instability theory . especially , skewness parameters with gaussian smoothing are detailed . these quantities are particularly important to perturbatively evaluate the statistics of smoothed field . * genus curve against the scaled threshold in perturbation theory is discussed and a theory - motivated new asymmetry parameter in genus curve is introduced . this paper is structured in the following way . [ sec2 ] and [ sec3 ] are carrying out a mathematical exercise of expressing high - order terms in terms of skewness parameters ; the physics of gravity only enters once one actually calculates the skewness parameters in [ sec4 ] . thus [ sec2 ] and [ sec3 ] are a derivation of an extension of the edgeworth expansion , and the basic results are equations ( [ eq1 - 13 ] ) and ( [ eq1 - 17 ] ) . in [ sec3 ] , popular statistics of smoothed fields are examined . the second - order expressions in terms of the skewness parameters are given for pdf , level - crossing statistics , 2-dimensional and 3-dimensional genus statistics , 2-dimensional weighted extrema , and minkowski functionals in [ sec3 - 1][sec3 - 6 ] . these quantities are re - expressed as functions of the volume - fraction threshold in [ sec3 - 7 ] . the skewness parameters for several cosmic fields are calculated in [ sec4 ] , applying the second order perturbation theory . detailed calculations of the simple hierarchical model , the 3-dimensional density field , the velocity field , the 2-dimensional projected density field , and the weak lensing field are given in [ sec4 - 1][sec4 - 5 ] . the effect of biasing on the skewness parameters are discussed in [ sec4 - 6 ] . we discuss implications of our second - order results in we introduce a theory - motivated asymmetry factor to characterize the shift of the genus curve in this section . the conclusions are given in useful gaussian integrals including hermite polynomials are given in appendix [ app1 ] . the bispectrum in a two - dimensional projected field is reviewed in appendix [ app2 ] . the symbols in this paper are summarized in appendix [ app3 ] . we consider a cosmic random field @xmath7 which represents any field constructed from observable quantities of the universe , such as three - dimensional density field , velocity field , or two - dimensional projected density field , shear or convergence field of weak lensing , temperature fluctuations of cmb , etc . the coordinates @xmath8 can be either three- , two- , or one - dimensions . we assume the field @xmath9 is already smoothed by a smoothing function @xmath10 with smoothing length @xmath2 which cuts the high frequency fluctuations which suffer strongly nonlinear effects : @xmath11 where @xmath12 is the dimension of the space @xmath8 , and @xmath13 is a raw , unsmoothed field . two of the most popular 3d smoothing functions are the tophat smoothing function @xmath14 and the gaussian smoothing function @xmath15 we also assume that the mean value of the field @xmath9 is zero , and that the existence the variance @xmath16 : @xmath17 it is convenient to introduce a normalized field @xmath18 which has a unit variance as follows : @xmath19 the statistics of a smoothed cosmic field we are interested in are the functions of the field @xmath18 and its spatial derivatives as we will see in the following sections . we denote the series of spatial derivatives by a set of variables @xmath20 which is defined by , for example , in 3d case , @xmath21 where @xmath22 . for convenience , the index is denoted as @xmath23 , 1 , 2 , 3 , ( 11 ) , ( 22 ) , ( 33 ) , ( 12 ) , ( 13 ) , ( 23 ) , @xmath24 in such cases . in this example of equation ( [ eq1 - 4 ] ) , only the value of the field and of the derivatives on a single point is considered , but in general , more than two points can be considered . the set @xmath25 forms multivariate random fields , which is denoted as an @xmath26-dimensional vector @xmath27 in the following . the dimension @xmath26 is the total number of derivatives which appear in the definition of the statistics we are interested in . the statistical information is described by the multivariate pdf , @xmath28 . the fourier transform of the pdf is the partition function : @xmath29 at this point , the cumulant expansion theorem ( e.g. , * ? ? ? * ) is very useful . this theorem states that @xmath30 is the generating function of the cumulants , @xmath31 : @xmath32 it follows from @xmath33 that @xmath34 , and first several cumulants are given by @xmath35 and so forth . from equations ( [ eq1 - 6 ] ) and ( [ eq1 - 7a ] ) , the partition function is given by @xmath36 where @xmath37 is an @xmath38 matrix whose components are given by @xmath39 . on the other hand , the equation ( [ eq1 - 5 ] ) is inverted as @xmath40 substituting @xmath41 in the last term of equation ( [ eq1 - 8 ] ) , the distribution function of equation ( [ eq1 - 9 ] ) can be transformed in a form @xmath42 where @xmath43 is the multivariate gaussian distribution function characterized by the correlation matrix @xmath37 . any statistical quantity of a smoothed cosmic field is expressed by an average @xmath44 of a certain function @xmath45 as we will see in the following sections . thus , from equation ( [ eq1 - 10 ] ) , we obtain @xmath46 where @xmath47 denotes the averaging by the gaussian distribution function of equation ( [ eq1 - 11b ] ) . this form , equation ( [ eq1 - 12b ] ) , is useful when the deviation from the gaussian distribution is not large . in principle , this equation reduces the general statistical averaging procedure to gaussian integrations . however , it contains the infinite series , thus we have to truncate this expression by some criteria . in most cases of interest , the weakly nonlinear evolution of cosmic fields satisfy @xmath48 , as we will see in [ sec4 ] . when this relation holds , we can expand the distribution function to arbitrary order in @xmath49 . in the following , we assume this relation , and introduce the normalized cumulants : @xmath50 which are assumed to be of order one in terms of @xmath49 . in this case , the equation ( [ eq1 - 12b ] ) is expanded as , up to @xmath51 , @xmath52 { \sigma_0}^2 + { \cal o}({\sigma_0}^3 ) , \label{eq1 - 13}\end{aligned}\ ] ] where we introduce the notation , @xmath53 etc . the calculation of the factors @xmath54 is performed for individual statistic which gives the explicit form of the function @xmath55 . in the expansion of equation ( [ eq1 - 13 ] ) , we need to calculate the gaussian average of derivatives of the function @xmath55 : @xmath56 in most cases , the evaluation is analytically feasible because only gaussian integration is needed . for the evaluation of such integration , we need the correlation matrix @xmath37 . throughout this paper , we consider statistically homogeneous , and isotropic fields , in which case , the correlation matrix @xmath37 is simplified because of the symmetry . in fact , the correlation matrix takes the following forms : @xmath57 where @xmath58 etc . and we define the following quantities : @xmath59 in most cases of interest , the derivatives of order higher than two do not appear in the definition of the statistics , so we do not give explicit form of those correlations . in the following , we use the notation @xmath60 and @xmath61 following @xcite . as is often the case , when the second order derivatives @xmath62 appears in @xmath55 as simple polynomials , the following transform is particularly useful : @xmath63 this transform erase the correlation between @xmath18 and @xmath64 , and the non - zero correlations are only @xmath65 where @xmath66 the gaussian integration is straightforward if the function @xmath55 is expressed by polynomials of @xmath64 . if the function @xmath55 is more complicated in which @xmath62 are not simply given by polynomials , it is useful to completely diagonalize the correlation matrix @xmath37 of equations ( [ eq1 - 13 - 2a])([eq1 - 13 - 2a ] ) . we introduce the following transform , which is quite similar one in @xcite : @xmath67 if @xmath68 , we ignore the variables @xmath69 , and similarly , if @xmath70 , we ignore the variable @xmath71 in the above equations and in the following . the above transform is similar to @xcite but notice it is not identical . this transform completely diagonalize the correlation matrix : @xmath72 and all non - diagonal correlations are zero . for later convenience , we write the inverse relation of the transform of @xmath73 , @xmath74 , and @xmath71 : @xmath75 if @xmath76 does not appear in the function @xmath55 , the variable @xmath71 should be omitted in the above equations . if both @xmath77 and @xmath76 do not appear in the function @xmath55 , the variables @xmath74 and @xmath71 should be omitted . after expressing the function @xmath55 in terms of the diagonalized variables @xmath18 , @xmath78 , @xmath73 , @xmath74 , @xmath71 and @xmath79 , the calculation of the gaussian integration of equation ( [ eq1 - 13 - 1 ] ) is performed . in this paper , only first two terms in equation ( [ eq1 - 13 ] ) are considered . thus we evaluate @xmath80 here . when the spatial symmetry is taken into account , the quantity @xmath80 reduces to the following expressions : @xmath81 , \label{eq1 - 14e}\\ & & \langle\alpha_{,i}\alpha_{,j}\alpha_{,k}\rangle = 0 , \label{eq1 - 14f}\\ & & \langle\alpha_{,i}\alpha_{,j}\alpha_{,kl}\rangle = \frac{\langle(\nabla\alpha\cdot\nabla\alpha)\nabla^2\alpha\rangle } { d(d-1 ) } \left [ \delta_{ij}\delta_{kl } - \frac12 \left(\delta_{ik}\delta_{jl } + \delta_{il}\delta_{jk}\right ) \right ] , \label{eq1 - 14g}\\ & & \langle\alpha_{,i}\alpha_{,jk}\alpha_{,lm}\rangle = 0 , \label{eq1 - 14h}\end{aligned}\ ] ] etc . to prove the above equations , the following identities for isotropic fields are useful : @xmath82 although the above equations are valid for @xmath83 , the case @xmath68 is obtained by just ignoring the terms with @xmath84 . generally , more complicated quantities can be appeared for @xmath80 , but the above relations are sufficient for our applications in this paper . according to the spatial symmetry of the above equations , the third - order correlations @xmath85 are explicitly given . at this point , it is useful to define the following quantities : @xmath86 we call quantities @xmath87 skewness parameters . the first one @xmath88 is usually called as skewness , and the others are its derivatives . they are to be calculated from usual perturbation theories in [ sec4 ] . using these quantities , the third - order correlations are given by @xmath89 , \qquad ( i \ne j ) , \label{eq1 - 16i}\\ & & { \widehat{m}^{(3)}}_{0(ii)(jk ) } = 0 , \qquad ( j < k ) , \label{eq1 - 16j}\\ & & { \widehat{m}^{(3)}}_{0(ij)(ij ) } = - \frac{1}{d(d+2)}\frac{{\sigma_1}^4}{{\sigma_0}^4 } \left [ s^{(2 ) } - s^{(2)}_2 \right ] \qquad ( i < j ) , \label{eq1 - 16k}\\ & & { \widehat{m}^{(3)}}_{0(ij)(kl ) } = 0 , \qquad ( i < j , k < l , i \ne k ) , \label{eq1 - 16l}\\ & & { \widehat{m}^{(3)}}_{ijk } = 0 , \label{eq1 - 16m}\\ & & { \widehat{m}^{(3)}}_{ii(ii ) } = 0 , \label{eq1 - 16n}\\ & & { \widehat{m}^{(3)}}_{ii(jj ) } = - \frac{2}{3 d^2}\frac{{\sigma_1}^4}{{\sigma_0}^4 } s^{(2 ) } , \qquad ( i \ne j ) , \label{eq1 - 16o}\\ & & { \widehat{m}^{(3)}}_{ii(jk ) } = 0 , \qquad ( j < k ) , \label{eq1 - 16p}\\ & & { \widehat{m}^{(3)}}_{ij(kk ) } = 0 , \qquad ( i \ne j ) , \label{eq1 - 16q}\\ & & { \widehat{m}^{(3)}}_{ij(ij ) } = \frac{1}{3 d^2}\frac{{\sigma_1}^4}{{\sigma_0}^4 } s^{(2 ) } , \qquad ( i < j ) , \label{eq1 - 16r}\\ & & { \widehat{m}^{(3)}}_{ij(kl ) } = 0 , \qquad ( i < j , k < l , i \ne k ) , \label{eq1 - 16s}\\ & & { \widehat{m}^{(3)}}_{i(jk)(lm ) } = 0 , \label{eq1 - 16t}\end{aligned}\ ] ] and so forth , where repeated indices are _ not _ summed over in the above equations , and @xmath90 is symmetric under permutation of its indices . thus , denoting @xmath91 for simplicity , @xmath92 \nonumber\\ & & \quad\qquad + \ , \frac{3 s^{(2)}_2}{d(d+2 ) } \frac{{\sigma_1}^4}{{\sigma_0}^4 } \left ( 3 \sum_{i } f_{0(ii)(ii ) } + 2 \sum_{i < j } f_{0(ii)(jj ) } + \sum_{i < j } f_{0(ij)(ij ) } \right ) + \cdots . \label{eq1 - 17}\end{aligned}\ ] ] this equation gives the second - order correction term of the statistical quantity @xmath44 through equation ( [ eq1 - 13 ] ) . once the field @xmath9 is specified , the skewness parameters @xmath87 are calculated by dynamical perturbation theory of the field @xmath9 . the remaining factors in the above equation are the gaussian integrations of the derivatives of the function @xmath55 , i.e. , @xmath93 . these factors can be calculated once the function @xmath55 is given . in the next section , we calculate the latter factors for individual statistics . in this section we calculate the factor @xmath94 for each statistic . some of the results in this section have been previously presented . the edgeworth expansion of the pdf in [ sec3 - 1 ] is a familiar result . the result of 3d genus statistic in [ sec3 - 4 ] was already given by @xcite . we include these old results for completeness . other subsections present new results . perhaps the simplest yet non - trivial statistic is the pdf , @xmath95 . the perturbative expansion of the pdf is known as the edgeworth expansion @xcite . as the simplest example , we re - derive the known edgeworth expansion from the point of view of our general formalism above ( see also * ? ? ? * ) . since the pdf is simply given by @xmath96 , where @xmath97 is the dirac s delta function , the function @xmath55 in the previous section for pdf @xmath95 is given by @xmath98 since this form of @xmath55 does not depend on derivatives of @xmath18 , only @xmath99 survives in the equation ( [ eq1 - 17 ] ) . from equation ( [ eqa1 - 4a ] ) with @xmath100 and @xmath101 , @xmath102 , where @xmath103 . thus , the pdf is derived from the equation ( [ eq1 - 13 ] ) : @xmath104 , \label{eq2 - 0 - 2}\end{aligned}\ ] ] which reproduces the well - known result . there is less advantage of applying our formalism to this simple statistic which can be treated by standard methods . our formalism has advantages when more non - trivial statistics are considered as shown below . next three statistics we consider here are the level - crossing statistic @xmath105 , the length statistic @xmath106 and the area statistic @xmath107 . the level - crossing statistic is defined by the mean number of intersection of a straight line and threshold contours of the field . the length statistic is defined by the mean length of intersection of a 2d surface and the threshold contours of the field . the area statistic is defined by the mean area of the contour surface in a 3d space @xcite . the level - crossing statistic is defined for 1d , 2d , and 3d cosmic fields , the length statistic is defined for 2d and 3d cosmic fields , while the area statistic is defined only for 3d cosmic fields . for statistically isotropic fields , those three statistics are proportional to each other . in general , a statistic of the smoothed field @xmath9 is a function of the threshold @xmath108 , or of the normalized threshold @xmath109 . the explicit expressions of statistics @xmath105 , @xmath106 , and @xmath107 are given by @xcite @xmath110^{1/2 } \right\rangle , \label{eq2 - 1b}\\ & & n_3(\nu ) = \left\langle \delta(\alpha - \nu ) \left [ ( \eta_1)^2 + ( \eta_2)^2 + ( \eta_3)^2 \right]^{1/2 } \right\rangle . \label{eq2 - 1c}\end{aligned}\ ] ] for isotropic fields , these statistics are actually equivalent @xcite . in fact , the distribution function of @xmath111 for fixed @xmath112 is the function of only the magnitude @xmath113 . thus , using spherical coordinates for @xmath114 , one can see @xmath115 i.e. , @xmath116 thus we only need to consider @xmath105 which has the simplest expression and the rests of the statistics are automatically given by equation ( [ eq2 - 2 ] ) . however , if the field is anisotropic , such as the density field in redshift space @xcite , the equation ( [ eq2 - 2 ] ) no longer holds , and equations ( [ eq2 - 1a])([eq2 - 1c ] ) should be used for each statistic . the equation ( [ eq1 - 17 ] ) only holds for statistically isotropic fields , but the equation ( [ eq1 - 13 ] ) is applicable even for anisotropic fields . now we calculate the factor @xmath117 for the particular statistic @xmath105 . the indices @xmath118 only take @xmath119 and @xmath120 for @xmath105 statistic . let the number of @xmath119 be @xmath121 and the number of @xmath120 be @xmath122 . then the factor is given by @xmath123 since the variables @xmath18 and @xmath78 are uncorrelated in the gaussian averaging , we can use equations ( [ eqa1 - 4a ] ) and ( [ eqa1 - 5a ] ) in appendix [ app1 ] . thus the above gaussian integration results in @xmath124 where @xmath125 is given by equation ( [ eqa1 - 3 ] ) . now , calculation of equation ( [ eq1 - 17 ] ) is straightforward : @xmath126 . \label{eq2 - 5}\end{aligned}\ ] ] on the other hand , the gaussian contribution is simply given by @xmath127 thus , the perturbative expansion of equation ( [ eq1 - 13 ] ) up to second order is finally given by @xmath128 \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\}. \label{eq2 - 7}\end{aligned}\ ] ] the second order formulas for area and length statistics are given by equation ( [ eq2 - 2 ] ) with the above equation . to evaluate the above formula , the factor @xmath88 and @xmath129 should be known . those factors are evaluated by usual perturbation theory in the following section . the next statistics we consider are the genus statistics . the genus statistics have been attractive because it has a geometrical meaning of the clustering as well as the cosmological significance . the 2d genus statistic @xmath130 is defined in a two - dimensional plane @xmath131 in a @xmath6-dimensional space , so that @xmath132 is required . in this plane @xmath131 , there are contours corresponding to each threshold @xmath133 . the 2d genus statistic is defined by the number of contours surrounding regions higher than the threshold value minus the number of contours surrounding regions lower than the threshold value @xcite . this definition is intuitive , but an alternative , equivalent definition is more useful : first we set an arbitrary , fixed direction on the plane @xmath131 . then there are maxima and minima of contours according to that direction . these points are classified into upcrossing points and downcrossing points with respect to that chosen direction . the number of those points are used to define the 2d genus statistic as the following way : @xmath134 , \label{eq2 - 8}\end{aligned}\ ] ] per unit area of the surface . according to this definition , the explicit expression of the 2d genus statistic is given by @xmath135 for this statistic , the indices @xmath136 only take @xmath119 , @xmath120 , @xmath137 and @xmath138 . let the number of @xmath119 be @xmath121 , @xmath120 be @xmath139 , @xmath137 be @xmath140 , and @xmath138 be @xmath141 . then we need to calculate the following quantity : @xmath142 since the second derivative @xmath143 appears as a polynomial , we just use the transform of equation ( [ eq1 - 13 - 3 - 1 ] ) . then , from equations ( [ eqa1 - 4a ] ) , ( [ eqa1 - 5a ] ) and ( [ eqa1 - 5b ] ) , the above equation reduces to @xmath144 . \label{eq2 - 11}\end{aligned}\ ] ] thus , from equation ( [ eq1 - 17 ] ) , @xmath145 - 2 s^{(2 ) } \left ( \frac{\sigma_1}{\sqrt{d}\sigma_0 } \right)^4 r(0;0;2,1 ) \nonumber\\ & = & \frac{1}{(2\pi)^{3/2 } } \left ( \frac{\sigma_1}{\sqrt{d}\sigma_0 } \right)^2 e^{-\nu^2/2 } \left [ s^{(0 ) } h_4(\nu ) + 4 s^{(1 ) } h_2(\nu ) + 2 s^{(2 ) } \right ] . \label{eq2 - 12}\end{aligned}\ ] ] on the other hand , the gaussian contribution is simply given by @xmath146 thus , the perturbative expansion of equation ( [ eq1 - 13 ] ) up to second order is finally given by @xmath147 \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\}. \label{eq2 - 14}\end{aligned}\ ] ] the second order formula for 3d genus statistic was already derived by @xcite , but the detailed derivation was omitted . for completeness , we revisit the same quantity from our general point of view here . while the 2d genus statistic is defined by the number of contour lines in 2d surface , the 3d genus statistic @xcite is defined by the number of contour surfaces and the number of handles in 3d space . thus the 3d genus is defined only for cosmic fields of @xmath148 . the 3d genus statistic @xmath149 is defined by @xmath150 , \label{eq2 - 15}\end{aligned}\ ] ] per unit volume of the 3d space . this quantity is mathematically equivalent to @xmath151 times euler characteristic of the contour surfaces , and thus is proportional to the total surface integral of local curvature of contours from the gauss - bonnet theorem . although those definition is intuitive , an alternative , equivalent definition is more useful as in the 2d genus case . we set an arbitrary direction in the 3d space . then there are maxima , minima and saddle points according to that direction . from the number of these points , the 3d genus is defined by @xmath152 , \label{eq2 - 16}\end{aligned}\ ] ] per unit volume . according to this definition , the explicit expression of the 3d genus statistic is given by @xcite @xmath153 we need to calculate the following quantity : @xmath154 after applying the transform of equation ( [ eq1 - 13 - 3 - 1 ] ) , the gaussian integration of @xmath64 fixing @xmath18 is given by @xmath155 , \label{eq2 - 19}\end{aligned}\ ] ] where @xmath156 is defined in table [ tab1 ] . ccc|ccc @xmath157 & @xmath158 & @xmath159 & @xmath160 & @xmath161 & @xmath162 + @xmath119 & @xmath119 & @xmath119 & @xmath120 & @xmath119 & @xmath163 + @xmath120 & @xmath119 & @xmath119 & @xmath119 & @xmath120 & @xmath163 + @xmath119 & @xmath120 & @xmath119 & @xmath119 & @xmath120 & @xmath163 + @xmath120 & @xmath120 & @xmath119 & @xmath119 & @xmath119 & @xmath164 + @xmath119 & @xmath119 & @xmath137 & @xmath119 & @xmath119 & @xmath165 + then , from equations ( [ eqa1 - 4a ] ) , ( [ eqa1 - 5a ] ) and ( [ eqa1 - 5b ] ) , the equation ( [ eq2 - 18 ] ) reduces to @xmath166 . \label{eq2 - 20}\end{aligned}\ ] ] since the 3d genus is only defined for @xmath167 and our universe has the spatial dimension 3 , only @xmath148 is meaningful for actual cosmic fields . nevertheless , we preserve the general dimension @xmath6 for some flavor of generality . thus , from equation ( [ eq1 - 17 ] ) , @xmath168 . \label{eq2 - 21}\end{aligned}\ ] ] the gaussian contribution is given by @xcite @xmath169 thus , the perturbative expansion of equation ( [ eq1 - 13 ] ) up to second order is finally given by @xmath170 \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\}. \label{eq2 - 23}\end{aligned}\ ] ] the above equation with @xmath148 agrees with @xcite , @xmath171 , and @xmath172 in @xcite are related to the notations here by @xmath173 , @xmath174 , @xmath175 . ] next , we consider the weighted extrema density above a threshold @xmath133 . the field extrema is defined to be points where all the first - order spatial derivatives of the field @xmath9 vanish : @xmath176 . the weight @xmath177 is associated to each extrema according to the number of negative eigenvalues of spatial second - order derivatives of the field . when the threshold is high enough , the weighted extrema are approximately identified with the field peaks . mathematically , this statistic is equivalent to the 2d genus statistics , according to the morse s theorem @xcite . therefore , we do not need to calculate separately to obtain the result on this statistic . however , we present the derivation of the extrema density as an alternative calculation to the 2d genus . the weighted extrema in 2d field is given by @xmath178 following the similar calculation of previous examples , and using equations in appendix [ app1 ] , we obtain @xmath179 , \label{eq2 - 25}\end{aligned}\ ] ] and the perturbative expansion of equation ( [ eq1 - 13 ] ) up to second order is finally given by @xmath180 \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\}. \label{eq2 - 26}\end{aligned}\ ] ] this result is in fact equivalent to the 2d genus statistics given by equation ( [ eq2 - 14 ] ) . most of the minkowski functionals are closely related to statistics discussed above . in this subsection , we comprehensively describe the exact relation between the minkowski functionals @xmath181 of a smoothed field and the statistical quantities considered above . the minkowski functional of @xmath182 is simply the volume fraction of the excursion set @xmath183 which is defined by high - density regions above a given threshold @xmath133 : @xmath184 the other functionals with @xmath185 are defined by the integral of the curvatures on isodensity surfaces of the threshold @xmath133 @xcite . in 3-dimensions , @xmath148 , they are evaluated by a surface integration averaged over whole system of volume @xmath186 @xcite , i.e. , @xmath187 where the local minkowski functionals , @xmath188 are defined by the principal curvatures @xmath189 and @xmath190 of the surface oriented toward lower density values . in 2-dimensions , @xmath70 , the minkowski functionals of @xmath191 are evaluated by a line integration averaged over whole system of 2d volume ( surface ) @xmath186 @xcite , i.e. , @xmath192 where the local minkowski functionals , @xmath193 are defined by the principal curvature @xmath189 of the line oriented toward lower density values . all the minkowski functionals for a gaussian random field are analytically derived by @xcite : @xmath194 where the factor @xmath195 is the volume of the unit ball in @xmath121 dimensions , so that @xmath196 , @xmath197 , @xmath198 , @xmath199 @xcite . it turns out that the minkowski functionals in 2- and 3-dimensions are identical to the statistics @xmath105 , @xmath130 and @xmath149 in each dimensions , except for normalization factors . in fact , from the crofton s formula @xcite , the @xmath121-th minkowski functional is given by @xmath200 in this formula for body @xmath183 in @xmath6 dimensions , we consider an arbitrary @xmath121-dimensional hypersurface @xmath201 and calculate the euler characteristic @xmath202 of the intersection @xmath203 in @xmath121 dimensions . this quantity is integrated over the space @xmath204 of all conceivable hypersurfaces . the integration measure @xmath205 is normalized to give @xmath206 . from this formula , we can see the statistics @xmath149 , @xmath130 , @xmath105 are proportional to the minkowski functionals of @xmath207 , @xmath208 , @xmath209 . in fact , @xmath210 is given by @xmath211 times the 3d genus ( or , equivalently , @xmath212 times the euler number of boundaries , @xmath213 ) , @xmath214 is identical to the 2d genus , and @xmath215 is just @xmath212 times the number of level - crossing points @xcite . thus , minkowski functionals are given by @xmath216 where the boundary of the body @xmath183 is identified with the isodensity contours of threshold @xmath133 and the statistics on right hand sides are defined in @xmath6-dimensions . therefore , we have already obtained the weakly non - gaussian expressions for minkowski functionals , i.e. , the minkowski functionals of @xmath217 are given by the above equations and equations ( [ eq2 - 7 ] ) , ( [ eq2 - 14 ] ) , and ( [ eq2 - 23 ] ) . the gaussian parts of the above equations exactly reproduce the tomita s formula ( [ eq2 - 54 ] ) . the remaining minkowski functional is the volume functional @xmath218 in this case , from equation ( [ eqa1 - 4a ] ) in appendix [ app1 ] , @xmath219 so that equation ( [ eq1 - 17 ] ) reduces to @xmath220 the equivalent form can be obtained by integrating the edgeworth expansion of pdf , equation ( [ eq2 - 0 - 2 ] ) . all the formulas of minkowski functionals derived above for @xmath221 fit into a single expression : @xmath222 \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\}. \label{eq2 - 60}\end{aligned}\ ] ] the density threshold @xmath133 so far is simply defined so that isodensity surface is identified by @xmath223 . however , the horizontal shift of the nonlinear genus curve etc . is considerably attributed to the nonlinear shift of probability distribution of the density field ( e.g. , * ? ? ? * ; * ? ? ? in order to cancel the latter shift , the threshold @xmath224 is defined so that the volume fraction @xmath225 on the high - density side of the isodensity surface equals to @xmath226 in fact , most of the work on genus analysis uses the genus curve plotted against the volume - fraction threshold @xmath224 . recently , @xcite re - expressed the weakly non - gaussian formula of genus curve @xcite in terms of @xmath224 , using perturbative expansion of the probability distribution function of the density field @xcite . we follow this method to re - express the weakly non - gaussian formulas of various statistical quantities derived above in terms of the volume - fraction threshold @xmath224 . the relation of @xmath133 and @xmath224 of weakly non - gaussian field is simply given by equating the two equations ( [ eq2 - 104 ] ) and ( [ eq2 - 105 ] ) . up to first order in @xmath49 , the relation reduces to @xmath227 it is straightforward to rewrite the various analytical formulas we derived above . the results for level - crossing statistic , 2d and 3d genus are respectively given by @xmath228 \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\ } , \label{eq2 - 207a}\\ g_2({{\widetilde{\nu } } } ) & = & \rho_{\rm e2}({{\widetilde{\nu } } } ) = \frac{1}{(2\pi)^{3/2 } } \left(\frac{\sigma_1}{\sqrt{d}\sigma_0}\right)^2 e^{-{{\widetilde{\nu}}}^2/2 } \nonumber\\ & & \qquad\qquad\quad\times\ , \left\ { h_1({{\widetilde{\nu } } } ) + \left [ \frac23 \left(s^{(1 ) } - s^{(0)}\right ) h_2({{\widetilde{\nu } } } ) + \frac13 \left(s^{(2 ) } - s^{(0)}\right ) \right ] \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\ } , \label{eq2 - 207b}\\ g_3({{\widetilde{\nu } } } ) & = & - \frac{1}{(2\pi)^2 } \left(\frac{\sigma_1}{\sqrt{d}\sigma_0}\right)^3 e^{-{{\widetilde{\nu}}}^2/2 } \nonumber\\ & & \qquad\qquad\times\ , \left\ { h_2({{\widetilde{\nu } } } ) + \left [ \left(s^{(1 ) } - s^{(0)}\right ) h_3({{\widetilde{\nu } } } ) + \left(s^{(2 ) } - s^{(0)}\right ) h_1({{\widetilde{\nu } } } ) \right ] \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\ } , \label{eq2 - 207c}\end{aligned}\ ] ] the results for minkowski functionals of @xmath217 are again given by these equations and equations ( [ eq2 - 101a])([eq2 - 101c ] ) . all the minkowski functionals for @xmath229 fit into a single expression : @xmath230 \sigma_0 + { \cal o}({\sigma_0}^2 ) \right\}. \label{eq2 - 208}\end{aligned}\ ] ] remarkably , the highest - order hermite polynomial in the non - gaussian correction terms vanishes in each case . in addition , the skewness parameters only appear as combinations of the form , @xmath231 , which makes the result simpler compared with the original form with the direct threshold @xmath133 . as we will see in the following sections , the numerical values of @xmath232 are quite close , or even identical in some special models . this means that the non - gaussian corrections of the above statistics are smaller with the rescaled threshold @xmath224 than with the original threshold @xmath133 . this is one of the central results in this paper . this tendency is in agreement with the analyses of numerical simulations . we need to know the skewness parameters , @xmath87 for the evaluation of the second - order perturbative terms of equation ( [ eq1 - 17 ] ) . these quantities can be calculated by usual perturbation theory , once we specify the cosmic field , @xmath9 . in each example of the previous section , the quantity @xmath233 does not appear so that we evaluate @xmath88 , @xmath129 , and @xmath234 in this section . the other kinds of skewness parameters like @xmath233 can be similarly evaluated without difficulty . before we explore actual cosmic fields , we consider a simple , phenomenological statistical model , i.e. , the hierarchical model of higher order correlation functions ( e.g. , * ? ? ? * ) . in this model , @xmath26-point correlation function is a sum of @xmath235 products of the two - point correlation function . specifically , the three - point correlation function is given by @xmath236 . \label{eq4 - 1}\end{aligned}\ ] ] we assume the field @xmath9 in the above equation is already smoothed . in this case , skewness parameters , equations ( [ eq1 - 15a])([eq1 - 15d ] ) are given by straightforward calculations @xcite : @xmath237 these values depend on a hierarchical amplitude , @xmath238 , which is a free parameter of this model . the relative amplitudes among @xmath87 are not freely adjusted in the above equation . in the case of volume - fraction threshold , the first non - gaussian correction of the various statistics considered in the previous section is absent , since they depend only on @xmath239 . next , we consider the three - dimensional density field . the skewness parameters of this field in perturbation theory are already calculated by @xcite and @xcite , using fourier transforms of the field . we comprehensively review this calculation here for completeness . there is an alternative way to calculate the skewness not depending on fourier transforms @xcite . the cosmic field @xmath9 is identified with the 3d density contrast , @xmath240 , where @xmath241 is the density field . the dimension this field is defined in is three , @xmath148 . the fourier transform of the field is useful in the following : @xmath242 in this notation , two- and three - point correlations in fourier space have the forms , @xmath243 the above forms are the consequence of the statistical homogeneity of the space , where the functions @xmath244 and @xmath245 are the power spectrum and the bispectrum , respectively . thus , the variance and its variants [ eqs . ( [ eq1 - 2 ] ) , ( [ eq1 - 13 - 3a ] ) , and ( [ eq1 - 13 - 3b ] ) ] , in their fourier representation , are given by @xmath246 and the skewness parameters of equations ( [ eq1 - 15a])([eq1 - 15d ] ) are given by @xmath247 for initial random gaussian density field , the second order perturbation theory predicts the power spectrum and the bispectrum as follows ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) : @xmath248 { p_{\rm lin}}(k_1 ) { p_{\rm lin}}(k_2 ) w(k_1r ) w(k_2r ) w(k_3r ) \nonumber\\ & & \qquad\qquad\qquad\quad + \ , { \rm cyc.(1,2,3 ) } + { \cal o}({\sigma_0}^6 ) , \label{eq4 - 7b}\end{aligned}\ ] ] where @xmath249 is the linear power spectrum , and @xmath201 is a weak function of cosmology @xcite . the field smoothing corresponds to the multiplication of the window function @xmath250 which is the three - dimensional fourier transform of the smoothing function @xmath10 . it is a good approximation to use the value of @xmath201 for an einstein - de sitter universe , @xmath251 , in most cases . the explicit form of the function @xmath201 in terms of cosmological parameters @xmath252 and @xmath253 is given by @xcite , which is accurately fitted by @xmath254 the perturbation theory is considered to be an expansion by a parameter @xmath49 . in this respect , the power spectrum @xmath244 and the bispectrum @xmath245 is of order @xmath16 and @xmath255 , respectively . substituting equation ( [ eq4 - 7a ] ) into equation ( [ eq4 - 5 ] ) , we obtain @xmath256 up to the lowest order . similarly , substituting equation ( [ eq4 - 7b ] ) into equations ( [ eq4 - 6a])([eq4 - 6c ] ) , and introducing new integration variables , @xmath257 , @xmath258 , and @xmath259 , we obtain @xmath260 where @xmath261 and @xmath262 \frac{w\left(\sqrt{{l_1}^2 + { l_2}^2 + 2 l_1 l_2 \mu } \right ) } { w(l_1 ) w(l_2 ) } \nonumber\\ & & \qquad\qquad\qquad \times\ , \left\ { \begin{array}{ll } \displaystyle 2 , & a=0,\\ { l_1}^2 + { l_2}^2 + l_1 l_2 \mu , & a=1,\\ 3\ , { l_1}^2 { l_2}^2 \left ( 1 - \mu^2 \right ) , & a=2 , \end{array } \right . \label{eq4 - 9}\end{aligned}\ ] ] so far the smoothing function is arbitrary . for a general smoothing function , the above equations can be numerically integrated to obtain the skewness parameters for each model of the power spectrum . for some smoothing functions , further analytical reductions of the above equations are possible . as a popular example , we consider the gaussian smoothing , @xmath263 , which is frequently adopted for practical purposes . we follow the similar technique of @xcite , in which they derived the skewness of the density field with gaussian smoothing . for the gaussian smoothing , @xmath264 in this case , the following formula of the modified bessel function @xmath265 is useful : @xmath266 where @xmath267 is the @xmath122-th legendre polynomial . from this formula , the angular integration of @xmath268 in equation ( [ eq4 - 9 ] ) can be analytically performed and the result is @xmath269 , \label{eq4 - 12a}\\ & & \widetilde{s}^{(1 ) } = \sqrt{2\pi l_1l_2 } \left\ { \frac{5 + 2e}{4 } \left(\frac{l_2}{l_1 } + \frac{l_1}{l_2}\right ) i_{1/2}(l_1l_2 ) - \left [ \frac{3(9 + e)}{10 } + \frac{{l_2}^2}{{l_1}^2 } + \frac{{l_1}^2}{{l_2}^2 } \right ] i_{3/2}(l_1l_2 ) \right . \nonumber\\ & & \qquad\qquad\qquad\qquad\qquad\qquad + \ , \left . \frac{2 - e}{2 } \left(\frac{l_2}{l_1 } + \frac{l_1}{l_2}\right ) i_{5/2}(l_1l_2 ) - \frac{3(1 - e)}{10 } i_{7/2}(l_1 l_2 ) \right\ } , \label{eq4 - 12b}\\ & & \widetilde{s}^{(2 ) } = \sqrt{2\pi}(l_1l_2)^{3/2 } \left [ \frac{3(3 + 2e)}{5 } i_{1/2}(l_1l_2 ) - \frac{9}{10 } \left(\frac{l_2}{l_1 } + \frac{l_1}{l_2}\right ) i_{3/2}(l_1l_2 ) \right . \nonumber\\ & & \qquad\qquad\qquad\qquad\qquad -\ , \left . \frac{3(3 + 4e)}{7 } i_{5/2}(l_1l_2 ) + \frac{9}{10 } \left(\frac{l_2}{l_1 } + \frac{l_1}{l_2}\right ) i_{7/2}(l_1l_2 ) - \frac{18(1 - e)}{35 } i_{9/2}(l_1l_2 ) \right ] . \label{eq4 - 12c}\end{aligned}\ ] ] at this point , it is useful to define the following quantity : @xmath270 in the above equation , variance parameters @xmath49 , @xmath271 are given by equation ( [ eq4 - 7 - 1 ] ) . the nonlinear correction for @xmath272 is not needed because our estimate of @xmath87 is only lowest order in @xmath49 . when the higher order corrections , e.g. , third - order perturbation corrections , are estimated , one should be sure that all the necessary nonlinear corrections are properly taken into account . with this quantity , the skewness parameters are given by @xmath273 , \label{eq4 - 14b}\\ & & s^{(2)}(r ) = 9 \left [ \frac{3 + 2e}{15 } s^{33}_0 - \frac15 s^{24}_1 - \frac{3 + 4e}{21 } s^{33}_2 + \frac15 s^{24}_3 - \frac{2(1-e)}{35 } s^{33}_4 \right ] . \label{eq4 - 14c}\end{aligned}\ ] ] thus the lowest order estimates of skewness parameters are given by the above equations ( [ eq4 - 13])([eq4 - 14c ] ) . for each given power spectrum , the integration of equation ( [ eq4 - 13 ] ) is straightforward . the resulting skewness parameters are independent on the amplitude of the power spectrum . when the power spectrum is given by a cdm - like model , @xmath274 , where @xmath275^{-1/4 } , \label{eq4 - 15}\end{aligned}\ ] ] is the cdm - like transfer function fitted by @xcite , and @xmath276 is the shape parameter of this model , then the skewness parameters are functions of @xmath277 . in table [ tab2 ] , we give the values of skewness parameters @xmath87 for cdm - like models for several values of @xmath277 . c|cccccc @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282 & @xmath283 + @xmath88 & @xmath284 & @xmath285 & @xmath286 & @xmath287 & @xmath288 & @xmath289 + @xmath129 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath294 & @xmath295 + @xmath234 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath300 & @xmath301 + in this table , the value of @xmath201 is approximated by @xmath302 . since the skewness parameters are weak functions of @xmath277 as seen from the table , one can interpolate the values in this table to obtain the values of arbitrary scales for practical purposes . when the power spectrum is given by a power - law form , @xmath303 the integration of equation ( [ eq4 - 13 ] ) can be analytically performed . first , the simple gaussian integration gives @xmath304 second , we expand the modified bessel function as @xmath305 then the equation ( [ eq4 - 13 ] ) reduces to @xmath306 where @xmath307 , and @xmath55 is the gauss hypergeometric function : @xmath308 the equations ( [ eq4 - 14a])([eq4 - 14c ] ) and equation ( [ eq4 - 19 ] ) give the skewness parameters @xmath87 . the following recursion relations for hypergeometric function @xmath309 , \label{eq4 - 21b}\\ & & ( \gamma-\alpha)(\gamma-\beta)z f(\alpha,\beta,\gamma+1;z ) + \gamma[(\alpha+\beta-2\gamma+1)z+\gamma-1 ] f(\alpha,\beta,\gamma;z ) \nonumber\\ & & \qquad\qquad\qquad \ , + \gamma(\gamma-1)(z-1 ) f(\alpha,\beta,\gamma-1;z ) = 0 . \label{eq4 - 21c}\end{aligned}\ ] ] simplify the result : @xmath310 in this power - law case , skewness parameters do not depend on scales @xmath2 but only on power - law index , @xmath311 . incidentally , @xmath312 and @xmath313 are identical . this is just the coincidence and is not generally the case when the power spectrum is not given by power - law . several numerical values are shown in table [ tab3 ] , where @xmath251 is assumed . c|ccccccccc @xmath314 & @xmath315 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath278 + @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath330 & @xmath331 & @xmath332 + @xmath234 & @xmath333 & @xmath334 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 + next , we consider the 3d velocity field as the cosmic field @xmath9 . since the rotational components of the velocity field are decaying modes of gravitational evolution in perturbation theory , we only consider the rotation - free component . the cosmic field @xmath9 is identified with the dimensionless scalar field , @xmath342 where @xmath343 is the hubble parameter . one can also consider other quantities like radial component of the velocity field , @xmath344 , where @xmath345 is the line - of - sight normal vector . those quantities are more complicated than a simple divergence . we illustrate only the simplest case in this paper . the second order perturbation theory predicts power spectrum and bispectrum of the velocity field as follows ( e.g. , * ? ? ? * ) : @xmath346 { p_{\rm lin}}(k_1 ) { p_{\rm lin}}(k_2 ) w(k_1r ) w(k_2r ) w(k_3r ) \nonumber\\ & & \qquad\qquad\qquad\quad + \ , { \rm cyc.(1,2,3 ) } + { \cal o}({\sigma_0}^6 ) . \label{eq4 - 24b}\end{aligned}\ ] ] in the above equation , the factor @xmath347 is a logarithmic derivative of the growth factor , given by @xmath348 @xcite and @xmath349 is the linear growth factor and @xmath350 is the expansion factor . the logarithmic derivative is evaluated at present . the factor @xmath351 is a weak function of cosmology . it is a good approximation to use the value for an einstein - de sitter universe , @xmath352 . the explicit form of the function @xmath351 in terms of @xmath353 and @xmath354 is given by @xcite , which is accurately fitted by @xmath355 the similarity of the equations ( [ eq4 - 24a ] ) and ( [ eq4 - 24b ] ) for velocity field to the equations ( [ eq4 - 7a ] ) , ( [ eq4 - 7b ] ) for density field is obvious . we can easily see the skewness parameters for the velocity field are obtained by similar equations as ( [ eq4 - 14a])([eq4 - 14c ] ) : @xmath356 , \label{eq4 - 26a}\\ & & s^{(1)}(r ) = - \frac{3}{2 { g_f } } \left [ \frac{5 + 2e_v}{3 } s^{13}_0 - \frac{9 + e_v}{5 } s^{22}_1 - s^{04}_1 + \frac{2(2-e_v)}{3 } s^{13}_2 - \frac{1-e_v}{5 } s^{22}_3 \right ] , \label{eq4 - 26b}\\ & & s^{(2)}(r ) = - \frac{9}{{g_f } } \left [ \frac{3 + 2e_v}{15 } s^{33}_0 - \frac15 s^{24}_1 - \frac{3 + 4e_v}{21 } s^{33}_2 + \frac15 s^{24}_3 - \frac{2(1-e_v)}{35 } s^{33}_4 \right ] , \label{eq4 - 26c}\end{aligned}\ ] ] where @xmath357 is given by equation ( [ eq4 - 13 ] ) without any modification . in table [ tab4 ] , we show the values of skewness parameters of the velocity field for the cdm - like models for several values of @xmath277 . c|cccccccc @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282 & @xmath283 + @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath362 & @xmath363 & @xmath364 + @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & @xmath370 & @xmath371 + @xmath372 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath377 & @xmath378 + in this table , the value of @xmath351 is set as @xmath379 . for the power spectrum of the power - law form , @xmath380 , \label{eq4 - 27a}\\ & & s^{(2)}({{n_{\rm s } } } ) = -\frac{3}{{g_f } } \left [ f\left(\frac{{{n_{\rm s}}}+5}{2},\frac{{{n_{\rm s}}}+5}{2},\frac52;\frac14\right ) - \frac15 ( { { n_{\rm s}}}+ 4 - 4e_v ) f\left(\frac{{{n_{\rm s}}}+5}{2},\frac{{{n_{\rm s}}}+5}{2},\frac72;\frac14\right ) \right ] . \label{eq4 - 28c}\end{aligned}\ ] ] several numerical values are shown in table [ tab5 ] , where @xmath381 is assumed . c|ccccccccc @xmath314 & @xmath315 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath278 + @xmath382 & @xmath327 & @xmath383 & @xmath384 & @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath389 & @xmath390 + @xmath372 & @xmath391 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath396 & @xmath397 & @xmath398 & @xmath399 + the projection of the density field @xmath400 defines the 2d cosmic fields on the sky ( @xmath70 ) . here , we derive the skewness parameters for this field . the 2d projected density field with top - hat kernel is investigated by @xcite . here , we are interested in gaussian kernel for our purpose . in a friedman - lematre universe , the comoving angular diameter distance at a comoving distance @xmath401 is given by @xmath402 depending on the sign of the spatial curvature @xmath403 . thus , in spherical coordinates , projected density field in 2d is given by @xmath404 where @xmath405 is the selection function without volume factor , normalized as @xmath406 , and @xmath407 is the 3d comoving density field constant . ] . the present value of the conformal time @xmath408 is @xmath409 . the projected density contrast is defined by @xmath410 , where one can see @xmath411 . we identify the projected density contrast with the 2d ( @xmath70 ) field @xmath9 . since the smoothing angle @xmath412 is much smaller than @xmath413 in most of the interested cases , we consider the small patch of the sky of the vicinity of the polar axis , @xmath414 . with this approximation , we introduce the variables @xmath415 , and @xmath416 , which are considered as 2d euclidean coordinates , @xmath417 . therefore , the projection equation is given by @xmath418 where @xmath419 is the density contrast at comoving coordinates @xmath8 and conformal time @xmath420 . the power spectrum and the bispectrum for the above projected field are given by the limber s equations ( [ eqa2 - 2 ] ) and ( [ eqa2 - 7 ] ) in appendix [ app2 ] , with @xmath421 : @xmath422 where @xmath244 , @xmath245 are 2d projected power spectrum and bispectrum , respectively , of the field @xmath9 , and @xmath423 , @xmath424 are 3d power spectrum and bispectrum , respectively . the 3d power spectrum and the 3d bispectrum are evaluated by the second order perturbation theory . they are similar to equations ( [ eq4 - 7a ] ) and ( [ eq4 - 7b ] ) , but we have to take into account the time - dependence here . they are given by @xmath425 { p_{\rm lin}}(k_1 ) { p_{\rm lin}}(k_2 ) \nonumber\\ & & \qquad\qquad\qquad\qquad\qquad\quad + \ , { \rm cyc.(1,2,3 ) } , \label{eq4 - 34b}\end{aligned}\ ] ] where @xmath426 is the linear growth factor at conformal lookback time @xmath401 , ( i.e. , at conformal time @xmath427 ) , which is normalized as @xmath428 . the following fitting formula @xcite is useful : @xmath429 where @xmath353 and @xmath354 are time - dependent cosmological parameters at conformal lookback time @xmath401 . the variable @xmath430 is a weak function of time and cosmology , and for einstein - de sitter universe , @xmath251 . this quantity @xmath201 is the same we used in 3d density field , but here we also take into account the time - dependence . it is accurately approximated by @xmath431 the variance parameters of the smoohted projected field are given by @xmath432 , \label{eq4 - 36}\end{aligned}\ ] ] where @xmath433 the skewness parameters of the smoothed projected field are given by @xmath434 { \sigma_1}^{2a}\left[s_k(\chi){{\theta_{\rm f}}}\right ] c^{(a)}\left[s_k(\chi){{\theta_{\rm f}}}\right ] , \label{eq4 - 38}\end{aligned}\ ] ] where @xmath435 and @xmath436 \frac{w\left(\sqrt{l_1^{\,2 } + l_2^{\,2 } + 2 l_1 l_2 \mu } \right ) } { w(l_1 ) w(l_2 ) } \nonumber\\ & & \qquad\qquad\quad \times\ , \left\ { \begin{array}{ll } \displaystyle 2 , & a=0,\\ { l_1}^2 + { l_2}^2 + l_1 l_2 \mu , & a=1,\\ 4\ , { l_1}^2 { l_2}^2 \left ( 1 - \mu^2 \right ) , & a=2 , \end{array } \right . \label{eq4 - 40}\end{aligned}\ ] ] so far the smoothing function is arbitrary . for a general smoothing function , the above equations can be numerically integrated to obtain the skewness parameters for each model of the power spectrum . in the following , we adopt the gaussian smoothing function , @xmath437 . for this smoothing function , the equation ( [ eq4 - 10 ] ) holds even for this 2d case . in this case , the following integral representation of the modified bessel function @xmath265 for @xmath438 is useful : @xmath439 actually , the derivatives of the above equation , @xmath440 are sufficient to perform the angular integration in equation ( [ eq4 - 40 ] ) . moreover , one can use the property of the bessel function , @xmath441 , and @xmath442 to obtain formulas , @xmath443 from these formulas , equation ( [ eq4 - 40 ] ) reduce to @xmath444 , \label{eq4 - 45a}\\ & & \widetilde{c}^{(1 ) } = \frac34 \left\ { ( 4+e ) ( l_1^{\,2 } + l_2^{\,2 } ) i_0(l_1 l_2 ) - \left [ \frac{15 + e}{2 } l_1 l_2 + 2 \left ( \frac{l_1^{\,3}}{l_2 } + \frac{l_2^{\,3}}{l_1 } \right ) \right ] i_1(l_1 l_2 ) \right . \nonumber\\ & & \qquad\qquad\qquad \left . + \ ; ( 2-e ) ( l_1^{\,2 } + l_2^{\,2 } ) i_2(l_1 l_2 ) - \frac{1 - e}{2 } l_1 l_2 i_3(l_1 l_2 ) \right\ } , \label{eq4 - 45b}\\ & & \widetilde{c}^{(2 ) } = \frac34 \left [ ( 5 + 3e ) l_1^{\,2 } l_2^{\,2 } i_0(l_1 l_2 ) - 2 l_1 l_2(l_1^{\,2 } + l_2^{\,2 } ) i_1(l_1 l_2 ) - 4(1+e ) l_1^{\,2 } l_2^{\,2 } i_2(l_1 l_2 ) \right . \nonumber\\ & & \qquad\qquad \left . + \ ; 2 l_1 l_2 ( l_1^{\,2 } + l_2^{\,2 } ) i_3(l_1 l_2 ) - ( 1-e ) l_1^{\,2 } l_2^{\,2 } i_4(l_1 l_2 ) \right ] . \label{eq4 - 45c}\end{aligned}\ ] ] at this point , defining @xmath445 the equation ( [ eq4 - 39 ] ) reduces to @xmath446 , \label{eq4 - 47a}\\ & & c^{(1 ) } = \frac34 \left [ 2(4+e ) c^{13}_0 - \frac{15 + e}{2 } c^{22}_1 - 4 c^{04}_1 + 2(2-e ) c^{13}_2 - \frac{1 - e}{2 } c^{22}_3 \right ] , \label{eq4 - 47b}\\ & & c^{(2 ) } = \frac34 \left [ ( 5 + 3e ) c^{33}_0 - 4 c^{24}_1 - 4(1 + e ) c^{33}_2 + 4c^{24}_3 - ( 1 - e ) c^{33}_4 \right ] . \label{eq4 - 47c}\end{aligned}\ ] ] the two - dimensional integration of equation ( [ eq4 - 46 ] ) is performed only once as a function of @xmath2 . the result is stored as a table , and is used in one - dimensional numerical integration of equation ( [ eq4 - 38 ] ) for finally obtaining the skewness parameters in 2d projected density fields . functions @xmath447 for cdm - like models are given in table [ tab6 ] which are functions of @xmath448 . c|cccccccccccc @xmath277 & @xmath449 & @xmath322 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282 & @xmath283 & @xmath450 + @xmath451 & @xmath452 & @xmath453 & @xmath454 & @xmath455 & @xmath456 & @xmath457 & @xmath458 & @xmath459 & @xmath460 + @xmath461 & @xmath462 & @xmath463 & @xmath464 & @xmath465 & @xmath466 & @xmath467 & @xmath468 & @xmath469 & @xmath470 + @xmath471 & @xmath472 & @xmath473 & @xmath474 & @xmath475 & @xmath476 & @xmath477 & @xmath478 & @xmath479 & @xmath480 + when the 3d power spectrum is given by a power - law form of equation ( [ eq4 - 16 ] ) , the integration by wave length @xmath139 , @xmath140 can be analytically performed as in the 3d case . in fact , the parameter @xmath481 of equation ( [ eq4 - 37 ] ) with gaussian smoothing @xmath482 is given by @xmath483 and the variance parameters are given by @xmath484 with similar technique used in 3d density field , @xmath485 the equations ( [ eq4 - 47a])([eq4 - 47c ] ) and the above equation finally give the values of @xmath447 . the recursion relations of equations ( [ eq4 - 21a ] ) and ( [ eq4 - 21b ] ) simplify the result : @xmath486 these functions are independent on scale @xmath2 in the power - law case , but dependent on power - law index , @xmath311 . again , @xmath487 and @xmath488 are identical only in the power - law case . numerical values are given in table [ tab7 ] , where @xmath489 is assumed . c|ccccccccc @xmath311 & @xmath315 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath278 + @xmath490 & @xmath491 & @xmath492 & @xmath493 & @xmath494 & @xmath495 & @xmath496 & @xmath497 & @xmath498 & @xmath499 + @xmath471 & @xmath500 & @xmath501 & @xmath502 & @xmath503 & @xmath504 & @xmath505 & @xmath506 & @xmath475 & @xmath507 + for the power - law case , the skewness parameters of equation ( [ eq4 - 38 ] ) reduces to the following simple form : @xmath508^{2 - 2{{n_{\rm s } } } } n^3(\chi ) d^4(\chi ) } { \displaystyle \left\ { \int d\chi \left[s_k(\chi)\right]^{2-{{n_{\rm s } } } } n^2(\chi ) d^2(\chi ) \right\}^2}\ , c^{(a)}({{n_{\rm s } } } ) . \label{4 - 50 - 2}\end{aligned}\ ] ] it is interesting to compare the results with those of top - hat smoothing . according to @xcite , the top - hat smoothing gives @xmath509 for power - law case . comparing this expression with our table [ tab7 ] , they roughly agree with each other , and have similar behavior with spectral index . in detail , the gaussian smoothing gives more or less larger values . for example , top - hat smoothing gives @xmath510 for @xmath511 , respectively , while gaussian smoothing gives @xmath512 , respectively . it is reasonable that the @xmath451 of top - hat smoothing differs from that of gaussian smoothing for large spectral index , because the top - hat smoothing gathers more power on small scales than gaussian smoothing . the local convergence field of the weak lensing is commonly used for studying the large - scale structure of the universe ( e.g. , * ? ? ? * ; * ? ? ? assuming the situation where limber s equation ( see appendix [ app2 ] ) and also the born approximation @xcite hold , the following correspondence between the convergence field @xmath513 and the 3d density contrast @xmath514 is useful ( e.g. , * ? ? ? * ) : @xmath515 , \label{eq4 - 60}\end{aligned}\ ] ] where @xmath516 is the scale factor at conformal lookback time @xmath517 . the above equation is reduced to exactly the same form as the projected field of equation ( [ eq4 - 31 ] ) , but with the substitution , @xmath518 however , one should note that this substitution is valid only under the assumption of the born approximation . although the effect of the born approximation on the skewness is known to be weak @xcite , the validity of the born approximation in general situation has not been tested in detail @xcite . there is some subtlety which could arise when various combinations of skewness are considered . the above expressions of the skewness parameters are for unbiased fields . the skewness parameters for biased fields are non - trivial . they depend on the details of the biasing scheme in the real universe which is poorly known so far . however , the skewness parameters are well - defined quantities , so that they are calculated from the first principle once the biasing scheme is given . perhaps , one of the simplest , yet non - trivial case is the local , deterministic biasing . in this case , we can follow @xcite to obtain the perturbative expansion of the biasing : @xmath519 where @xmath520 and @xmath97 are the galaxy and mass density contrast , respectively . the spatial dimension is arbitrary , so that @xmath520 , @xmath97 can be functions in either 1d , 2d , or 3d space . the biased variance in lowest order is given by @xmath521 it is also straightforward to calculate skewness parameters . after some algebra , all skewness parameters are shown to transform in a same way : @xmath522 irrespective of spatial dimensions , @xmath6 , and of kinds of skewness parameters , @xmath523 . in this framework , the parameters @xmath524 and @xmath525 are needed . a possible way to determine these parameters is to measure the variance @xmath526 and the skewness @xmath527 from the observation . theoretical models predict the variance @xmath528 and the skewness @xmath529 of the mass distribution . we obtain biasing parameters by @xmath530 and @xmath531 . the derivatives of the skewness , @xmath532 , @xmath533 , are then obtained from equation ( [ eq4 - 71 ] ) . in case of level - crossing , and genus statistics as functions of scaled threshold @xmath224 , the second - order corrections in equation ( [ eq2 - 207a])([eq2 - 207c ] ) only depends on the difference of the skewness parameters times the variace . quite remarkably , this type of second - order correction term does not depend on bias parameter at all : @xmath534 \sigma_{\rm 0,g } = \left[s^{(a ) } - s^{(0)}\right ] \sigma_{\rm 0 } \label{eq4 - 72}\end{aligned}\ ] ] thus , the second - order nonlinear corrections for level - crossing and genus statistics of locally biased field are exactly the same as that of unbiased mass density field . this can be considered as an advantage of the scaled threshold @xmath224 in these statistics , because the biasing is one of the most embarrasing uncertainty in the analysis of galaxy distribution . however , one should note this result is derived under the local biasing scheme . so far the locality of the bias is not guaranteed in general . stochastic argument @xcite of the biasing is more complicated for derivative skewness @xmath129 and @xmath234 than for usual skewness @xmath88 , because they involve the correlation between field derivatives . the phenomenological nature of the stochastic biasing requires many parameters which is not calculable from first principles . therefore , the stochastic biasing sheme does not effectively work for our problem . more physical treatment of the biasing schemes of galaxy formation is needed , in which case the nonlocality of the bias could also be important @xcite . the perturbative calculations offer valuable aspects of weakly non - linear evolution of the various statistics without laborious parameter survey by numerical simulations . in this paper , we obtain the lowest nonlinear corrections to relatively popular statistics of smoothed cosmic fields . using these results , it is interesting to see how the weakly nonlinear effect tends to distort the gaussian prediction for those exemplified statistics . in figure [ fig1 ] , various statistics for 3d density field is shown . the amplitude of each statistic is appropriately normalized as we are interested in the deviation from the gaussian prediction . if we neglect the normalization , minkowski functionals of @xmath217 are equivalent to the statistics @xmath105 , @xmath130 , and @xmath149 , respectively , so that they degenerate in this figure ( we should note the sign of @xmath535 is inverted ) . the rms @xmath49 , which is considered as a weakly nonlinear parameter , is set @xmath536 . a limit @xmath537 corresponds to the prediction of the linear theory , which is given by thin solid lines in the figure . this linear prediction is equivalent to the gaussian fluctuations , because we assume the initial density field is random gaussian . in general , the curves of statistics plotted against the direct density threshold , @xmath133 ( dotted lines ) , exhibit considerable deviations from gaussian predictions . the overall tendency does not depend much on the shape of the spectrum we consider here , i.e. , power - law spectrum with index @xmath165 , @xmath211 , @xmath119 , and cdm model with smoothing length @xmath538 , where @xmath276 is the shape parameter of the cdm spectrum . they are consistent with the so - called meat - ball shift , which means that there are more isolated regions in a nonlinear field than in a gaussian field for a fixed threshold . in fact , @xmath105 , @xmath130 , @xmath539 of high value of threshold , e.g. , @xmath540 , virtually correspond to the number of isolated regions , and each figure shows that the number is indeed increased by weakly nonlinear evolution . the weakly nonlinear formula for the genus curve against the density threshold , @xmath541 , which was first derived by @xcite has been compared with numerical simulations in literatures . @xcite shows the good agreement of the analytic prediction with the simulations for various spectra . @xcite compared the prediction of the genus curve against direct @xmath133 with the simulated sdss data . unfortunately , in their published paper , they have transcribing errors , which incorrectly made the perturbation theory considerably disagree with their data . there are also ambiguity in their comparison on biasing which alter the values of skewness parameters . one can guess the biasing effect on skewness parameters by equation ( [ eq4 - 70 ] ) . they chose the peak particles as galaxies , and the linear biasing parameter is inferred as @xmath542 , but the nonlinear parameter @xmath525 is not obvious in their work . some literatures indicate the skewness @xmath88 of peaks are roughly given by 12 @xcite but for highly biased peaks @xmath543 . if we adopt @xmath544 , the skewness is given by @xmath545 which is not unreasonable . if it is the case for their simulation , the perturbative prediction and their data completely agree with each other . the @xmath546-value per degrees of freedom reduces to only 1.03 ( private communication with w. n. colley & d. h. weinberg ) . obviously , we have to further investigate the biasing effects in numerical simulations to obtain a conclusive result . most of the topological analyses of the previous work use the scaled threshold @xmath224 . the dashed lines in fig . [ fig1 ] shows the corresponding curves . the deviations from the linear theory is dramatically reduced . this fact is empirically known by the analyses of numerical simulations @xcite . the reason for this reduction is mathematically due to the closeness of the values of skewness parameters @xmath87 , ( @xmath523 ) , since all the terms of the nonlinear corrections in equations ( [ eq2 - 207a])([eq2 - 207c ] ) depend only on @xmath239 ( @xmath547 ) . for the hierarchical model of equation ( [ eq4 - 1 ] ) , they are exactly zero , that means there is not any ( second order ) nonlinear correction for the hierarchical model . since the hierarchical model is known to roughly approximate the nonlinear evolution , it is not surprising that more realistic fields have only small corrections of nonlinearity if they are plotted against the volume - fraction threshold , @xmath224 . for power - law models , @xmath88 and @xmath129 are exactly the same . that makes the nonlinear correction for @xmath105 or @xmath548 exactly vanishes . thus , the nonlinear corrections for other statistics are arisen by the difference between @xmath88 and @xmath234 . for the cdm model , there still is a difference between @xmath88 and @xmath129 , but it is relatively small as seen from the table [ tab2 ] . as for the topological statistics , @xmath130 , @xmath149 , @xmath549 and @xmath535 , deviations of curves against @xmath224 from the linear theory prediction depend on the underlying spectrum through differences of skewness parameters . the weakly nonlinear effect for redder spectrum of @xmath550 induces a sponge - like shift , that means the number of holes in isolated regions increases . on the other hand , the bluer spectrum of @xmath551 indicates a meat - ball shift . these tendencies are qualitatively in agreement with the numerical results ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? it is not trivial to estimate errors expected in observationally estimating statistics of smoothed cosmic fields . the observational errors mainly consists of the shot noise and the cosmic variace . the systematic comparison with simulations should be used for the error estimates . unfortunately , previous earlier simulations does not have enough resolution to be quantitatively compared in weakly nonlinear regime . @xcite use much larger simulations than those earlier work , and gives the genus curve in weakly nonlinear regime in their fig . 9 . relative depths of left and right troughs in the genus curve show slightly meat - ball shift , which is consistent with our prediction . although their variation of smoothing length is rather limited , their plots ensures the slight meat - ball shift predicted by the perturbation theory can definitely be observed . @xcite plot the 2d genus curve , but they use the thin slice which suffers strongly nonlinear effects so that the quantitative comparison is not appropriate with weakly nonlinear results . they report the meat - ball shift in 2d genus , which is the same direction of the weakly nonlinear correction besides amplitude . we find there is stll neccesity of proper , systematic comparison between the perturbation theory and numerical simulations in weakly nonlinear regime . the decisive factor between sponge - like shift and meat - ball shift is the sign and amplitude of @xmath552 , since @xmath553 is exactly or approximately zero for wide range of models like power - law , or cdm - like models . if the factor @xmath552 is positive , the meat - ball shift takes place . if this factor is negative , the sponge - like shift occurs . for power - law case , that factor is positive for @xmath554 , and is negative for @xmath555 . the amplitude of this factor times the amplitude of the nonlinearity parameter , @xmath556 determines the amplitude of the meat - ball shift in the genus curve . prominent shifts are seen around the two troughs of the genus curve at @xmath557 . at these troughs , the factor @xmath553 does not contribute to the genus curve since the accompanying factor @xmath558 vanishes . therefore , one can almost completely characterize the shifts by a factor @xmath559 , which we call the genus asymmetry parameter , around the troughs . we propose this genus asymmetry parameter as a theory - motivated parameter of the genus asymmetry . this factor is observationally determined by fitting the shape of the genus curve by a form , @xmath560 where @xmath245 is the genus asymmetry parameter . in this fitting , @xmath561 is much smaller if the non - gaussian features are from purely gravitational evolution of the initial gaussian field , and if the power spectrum is smooth enough as cdm models . the genus asymmetry parameter for cdm models with shape parameter @xmath276 , normalized by @xmath562 is listed for various smoothing length in table [ tab8 ] . cccccccccc @xmath563 & ... & @xmath564 & @xmath565 & @xmath566 & @xmath567 & @xmath568 & @xmath569 & @xmath570 & @xmath571 + @xmath572 & ... & @xmath573 & @xmath574 & @xmath575 & @xmath576 & @xmath577 & @xmath578 & @xmath579 & @xmath580 + @xmath581 & ... & @xmath582 & @xmath583 & @xmath584 & @xmath585 & @xmath586 & @xmath587 & @xmath588 & @xmath589 + @xmath590 & ... & @xmath591 & @xmath592 & @xmath593 & @xmath594 & @xmath595 & @xmath586 & @xmath596 & @xmath597 + @xmath322 & ... & @xmath598 & @xmath599 & @xmath600 & @xmath601 & @xmath602 & @xmath603 & @xmath604 & @xmath597 + since the factor @xmath552 increase with the spectral index , as seen from table [ tab3 ] , this factor also increase with the smoothing length for cdm model on scales of interest . on the other hand , the factor @xmath49 is a decreasing function of the smoothing length . as a result , the genus asymmetry parameter is not a monotonic function of the smoothing length . for example , the genus asymmetry parameter in the @xmath605 , @xmath606 cdm model is approximately @xmath607 on wide range of smoothing length of @xmath608 . from the second order formula for the genus of equation ( [ eq2 - 207c ] ) , the genus at the troughs is proportional to @xmath609 . therefore , the fraction of the deviation from the gaussian prediction at @xmath557 is @xmath610 . when the genus asymmetry parameter is @xmath607 , this fraction is about @xmath611 . thus , the perturbation theory predicts that the values of genus at positive and negative troughs differ @xmath612 for the @xmath613 , @xmath606 cdm model , and so on . this difference increases with the shape parameter @xmath276 . the degree of deviation is qualitatively consistent with the numerical result in fig . 9 of @xcite , although quantitative comparison is still difficult because of the noise in the simulation . the statistics for the velocity field is plotted in figure [ fig2 ] . since the the sign of the skewness parameters is negative in this case , the weakly nonlinear evolution of statistics against the density threshold @xmath133 indicates the sponge - like shift . in terms of the volume - fraction threshold @xmath224 , on the other hand , meat - ball shifts are observed even for relatively bluer spectrum with @xmath614 . the 2d projected galaxy statistics are dependent on the selection function of galaxies and cosmological models . as an example , we assume the apm luminosity function @xcite for galaxies with b band magnitude limit @xmath615 . the differential number count @xmath616 for this sample is plotted in figure [ fig3 ] . the resulting mean redshift is @xmath617 . the relation between the differential number count and the normalized mean number density @xmath405 in comoving coordinates are given by @xmath618 = \frac{\displaystyle h(z ) \frac{dn}{dz}(z ) } { \displaystyle \left\{s_k\left[\chi(z)\right]\right\}^2 \int_0^\infty \frac{dn}{dz } dz } , \label{eq5 - 1}\end{aligned}\ ] ] where @xmath619 are the hubble parameter and the comoving distance at redshift @xmath71 , respectively . once the selection function @xmath405 is given by equation ( [ eq5 - 1 ] ) , the integration of equations ( [ eq4 - 36])([eq4 - 38 ] ) , and the interpolation of tabulated values of @xmath447 in table [ tab6 ] give the skewness parameters . for power - law power spectra , the skewness parameters are given by the simpler integration of equation ( [ 4 - 50 - 2 ] ) and the values of @xmath447 in table [ tab7 ] . in figure [ fig4 ] , 2-dimensional statistics , @xmath105 , @xmath130 , @xmath620 and @xmath621 are plotted with assumed cosmological parameters , @xmath622 , @xmath623 , and apm luminosity function with limiting magnitude 19 . the nonlinear parameter is assumed as @xmath624 . for the cdm model , the shape parameter is assumed as @xmath625 , and we take the smoothing angle as @xmath626 . with this smoothing angle the value @xmath624 corresponds to the normalization @xmath627 . basic features for these 2-dimensional statistics are the same as for 3-dimensional density field , except that the smoothing scale in cdm model we adopt corresponds to smaller scale than in the example of 3-dimensional density field of the figure [ fig1 ] . in this work , we comprehensively presented a basic formalism to treat the statistics of smoothed cosmic fields in perturbation theory . this formalism provides a methodology on evaluating how various statistics deviate from the prediction of simple random gaussian fields . as long as the non - gaussianity is weak , the behavior of the statistics caused by non - gaussianity is predicted by our formalism , which enable us to quantitatively compare the statistical quantities and the source of the non - gaussianity . this method is considered as an extension of the edgeworth expansion , which has been proven to be useful in various fields of research , as long as the non - gaussianity is weak . in this paper , we derive useful formulas and relations focusing on application of the second - order perturbation theory to various cosmic fields . several examples of statistics of cosmic fields in second - order perturbation theory are investigated in datail , including level - crossing statistics , 2d and 3d genus statistics , 2d extrema statistics , and the minkowski functionals , which are extensively used in cosmology . more complicated statistics , such as 2d and 3d density peaks , can also be calculated , although they are more tedius . a particular interest in cosmology of our method is in the application to the cosmic fields . even if the cosmic field was random gaussian at the initial stage , the gravitational evolution induces the non - gaussianity . the gravitational instability is a well - defined process , so that we can evaluate the non - gaussianity without any ambiguity when the evolution remains in the quasi - linear regime provided that the biasing from the non - gravitational process is simple enough on large scales . therefore , we performed the perturbative analysis to obtain the necessary skewness parameters based on the gravitational instability theory . we considered the 3d density field , the 3d velocity field , the 2d projected density field . in the application of second order theory to various statistics of smoothed cosmic fields , three types of skewness parameters are commonly useful , i.e. , @xmath88 , @xmath129 , an @xmath234 . extensive calculations of these parameters for various cosmic fields are one of the new results of this paper . it would be true that other skewness parameters are needed when other complex statistics are considered . such other parameters , if needed , are similarly calculated by the method we outlined in this paper . we find the lowest order deviations from the gaussian predictions of various statistics of smoothed cosmic fields depend only on the differences of the skewness parameters when we use a threshold @xmath224 , which is rescaled by a volume - fraction of the smoothed field . this rescaling makes the lowest deviations much smaller than in the case of direct threshold @xmath133 . this is because three skewness parameters @xmath87 take similar values if it is arisen from the gravitational evolution . for the phenomenological hierarchical model , these parameters are identical . in this case , the weakly nonlinear correction of the statistical quantities in terms of the volume - fraction threshold vanishes . when evaluated by the second - order perturbation theory of density fluctuations , those three types of skewness parameter are still close to each other this fact explains the smallness of the deviations from gaussian predictions of statistical quantities like genus , level - crossing , or minkowski functionals when the rescaled threshold by volume - fraction is used . we discussed small , but detectable deviations from gaussian prediction of the 3d genus curve against rescaled threshold in detail . in the framework of the second - order perturbation theory , a prominent deviation of the genus curve occurs at the two troughs of the curve . relative depths of left and right troughs in the genus curve show slightly meat - ball shift for cdm - like models . we found the degree of this asymmetry is proportional to the combination @xmath559 . we call this factor as a genus asymmetry parameter , which we propose as theory - motivated parameter that characterize the asymmetry of the genus curve . the genus asymmetry parameter can observationally be obtained by fitting the genus curve by hermite polynomials as equation ( [ eq5 - 0 ] ) . qualitative comparison with the numerical simulations in literatures suggests that such asymmetry can actually be observable . we have not estimated over what dynamic range in nonlinearity parameter like @xmath49 , @xmath628 , or @xmath629 are each perturbative expressions valid . this estimation requires a systematic comparison with large n - body simulations , which is beyond the scope of this paper , and will be given in a subsequent paper of the series . in principle , any order in the perturbation theory can be calculated as further as one would like . although the computation of the higher - order theory becomes more and more tedious , the necessity of the comparison with large - scale cosmological observations is a good reason to perform such computation as further as we can . one of the spectacular example of the detailed comparison between perturbation theory and observations is the fine structure constant in quantum electrodynamics ( e.g. , * ? ? ? our analysis in this paper will be extended to the third - order perturbation theory in a subsequent paper of the series . the present time is in an unique decade when the observations of cosmic fields are in unforeseen progress , like large - scale redshift surveys , detailed mapping of cmb fluctuations , gravitational lensing surveys , and so forth . statistics of smoothed density field with higher - order perturbation theory will provide an unique method to analyze those high - precision data . the precision cosmology is undoubtedly providing clues to unlock the door to the origin of the universe . i would like to thank m. kerscher and b. jain for discussions . i wish to acknowledge support from jsps postdoctoral fellowships for research abroad , and from the ministry of education , culture , sports , science , and technology , grant - in - aid for encouragement of young scientists , 13740150 , 2001 . in this appendix , we give gaussian integrals which are useful in this paper . in the following , @xmath630 is the hermite polynomials , @xmath631 and we further employ the notation @xmath632 several hermite polynomials are @xmath633 the hermite polynomial at zero is given by @xmath634 we generalize the above definition of @xmath635 to the case @xmath636 by interpreting @xmath637 as the appropriate gamma function so that @xmath638 , @xmath639 , etc . for example , @xmath640 , @xmath641 , @xmath642 , @xmath643 , and so forth . in this appendix , we give useful gaussian averages @xmath644 of equation ( [ eq1 - 12 - 1 ] ) for normalized cosmic field @xmath18 defined by equation ( [ eq1 - 3 ] ) and its spatial derivatives @xmath645 . in the following , @xmath646 represents any one of the components @xmath78 . first , concerning @xmath18 , @xmath647 second , concerning @xmath646 , in the notation of equation ( [ eqa1 - 3 ] ) , @xmath648 the correlation functions on projected sky is expressible by the 3-dimensional correlation functions . the explicit relation for the two - point correlation function is given by limber s equation @xcite . the fourier - space version of limber s equation is given by kaiser @xcite and is somewhat simpler . his argument was generalized to higher - order correlation functions and their fourier transforms @xcite . since the higher - order limber s equation in fourier space was discussed in the context of weak lensing field in literatures , here we review the derivation in a way more useful in this paper . let 2d projected field @xmath9 be the projection of a time - dependent 3d field @xmath649 along a light - cone : @xmath650 where @xmath651 is some radial weighting function and @xmath401 is the radial comoving distance , and @xmath409 is the conformal time at the observer . the past light - cone of the observer is specified by the equation @xmath652 . the comoving angular distance @xmath653 is defined by equation ( [ eq4 - 29 ] ) . in the following , we explicitly derive the relation for 3-point correlation function , and bispectrum . from the limber s equation , power spectrum is already derived by @xcite : @xmath654 where @xmath655 and @xmath656 is the power spectrum of fields @xmath9 and @xmath55 , respectively . we generalize this equation to the one for 3-point statistics below . the generalization of the following derivation to higher - order statistics is straightforward . the angular 3-point correlation function @xmath657 of @xmath9 is @xmath658 where @xmath659 is the spatial 3-point correlation function of the field @xmath55 with the 3-point configuration @xmath660 at conformal time @xmath420 . according to the spirit of the limber s equation , we assume that @xmath661 is slowly varying compared to the scale of the fluctuations of interest and also that these fluctuations occur on a scale much smaller than the curvature scale . the equation ( [ eqa2 - 3 ] ) is the generalization of the limber s equation to higher - order correlation functions . now we transform equation ( [ eq2 - 13 ] ) to obtain the 2d bispectrum . we use the following convention of the fourier transforms @xmath662 the bispectrum @xmath663 of the 2d field @xmath9 , and @xmath664 of the 3d field @xmath55 are defined by @xmath665 where @xmath666 and @xmath667 . the dirac s delta function comes from the translational invariance of statistics . from the relations , @xmath668 the fourier transform of the equation ( [ eqa2 - 3 ] ) reduces to a simple equation , @xmath669 this is a bispectrum version of the limber s equation and the generalization of the kaiser s equation for power spectrum ( [ eqa2 - 2 ] ) . in table [ tab9 ] , the quantites used in this paper are listed . ccc @xmath670 & ensemble average & ( [ eq1 - 12a ] ) + @xmath644 & ensemble average by gaussian field & ( [ eq1 - 12 - 1 ] ) + @xmath481 & 2d spectral parameter & ( [ eq4 - 36 ] ) + @xmath353 & time - dependent density parameter & ( [ eq4 - 34 - 1 ] ) + @xmath252 & density parameter & ( [ eq4 - 34 - 1 ] ) + @xmath18 & normalized field of @xmath9 & ( [ eq1 - 3 ] ) + @xmath671 , etc . & @xmath672 , etc . & ( [ eq1 - 4 ] ) + @xmath673 & spectral parameter & ( [ eq1 - 13 - 5 ] ) + @xmath97 & mass density contrast & ( [ eq4 - 70 ] ) + @xmath674 & 3d density contrast & ( [ eq4 - 31 ] ) + @xmath520 & galaxy number density contrast & ( [ eq4 - 70 ] ) + @xmath62 & @xmath675 , tensor notation & + @xmath64 & linear combination of @xmath18 and @xmath62 & ( [ eq1 - 13 - 3 - 1 ] ) + @xmath78 & @xmath676 , vector notation & + @xmath417 & angular coordinates in a small patch of the sky & ( [ eq4 - 31 ] ) + @xmath513 & local convergence field & ( [ eq4 - 60 ] ) + @xmath354 & time - dependent scaled cosmological constant & ( [ eq4 - 34 - 1 ] ) + @xmath253 & scaled cosmological constant & ( [ eq4 - 34 - 1 ] ) + @xmath133 & threshold by variance & + @xmath677 & threshold by volume fraction & ( [ eq2 - 105 ] ) + @xmath241 & 3d comoving density field & ( [ eq4 - 30 ] ) + @xmath678 & 2d weighted extrema & ( [ eq2 - 24 ] ) + @xmath400 & 2d projected density field & ( [ eq4 - 30 ] ) + @xmath49 & variance of @xmath9 & ( [ eq1 - 2 ] ) + @xmath271 , @xmath679 & spectral parameters & ( [ eq1 - 13 - 3a ] ) , ( [ eq1 - 13 - 3b ] ) + @xmath409 & conformal time at the present & ( [ eq4 - 30 ] ) + @xmath401 & comoving distance & ( [ eq4 - 29 ] ) + @xmath680 & volume of the unit ball in @xmath121 dimensions & ( [ eq2 - 54 ] ) + @xmath681 & vector of spatial derivatives of @xmath18 & ( [ eq1 - 4 ] ) + @xmath682 & bispectrum at the present time & ( [ eq4 - 4b ] ) + @xmath683 & 2d bispectrum & ( [ eq4 - 33b ] ) + @xmath684 & 3d bispectrum at conformal time @xmath420 & ( [ eq4 - 33b ] ) + @xmath685 & integrand of the 2d skewness parameters & ( [ eq4 - 39 ] ) + @xmath686 & integrand of the 2d skewness parameters & ( [ eq4 - 40 ] ) + @xmath687 & contributions to the 2d skewness parameters & ( [ eq4 - 46 ] ) + @xmath349 & linear growth rate & ( [ eq4 - 34a ] ) + @xmath201 & parameter of the perturbation theory & ( [ eq2 - 21 - 1 ] ) + @xmath351 & parameter of the perturbation theory & ( [ eq4 - 24b ] ) + @xmath55 & cosmic field , in general & ( [ eq1 - 12a ] ) + @xmath688 , etc . & @xmath689 , etc . & ( [ eq1 - 13 ] ) + @xmath690 & gauss hypergeometric function @xmath691 & ( [ eq4 - 20 ] ) + @xmath130 & 2d genus statistic & ( [ eq2 - 8 ] ) , ( [ eq2 - 9 ] ) + @xmath149 & 3d genus statistic & ( [ eq2 - 15 ] ) , ( [ eq2 - 17 ] ) + @xmath343 & time - dependent hubble parameter & ( [ eq4 - 23 ] ) + @xmath692 & hubble constant & ( [ eq5 - 2 ] ) + @xmath693 & modified bessel function & ( [ eq4 - 11 ] ) + @xmath694 & fourier counterpart of @xmath25 & ( [ eq1 - 5 ] ) + @xmath695 & integers defined by table [ tab1 ] & ( [ eq2 - 19 ] ) + @xmath183 & spatial curvature & ( [ eq4 - 29 ] ) + @xmath696 & hermite polynomials & ( [ eqa1 - 1 ] ) + @xmath697 & extended hermite polynomial of order @xmath211 & ( [ eqa1 - 2 ] ) + @xmath698 & @xmath314-th order cumulants of @xmath25 & ( [ eq1 - 6])([eq1 - 7d ] ) + @xmath699 & normalized cumulants of @xmath25 & ( [ eq1 - 12 ] ) + @xmath37 & second - order cumulant of @xmath25 & ( [ eq1 - 8 ] ) + @xmath105 & level - crossing statistic & ( [ eq2 - 1a ] ) + @xmath106 & length statistic & ( [ eq2 - 1b ] ) + @xmath107 & area statistic & ( [ eq2 - 1c ] ) + @xmath700 & multivariate gaussian distribution function & ( [ eq1 - 5 ] ) + @xmath244 & multivariate probablitiy distribution function & ( [ eq1 - 5 ] ) + @xmath701 & power spectrum at present & ( [ eq4 - 4a ] ) + @xmath702 & 2d power spectrum & ( [ eq4 - 33a ] ) + @xmath703 & 3d power spectrum at conformal time @xmath420 & ( [ eq4 - 33a ] ) + @xmath249 & linear power spectrum & ( [ eq4 - 7a ] ) + @xmath238 & hierarchical amplitude of the 3-point function & ( [ eq4 - 1 ] ) + @xmath87 , @xmath704 & skewness parameters & ( [ eq1 - 15a])([eq1 - 15d ] ) + @xmath705 & integrand of the skewness parameters & ( [ eq4 - 9 ] ) + @xmath706 & contributions to the skewness parameters & ( [ eq4 - 13 ] ) + @xmath707 & galaxy skewness parameters & ( [ eq4 - 29 ] ) + @xmath708 & comoving angular diameter distance & ( [ eq4 - 29 ] ) + @xmath709 & cdm transfer function & ( [ eq4 - 15 ] ) + @xmath181 & minkowski functionals & ( [ eq2 - 50 - 0 ] ) , ( [ eq2 - 50 ] ) , ( [ eq2 - 52 ] ) , ( [ eq2 - 54 ] ) + @xmath710 & smoothing function in real space & ( [ eq1 - 1 ] ) + @xmath10 & smoothing function in real space & ( [ eq1 - 1 ] ) + @xmath711 & generating function & ( [ eq1 - 5 ] ) + @xmath524 & bias parameter & ( [ eq4 - 70 ] ) + @xmath525 & nonlinear bias parameter & ( [ eq4 - 70 ] ) + @xmath6 & dimension of the sample space & ( [ eq1 - 1 ] ) + @xmath712 & differential number count & ( [ eq5 - 1 ] ) + @xmath9 & smoothed cosmic field , in general & ( [ eq1 - 1 ] ) + @xmath225 & volume fraction on the high - density side & ( [ eq2 - 105 ] ) + @xmath347 & logarithmic derivative of the growth factor & ( [ eq4 - 25 ] ) + @xmath125 & hermite polynomial at zero , @xmath713 , extended to @xmath714 & ( [ eqa1 - 3 ] ) + @xmath314 & selection function per unit comoving volume & ( [ eq4 - 30 ] ) + @xmath311 & index for the power - law power spectrum & ( [ eq4 - 16 ] ) + @xmath73 , @xmath74 , @xmath71 & linear combinations of @xmath18 and @xmath62 s & ( [ eq1 - 13 - 4a])([eq1 - 13 - 4c ] ) + adler , r. j. 1981 , r. k. 1986 , the geometry of random fields ( chichester ; wiley ) balian , r. & schaeffer , r. 1989 , , 220 , 1 bardeen , j. m. , bond , j. r. , kaiser , n. , & szalay , a. s. , 1986 , , 304 , 15 bartelmann , m. & schneider , p. 1999 , to be submitted to physics reports ( astro - ph/9912508 ) baugh , c. m. & efstathiou , g. 1994 , , 270 , 183 beaky , m. m. , scherrer , r. j. & villumsen , j. v. 1992 , , 387 , 443 bernardeau , f. 1992 , , 392 , 1 bernardeau , f. 1994a , , 433 , 1 bernardeau , f. 1994b , , 291 , 697 bernardeau , f. 1995 , , 301 , 309 bernardeau , f. , juszkiewicz , r. , dekel , a. & bouchet , f. r. 1995 , , 274 , 20 bernardeau , f. & kofman , l. 1995 , , 443 , 479 bernardeau , f. , van waerbeke , l. & mellier , y. 1997 , , 322 , 1 bernardeau , f. , colombi , s. , gaztanaga , e. , & scoccimarro , r. 2002 , , 367 , 1 bharadwaj , s. , sahni , v. , sathyaprakash , b. s. , shandarin , s. f. , & yess , c. 2000 , , 528 , 21 bond , j. r. & efstathiou , g. 1987 , , 226 , 655 bouchet , f. r. , juszkiewicz , r. , colombi , s. & pellat , r. 1992 , , 394 , l5 buchalter , a. , kamionkowski , m. , & jaffe , a. h. 2000 , , 530 , 36 buchalter , a. & kamionkowski , m. 1999 , , 521 , 1 canavezes , a. et al . 1998 , , 297 , 777 carroll , s. m. , press , w. h. & turner , e. l. 1992 , , 30 , 499 cayon , l. & smoot , g. 1995 , , 452 , 487 cen , r. 1998 , , 509 , 494 chodorowski , m. j. & lokas , e. l. 1997 , , 287 , 591 coles , p. 1988 , , 234 , 509 coles , p. 1990 , , 243 , 171 coles , p. & barrow , j. d. 1987 , , 228 , 407 coles , p. & plionis , m. 1991 , , 250 , 75 coles , p. , moscardini , l. , plionis , m. , lucchin , f. , matarrese , s. & messina , a. 1993 , , 260 , 572 colley , w. n. 1997 , , 489 , 471 colley , w. n. , gott , j. r. i. , weinberg , d. h. , park , c. and berlind , a. a. 2000 , , 529 , 795 croft , r. a. c. & efstathiou , g. 1994 , , 267 , 390 crofton , m. w. 1868 , phil . lond . a , 158 , 181 davies , a. & coles , p. 1993 , , 260 , 553 dekel , a. & lahav , o. 1999 , , 520 , 24 doroshkevich , a. g. 1970 , astrophysica , 6 , 320 fabbri , r. & torres , s. 1996 , , 307 , 703 fry , j. n. 1984 , , 279 , 499 fry , j. n. & gaztanaga , e. 1993 , , 413 , 447 gabrielli , a. , labini , f. s. & durrer , r. 2000 , , 531 , l1 goroff , m. h. , grinstein , b. , rey , s .- j . & wise , m. b. 1986 , , 311 , 6 gott , j. r. , melott , a. l. & dickinson , m. 1986 , , 306 , 341 gott , j. r. , weinberg , d. h. & melott , a. l. 1987 , , 319 , 1 gott , j. r. , mao , s. , park , c. & lahav , o. 1992 , , 385 , 26 gott , j. r. i. et al . 1989 , , 340 , 625 gott , j. r. , park , c. , juszkiewicz , r. , bies , w. e. , bennett , d. p. , bouchet , f. r. & stebbins , a. 1990 , , 352 , 1 hamilton , a. j. s. , gott , j. r. & weinberg , d. 1986 , , 309 , 1 heavens , a. f. & sheth , r. k. 1999 , , 310 , 1062 jain , b. & bertschinger , e. 1994 , , 431 , 495 jain , b. & van waerbeke , l. 2000 , , 530 , l1 juszkiewicz , r. 1981 , , 197 , 931 juszkiewicz , r. , sonoda , d. h. & barrow , j. d. 1984 , , 209 , 139 juszkiewicz , r. , bouchet , f. r. & colombi , s. 1993 , , 412 , l9 juszkiewicz , r. , weinberg , d. h. amsterdamski , p. , chodorowski , m. & bouchet , f. r. 1995 , , 442 , 39 kaiser , n. 1984 , , 284 , l9 kaiser , n. 1998 , , 498 , 26 kerscher , m. , schmalzing , j. , buchert , t. & wagner , h. 1998 , , 333 , 1 kinoshita , t. 1996 , rep . phys . , 59 , 1459 kogut , a. , banday , a. j. , bennett , c. l. , hinshaw , g. , lubin , p. m. & smoot , g. f. 1995 , , 439 , l29 lahav , o. , lilje , p. b. , primack , j. r. & rees , m. j. 1991 , , 251 , 128 laveday , j. , peterson , b. a. , efstathiou , g. & maddox , s. j. 1992 , , 390 , 338 lightman , a. p. & schechter , p. l. 1990 , , 74 , 831 limber , d. n. 1954 , , 119 , 655 lokas , e. l. 1998 , , 296 , 491 okas , e. l. , juszkiewicz , r. , weinberg , d. h. & bouchet , f. r. 1995 , , 274,730 ma , s .- k . 1985 , statistical mechanics ( philadelphia ; world scientific ) makino , n. , sasaki , m. & suto , y. 1992 , , 46 , 585 & suto , y. 1992 , mann , r. g. , heavens , a. f. & peacock , j. a. 1993 , , 263 , 798 martel , h. 1991 , , 377 , 7 matsubara , t. 1994 , , 434 , l43 matsubara , t. 1995a , , 101 , 1 matsubara , t. 1995b , prog . ( lett . ) , 94 , 1151 matsubara , t. 1996 , , 457 , 13 matsubara , t. 1999 , , 525 , 543 matsubara , t. & suto , y. 1996 , , 460 , 51 matsubara , t. & yokoyama , j. 1996 , , 463 , 409 mecke , k. r. , buchert , t. & wagner , h. 1994 , , 288 , 697 mellier , y. 1999 , ara&a , 37 , 127 melott , a. l. , cohen , a. p. , hamilton , a. j. s. , gott , j. r. & weinberg , d. h. 1989 , , 345 , 618 moore , b. et al . 1992 , , 256 , 477 morse , m. & cairns , s. 1969 , critical point theory in global analysis and differential topology ( new york ; academic press ) naselsky , p. d. & novikov , d. i. 1998 , , 507 , 31 park , c. & gott , j. r. 1991 , , 378 , 457 park , c. , gott , j. r. & da costa , l. n. 1992 , , 392 , l51 park , c. , colley , w. n. , gott , j. r. i. , ratra , b. , spergel , d. n. and sugiyama , n. 1998 , , 506 , 473 peebles , p. j. e. 1980 , the large - scale structure of the universe ( princeton ; princeton university press ) plionis , m. , valdarnini , r. & coles , p. 1992 , , 258 , 114 plionis , m. & valdarnini , r. 1995 , , 272 , 869 protogeros , z. a. m. & weinberg , d. h. 1997 , , 489 , 457 rhoads , j. e. , gott , j. r. i. & postman , m. 1994 , , 421 , 1 ryden , b. s. 1988 , , 333 , l41 ryden , b. s. 1988 , , 100 , 1360 ryden , b. s. , melott , a. l. , craig , d. a. , gott , j. r. , iii , wenberg , d. h. , scherrer , r. j. , bhavsar , s. p. & miller , j. m. 1989 , , 340 , 647 sahni , v. , sathyaprakash , b. s. , & shandarin , s. f. 1998 , , 495 , l5 sahni , v. , sathyaprakash , b. s. , & shandarin , s. f. 1997 , , 476 , l1 sathyaprakash , b. s. , sahni , v. , & shandarin , s. 1998 , , 508 , 551 sazhin , m. v. 1985 , , 216 , 25p scherrer , r. j. & bertschinger , e. 1991 , , 381 , 349 schmalzing , j. & buchert , t. 1997 , , 482 , l1 schmalzing , j. & gorski , k. m. 1998 , , 297 , 355 schmalzing , j. , buchert , t. , melott , a. l. , sahni , v. , sathyaprakash , b. s. & shandarin , s. f. 1999 , , 526 , 568 schmalzing , j. and diaferio , a. 2000 , , 312 , 638 schneider , p. , ehlers , j. , falco , . e. e. 1992 , gravitational lenses , ( springer ) scoccimarro , r & frieman , j. a. 1996a , , 105 , 37 scoccimarro , r & frieman , j. a. 1996b , , 473 , 620 scoccimarro , r. ; . , zaldarriaga , m. , & hui , l. 1999 , , 527 , 1 seto , n. 2000 , , 537 , in press ( astro - ph/0002315 ) smoot , g. f. , tenorio , l. , banday , a. j. , kogut , a. , wright , e. l. , hinshaw , g. & bennett , c. l. 1994 , , 437 , 1 springel , v. , and 10 colleagues 1998 , , 298 , 1169 suto , y. & sasaki , m. 1991 , physical review letters , 66 , 264 taruya , a. & soda , j. 1999 , , 522 , 46 tomita , h. 1986 , progr . , 76 , 952 torres , s. 1994 , , 423 , l9 torres , s. , cayon , l. , martinez - gonzalez , e. & sanz , j. l. 1995 , , 274 , 853 van waerbeke , l. 2000 , , 313 , 524 vishniac , e. t. 1983 , , 203 , 345 vittorio , n. & juszkiewicz , r. 1987 , , 314 , l29 vogeley , m. s. , park , c. , geller , m. j. , huchra , j. p. & gott , j. r. i. 1994 , , 420 , 525 watanabe , t. , matsubara , t. & suto , y. 1994 , , 432 , 17 winitzki , s. & kosowsky , a. 1997 , new astronomy , 3 , 75

in the _ tree compatibility problem _ , we are given a collection @xmath4 of rooted phylogenetic trees with partially overlapping taxon sets . @xmath0 is called a _ profile _ and the trees in @xmath0 are the _ input trees_. the question is whether there exists a tree @xmath1 whose taxon set is the union of the taxon sets of the input trees , such that @xmath1 exhibits the clusterings implied by the input trees . that is , if two taxa are together in a subtree of some input tree , then they must also be together in some subtree of @xmath1 . the tree compatibility problem has been studied for over three decades @xcite . in the original version of the tree compatibility problem , only the leaves of the input trees are labeled . here we study a generalization , called _ ancestral compatibility _ , in which taxa may be _ nested_. that is , the internal nodes may also be labeled ; these labels represent _ higher - order taxa _ , which are , in effect , sets of taxa . thus , for example , an input tree may contain the taxon _ glycine max _ ( soybean ) nested within a subtree whose root is labeled fabaceae ( the legumes ) , itself nested within an angiosperm subtree . note that leaves themselves may be labeled by higher - order taxa . the question now is whether there is a tree @xmath1 whose taxon set is the union of the taxon sets of the input trees , such that @xmath1 exhibits not only the clusterings among the taxa , but also the ancestor / descendant relationships among taxa in the input trees . our main result is a @xmath2 algorithm for the compatibility problem for trees with nested taxa , where @xmath3 is the total number of nodes and edges in the trees in @xmath0 . [ [ background . ] ] background . + + + + + + + + + + + the tree compatibility problem is a basic special case of the _ supertree problem_. a supertree method is a way to synthesize a collection of phylogenetic trees with partially overlapping taxon sets into a single supertree that represents the information in the input trees . the supertree approach , proposed in the early 90s @xcite , has been used successfully to build large - scale phylogenies @xcite . the original supertree methods were limited to input trees where only the leaves are labeled . page @xcite was among the first to note the need to handle phylogenies where internal nodes are labeled , and taxa are nested . a major motivation is the desire to incorporate _ taxonomies _ as input trees in large - scale supertree analyses , as way to circumvent one of the obstacles to building comprehensive phylogenies : the limited taxonomic overlap among different phylogenetic studies @xcite . taxonomies group organisms according to a system of taxonomic rank ( e.g. , family , genus , and species ) ; two examples are the ncbi taxonomy @xcite and the angiosperm taxonomy @xcite . taxonomies spanning a broad range of taxa provide structure and completeness that might be hard to obtain otherwise . a recent example of the utility of taxonomies is the open tree of life , a draft phylogeny for over 2.3 million species @xcite . taxonomies are not , strictly speaking , phylogenies . in particular , their internal nodes and some of their leaves are labeled with higher - order taxa . nevertheless , taxonomies have many of the same mathematical characteristics as phylogenies . indeed , both phylogenies and taxonomies are _ semi - labeled trees _ @xcite . we will use this term throughout the rest of the paper to refer to trees with nested taxa . the fastest previous algorithm for testing ancestral compatibility , based on earlier work by daniel and semple @xcite , is due to berry and semple @xcite . their algorithm runs in @xmath5 time using @xmath6 space . here , @xmath7 is the number of distinct taxa in @xmath0 and @xmath8 , where @xmath9 is the set of internal nodes of @xmath10 , for each @xmath11 , and @xmath12 is the degree of node @xmath13 . while the algorithm is polynomial , its dependence on node degrees is problematic : semi - labeled trees can be highly unresolved ( i.e. , contain nodes of high degree ) , especially if they are taxonomies . [ [ our - contributions . ] ] our contributions . + + + + + + + + + + + + + + + + + + the @xmath2 running time of our ancestral compatibility algorithm is independent of the degrees of the nodes of the input trees , a valuable characteristic for large datasets that include taxonomies . to achieve this time bound , we extend ideas from our recent algorithm for testing the compatibility of ordinary phylogenetic trees @xcite . as in that algorithm , a central notion in the current paper is the _ display graph _ of profile @xmath0 , denoted @xmath14 . this is the graph obtained from the disjoint union of the trees in @xmath0 by identifying nodes that have the same label ( see section [ sec : testac ] ) . the term `` display graph '' was introduced by bryant and lagergren @xcite , but similar ideas have been used elsewhere . in particular , the display graph is closely related to berry and semple s _ restricted descendancy graph _ @xcite , a mixed graph whose directed edges correspond to the ( undirected ) edges of @xmath14 and whose undirected edges have no correspondence in @xmath14 . the second kind of edges are the major component of the @xmath15 term in the time and space complexity of berry and semple s algorithm . the absence of such edges makes @xmath14 significantly smaller than the restricted descendancy graph . display graphs also bear some relation to _ tree alignment graphs _ @xcite . here , we exploit the display graph more extensively and more directly than our previous work . although the display graph of a collection of semi - labeled trees is more complex than that of a collection of ordinary phylogenies , we are able to extend several of the key ideas notably , that of a semi - universal label to the general setting of semi - labeled trees . as in @xcite , the implementation relies on a dynamic graph data structure , but it requires a more careful amortized analysis based on a weighing scheme . [ [ contents . ] ] contents . + + + + + + + + + section [ sec : prelims ] presents basic definitions regarding semi - labeled trees and ancestral compatibility . . section [ sec : dispgraph ] introduces the display graph and discusses its properties . section [ sec : testac ] presents @xmath16 , our algorithm for testing ancestral compatibility . section [ sec : implementation ] gives the implementation details for @xmath16 . section [ sec : discussion ] gives some concluding remarks . for each positive integer @xmath17 , @xmath18 $ ] denotes the set @xmath19 . let @xmath20 be a graph . @xmath21 and @xmath22 denote the node and edge sets of @xmath20 . the _ degree _ of a node @xmath23 is the number of edges incident on @xmath13 . a _ tree _ is an acyclic connected graph . in this paper , all trees are assumed to be rooted . for a tree @xmath24 , @xmath25 denotes the root of @xmath24 . suppose @xmath26 . then , @xmath27 is an _ ancestor _ of @xmath13 in @xmath24 , denoted @xmath28 , if @xmath27 lies on the path from @xmath13 to @xmath25 in @xmath24 . if @xmath28 , then @xmath13 is a _ descendant _ of @xmath27 . node @xmath27 is a _ proper descendant _ of @xmath13 if @xmath27 is a descendant of @xmath13 and @xmath29 . if @xmath30 and @xmath28 , then @xmath27 is the _ parent _ of @xmath13 and @xmath13 is a _ child _ of @xmath27 . if neither @xmath28 nor @xmath31 hold , then we write @xmath32 and say that @xmath27 and @xmath13 are _ not comparable _ in @xmath24 . [ [ semi - labeled - trees . ] ] semi - labeled trees . + + + + + + + + + + + + + + + + + + + a _ semi - labeled tree _ is a pair @xmath33 where @xmath24 is a tree and @xmath34 is a mapping from a set @xmath35 to @xmath36 such that , for every node @xmath37 of degree at most two , @xmath38 . @xmath35 is the _ label set _ of @xmath1 and @xmath34 is the _ labeling function _ of @xmath1 . for every node @xmath37 , @xmath39 denotes the ( possibly empty ) subset of @xmath35 whose elements map into @xmath13 ; these elements as the _ labels of @xmath13 _ ( thus , each label is a taxon ) . if @xmath40 , then @xmath13 is _ labeled _ ; otherwise , @xmath13 is _ unlabeled_. note that , by definition , every leaf in a semi - labeled tree is labeled . further , any node , including the root , that has a single child must be labeled . nodes with two or more children may be labeled or unlabeled . a semi - labeled tree @xmath33 is _ singularly labeled _ if every node in @xmath24 has at most one label ; @xmath1 is _ fully labeled _ if every node in @xmath24 is labeled . semi - labeled trees , also known as _ @xmath41-trees _ , generalize ordinary phylogenetic trees , also known as _ phylogenetic @xmath41-trees _ @xcite . an ordinary phylogenetic tree is a semi - labeled tree @xmath33 where @xmath25 has degree at least two and @xmath34 is a bijection from @xmath35 into leaf set of @xmath24 ( thus , internal nodes are not labeled ) . let @xmath33 be a semi - labeled tree and let @xmath42 and @xmath43 be two labels in @xmath35 . if @xmath44 , then we write @xmath45 , and say that @xmath43 is a _ descendant _ of @xmath42 in @xmath1 and that @xmath42 is an _ ancestor _ of @xmath43 . we write @xmath46 if @xmath47 is a proper descendant of @xmath48 . if @xmath49 , then we write @xmath50 and say that @xmath42 and @xmath43 are _ not comparable _ in @xmath1 . if @xmath1 is fully labeled and @xmath48 is the parent of @xmath47 in @xmath24 , then @xmath42 is the _ parent _ of @xmath43 in @xmath1 and @xmath43 is a _ child _ of @xmath42 in @xmath1 ; two labels with the same parent are _ siblings_. two semi - labelled trees @xmath33 and @xmath51 are _ isomorphic _ if there exists a bijection @xmath52 such that @xmath53 and , for any two nodes @xmath26 , @xmath54 if and only @xmath55 . let @xmath33 be a semi - labeled tree . for each @xmath56 , @xmath57 denotes the set of all labels in the subtree of @xmath24 rooted at @xmath27 ; that is , @xmath58 . @xmath57 is called a _ cluster _ of @xmath24 . @xmath59 denotes the set of all clusters of @xmath1 . it is well known ( * ? ? ? * theorem 3.5.2 ) that a semi - labeled tree @xmath1 is completely determined by @xmath59 . that is , if @xmath60 for some other semi - labeled tree @xmath61 , then @xmath1 is isomorphic to @xmath61 . suppose @xmath62 for a semi - labeled tree @xmath33 . the _ restriction _ of @xmath1 to @xmath63 , denoted @xmath64 , is the semi - labeled tree whose cluster set is @xmath65 intuitively , @xmath66 is obtained from the minimal rooted subtree of @xmath24 that connects the nodes in @xmath67 by suppressing all vertices of degree two that are not in @xmath67 . let @xmath33 and @xmath51 be semi - labeled trees such that @xmath68 . @xmath1 _ ancestrally displays _ @xmath61 if @xmath69 . equivalently , @xmath1 ancestrally displays @xmath61 if @xmath61 can be obtained from @xmath70 by contracting edges , and , for any @xmath71 , ( i ) if @xmath72 , then @xmath73 , and ( ii ) if @xmath74 , then @xmath75 . the notion of `` ancestrally displays '' for semi - labeled trees generalizes the well - known notion of `` displays '' for ordinary phylogenetic trees @xcite . for a semi - labelled tree @xmath1 , let @xmath76 note that @xmath77 consists of _ ordered _ pairs , while @xmath78 consists of _ unordered _ pairs . [ lm : dn ] let @xmath1 and @xmath61 be semi - labelled trees such that @xmath68 . then @xmath1 ancestrally displays @xmath61 if and only if @xmath79 and @xmath80 . [ [ profiles - and - ancestral - compatibility . ] ] profiles and ancestral compatibility . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + throughout the rest of this paper @xmath81 denotes a set where , for each @xmath82 $ ] , @xmath83 is a semi - labeled tree . we refer to @xmath0 as a _ profile _ , and write @xmath84 to denote @xmath85 } l({\ensuremath{\mathcal{t}}}_i)$ ] , the _ label set _ of @xmath0 . figure [ fig : profile ] shows a profile where @xmath86 . we write @xmath87 for @xmath85 } v(t_i)$ ] and @xmath88 for @xmath85 } e(t_i)$ ] , the _ size _ of @xmath0 is @xmath89 . for the profile of figure [ fig : profile ] . ] for the profile of figure [ fig : profile ] . ] for the profile of figure [ fig : profile ] . ] @xmath0 is _ ancestrally compatible _ if there is a rooted semi - labeled tree @xmath1 that ancestrally displays each of the trees in @xmath0 . if @xmath1 exists , we say that @xmath1 _ ancestrally displays _ @xmath0 ( see figure [ fig : supertree ] ) . given a subset @xmath41 of @xmath84 , the _ restriction _ of @xmath0 to @xmath41 , denoted @xmath90 , is the profile @xmath91 . the proof of the following lemma is straightforward . [ lm : compatsubprofile ] suppose @xmath0 is ancestrally compatible and let @xmath1 be a tree that ancestrally displays @xmath0 . then , for any @xmath92 , @xmath93 ancestrally displays @xmath94 . a semi - labeled tree @xmath95 is _ fully labeled _ if every node in @xmath24 is labeled . suppose @xmath0 contains trees that are not fully labeled . we can convert @xmath0 into an equivalent profile @xmath96 of fully - labeled trees as follows . for each @xmath82 $ ] , let @xmath97 be the number of unlabeled nodes in @xmath98 . create a set @xmath99 of @xmath100 } l_i$ ] labels such that @xmath101 . for each @xmath82 $ ] and each @xmath102 such that @xmath103 , make @xmath104 , where @xmath42 is a distinct element from @xmath99 . we refer to @xmath96 as the _ profile obtained by adding distinct new labels to @xmath0 _ ( see figure [ fig : profile ] ) . [ lm : fullyl ] let @xmath96 be the profile obtained by adding distinct new labels to @xmath0 . then , @xmath0 is ancestrally compatible if and only if @xmath96 is ancestrally compatible . further , if @xmath1 is a semi - labeled phylogenetic tree that ancestrally displays @xmath96 , then @xmath1 ancestrally displays @xmath0 . from this point forward , we shall assume that , for each @xmath82 $ ] , @xmath10 is fully and singularly labeled . by lemma [ lm : fullyl ] , no generality is lost in assuming that all trees in @xmath0 are fully labeled . the assumption that the trees are singularly labeled is inessential ; it is only for clarity . note that , even with the latter assumption , a tree that ancestrally displays @xmath0 is not necessarily singularly labeled . figure [ fig : supertree ] illustrates this fact . the _ display graph _ of a profile @xmath0 , denoted @xmath14 , is the graph obtained from the disjoint union of the underlying trees @xmath105 by identifying nodes that have the same label . multiple edges between the same pair of nodes are replaced by a single edge . see figure [ fig : displaygraph ] . @xmath14 has @xmath106 nodes and edges , and can be constructed in @xmath106 time . by our assumption that all the trees in @xmath0 are fully and singularly labeled , there is a bijection between the labels in @xmath84 and the nodes of @xmath14 . thus , from this point forward , we refer to the nodes of @xmath14 by their labels . it is easy to see that if @xmath14 is not connected , then @xmath0 decomposes into label - disjoint sub - profiles , and that @xmath0 is compatible if and only if each sub - profile is compatible . thus , we shall assume , without loss of generality , that @xmath14 is connected . [ [ positions . ] ] positions . + + + + + + + + + + a _ position _ ( for @xmath0 ) is a vector @xmath107 , where @xmath108 , for each @xmath82 $ ] . since labels may be shared among trees , we may have @xmath109 , for @xmath110 $ ] with @xmath111 . for each @xmath82 $ ] , let @xmath112 , and let @xmath113 } { \ensuremath{\mathrm{desc}}}_i(u)$ ] . a position @xmath114 is _ valid _ if , for each @xmath82 $ ] , 1 . [ item : v1 ] if @xmath115 , then the elements of @xmath116 are siblings in @xmath10 and 2 . [ item : v2 ] @xmath117 . [ lem : v2 ] for any valid position @xmath114 , @xmath118 . by ( v[item : v2 ] ) , we have that @xmath119 and @xmath120 are isomorphic , for each @xmath82 $ ] . the lemma then follows from the definition of @xmath121 . for any valid position @xmath114 , @xmath122 denotes the subgraph of @xmath14 induced by @xmath123 . [ obs : update1 ] for any valid position @xmath114 , @xmath122 is the subgraph of @xmath14 obtained by deleting all labels in @xmath124 , along with all incident edges . a valid position of special interest to us is @xmath125 , where @xmath126 , for each @xmath82 $ ] . that is , @xmath127 is a singleton containing only the label of @xmath128 . thus , in figure [ fig : displaygraph ] , @xmath129 . it is straightforward to verify that @xmath125 is indeed valid , that @xmath130 , and that @xmath131 . [ [ semi - universal - labels . ] ] semi - universal labels . + + + + + + + + + + + + + + + + + + + + + + let @xmath114 be a valid position , and let @xmath42 be a label in @xmath114 . then , @xmath42 is _ semi - universal in @xmath114 _ if @xmath132 , for every @xmath82 $ ] such that @xmath133 . it can be verified that in figure [ fig : displaygraph ] , labels @xmath134 and @xmath135 are semi - universal in @xmath125 , but @xmath136 is not , since @xmath136 is in both @xmath137 and @xmath138 , but @xmath139 . the term `` semi - universal '' , borrowed from peer et al . @xcite , derives from the following fact . suppose that @xmath0 is ancestrally compatible , that @xmath1 is a tree that ancestrally displays @xmath0 , and that @xmath42 is a semi - universal label for some valid position @xmath114 . then , as we shall see , @xmath42 must label the root @xmath140 of a subtree of @xmath1 that contains all the descendants of @xmath42 in @xmath10 , for every @xmath141 such that @xmath133 . the qualifier `` semi '' is because this subtree may also contain labels that do not descend from @xmath42 in any input tree , but descend from some other semi - universal label @xmath43 in @xmath114 instead . in this case , @xmath43 also labels @xmath140 . this property of semi - universal labels is exploited in both our ancestral compatibility algorithm and its proof of correctness ( see section [ sec : testac ] ) . for each label @xmath142 , let @xmath143 denote the number of input trees that contain label @xmath42 . we can obtain @xmath143 for every @xmath142 in @xmath106 time during the construction of @xmath14 . [ lem : semiu ] let @xmath144 be a valid position . then , label @xmath42 is semi - universal in @xmath114 if the cardinality of the set @xmath145 : u(i ) = \{\ell\}\}$ ] equals @xmath143 . by definition , @xmath132 , for every @xmath146 . since @xmath147 , the lemma follows . [ [ successor - positions . ] ] successor positions . + + + + + + + + + + + + + + + + + + + + for every @xmath82 $ ] and every @xmath133 , let @xmath148 denote the set of children of @xmath42 in @xmath149 . for a subset @xmath63 of @xmath149 , let @xmath150 . let @xmath114 be a valid position , and @xmath151 be the set of semi - universal labels in @xmath114 . the _ successor of @xmath114 with respect to @xmath151 _ is the position @xmath152 defined as follows . for each @xmath153 and each @xmath82 $ ] , if @xmath132 , then @xmath154 ; otherwise , @xmath155 . in figure [ fig : displaygraph ] , the set of semi - universal labels in @xmath125 is @xmath156 . since @xmath157 and @xmath158 , the successor of @xmath125 is @xmath159 . [ obs : update2 ] let @xmath114 be a valid position , and let @xmath152 be the successor of @xmath114 with respect to the set @xmath151 of semi - universal labels in @xmath114 . then , @xmath160 can be obtained from @xmath122 by doing the following for each @xmath153 : ( 1 ) for each @xmath82 $ ] such that @xmath132 , delete all edges between @xmath42 and @xmath148 ; ( 2 ) delete @xmath42 . let @xmath114 be a valid position , and @xmath161 be a subset of @xmath123 . then , @xmath162 denotes the position @xmath163 . in figure [ fig : displaygraph ] , the components of @xmath160 , where @xmath152 is the successor of @xmath125 , are @xmath164 and @xmath165 . thus , @xmath166 and @xmath167 . we have the following result . [ lm : childofu ] let @xmath114 be a valid position , and @xmath151 be the set of all semi - universal labels in @xmath114 . let @xmath152 be the successor of @xmath114 with respect to @xmath151 , and let @xmath168 be the label sets of the connected components of @xmath160 . then , @xmath169 is a valid position , for each @xmath170 $ ] . it suffices to argue that @xmath152 satisfies conditions ( v[item : v1 ] ) and ( v[item : v2 ] ) . the lemma then follows from the fact that the connected components of @xmath160 are label - disjoint . @xmath152 must satisfy condition ( v[item : v1 ] ) , since @xmath114 does . suppose @xmath153 . then , for each @xmath82 $ ] such that @xmath133 , @xmath171 and @xmath172 . thus , since ( v[item : v2 ] ) holds for @xmath114 , it also holds for @xmath152 . @xmath16 ( algorithm [ alg : slbuild ] ) is our algorithm for testing compatibility of semi - labeled trees . its argument , @xmath114 , is a valid position in @xmath0 such that @xmath173 is connected . line [ alg : findsemi ] computes the set @xmath151 of semi - universal labels in @xmath114 . if @xmath151 is empty , then , as argued in theorem [ thm : buildcorrect ] below , @xmath174 is incompatible , and , thus , so is @xmath0 . this fact is reported in line [ alg : s0emptyincompat ] . line [ alg : s0singleton ] checks if @xmath151 contains exactly one label @xmath42 , with no proper descendants . if so , by the connectivity assumption , @xmath42 must be the only element in @xmath123 . therefore , line [ alg : s0singletonreturn ] simply returns the tree with a single node , labeled @xmath42 . line [ alg : updateu ] updates @xmath114 , replacing it by its successor with respect to @xmath151 . let @xmath175 be the connected components of @xmath122 after updating @xmath114 . by lemma [ lm : childofu ] , @xmath176 is a valid position , for each @xmath170 $ ] . lines [ alg : connected][alg : recurseend ] recursively invoke @xmath177 on @xmath176 for each @xmath170 $ ] , to determine if there is a tree @xmath178 that ancestrally displays @xmath179 . if any subproblem is incompatible , line [ alg : recurseend ] reports that @xmath0 is incompatible . otherwise , lines [ alg : ru][alg : finalreturn ] assemble the @xmath178s into a single tree that displays @xmath174 , whose root is labeled by the semi - universal labels in the set @xmath151 of line [ alg : findsemi ] . let @xmath180 [ alg : findsemi ] replace @xmath114 by the successor of @xmath114 with respect to @xmath151.[alg : updateu ] let @xmath168 be the connected components of @xmath122 [ alg : connected ] create a node @xmath181 , whose label set is @xmath151 [ alg : ru ] the tree with root @xmath181 and subtrees @xmath182 [ alg : finalreturn ] next , we argue the correctness of @xmath177 . [ thm : buildcorrect ] let @xmath183 be a profile and let @xmath184 , where , for each @xmath82 $ ] , @xmath126 . then , @xmath185 returns either ( i ) a semi - labeled tree @xmath1 that ancestrally displays @xmath0 , if @xmath0 is ancestrally compatible , or ( ii ) ` incompatible ` otherwise . \(i ) suppose that @xmath185 outputs a semi - labeled tree @xmath1 . we prove that @xmath1 ancestrally displays @xmath0 . by lemma [ lm : dn ] , it suffices to show that @xmath186 and @xmath187 , for each @xmath82 $ ] . consider any @xmath188 . then , @xmath42 has a child @xmath189 in @xmath10 such that @xmath190 . there must be a recursive call to @xmath191 , for some valid position @xmath114 , where @xmath42 is the set @xmath151 of semi - universal labels obtained in line [ alg : findsemi ] . by observation [ obs : update2 ] , label @xmath189 , and thus @xmath43 , both lie in one of the connected components of the graph obtained by deleting all labels in @xmath151 , including @xmath42 , and their incident edges from @xmath122 . it now follows from the construction of @xmath1 that @xmath192 . thus , @xmath186 . now , consider any @xmath193 . let @xmath13 be the lowest common ancestor of @xmath194 and @xmath195 in @xmath10 and let @xmath196 be the label of @xmath13 . then , @xmath196 has a pair of children , @xmath197 and @xmath198 say , in @xmath10 such that @xmath199 , and @xmath200 . because @xmath185 returns a tree , there are recursive calls @xmath201 and @xmath202 for valid positions @xmath203 and @xmath204 such that @xmath197 is semi - universal for @xmath203 and @xmath198 is semi - universal for @xmath204 . we must have @xmath205 ; otherwise , @xmath206 , and , thus , neither @xmath197 nor @xmath198 is semi - universal , a contradiction . further , it follows from the construction of @xmath1 that we must have @xmath207 . hence , @xmath50 , and , therefore , @xmath208 . \(ii ) asssume , by way of contradiction , that @xmath185 returns ` incompatible ` , but that @xmath0 is ancestrally compatible . by assumption , there exists a semi - labeled tree @xmath1 that ancestrally displays @xmath0 . since @xmath185 returns ` incompatible ` , there is a recursive call to @xmath191 for some valid position @xmath114 such that @xmath114 has no semi - universal label , and the set @xmath151 of line [ alg : findsemi ] is empty . by lemma [ lm : compatsubprofile ] , @xmath209 ancestrally displays @xmath174 . thus , by lemma [ lem : v2 ] , @xmath209 ancestrally displays @xmath210 , for every @xmath82 $ ] . let @xmath42 be any label in the label set of the root of @xmath209 . then , for each @xmath82 $ ] such that @xmath211 , @xmath42 must be the label of the root of @xmath119 . thus , for each such @xmath141 , @xmath132 . hence , @xmath42 is semi - universal in @xmath114 , a contradiction . here we describe an efficient implementation of @xmath177 . we focus on two key aspects : finding semi - universal labels in line [ alg : findsemi ] , and updating @xmath114 and @xmath122 in lines [ alg : updateu ] and [ alg : connected ] . by observation [ obs : update1 ] , at each recursive call , @xmath177 deals with a graph obtained from @xmath14 through edge and node deletions . to handle these deletions efficiently , we represent @xmath14 using the dynamic graph connectivity data structure of holm et al . @xcite , which we refer to as _ hdt_. hdt allows us to maintain the list of nodes in each component , as well as the number of these nodes so that , if we start with no edges in a graph with @xmath212 nodes , the amortized cost of each update is @xmath213 . since @xmath14 has @xmath106 nodes , each update takes @xmath214 time . the total number of edge and node deletions performed by @xmath185 including all deletions in the recursive calls is at most the total number of edges and nodes in @xmath14 , which is @xmath106 . hdt allows us to maintain connectivity information throughout the entire algorithm in @xmath215 time . as deletions are performed on @xmath14 , @xmath177 maintains three data fields for each connected component @xmath216 that is created : @xmath217 , @xmath218 , and @xmath219 . it also maintains a field @xmath220 , for each @xmath142 . 1 . @xmath217 equals @xmath221 . 2 . @xmath218 is a map from a set @xmath222 $ ] to a set of nonempty subsets of @xmath223 . for each @xmath224 , @xmath225 denotes the set associated with @xmath141 . 3 . @xmath220 equals the cardinality of the set @xmath226 : y.{\ensuremath{\mathtt{map}}}(i ) \text { is defined and } y.{\ensuremath{\mathtt{map}}}(i ) = \{\ell\}\}$ ] . ( recall that @xmath143 is the number of input trees that contain @xmath42 . ) @xmath227 is a set containing all labels @xmath228 such that @xmath229 . informally , each set @xmath225 corresponds to a non - empty @xmath116 ; @xmath227 corresponds to the semi - universal labels in @xmath216 . next , we formalize these ideas . at the start of the execution of @xmath191 for any valid position @xmath114 , @xmath122 has a single connected component , @xmath230 . our implementation maintains the following invariant . inv : : : at the beginning of the execution of @xmath191 , @xmath231 for each @xmath82 $ ] such that @xmath232 , and @xmath233 is undefined for each @xmath82 $ ] such that @xmath234 . thus , @xmath220 equals the number of indices @xmath82 $ ] such that @xmath132 . along with lemma [ lem : semiu ] , inv implies that , at the beginning of the execution of @xmath191 , @xmath235 contains precisely the semi - universal labels of @xmath114 . thus , the set @xmath151 of line [ alg : findsemi ] of @xmath191 can be retrieved in @xmath236 time . to establish inv for the initial valid position @xmath125 , we proceed as follows . by assumption , @xmath237 has a single connected component , @xmath238 . since @xmath237 equals @xmath14 , we initialize data fields 14 for @xmath239 during the construction of @xmath14 . @xmath240 is simply @xmath241 . for each @xmath82 $ ] , @xmath242 is @xmath243 , where @xmath42 is the label of the root of @xmath98 . we initialize the @xmath244 fields as follows . first , set @xmath220 to @xmath245 for all @xmath142 . then , iterate through each @xmath82 $ ] , incrementing @xmath220 by one if @xmath246 . finally , @xmath247 consists of all @xmath248 such that @xmath229 . all data fields can be initialized in @xmath106 time . we now focus on lines [ alg : updateu ] and [ alg : connected ] of @xmath177 . by observation [ obs : update2 ] , we can update @xmath114 and @xmath122 jointly as follows . we use a temporary variable @xmath249 . prior to executing line [ alg : updateu ] , we set @xmath250 . then , we successively consider each label @xmath153 , and perform two steps : ( i ) initialize data fields 14 in preparation for the deletion of @xmath42 and ( ii ) delete from @xmath249 the edges incident on @xmath42 and then @xmath42 itself , updating data fields 14 as necessary , to maintain inv . after these steps are executed , @xmath249 will equal @xmath122 for the new set @xmath114 created by line [ alg : updateu ] . steps ( i ) and ( ii ) are done by @xmath251 ( algorithm [ alg : initmarks ] ) and @xmath252 ( algorithm [ alg : dellabel ] ) , respectively . lines [ alg : delsemi][alg : addpairmap ] of @xmath251 initialize @xmath218 and @xmath227 to reflect the fact that label @xmath153 is leaving @xmath116 , for each @xmath82 $ ] such that @xmath253 , to be replaced by its children in @xmath10 , and will no longer be semi - universal . lines [ alg : initsingleton][alg : foraddalphasemi ] are needed to update certain @xmath254 fields due to the possibility that singleton sets @xmath225 may be created in the preceding steps . the number of operations on @xmath218 performed by @xmath251 is @xmath255 : \ell \in l({\ensuremath{\mathcal{t}}}_i ) } |{{\ensuremath{{\mathrm{ch}}}}}_i(\ell)|)$ ] ; i.e. , it is proportional to the total number of children of @xmath42 in all the input trees . since @xmath42 is considered only once , the total number of operations on @xmath256 fields of the various sets @xmath216 considered over the entire execution of @xmath185 is @xmath106 . the number of updates of @xmath218 done by @xmath251 is at most @xmath143 ; the total over all @xmath42 considered by @xmath185 over its entire execution is @xmath106 . delete @xmath42 from @xmath227 [ alg : delsemi ] @xmath252 begins by consulting hdt to identify the connected component @xmath216 that currently contains @xmath42 . the loop in lines [ alg : forscanedges][alg : endforscanedges ] successively deletes each edge between @xmath42 and a child @xmath257 of @xmath42 , updating the appropriate data fields for the resulting connected components . line [ alg : testconn ] queries the hdt data structure to determine whether deleting @xmath258 splits @xmath216 into two components . if @xmath216 remains connected , no updates are needed the * continue * statement skips the rest of the current iteration and proceeds directly to the next . otherwise , @xmath216 is split into two parts @xmath259 and @xmath260 . @xmath261 uses a weighted version of the technique of scanning the smaller component @xcite . line [ alg : smallerweight ] identifies which of the two components has the smaller @xmath262 field ; without loss of generality , it assumes that @xmath263 . lines [ alg : delinity1 ] and [ alg : delinity2 ] initialize @xmath264 and @xmath265 to @xmath266 and @xmath267 and @xmath268 to the corresponding fields of @xmath216 . lines [ alg : forbetay1][alg : addtoj ] , scan each label @xmath269 in @xmath259 , from @xmath270 to @xmath271 , for every @xmath141 such that @xmath272 . set @xmath273 , updated in line [ alg : addtoj ] , keeps track of the indices @xmath141 such that @xmath271 and @xmath270 are modified . lines [ alg : forupdatesemi][alg : endforscanedges ] iterate through @xmath273 to determine if any new singleton sets were created in either @xmath259 or @xmath260 . this operation requires at most one update in each of @xmath265 and @xmath268 ; each update takes @xmath236 time . after all edges incident on @xmath42 are deleted , @xmath42 itself is deleted ( line [ alg : delell ] ) . the preceding description of @xmath252 omits the updating of the @xmath274 fields of the connected components created by an edge deletion . this is done before line [ alg : smallerweight ] , by ( again ) using the technique of scanning the smaller component . we consult hdt to determine which of @xmath259 and @xmath260 has fewer labels . assuming , without loss of generality , that @xmath275 , compute @xmath276 in a sequential scan of @xmath259 . then , @xmath277 . let @xmath216 be the connected component of @xmath249 that contains @xmath42 [ alg : findy ] delete @xmath42 from @xmath249 [ alg : delell ] let us track the number of operations on @xmath256 fields in lines [ alg : forbetay1][alg : addtoj ] of @xmath252 that can be attributed to some specific label @xmath278 over the entire execution of @xmath185 . each execution of lines [ alg : forbetay1][alg : addtoj ] for @xmath269 performs @xmath279 operations on @xmath256 fields . let @xmath280 be the weight of the connected component containing @xmath269 at the beginning of the loop of lines [ alg : forbetay1][alg : addtoj ] , at the @xmath17th time that @xmath269 is considered in those lines ; thus , @xmath281 . then , @xmath282 . the reason is that we only consider @xmath269 if ( i ) @xmath269 is contained in one of the two components that result from deleting an edge in line [ alg : deleteedge ] and ( ii ) the component containing @xmath269 has the smaller weight of the two components . thus , the number of times @xmath269 is considered in lines [ alg : forbetay1][alg : addtoj ] over the entire execution of @xmath185 is @xmath283 , which is @xmath284 , since @xmath285 . therefore , the total number of updates of @xmath256 fields over all labels is @xmath286 , which is @xmath287 . it can be verified that the number of updates to @xmath244 and @xmath288 fields is also @xmath287 . a similar analysis shows that the total time to update @xmath274 fields over all edge deletions performed by @xmath185 is @xmath287 . to summarize , the work done by @xmath185 consists of three parts : ( i ) initialization , ( ii ) maintaining connected components , and ( iii ) maintaining the @xmath262 , @xmath289 , and @xmath288 , and fields for each connected component , as well as @xmath220 for each label @xmath42 . part ( i ) takes @xmath106 time . part ( ii ) involves @xmath106 edge and node deletions on the hdt data structure , at an amortized cost of @xmath214 per deletion . part ( iii ) requires a total of @xmath287 updates to the various fields . using data structures that take logarithmic time per update , leads to our main result . [ thm : slbuilda ] @xmath16 can be implemented so that @xmath16@xmath290 runs in @xmath215 time . like our earlier algorithm for compatibility of ordinary phylogenetic trees , the more general algorithm presented here , @xmath16 , is a polylogarithmic factor away from optimality ( a trivial lower bound is @xmath291 , the time to read the input ) . @xmath16 has a linear - space implementation , using the results of thorup @xcite . a question to be investigated next is the performance of the algorithm on real data . another important issue is integrating our algorithm into a synthesis method that deals with incompatible profiles . o. r. p. bininda - emonds , m. cardillo , k. e. jones , r. d. e. macphee , r. m. d. beck , r. grenyer , s. a. price , r. a. vos , j. l. gittleman , and a. purvis . the delayed rise of present - day mammals . , 446:507512 , 2007 . p. daniel and c. semple . supertree algorithms for nested taxa . in o. r. p. bininda - emonds , editor , _ phylogenetic supertrees : combining information to reveal the tree of life _ , pages 151171 . kluwer , dordrecht , 2004 . c. e. hinchliff , s. a. smith , j. f. allman , j. g. burleigh , r. chaudhary , l. m. coghill , k. a. crandall , j. deng , b. t. drew , r. gazis , k. gude , d. s. hibbett , l. a. katz , h. d. laughinghouse iv , e. j. mctavish , p. e. midford , c. l. owen , r. h. reed , j. a. reesk , d. e. soltis , t. williams , and k. a. cranston . synthesis of phylogeny and taxonomy into a comprehensive tree of life . , 112(41):1276412769 , 2015 . r. m. page . taxonomy , supertrees , and the tree of life . in o. r. p. bininda - emonds , editor , _ phylogenetic supertrees : combining information to reveal the tree of life _ , pages 247265 . kluwer , dordrecht , 2004 .

the present article is a continuation of our works @xcite . by using the lax operator formalism , in @xcite we constructed a family of pairwise commuting operators such that the jack symmetric functions of the infinitely many variables @xmath2 are their eigenfunctions . we expressed these commuting operators in terms of the power sum symmetric functions @xmath3 where @xmath4 . the jack symmetric functions can be regarded as limiting cases of macdonald symmetric functions . the latter depend on the same variables @xmath2 and also on two parameters @xmath5 . in @xcite we extended the results of @xcite to the latter setting . in particular , by again using the lax operator formalism we constructed a family of pairwise commuting operators such that the macdonald symmetric functions are their eigenfunctions . in @xcite we expressed these commuting operators in terms of the hall - littlewood symmetric functions of the variables @xmath6 and of the parameter @xmath7 . these expressions involve only the hall - littlewood symmetric functions corresponding to the partitions with one part , see subsection [ sec:111 ] here . shortly after @xcite was published , a.n.sergeev and a.p.veselov communicated to us their remarkable works @xcite where in particular they found essentially the same commuting operators as we did in @xcite . their approach was different however . they first computed the limits at @xmath8 of the heckman operators @xcite acting on all polynomials in the variables @xmath9 . these @xmath0 operators do not commute in general . but the restrictions of the power sums of these @xmath0 operators to the space of symmetric polynomials do commute . moreover , jack symmetric polynomials are eigenfunctions of these restrictions . jack symmetric functions are then eigenfunctions of the limits of these restrictions at @xmath1 . -5 mm here we extend this approach from jack to macdonald symmetric functions . it has been discovered by i.v.cherednik @xcite that the macdonald polynomials in the variables @xmath10 are eigenfunctions of power sums of some @xmath0 commuting operators , acting on all polynomials in these variables . these operators are called the cherednik operators , see our subsection [ sec:13 ] for their definition . it has been also known @xcite that the cherednik operators have limits at @xmath1 . however , explicit expressions for these limits are unknown . we offer a solution to this open problem by firstly introducing for the macdonald polynomials the analogues of non - commuting heckman operators , see our subsection [ sec:15 ] . these analogues act on the rational functions of @xmath10 and are denoted by @xmath11 . they are related to the cherednik operators by proposition [ sec:15 ] . the principal property of the operators @xmath12 is stated as theorem [ sec:16 ] , see also proposition [ sec:13 ] . an explanation is needed here regarding our scheme of referring to lemmas , propositions , theorems and corollaries . when referring to these , we indicate the subsections where they respectively appear . there is no more than one of each of these in every subsection , so our scheme should cause no confusion . for example , proposition [ sec:13 ] is the only proposition that appears in subsection [ sec:13 ] . in subsection [ sec:17 ] we reformulate theorem [ sec:16 ] by introducing a certain @xmath13 matrix @xmath14 with operator entries acting on the rational functions of the variables @xmath9 . it is closely related to the classical lax matrix of the trigonometric ruijsenaars - schneider model @xcite . to be precise , let @xmath15 be the operator defined by . in the classical limit @xmath16 , when the canonical commutation relation @xmath17 degenerates to the poisson bracket @xmath18 , our @xmath14 degenerates to this lax matrix up to a change of variables and up to conjugation by a diagonal matrix . in the same classical limit , the macdonald determinant degenerates to the characteristic determinant of the lax matrix . thus we have found a way to derive a quantum analogue of this lax matrix directly from the cherednik operators . it is well known how the quantum hamiltonians @xcite of the trigonometric ruijsenaars - schneider model are related to the macdonald operators , see for instance @xcite . but our generating series for quantum hamiltonians differs from the determinant and is new . a similar resolvent - type expression was used for the quantum calogero - sutherland model in @xcite . the eigenstates of the latter model are the jack symmetric polynomials . the limit at @xmath8 of the generalisation of that model to particles with spin has been studied in @xcite as another extension of our work @xcite . following the approach of @xcite in subsection [ sec:31 ] of our article we compute the limits at @xmath8 of the operators @xmath11 . then we also compute the limit of the restriction of the operator sum appearing in theorem [ sec:16 ] to the space of symmetric polynomials in @xmath9 . this limit is a formal power series in another variable @xmath19 with operator coefficients acting on the symmetric functions of @xmath6 . after renormalisation and a change of the variable @xmath20 , this limit becomes the same generating series of the pairwise commuting operators as we constructed in @xcite . for details , see subsections [ sec:32 ] and [ sec:33 ] here . in this article we generally keep to the notation of the book @xcite for symmetric functions . when using results from @xcite we simply indicate their numbers within the book . for example , the statement ( 6.9 ) from chapter i of the book will be referred to as [ i.6.9 ] assuming it is from @xcite . fix a field @xmath21 . for any positive integer @xmath22 denote by @xmath23 the @xmath24-algebra of symmetric polynomials in @xmath0 variables @xmath9 . the algebra @xmath23 is graded by the polynomial degree . the substitution @xmath25 defines a homomorphism @xmath26 preserving the degree . here the inverse limit of the sequence @xmath28 in the category of graded algebras is denoted by @xmath29 . note that we get a canonical homomorhism @xmath30 . the elements of the algebra @xmath31 are called _ symmetric functions_. following @xcite we now will introduce some standard bases of @xmath29 . let @xmath32 be any partition of @xmath33 . the number of non - zero parts is called the _ length _ of @xmath34 and is denoted by @xmath35 . if @xmath36 then the sum of all distinct monomials obtained by permuting the @xmath0 variables in @xmath37 is denoted by @xmath38 . the symmetric polynomials @xmath39 with @xmath36 form a basis of the vector space @xmath40 . by definition , for @xmath36 @xmath41 where we write @xmath42 instead of @xmath35 . here @xmath43 is the symmetric group permuting the numbers @xmath44 and @xmath45 if @xmath46 are the respective multiplicites of the parts @xmath47 of @xmath48 . further , @xmath49 0 & \quad\textrm{if}\quad\,\ell({\lambda})=n\ , . \end{array } \right.\ ] ] hence for any fixed partition @xmath34 the sequence of polynomials @xmath39 with @xmath50 has a limit in @xmath29 . this limit is called the _ monomial symmetric function _ corresponding to @xmath48 . simply omitting the variables , we will denote the limit by @xmath51 . with @xmath34 ranging over all partitions of @xmath52 the symmetric functions @xmath53 form a basis of the vector space @xmath29 . note that if @xmath54 then we set @xmath55 . we will be also using another standard basis of the vector space @xmath29 . for each @xmath56 denote @xmath57 . when the index @xmath58 is fixed the sequence of symmetric polynomials @xmath59 with @xmath60 has a limit in @xmath29 , called the _ power sum symmetric function _ of degree @xmath61 . we will denote it by @xmath62 . more generally , for any partition @xmath34 put @xmath63 where @xmath64 as above . the elements @xmath65 form another basis of @xmath29 . equivalently , the elements @xmath66 are free generators of the commutative algebra @xmath31 over @xmath21 . in this article we will be using the _ natural ordering _ of partitions . by definition , here @xmath67 if @xmath34 and @xmath68 are partitions of the same number and @xmath69 this is a partial ordering . note that by [ i.6.9 ] any monomial symmetric function @xmath70 is a linear combination of the symmetric functions @xmath65 where @xmath71 . define a bilinear form @xmath72 on @xmath31 by setting for any two partitions @xmath34 and @xmath68 @xmath73 in the above notation . this form is obviously symmetric and non - degenerate . we will indicate by the superscript @xmath74 the operator conjugation relative to this form . in particular , by for the operator conjugate to the multiplication in @xmath31 by @xmath75 with @xmath76 we have @xmath77 next put @xmath78 for any fixed @xmath58 the sequence of the symmetric polynomials @xmath79 with @xmath60 has a limit in @xmath29 , denoted by @xmath80 and called the _ elementary symmetric function _ of degree @xmath81 we will also use a formal power series in another variable @xmath82 , @xmath83 by taking logarithms of the left and right hand side of the above display and then exponentiating , @xmath84 also put @xmath85 where the sum is taken over partitions @xmath34 of @xmath81 then the sequence of symmetric polynomials @xmath86 with @xmath60 has a limit in @xmath29 , denoted by @xmath87 and called the _ complete symmetric function _ of degree @xmath61 . by [ i.2.6 ] for the series @xmath88 we have the relation @xmath89 hence implies @xmath90 the elements @xmath91 as well as the elements @xmath92 are free generators of the commutative algebra @xmath31 over the field @xmath21 . we will also use the _ vertex operator _ @xmath93 it follows from and that for any @xmath56 we have the equality @xmath94 it also follows from and that @xmath95 $ ] is a homomorphism of @xmath24-algebras . see [ i.5 , example 29 ] for both of the last two statements . hence by applying @xmath96 to any symmetric function in the variables @xmath97 we get the same symmetric function but in the variables @xmath98 . let @xmath24 be the field @xmath99 with @xmath100 a parameter . the hall - littlewood symmetric functions [ iii.2.11 ] are labelled by all partitions of @xmath52 and constitute another remarkable basis of the vector space @xmath31 over @xmath21 . in the present article we will use only the elements of this basis corresponding to the single part partitions @xmath101 . these elements will be denoted by @xmath102 respectively . their generating series is @xmath103 by using and we get the relation @xmath104 while by using and we get the relation @xmath105 now let @xmath24 be the field @xmath106 where @xmath107 and @xmath100 are formal parameters . then define a bilinear form @xmath108 on @xmath31 by setting @xmath109 for any partitions @xmath34 and @xmath110 . this form is symmetric and non - degenerate . if @xmath111 , it specializes to the form defined by . we will indicate by the superscript @xmath112 the operator conjugation relative to @xmath113 . in particular , by and @xmath114 for any @xmath115 . hence by using we get @xmath116 note that by using , the latter identity can be rewritten as @xmath117 similarly to @xmath96 the map @xmath118 $ ] is a homomorphism of @xmath24-algebras . by [ vi.4.7 ] there exists a unique family of elements @xmath119 such that @xmath120 and such that any @xmath121 equals @xmath53 plus a linear combination of the elements @xmath122 with @xmath123 in the natural partial ordering . the elements @xmath124 are called the _ macdonald symmetric functions_. by [ vi.4.10 ] the canonical homomorphism @xmath125 maps @xmath126 if @xmath127 . if @xmath36 then the image of @xmath124 under the homomorphism @xmath125 is the _ macdonald symmetric polynomial _ usually denoted by @xmath128 . all these polynomials with @xmath129 form a basis of the vector space @xmath23 over @xmath130 . let @xmath131 as in subsection [ sec:1111 ] . for @xmath132 the _ inverse q - shift operator _ @xmath15 acts on any rational function @xmath133 by @xmath134 denote by @xmath135 the _ vandermonde polynomial _ of @xmath0 variables @xmath136{\phantom{\big[}\!\!}_{i , j=1}^n= \prod_{1{\leqslant}i < j{\leqslant}n}({\hspace{0.75pt}}x_i - x_j)\,.\ ] ] put @xmath137{\phantom{\big[}\!\!}_{i , j=1}^n \hspace{-8pt}\ ] ] where @xmath19 is another variable . the last determinant is defined as the alternated sum @xmath138 where as usual @xmath139 denotes the sign of permutation @xmath140 . in every product over @xmath132 appearing in the alternated sum all the operator factors pairwise commute , hence their ordering does not matter . note that @xmath141 . by the @xmath142 is a polynomial of degree @xmath0 in the variable @xmath19 with operator coefficients . it also follows from that these coefficients map the space @xmath23 to itself . by [ vi.3.3 ] for any @xmath143 the coefficient of @xmath142 at @xmath144 equals @xmath145 where the sum is taken over all subsets @xmath146 of @xmath147 of size @xmath148 , whereas @xmath149 now for every @xmath143 consider the restriction of the operator to the space @xmath40 . by [ vi.4.16 ] all these restrictions to @xmath23 pairwise commute . they are called the _ macdonald operators_. the macdonald polynomials @xmath150 with @xmath36 make a common eigenbasis of these operators . by [ vi.4.15 ] the eigenvalue of @xmath142 corresponding to any such eigenvector @xmath150 is @xmath151 note that our definition of the @xmath142 differs from [ vi.3.2 ] by changing the parameters @xmath152 and @xmath153 . however , by [ vi.4.14 ] the macdonald polynomials @xmath150 are invariant under this change of their parameters . after this change , we also replaced the variable @xmath154 used in the definition [ vi.3.2 ] by @xmath155 . the reasons for these alterations will be explained in subsection [ sec:31 ] . for @xmath156 with @xmath157 introduce the operator acting on the vector space @xmath158 @xmath159 where @xmath160 acts by exchanging the variables @xmath161 and @xmath162 . it is immediately obvious from the definition that the operator @xmath163 maps polynomials in the variables @xmath10 to polynomials . further , one can check that @xmath164 the _ cherednik operators _ @xmath165 acting on @xmath166 $ ] are then defined by @xmath167 these operators pairwise commute . in general , they do not map the space @xmath23 to itself . however , any symmetric polynomial of the operators @xmath165 with the coefficients from the field @xmath24 does . moreover by ( * ? ? ? * subsection 1.3.5 ) we have the action of @xmath142 on @xmath23 coincides with that of the product @xmath168 in accord with the remark we made at the end of previous subsection , the operator @xmath169 differs from the operator defined by ( * ? ? ? * equation 1.3.32 ) by changing the parameters @xmath152 and @xmath153 . our normalisation of @xmath169 is also different . we will use the following property of operators . for @xmath170 let @xmath171 be the cherednik operators acting on @xmath172 $ ] instead of @xmath166\,$ ] . the action of @xmath173 on the space @xmath23 coincides with the action of @xmath174 first let us prove by the downward induction on @xmath175 that the action of on the space @xmath23 coincides with the action of the product @xmath176 where the factors corresponding to the indices @xmath177 are arranged from left to right . if @xmath178 then neither of the products and has any factors , so the statement to prove is trivial . now assume that our statement is already proved for some @xmath179 . consider the product obtained from by replacing the index @xmath42 by @xmath180 . by the induction assumption , the action on @xmath23 of the so obtained product coincides with that of @xmath181 the last @xmath182 factors of the cherednik operator @xmath183 appearing in @xmath184 commute with @xmath185 and @xmath15 for any @xmath186 . they also act trivially on @xmath40 . after removing these @xmath182 factors from @xmath183 in we get the product @xmath187 thus we are making the induction step , and our statement is proved for any @xmath148 . by using this statement in the particular case when @xmath188 , the action of on the space @xmath23 coincides with the action of the product @xmath189 by applying the latter result to the set of operators instead of @xmath165 we obtain that for @xmath190 the action of on @xmath23 coincides with that of @xmath191 the last displayed product equals by renaming @xmath192 to @xmath193 . but we had also proved that the action of on @xmath23 coincides with the action of . for @xmath156 with @xmath157 denote @xmath194 so that by @xmath195 define operators @xmath12 acting on @xmath196 by setting @xmath197 where @xmath198 in general , these operators do not map polynomials in the variables @xmath10 to polynomials . but by definition , these operators make a _ covariant set _ relative to the action of the symmetric group @xmath199 by permutations of the @xmath200 note that for @xmath201 the operators @xmath165 on @xmath202 $ ] do not enjoy the covariance property . on the other hand , our @xmath12 do not commute . for @xmath170 let @xmath203 $ ] be the subspace of polynomials symmetric in the variables @xmath204 . then @xmath205\,.\ ] ] now consider the cherednik operator @xmath206 acting on @xmath166\,$ ] . our definition of the operator @xmath207 originates from the following proposition . the actions of the operators @xmath206 and @xmath207 on @xmath208 coincide . we will prove that the action of @xmath206 on @xmath209 coincides with the action of @xmath210 we will use the downward induction on @xmath211 . our proposition will be then obtained when @xmath212 . if @xmath213 then by the definition we have @xmath214 as required . now assume that our statement is proved for some @xmath215 . since @xmath216 we then know in particular that the action of @xmath206 on the space @xmath217 coincides with the action of the product . the latter product can be rewritten as @xmath218 ( { \hspace{0.75pt}}a_{{\hspace{0.75pt}}1k}+b_{{\hspace{0.75pt}}1k}\,{\sigma}_{{\hspace{0.75pt}}1k}{\hspace{0.75pt}})\,\ , \bigl(\ \prod_{k<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n}a_{{\hspace{0.75pt}}1l}\,+\ ! \sum_{k < j{\leqslant}n}b_{{\hspace{0.75pt}}1j}\ , \bigl(\ \prod_{\substack{k<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n\\l\neq j}}a_{{\hspace{0.75pt}}jl}{\hspace{0.75pt}}\bigr)\,{\sigma}_{{\hspace{0.75pt}}1j}{\hspace{0.75pt}}\bigr)\,{\gamma}_{{\hspace{0.75pt}}1}\,.\end{gathered}\ ] ] in its turn , the expression in the last displayed line can be rewritten as @xmath219 \sum_{k < j{\leqslant}n}\ , ( { \hspace{0.75pt}}a_{{\hspace{0.75pt}}1k}\,b_{{\hspace{0.75pt}}1j}+b_{\,1k}\,b_{{\hspace{0.75pt}}kj}\,{\sigma}_{{\hspace{0.75pt}}1k}{\hspace{0.75pt}})\,\ , \bigl(\ \prod_{\substack{k<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n\\l\neq j}}a_{{\hspace{0.75pt}}jl}{\hspace{0.75pt}}\bigr)\,{\sigma}_{{\hspace{0.75pt}}1j}{\hspace{0.75pt}}\bigr)\,{\gamma}_{{\hspace{0.75pt}}1}\,.\end{gathered}\ ] ] here none of the indices of the factor @xmath220 can be equal to @xmath221 or @xmath148 , because @xmath222 and @xmath223 . further , here @xmath224 where the factor @xmath225 commutes with the operator @xmath226 on @xmath166 $ ] and acts trivially on the subspace . thus by the identity @xmath227 the action of the operator @xmath206 on @xmath217 coincides with the action of @xmath228 \bigl(\ \prod_{k{\leqslant}{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n}\!a_{{\hspace{0.75pt}}1l}\,+\ , b_{{\hspace{0.75pt}}1k}\ , \bigl(\ \prod_{k<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n}a_{{\hspace{0.75pt}}kl}{\hspace{0.75pt}}\bigr)\,{\sigma}_{{\hspace{0.75pt}}1k}\,+\ ! \sum_{k < j{\leqslant}n}b_{{\hspace{0.75pt}}1j}\ , \bigl(\ \prod_{\substack{k{\leqslant}{\hspace{0.75pt}}l{\leqslant}{\hspace{0.75pt}}n\\l\neq j}}a_{{\hspace{0.75pt}}jl}{\hspace{0.75pt}}\bigr)\,{\sigma}_{{\hspace{0.75pt}}1j}{\hspace{0.75pt}}\bigr)\,{\gamma}_{{\hspace{0.75pt}}1 } \\ \ = r_{{\hspace{0.75pt}}12}\ldots r_{{\hspace{0.75pt}}1,k-1}\,\ , \bigl(\ \prod_{k-1<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n}a_{{\hspace{0.75pt}}1l}\,+\ ! \sum_{k-1<j{\leqslant}n}b_{{\hspace{0.75pt}}1j}\ , \bigl(\ \prod_{\substack{k-1<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n\\l\neq j}}a_{{\hspace{0.75pt}}jl}{\hspace{0.75pt}}\bigr)\,{\sigma}_{{\hspace{0.75pt}}1j}{\hspace{0.75pt}}\bigr)\,{\gamma}_{{\hspace{0.75pt}}1}\,.\end{gathered}\ ] ] thus we have made the downward induction step . we have already noted that for any @xmath132 the cherednik operator @xmath169 maps the polynomials in the variables @xmath10 to polynomials . on the other hand , the operator @xmath229 commutes with those permutations of the variables that preserve @xmath230 . by using these two observations when @xmath231 , our proposition implies both operators @xmath206 and @xmath207 map the space @xmath208 to itself . our main result of the current section is the theorem below . define operators @xmath232 acting on @xmath233 by setting @xmath234 similarly to @xmath12 the operators @xmath232 make a covariant set relative to the action of the group @xmath199 by permutations of the variables @xmath9 : @xmath235 the action of the ratio @xmath236 on @xmath23 coincides with the action of the sum @xmath237 we will relate operators on the space @xmath233 by the symbol @xmath238 if their actions on the subspace @xmath23 coincide . in subsection [ sec:12 ] we already noted that the coefficients of the polynomial @xmath142 map the space @xmath23 to itself . let us multiply by @xmath142 on the right both the ratio and the sum appearing in our theorem , and then subtract @xmath142 from the results . we get to prove the relation @xmath239 in the notation of subsection [ sec:12 ] the left hand side of the relation equals @xmath240 now consider the summand at the right hand side of with the index @xmath241 . by proposition [ sec:13 ] the action of this summand on @xmath23 coincides with that of @xmath242 where we used proposition [ sec:15 ] and then lemma [ sec:14 ] in the particular case @xmath212 . hence by applying proposition [ sec:13 ] once again , but to the cherednik operators @xmath243 instead of @xmath165 we obtain that the summand at the right hand side of the relation with the index @xmath244 acts on @xmath23 as @xmath245 here @xmath246 ranges over all subsets of @xmath247 of size @xmath180 . it follows that the summands at the right hand side of with @xmath248 act on @xmath23 as the operators obtained from via conjugation by @xmath249 respectively . thus the right hand side of acts on @xmath23 as the operator sum of the form @xmath250 where @xmath146 ranges over all subsets of @xmath147 of size @xmath148 , and each @xmath251 is a certain rational function of the variables @xmath10 over the field @xmath252 . to prove the relation it now suffices to demonstrate that for each @xmath146 @xmath253 moreover , because both sides of are invariant under conjugation by the elements of @xmath254 , it suffices to verify only in the case when @xmath255 . note that in the latter case the left hand side of equals @xmath256 now consider the right hand side of in the case when @xmath255 . let us denote it by @xmath257 for short . the contribution to @xmath257 from corresponds to the set @xmath258 and hence equals @xmath259 \label{idtwo } ( \,t-1{\hspace{0.75pt}})\ , \bigl(\ \prod_{1<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}k}\,a_{{\hspace{0.75pt}}1l}\ , \bigr)\,{\hspace{0.75pt}}&t^{\,k{\hspace{0.75pt}}(k-1)/2 } \!\!\prod_{\substack{1{\leqslant}{\hspace{0.75pt}}j{\hspace{0.75pt}}{\leqslant}k\\k<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n}}\!\!a_{{\hspace{0.75pt}}jl}\,.\end{aligned}\ ] ] if we conjugate by any @xmath260 with @xmath261 , the result will make contribution to @xmath262 only when @xmath263 , and this contribution will correspond to @xmath264 . indeed , then we will need @xmath246 in such that @xmath265 . hence for each index @xmath266 we get a contribution to @xmath257 @xmath267 \label{three } ( \,t-1{\hspace{0.75pt}})\ , \bigl(\ \prod_{\substack{1{\leqslant}{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}k\\l\neq i}}\,a_{{\hspace{0.75pt}}il}\ , \bigr)\,{\hspace{0.75pt}}&t^{\,k{\hspace{0.75pt}}(k-1)/2 } \!\!\prod_{\substack{1{\leqslant}{\hspace{0.75pt}}j{\hspace{0.75pt}}{\leqslant}k\\k<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n}}\!\!a_{{\hspace{0.75pt}}jl}\,.\end{aligned}\ ] ] by dividing , , by @xmath268 and by cancelling there all the common factors @xmath220 the relation now reduces to the identity @xmath269 the latter identity is easy to verify , and we omit the details of verification . consider the operator sum over @xmath132 appearing in . denote by @xmath270 the restriction of this operator sum to the subspace @xmath271 . the @xmath270 expands as a formal power series in @xmath19 with coefficients acting on @xmath40 . our theorem means that the action of the coefficients of the series @xmath272 on @xmath23 coincides with the action of the respective coefficients of @xmath273 . the latter ratio should be also expanded as a formal power series in @xmath19 here . the coefficients of the series @xmath270 will be called the _ quantum hamiltonians _ corresponding to the basis of macdonald polynomials in the vector space @xmath40 . in the next subsection we give another expression for @xmath270 by using the resolvent of a certain @xmath13 matrix with operator entries which act on @xmath274 . take any @xmath275 and consider the column vector @xmath276 now define a @xmath13 matrix @xmath14 with operator entries acting on the vector space @xmath196 as follows . the @xmath277-entry @xmath278 of the matrix @xmath14 is defined by setting @xmath279 then by using the definition we have @xmath280 where for @xmath281 we also use the relation @xmath282 . it follows from that @xmath283 z_{{\hspace{0.75pt}}2}\,{\hspace{0.75pt}}{\sigma}_{{\hspace{0.75pt}}12}\,(f ) \\ \vdots \\[2pt ] \ { \hspace{0.75pt}}z_n\,{\sigma}_{{\hspace{0.75pt}}1n}{\hspace{0.75pt}}(f ) \ \end{bmatrix } \,=\,{\mathcal{z}}\,{\mathcal{f}}\,.\ ] ] indeed , in its first entry the vector equality holds by with @xmath241 . if @xmath284 then by using we get @xmath285 as required . here for three pairwise distinct indices @xmath286 we also use the relations @xmath287 by the the covariance property of the operators @xmath12 the column vector at the left hand side of has a form similar to @xmath288 . namely , it can be obtained by replacing the polynomial @xmath289 in @xmath290 by @xmath291 . here we use corollary [ sec:15 ] . by expanding @xmath292 for every @xmath132 as a formal power series in @xmath19 and by repeatedly using the above arguments , we get the equality @xmath293 ( \,1+u\,z_{{\hspace{0.75pt}}2}{\hspace{0.75pt}})^{{\hspace{0.75pt}}-1}\,{\sigma}_{{\hspace{0.75pt}}12}\,(f ) \\ \vdots \\[4pt ] \ ( \,1+u\,z_n{\hspace{0.75pt}})^{{\hspace{0.75pt}}-1}\,{\sigma}_{{\hspace{0.75pt}}1n}{\hspace{0.75pt}}(f)\ \end{bmatrix } \,=\,(\,1+u\,{\mathcal{z}}\,)^{{\hspace{0.75pt}}-1}\,{\mathcal{f}}\,.\ ] ] now suppose @xmath294 so that the polynomial @xmath289 is symmetric in all the variables @xmath295 . then @xmath296 so that @xmath297 where @xmath298 is the column vector of size @xmath0 with every entry being @xmath299 . let @xmath300 be the row vector of size @xmath0 where the @xmath301-entry is the @xmath302 defined by . by using and the definition of the series @xmath270 as given in subsection [ sec:16 ] @xmath303 thus we have proved that the action of @xmath270 on @xmath23 coincides with the action of @xmath304 hence we now obtain the following corollary to theorem [ sec:16 ] . the action of the ratio @xmath236 on @xmath23 coincides with that of @xmath305 let @xmath131 as before . we will find first the inverse limit at @xmath8 of the restriction of the operator @xmath207 to the subspace @xmath306 by proposition [ sec:15 ] the operator @xmath206 has the same restriction to @xmath208 . the limit will be an operator acting on the space @xmath307 $ ] and denoted simply by @xmath308 . to define the limit extend the canonical homomorphism @xmath125 to a homomorphism @xmath309\,\to\,{\lambda}_n^{{\hspace{0.75pt}}(1)}:\ , v\,\mapsto\,x_1\,.\ ] ] here @xmath310 we will now define an operator @xmath311 on the vector space @xmath307 $ ] explicitly . denote by @xmath312 and @xmath313 the automorphisms of the @xmath24-algebra @xmath307 $ ] which act trivially on the subalgebra @xmath31 but map the variable @xmath314 to @xmath315 and @xmath316 respectively . thus @xmath312 is the inverse @xmath317-shift of @xmath314 while @xmath313 is the usual @xmath7-shift . next equip the vector space @xmath318 $ ] with the standard inner product so that @xmath319 form an orthonormal basis . denote by @xmath320 the operator on @xmath318 $ ] conjugate to multiplication by @xmath82 . explicitly , @xmath321 0 & \quad\textrm{if}\quad\,n=0\ , . \end{array } \right.\ ] ] extend the operator @xmath320 from @xmath318 $ ] to @xmath307 $ ] by @xmath322-linearity . the extension will still be denoted by @xmath320 . for every @xmath323 extend from @xmath31 to @xmath307 $ ] by @xmath318{\hspace{0.75pt}}$]-linearity the operator of multiplication by @xmath289 and its conjugate operator @xmath324 . recall that the superscript @xmath74 here indicates conjugation relative to the inner product . the conjugate @xmath325 relative to the inner product extends from @xmath31 to @xmath307 $ ] in the same way as @xmath326 does . using these conventions , put @xmath327 where @xmath328 we have a commutative diagram of @xmath24-linear mappings @xmath329 { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{r}[above=1pt]{z } \arrow[xshift=0pt , yshift=0pt]{d}[left=1pt]{\pi_n } & { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{d}[right=2pt]{\pi_n } \\ { \lambda}_n^{{\hspace{0.75pt}}(1 ) } \arrow[xshift=0pt , yshift=0pt]{r}[below=2pt]{z_1 } & { \lambda}_n^{{\hspace{0.75pt}}(1 ) } \end{tikzcd } \vspace{2pt}\ ] ] we will verify commutativity of the two diagrams obtained from by replacing @xmath330 respectively by @xmath331 and @xmath332 . our theorem will then follow . firstly observe that the extended operator @xmath333\to{\lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt}}]$ ] appearing in the definition of @xmath334 is a homomorphism of @xmath24-algebras , and so is @xmath335\to { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt}}]\,$ ] . hence it suffices to show that the compositions @xmath336 and @xmath337 coincide on @xmath314 and on @xmath75 for @xmath338 . by applying the compositions to @xmath314 we get the same result : @xmath339 v \arrow[mapsto]{r}[below=2pt]{{\gamma } } & q^{{\hspace{0.75pt}}-1}{\hspace{0.75pt}}v \arrow[mapsto]{r}[below=2pt]{\pi_n } & q^{{\hspace{0.75pt}}-1}{\hspace{0.75pt}}x_1 \end{tikzcd } \quad\text{and}\quad \begin{tikzcd}[column sep=32pt ] v \arrow[mapsto]{r}[below=2pt]{\pi_n } & x_1 \arrow[mapsto]{r}[below=2pt]{{\gamma}_{{\hspace{0.75pt}}1 } } & q^{{\hspace{0.75pt}}-1}{\hspace{0.75pt}}x_1\ , . \end{tikzcd}\ ] ] to check the coincidence on any generator @xmath75 note that by the identity @xmath340 hence due to by applying @xmath336 and @xmath337 to @xmath75 we also get the same result : @xmath339 p_n \arrow[mapsto]{r}[below=2pt]{{\gamma } } & q^{{\hspace{0.75pt}}-n}{\hspace{0.75pt}}v^n - v^n+p_n \arrow[mapsto]{r}[below=2pt]{\pi_n } & q^{{\hspace{0.75pt}}-n}{\hspace{0.75pt}}x_1+x_2^n+\ldots+x_n^n \end{tikzcd } \vspace{-8pt}\ ] ] and @xmath339 p_n \arrow[mapsto]{r}[below=2pt]{\pi_n } & x_1^n+\ldots+x_n^n \arrow[mapsto]{r}[below=2pt]{{\gamma}_{{\hspace{0.75pt}}1 } } & q^{{\hspace{0.75pt}}-n}{\hspace{0.75pt}}x_1+x_2^n+\ldots+x_n^n\ , . \end{tikzcd } \vspace{2pt}\ ] ] now consider the compositions @xmath341 and @xmath342 . by definition , the extended operator @xmath343\to { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt}}]$ ] commutes with multiplication by any @xmath344 . but @xmath345 by . in particular , the restriction of the operator @xmath346 to the subspace commutes with multiplication by @xmath347 . hence it suffices to show that the compositions @xmath341 and @xmath348 coincide on the elements @xmath349\,$ ] . let us use the generating series of these elements @xmath350 in the other variable @xmath20 . by applying @xmath341 to the series we get @xmath351 \displaystyle \frac1{\,1-u\,v\ , } \arrow[mapsto]{r}[below=2pt]{{\small w } } & \displaystyle \frac{\,q(u)}{\,1-u\,t\,v\ , } \arrow[mapsto]{r}[below=2pt]{\pi_n } & \displaystyle \frac1{\,1-u\,t\,x_1 } \,\,\prod_{i=1}^n\,{\hspace{0.75pt}}\frac{\,1-u\,t\,x_i\,}{1-u\,x_i}\ . \end{tikzcd}\ ] ] here we employed the general fact that for any formal power series @xmath352 with the coefficients from @xmath24 @xmath353 we have also employed the relation with the variable @xmath314 replaced by @xmath20 . on the other hand , by applying @xmath348 to the series we get @xmath339 \displaystyle \frac1{\,1-u\,v\ , } \arrow[mapsto]{r}[below=2pt]{{\pi_n } } & \displaystyle \frac{1}{\,1-u\,x_1\ , } \arrow[mapsto]{r}[below=2pt]{w_1 } & \displaystyle \frac1{\,1-u\,x_1 } \prod_{1<{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n}\frac{\,x_1-t\,x_l}{x_1-x_l } \ \,+ \end{tikzcd } \vspace{-2pt}\ ] ] @xmath354 here we also used and . it easy to verify that the results obtained in and in the last two displayed lines are the same . consider them as rational functions of @xmath19 and assume that @xmath355 . then both rational functions vanish at @xmath356 and have poles only at @xmath357 . all these poles are simple , and the corresponding residues of the two functions coincide . note that for any index @xmath358 one can also consider the restriction of the operator @xmath229 to the subspace of @xmath233 consisting of the polynomials in @xmath10 symmetric in all the variables but @xmath230 . by the covariance property our corollary [ sec:15 ] implies that the operator @xmath229 preserves this subspace . we could have defined the extension @xmath359 of the homomorphism @xmath125 from @xmath31 to @xmath307 $ ] by mapping the variable @xmath314 to @xmath161 instead of @xmath360 . the image of @xmath359 would be then the latter subspace of @xmath361 . the inverse limit of the restriction of @xmath229 to that subspace would be then the same operator @xmath311 acting on @xmath307{\hspace{0.75pt}}$ ] . this coincidence follows immediately from the property . it is the change of parameters @xmath152 and @xmath362 in the original definition [ vi.3.2 ] that allowed us to state the last theorem in terms of the hall - littlewood symmetric functions @xmath363 . otherwise we would have to change @xmath362 in the definition of the latter symmetric functions . the change of the variable @xmath154 in [ vi.3.2 ] and the corresponding choice of normalization of the operator @xmath169 as in and as in proposition [ sec:13 ] ensure that every @xmath229 has a limit at @xmath1 . in this subsection we will find the inverse limits at @xmath8 of the quantum hamiltonians corresponding to the basis of macdonald polynomials in @xmath40 . these quantum hamiltonians are defined as the operator coefficients of the series @xmath270 acting on the vector space @xmath40 , see the end of subsection [ sec:16 ] . we will denote by @xmath364 the inverse limit of the series @xmath365 . the coefficients of the series @xmath364 will be certain operators @xmath366 $ ] where @xmath367 is yet another formal variable . we will then eliminate the dependence of the coefficients on @xmath367 by renormalising the series @xmath368 . hence the coefficients of the renormalised series will be operators acting on @xmath29 . consider the sum over @xmath132 appearing in . by the action of this sum on the subspace @xmath369 coincides with that of @xmath370 where we set @xmath371 here @xmath372 . we will demonstrate that the operator @xmath373 maps the subspace to @xmath40 . at the same time we will determine the inverse limit at @xmath8 of the restriction of the operator @xmath373 to the subspace . the latter limit will be an operator @xmath307\to{\lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}w{\hspace{0.75pt}}]$ ] denoted simply by @xmath374 . to determine this limit extend the canonical homomorphism @xmath125 to a homomorphism @xmath375\to{\lambda}_n:\ , w\,\mapsto\,t^{{\hspace{0.75pt}}n}\,.\ ] ] here @xmath376 this definition of the homomorphism @xmath377 goes back to ( * ? ? ? * section 6 ) . now define @xmath374 explicitly as the unique @xmath322-linear operator @xmath307\to{\lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}w{\hspace{0.75pt}}]$ ] such that @xmath378 w-1 & \quad\textrm{if}\quad\,n=0\ , . \end{array } \right.\ ] ] we have a commutative diagram of @xmath24-linear mappings @xmath379 { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{r}[above=1pt]{v } \arrow[xshift=0pt , yshift=0pt]{d}[left=1pt]{\pi_n } & { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}w{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{d}[right=2pt]{\tau_{n } } \\ { \lambda}_n^{{\hspace{0.75pt}}(1 ) } \arrow[xshift=0pt , yshift=0pt]{r}[below=2pt]{v_1 } & \,{\lambda}_n^{\phantom{(1 ) } } \end{tikzcd } \vspace{2pt}\ ] ] the operator @xmath380\to{\lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}w{\hspace{0.75pt}}]$ ] commutes with the multiplication by any @xmath344 . in turn , the restriction of the operator @xmath381 to the subspace commutes with multiplication by @xmath347 . so it suffices to show that the compositions @xmath382 and @xmath383 coincide on the elements @xmath349\,$ ] . let us again use the generating series of these elements . by applying @xmath382 to we get @xmath384 \displaystyle \frac1{\,1-u\,v\ , } \arrow[mapsto]{r}[below=2pt]{{\small v } } & w - q(u ) \arrow[mapsto]{r}[below=2pt]{\tau_{{\hspace{0.75pt}}n } } & \displaystyle t^{{\hspace{0.75pt}}n}-{\hspace{0.75pt}}\,\prod_{i=1}^n\,{\hspace{0.75pt}}\frac{\,1-u\,t\,x_i\,}{1-u\,x_i } \end{tikzcd}\ ] ] where we used . on the other hand , by applying @xmath383 to we get @xmath385 \displaystyle \frac1{\,1-u\,v\ , } \arrow[mapsto]{r}[below=2pt]{{\pi_n } } & \displaystyle \frac{1}{\,1-u\,x_1\ , } \arrow[mapsto]{r}[below=2pt]{v_1 } & \displaystyle \,\,\sum_{i=1}^n\,\ , \frac{t-1}{\,1-u\,x_i } \,\prod_{\substack{1{\leqslant}{\hspace{0.75pt}}l{\hspace{0.75pt}}{\leqslant}n\\l\neq i}}\frac{\,x_i - t\,x_l}{x_i - x_l } \ . \end{tikzcd}\ ] ] the results obtained in and are equal to each other . indeed , consider them as rational functions of @xmath19 and assume that @xmath355 . then both rational functions vanish at @xmath356 and have poles at @xmath357 . these poles are simple , and the corresponding residues of two functions coincide . by the surjectivity of @xmath359 the last proposition implies that the operator @xmath373 maps the subspace to @xmath40 . moreover , it implies that the inverse limit at @xmath8 of the restriction of the operator sum to the subspace equals @xmath386 by the definitions of @xmath387 and @xmath374 here for every @xmath388 the composition @xmath389 is an operator @xmath307\to{\lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}w{\hspace{0.75pt}}]\,$ ] . the above stated equality of the inverse limit follows from the commutativity of the diagram @xmath390 { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{r}[above=1pt]{z^n } \arrow[xshift=0pt , yshift=0pt]{d}[left=1pt]{\pi_n } & { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{r}[above=1pt]{{\gamma}\ } \arrow[xshift=0pt , yshift=0pt]{d}[left=1pt]{\pi_n } & { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{r}[above=1pt]{v } \arrow[xshift=0pt , yshift=0pt]{d}[right=1pt]{\pi_n } & { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}w{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{d}[right=2pt]{\tau_{n } } \\ { \lambda}_n^{{\hspace{0.75pt}}(1 ) } \arrow[xshift=0pt , yshift=0pt]{r}[below=2pt]{z_1^n } & \!{\lambda}_n^{{\hspace{0.75pt}}(1 ) } \arrow[xshift=0pt , yshift=0pt]{r}[below=2pt]{{\gamma}_{{\hspace{0.75pt}}1 } } & { \lambda}_n^{{\hspace{0.75pt}}(1 ) } \arrow[xshift=0pt , yshift=0pt]{r}[below=2pt]{v_1 } & \,{\lambda}_n^{\phantom{(1 ) } } \end{tikzcd } \vspace{2pt}\ ] ] here we use the commutativity of , and that of the diagram obtained from by replacing @xmath330 respectively by @xmath391 . the commutativity of the diagram so obtained has been established as a part of our proof of theorem [ sec:31 ] . denote by @xmath392 the embedding of @xmath31 to @xmath307 $ ] as the subspace of degree zero in @xmath82 . then we have a commutative diagram @xmath393 { \lambda}{\phantom{[]}}\!\ ! \arrow[xshift=0pt , yshift=0pt]{r}[above=2pt]{{\delta } } \arrow[xshift=0pt , yshift=0pt]{d}[left=0pt ] { } & { \lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt } } ] \arrow[xshift=0pt , yshift=0pt]{d}[right=2pt]{\pi_n } \\ { \lambda}_n^{\phantom{(1)}}\!\!\ ! \arrow[xshift=0pt , yshift=0pt]{r}[below=0pt ] { } & { \lambda}_n^{(1 ) } \end{tikzcd}\ ] ] where the left vertical arrow is the canonical projection . the bottom horizontal arrow is the natural embedding . it follows that the inverse limit of @xmath270 equals @xmath394 by the above definition , every coefficient in the formal power series expansion of @xmath364 in @xmath19 is a certain operator @xmath366\,$ ] . now consider the series @xmath395 where the summand @xmath221 in front of @xmath396 stands for the embedding of @xmath31 to @xmath397 $ ] as the subspace of degree zero in @xmath398 . this should cause no confusion . in the next subsection we will show that the series does not depend on @xmath398 . hence the coefficients of this series will be operators mapping the vector space @xmath31 to itself . note that by the definition of the homomorphism @xmath399 and by the above given arguments , the series is equal to the inverse limit at @xmath8 of @xmath400 but by using the multiplicative formula , the eigenvalue of @xmath236 on the trivial macdonald polynomial @xmath401 corresponding to @xmath402 is @xmath403 hence our theorem [ sec:16 ] implies that the eigenvalue of on @xmath401 equals @xmath299 . by taking the limit at @xmath8 the eigenvalue of on the trivial macdonald symmetric function @xmath404 also equals @xmath299 . this explains the definition of . here we will use the vector space decomposition @xmath405={\lambda}\oplus v{\hspace{0.75pt}}{\lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}v{\hspace{0.75pt}}]\,.\ ] ] the second direct summand in will be called the _ truncated space_. relative to this decomposition the operator @xmath334 on @xmath307 $ ] is represented by the @xmath406 matrix with operator entries @xmath407 \ { \beta}&{\alpha}\ \end{bmatrix}\ ] ] where @xmath408 denotes the composition of the restriction of @xmath334 to the first summand in with the projection to the second summand . the map @xmath334 preserves the second summand , and @xmath409 denotes the restriction of @xmath334 to it . similarly , the operator @xmath410 on @xmath307 $ ] is represented by the @xmath406 matrix with operator entries @xmath411 \ 0&\,x\ , \end{bmatrix}\ ] ] where @xmath154 and @xmath412 respectively denote the compositions of the restriction of @xmath410 to the second summand in with the projections to the first and to the second summands . note that the operator @xmath327 is then represented by the product @xmath413 \ 0&\,x\ , \end{bmatrix } \begin{bmatrix } \ 1&\,0\ \\[2pt ] \ { \beta}&{\alpha}\ \end{bmatrix } = \begin{bmatrix } \ 1+y{\beta}&\ y{\alpha}\ \\[2pt ] \ x{\hspace{0.75pt}}{\beta}&\,x{\hspace{0.75pt}}{\alpha}\ \end{bmatrix } .\ ] ] by definition the operator @xmath380\to{\lambda}{\hspace{0.75pt}}[{\hspace{0.75pt}}w{\hspace{0.75pt}}]$ ] acts on the first direct summand in as multiplication by @xmath414 . the restriction of @xmath374 to the second direct summand does not depend on @xmath398 . thus it maps @xmath415 $ ] to @xmath29 . moreover , by the definitions of @xmath374 and @xmath410 this restriction coincides with the operator @xmath416 . hence relative to the operator @xmath374 is represented by the row with operator entries @xmath417 finally denote @xmath418 . this is an operator on the truncated space @xmath415\,$ ] . we shall call it the _ lax operator _ for the macdonald symmetric functions in @xmath29 . this terminology is justified by the following theorem . the series is equal to @xmath419 where @xmath420 in particular , the series does not depend on the variable @xmath398 . relative to the operator @xmath421 $ ] is represented by the column with two operator entries @xmath422 therefore the operator product @xmath423 appearing in the definition of the series @xmath364 is represented by the first column of the @xmath406 matrix inverse to @xmath424 \ u\,x{\hspace{0.75pt}}{\beta}&\,1+u\,x{\hspace{0.75pt}}{\alpha}\ \end{bmatrix } .\ ] ] here we employ the matrix representation of the operator @xmath308 . to find that column we will use a well known formula for the inverse of a @xmath406 block matrix with invertible diagonal blocks , see ( * ? ? * lemma 3.2 ) . the block matrix is assumed to be invertible too . the first entry of the first column that we find in this way is @xmath425 the second entry of the first column of the inverse matrix that we find is then @xmath426 the product @xmath427 in is represented by the row with operator entries @xmath428 \ { \beta}&{\alpha}\ \end{bmatrix } = \begin{bmatrix } \,w-1-y{\hspace{0.75pt}}{\beta}\ , \ , -{\hspace{0.75pt}}y{\alpha}\ , \end{bmatrix } .\ ] ] the series @xmath364 is equal to the product of this row by the column representing @xmath429 . that column has the entries and . hence @xmath364 equals @xmath430 our theorem immediately follows from the last displayed expression for @xmath368 . note that by replacing in the series @xmath431 the variable @xmath19 by @xmath432 we get the same generating series for the limits of the quantum hamiltonians at @xmath8 as was denoted in ( * ? ? ? * section 2 ) by @xmath368 . but in the present article the notation @xmath364 was introduced in and has a meaning different from that in @xcite . we are grateful to i.v.cherednik for illuminating conversations . the first named author was supported by the epsrc grant ep / n023919 , and by the programme _ geometry and representation theory _ at the erwin schrdinger institute . the second named author was supported by leverhulme senior research fellowship .

the phase - space formulation of quantum mechanics has received a renewed attention in the last decades @xcite because it allows to employ notions and tools of classical physics in the quantum realm . the central object in this approach is _ wigner s function _ @xmath0 which is merely a phase - space @xmath1 representation of the density matrix @xmath2 . in a closed system , the dynamical equation for the wigner function is equivalent to _ schrdinger s ( or von neumann ) equation _ while , in the @xmath3 limit , it recovers _ liouville s equation _ for the phase - space distribution in classical mechanics . another remarkable property of the wignerian representation is the _ average property_. the expectation value of a quantum operator @xmath4 is obtained from the corresponding classical variable @xmath5 ( related with @xmath4 by weyl s rule ( * ? ? ? * ch . 1 ) ) by a prescription analogous to the classical one , namely @xmath6 thus , in spite of not being necessarily positive , the wigner function can be considered as a sort of quantum - mechanical `` distribution '' . besides , the marginal distributions of @xmath7 give the true quantum probabilities of @xmath8 , as @xmath9 , or of @xmath10 , as @xmath11 . therefore , the wigner formalism provides a natural quantum - classical connection . this has been specially exploited in the search for quantum analogues of various classical effects ( e.g. , chaos ) or in the problem of the fuzzy borders between the quantum and classical worlds @xcite . in contact with the surrounding medium , the ( open ) system experiences dissipation , fluctuations , and decoherence @xcite . typically one is interested in a system consisting of a few relevant degrees of freedom coupled to its environment , or bath , which has a very large number of degrees of freedom ( e.g. , phonons , photons , nuclear spins , etc . ) . the system is not necessarily microscopic , but it can be a mesoscopic system described by some collective variables which under appropriate conditions can behave quantum mechanically @xcite . because of the genericity of these situations and the fundamental issues raised ( e.g. , approach to thermal equilibrium , dissipation in tunnelling and coherence , measurement problem ) , the study of _ quantum dissipative systems _ is of interest in several areas of physics and chemistry @xcite . the dynamics of an open system can in many cases be formulated in terms of a _ quantum master equation _ for the ( reduced ) density matrix . in the wigner representation this can be compactly written as @xmath12 , with @xmath13 a certain evolution operator . taking the classical limit , the master equation for a particle of mass @xmath14 in a potential @xmath15 goes over the klein kramers equation @xcite @xmath16 \wf % % % % % % % % % % % % % % % % % % % \;,\ ] ] which is simply a _ planck equation _ in phase space . the first two terms ( poisson bracket ) generate the classical reversible evolution ( liouville equation ) . the last term ( `` collision operator '' ) accounts for irreversible effects due to the coupling to the environment ( dissipation and fluctuations ) . the damping parameter @xmath17 measures the strengh of the coupling to the bath , which is at temperature @xmath18 . it is worth recalling that phase - space dynamics also includes problems reducible to `` mechanical '' analogues : josephson junctions , certain electrical circuits , chemical reactions , etc . a suitable non - perturbative technique to solve classical fokker planck equations of systems with few degrees of freedom is the _ continued - fraction method _ @xcite . this is a special case of the expansion into complete sets ( grad s ) method to solve kinetic equations in statistical mechanics @xcite . in brief , the distribution is expanded into a series of orthogonal polynomials and , as a result , the kinetic equation is transformed into an _ infinite set of coupled equations _ for the expansion coefficients @xmath19 . approximate solutions are obtained by truncating the hierarchy of equations at various levels . to get manageable expressions , however , the truncation needs to be performed at a low level . in the continued - fraction method , instead of truncating directly , one seeks for a basis in which the range of index coupling is short ( ideally , the equation for @xmath19 involves @xmath20 , @xmath19 , and @xmath21 ) . then , the ( differential ) recurrence relations among the @xmath19 can be solved by iterating a simple algorithm , the structure of which is like that of a continued fraction ( [ app : rr - cf ] ) . for the klein kramers equation ( [ kk ] ) the tridiagonal chain of coupled equations for the @xmath19 is called the _ brinkman hierarchy _ it has been solved by continued - fraction methods to study classical particles subjected to dissipation and fluctuations @xcite . the method has been also extended to rotational brownian motion problems involving classical spins and dipoles ( vd . @xcite and references therein ) . to deal with quantum dissipative systems , however , is a more delicate task @xcite . to begin with , phenomenological quantisation of dissipative systems poses fundamental problems ( e.g. , with the uncertainty and superposition principles ) . a rigorous route is to model the bath in a simple way , quantise it together with the system , and eventually trace over the environmental variables . however , the resulting theoretical frameworks are technically involved . _ quantum master equations _ , except for simple cases , are difficult to solve . an alternative is provided by _ quantum state diffusion _ methods ( vd . @xcite ) , where stochastic evolution equations for state vectors in hilbert space are introduced as computational tools . nevertheless , its implementation seems to be restricted to systems with discrete spectrum ( e.g. , oscillators , 2-state systems , etc . ) . similarly , _ quantum langevin equations _ @xcite are of limited use beyond nearly harmonic systems . for an arbitrary system , there exist exact _ path - integral _ expressions for the evolution of the density matrix ( involving the so - called _ influence functional _ that incorporates environmental effects ) @xcite . however , those expressions are difficult to evaluate , even numerically , because the propagating function is highly oscillatory , rendering numerical methods unstable at long times . finally , _ quantum monte carlo _ simulations can in principle be used . nevertheless , in spite of ongoing progress ( vd . @xcite ) , they are computationally complex and suffer from the ( dynamical ) sign problem . this situation strongly motivates the development of _ alternative methods _ for quantum dissipative systems . inspired on their suitability for classical systems , continued - fraction techniques have been developed for a number of problems . shibata and co - workers @xcite applied them to solve a c - number quantum fokker planck equation for a spin in a dissipative environment . vogel and risken @xcite employed continued - fraction methods to solve master equations in quantum nonlinear optics ( vd.also @xcite ) . recently @xcite this method has been extended to study genuine phase - space problems within the wigner formalism . a generalisation of the brinkman hierarchy for quantum master equations of the caldeira leggett type was presented . it was shown that the continued - fraction method for the classical problem can , in principle , be adapted to solve this hierarchy , yielding a promising technique to study several classes of nonlinear quantum systems subjected to environmental effects . in this article we first present the details necessary for the derivation of the quantum hierarchies and discuss their solution by continued fractions . the approach is then implemented for the problem of quantum brownian motion in periodic potentials ( a demanding problem with ( partly ) continuous spectrum ) . both the cosine and simple ratchet potentials ( lacking inversion symmetry ) are considered . for the former a number of previous results are recovered ( so testing the method ) and extended , while the physical interpretations are revisited within the wigner formalism . for particles in ratchet potentials we study the effects of finite damping ( inertia ) on the rectified velocities . the phenomenology is interpreted in terms of the interplay of thermal hopping , overbarrier wave reflection , and tunnelling . results for non - equilibrium dynamics under oscillating forcing are also discussed , which show the combined effect of quantum phenomena , thermal activation , and dissipation in a nonlinear system . a number of technical issues are consigned to the appendices . we start from the following generic form for the quantum master equation in the wigner representation @xcite : @xmath22 \wf % % % % % % % % % % % % % % % % % % \;. \;\ ] ] the first three terms , identifying @xmath23 , give the classical klein kramers equation ( [ kk ] ) . the mixed diffusion term @xmath24 is heuristically related with the colour of the quantum noise @xcite . the series of derivatives @xmath25 ( wigner moyal term ) gives the quantum contribution to the unitary evolution of the closed system . conditions under which this series can be truncated are sometimes discussed , to avoid dealing with an infinite - order partial differential equation . however , we shall fully keep this term because it can be specially important in nonlinear systems @xcite . conditions of validity for the master - equation description are discussed in @xcite . equations of the type of ( [ wkk : dxp ] ) can be derived modelling the particle surroundings as a _ bath of oscillators _ representing the normal modes of the environment . however , a number of assumptions are typically involved like semiclassical or high - temperature bath ( rendering the kernels of the influence functional local in time ) , weak system - bath coupling ( born markov approximations ) , etc . then , one would heuristically expect equation ( [ wkk : dxp ] ) being valid for small enough @xmath26 . notwithstanding this , the _ structure _ of the equation seems to be quite generic . thus , a quantum master equation recently derived for strong coupling @xcite , and valid for all @xmath26 , involves terms similar to those in ( [ wkk : dxp ] ) [ with @xmath8-dependent coefficients and renormalised @xmath15 ] . besides , the phase - space representation of the celebrated lindblad master equation is obtained by simply adding to ( [ wkk : dxp ] ) terms of the form @xmath28 and @xmath29 @xcite . most of these terms can be readily incorporated in the treatment below ( indeed , the term @xmath24 , absent in the original caldeira leggett equation , is included to illustrate this ) . the same , in principle , would apply to possible extensions of the quantum master equation ( [ wkk : dxp ] ) . first it is convenient to introduce appropriate scaled units . these are based on a characteristic length @xmath30 ( e.g. , the distance between the minima in a double - well oscillator , the period in a periodic potential ) and a characteristic energy @xmath31 ( e.g. , a barrier height ) . then one scales frequencies by @xmath32 , time by @xmath33 , forces by @xmath34 , the action by @xmath35 , etc . on the other hand , @xmath36 ( @xmath37 in a high-@xmath18 bath ) is handled as an effective temperature @xmath38 and we introduce some convenient thermal rescalings via @xmath39 . thus , in the equations @xmath10 is scaled by @xmath40 , the potential and @xmath41 appear divided by @xmath42 , @xmath17 enters in the combination @xmath43 , time is multiplied by @xmath44 , while the thermal de broglie wave length @xmath45 enters divided by @xmath30 . omitting any marks for the scaled quantities ( but keeping in mind specially the thermal rescaling of @xmath46 , @xmath15 and @xmath10 ) , the master equation is simply written as @xmath47 \wf % % % % % % % % % % % % % % % % % % % \ ; , \quad\ ] ] with the coefficient in the quantum sum given by @xmath48 planck s constant @xmath49 is introduced in terms of the characteristic action @xmath35 via the quantum parameter @xmath50 ( denoted @xmath51 in @xcite ) @xmath52 the second relation gives @xmath53 in terms of the kondo parameter @xmath54 , related by the third one with our @xmath50 . the classical limit is approached as @xmath55 , i.e. , letting @xmath56 . let us now introduce a splitting of the evolution operator @xmath13 that will facilitate the calculation of the matrix elements required when expanding @xmath57 into complete sets . on inspecting equations ( [ wkk : scaled ] ) and ( [ kappa : s ] ) , we see that extending the quantum sum down to @xmath58 , we obtain the part @xmath59 of the classical term . thus , we can decompose @xmath13 into the irreversible , kinetic , and potential parts : @xmath60 % % % % % % % % % % % % % % % % % % \lkin & = & -\big(\p-\bdxp\pp\big)\ , \px \\[0.5ex ] % % % % % % % % % % % % % % % % % % \lv & = & \sumiqo \qcoef^{(\iq ) } \ , \bv^{(2\iq+1 ) } \ , \pp^{(2\iq+1 ) } \end{array } \right . % % % % % % % % % % % % % % % % % % \;.\ ] ] this natural splitting has the advantage of dealing with @xmath15 ( the most problem dependent part ) in a unified way . besides , we have grouped the term @xmath61 with the kinetic term @xmath62 because they are structurally similar . in the expansion into complete sets approach to solve kinetic equations , the distribution is expanded in an orthonormal basis @xmath63 and the coupled equations for the coefficients @xmath64 derived @xcite . for the klein kramers equation ( [ kk ] ) one uses hermite functions of @xmath10 as basis ( * ? ? ? * sec 4.4 ) @xmath65^{-1/2 } % % % % % % % % % % % % % % % % % % \;.\ ] ] one starts with the expansion in the momentum because , while the @xmath10 dependence of the hamiltonian is fully specified ( kinetic energy @xmath66 ) , the potential part @xmath15 depends on the problem . thus , explicit manipulations can be done on the parts involving @xmath10 , which are valid for any system . on the other hand , the hermite basis has a number of advantages @xcite ; not the least is the handling of the derivatives @xmath67 in the dynamical equation by means of the associated creation and annihilation operators @xmath68 and @xmath69 ( this will prove very convenient in the quantum case ) . for the klein kramers equation the resulting equations for the expansion coefficients @xmath64 are called the _ brinkman hierarchy _ @xcite . this plays an important rle in classical systems , both for analytical treatments derivations of @xmath70 expansions , etc. and to develope efficient non - perturbative numerical methods . in fact , it has the structure of a @xmath71-term recurrence relation which , after expansion in the position basis , can be solved by continued fractions . in what follows we shall proceed in the quantum case following as closely as possible the steps of the classical limit . we start expanding the wigner function as @xmath72 where we have extracted the boltzmann - type factor @xmath73 . this involves the auxiliary parameter @xmath74 and potential @xmath75 , which is frequently chosen proportional to @xmath76 , i.e. , @xmath77 . the results should not depend on the @xmath78 or @xmath79 used , but these may ( i ) put parts of the evolution operator @xmath13 in self - adjoint form and/or ( ii ) improve the stability and convergence of the eventual numerical implementation . from the orthonormality of the @xmath80 , the expansion `` coefficients '' can be written as @xmath81 . then , differentiating with respect to @xmath46 and using the evolution equation @xmath12 , we get the dynamical equations for @xmath64 @xmath82 to get the matrix elements @xmath83 , which are still operators on the @xmath8-dependence of @xmath84 , one needs the @xmath85 transform of @xmath13 . this can be split as @xmath86 , with @xmath87 , @xmath88 , and @xmath89 given by ( [ wkk : lir : lkin : lv ] ) . the transformation of these operators is done in [ app : lb ] , taking advantage of results for _ normal ordering _ of @xmath90 and @xmath91 ( which we use to express the @xmath67 and @xmath10 s in @xmath13 ) . then , with the @xmath92 , @xmath93 , and @xmath94 obtained we compute their matrix elements between the @xmath95 , getting the @xmath83 . the details of this part of the calculation are given in [ app : qnm ] . introducing the results obtained for @xmath83 into @xmath96 one gets the following _ quantum ( brinkman ) hierarchy _ @xmath97 } \ , \big [ \g_{\ip}^{\iq,- } \ , \bv^{(2\iq+1 ) } \big ] \ , \ec_{\ip-(2\iq+1 ) } + \sqrt{(\ip-1)\ip}\ ; \gammamm \ , \ec_{\ip-2 } \nonumber\\ % % % % % % % % % % % % % % % % % % & & { } + \sqrt{\ip}\ ; \dm \ec_{\ip-1 } + \gammao_{\ip } \ , \ec_{\ip } + \sqrt{\ip+1}\ ; \dp \ec_{\ip+1 } \\ % % % % % % % % % % % % % % % % % % & & { } + \sqrt{(\ip+1)(\ip+2)}\ ; \gammapp \ , \ec_{\ip+2 } + \sum_{\iq=0}^{\infty } \big [ \g_{\ip+(2\iq+1)}^{\iq,+ } \ , \bv^{(2\iq+1 ) } \big ] \ , \ec_{\ip+(2\iq+1 ) } \nonumber % % % % % % % % % % % % % % % % % % \;.\end{aligned}\ ] ] the auxiliary ( damping ) parameters introduced read @xmath98 % [ ( \etap-\etapro)+2\ip(\etab-\etapro ) ] % % % % % % % % % % % % % % % % % % % \;,\ ] ] which involve the following @xmath78-related parameters ( note the sign exchange ) @xmath99 in equation ( [ qbh ] ) , the operators on the @xmath8 dependences are @xmath100 here @xmath101 , @xmath102 is the confluent hypergeometric ( kummer ) function ( * ? ? ? * ch . 13.6 ) , and @xmath103 is given by equation ( [ kappa : s ] ) . in is important to note that the @xmath104 act only on @xmath15 not on the @xmath105 ( [ app : lb ] ) . the quantum hierarchy ( [ qbh ] ) is equivalent to the caldeira leggett master equation ( [ wkk : dxp ] ) . previously , several hierarchies had been derived @xcite , but the treatment was for closed ( hamiltonian ) systems @xcite or involved unsuitable bases like @xmath106 @xcite . in contrast , our hierarchy of _ hermitian moments _ constitutes a direct quantum generalisation of the brinkman hierarchy . to verify this , note that in the classical case only the @xmath58 terms survive in the sums of ( [ qbh ] ) ( [ app : qnm ] ) . then , setting the usual @xmath107 , equation ( [ qbh ] ) reduces to @xmath108 % \;,\ ] ] which is indeed the celebrated hierarchy associated to the klein kramers equation . for potentials obeying @xmath109 , @xmath110 , the quantum hierarchy ( [ qbh ] ) is a recurrence relation with coupling to a finite number of terms . examples include the harmonic oscillator @xmath111 and the double - well ( duffing ) oscillator @xmath112 . finite - coupling recurrences are always reducible to @xmath71-term vector recurrences ( [ app : rr - cf ] ) and solvable by continued fractions , as in the classical case . however , for non - polynomial @xmath15 this approach is prevented by the infinite coupling range in the index @xmath113 , consequence of the infinite series of @xmath10-derivatives @xmath114 in the wigner moyal term . in what follows we shall exploit the expansion in the position basis to show how this problem can be circumvented . recall that the coefficients @xmath64 are still functions of @xmath8 . inserting their expansion in an orthonormal basis @xmath115 into the quantum hierarchy ( [ dcdt : qtot : lbar ] ) , we get the generic form @xmath116 with matrix elements @xmath117 . to put ( [ qbh ] ) into the above form we just need to ( i ) substitute the operators ( and @xmath8-dependent coefficients ) by their matrix elements @xmath118 , ( ii ) attach position superscripts to the @xmath64 , and , finally , ( iii ) sum over them . on so doing we get the _ expanded quantum hierarchy _ @xmath119_{\ix\ixp}\ , \ec_{\ip-(2\iq+1)}^{\ixp } + \sqrt{(\ip-1)\ip } \sum_{\ixp } \gammamm \ , \delta_{\ix\ixp } \ , \ec_{\ip-2}^{\ixp } \nonumber\\ % % % % % % % % % % % % % % % % % % \fl & { } + \sqrt{\ip } \sum_{\ixp } \dm_{\ix\ixp}\ , \ec_{\ip-1}^{\ixp } + \sum_{\ixp } \gammao_{\ip}\ , \delta_{\ix\ixp } \ , \ec_{\ip}^{\ixp } + \sqrt{\ip+1 } \sum_{\ixp } \dp_{\ix\ixp}\ , \ec_{\ip+1}^{\ixp } \\ % % % % % % % % % % % % % % % % % % \fl & { } + \sqrt{(\ip+1)(\ip+2 ) } \sum_{\ixp } \gammapp \ , \delta_{\ix\ixp } \ , \ec_{\ip+2}^{\ixp } + \sumiqo \sum_{\ixp } \big [ \g_{\ip+(2\iq+1)}^{\iq,+ } \bv^{(2\iq+1 ) } \big]_{\ix\ixp}\ , \ec_{\ip+(2\iq+1)}^{\ixp } \nonumber % % % % % % % % % % % % % % % % % % \;.\end{aligned}\ ] ] considering that @xmath120 and @xmath77 , we see that to get all matrix elements involved we just need those of @xmath121 and of @xmath122 , namely @xmath123 because the @xmath104 only increase the order of the potential derivatives . the expanded hierarchy has the form of a system of ordinary differential equations , @xmath124 . if @xmath125 and @xmath126 are some large truncation indices for the bases @xmath127 and @xmath115 , we have a problem with @xmath128 equations . one may be tempted to try direct runge kutta integration or numerical diagonalization of the associated @xmath129 matrix . however , the dimensions involved are typically very large , so that it is worth to pursue a continued - fraction treatment . as mentioned before the @xmath10-recurrence has a finite coupling range for polynomial potentials . for other potentials , with an appropriate choice of the basis @xmath115 , the matrix elements ( [ matr - elem : der : v ] ) may vanish when the second index @xmath130 lays at a certain `` distance '' of the first @xmath131 . then , the expanded hierarchy would be a recurrence relation in the position indices with a finite coupling range , also amenable for a continued - fraction treatment . to proceed in these cases , we just need to transform the two - index extended hierarchies into ordinary recurrence problems ( i.e. , with one index ) . in order to apply the continued - fraction method when finite coupling range in one of the indices is attained , we need to convert first the two - index recurrences into ordinary recursions . in the resulting one - index recurrences the coefficients will be matrices acting on appropriate vectors formed with the @xmath132 . the generic expanded form ( [ dcdt : generic ] ) can be converted into an _ one - index _ recurrence relation in the momentum index introducing the following vectors and matrices @xmath133 where @xmath134 . that is , for a fixed @xmath113 one forms the column vector with all the @xmath132 with @xmath135 ( this index may take positive and negative values ) . similarly , for given @xmath136 we build the matrix @xmath137 with all indices @xmath138 . before giving explicit expressions for the matrix coefficients we introduce the following notation : @xmath139 , @xmath140 , @xmath141 , and @xmath142 . then , the matrix quantum hierarchy ( [ dcdt : vectors : matrices : p ] ) can be written as @xmath143 % % % % % % % % % % % % % % % % % % & & { } + \mq_{\ip}^{- } \mc_{\ip-1 } + \mq_{\ip } \mc_{\ip } + \mq_{\ip}^{+ } \mc_{\ip+1 } \\[+0.5ex ] % % % % % % % % % % % % % % % % % % & & { } + \mq_{\ip}^{++ } \mc_{\ip+2 } + { \textstyle\sum}_{\iq\ge1 } \mq_{\ip}^{+(2\iq+1 ) } \mc_{\ip+(2\iq+1 ) } \nonumber % % % % % % % % % % % % % % % % % % \;.\end{aligned}\ ] ] we have accounted for @xmath144 , @xmath145 [ vd . equation ( [ qbh : pre - matrix : i ] ) ] and incorporated the @xmath58 terms into the central @xmath146 ones . the matrix coefficients explicitly read @xmath147_{\ix\ixp}\ , \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ip}^{-- } \big)_{\ix\ixp } & = & -\sqrt{(\ip-1)\ip}\ ; \gammamm\ , \delta_{\ix\ixp } \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ip}^{- } \big)_{\ix\ixp } & = & - \sqrt{\ip}\ , \dm_{\ix\ixp } - \big [ \g_{\ip}^{0,- } \bv ' \big]_{\ix\ixp}\ , \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ip } \big)_{\ix\ixp } & = & - \gammao_{\ip}\,\delta_{\ix\ixp } \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ip}^{+ } \big)_{\ix\ixp } & = & - \sqrt{\ip+1}\ , \dp_{\ix\ixp } - \big [ \g_{\ip+1}^{0,+ } \bv ' \big]_{\ix\ixp}\ , \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ip}^{++ } \big)_{\ix\ixp } & = & -\sqrt{(\ip+1)(\ip+2)}\ ; \gammapp\ , \delta_{\ix\ixp } \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ip}^{+(2\iq+1 ) } \big)_{\ix\ixp } & = & - \big [ \g_{\ip+(2\iq+1)}^{\iq,+ } \bv^{(2\iq+1 ) } \big]_{\ix\ixp}\ , % % % % % % % % % % % % % % % % % % \end{array } \right . % \;.\ ] ] for finite coupling range in @xmath113 ( e.g. , in polynomial potentials ) , the hierarchy ( [ qbh : matrix : p ] ) is a recurrence relation to which continued - fraction methods can now be applied . this amounts to convert the underlying large @xmath129 problem into @xmath125 problems with associated @xmath148 matrices [ or @xmath149 . as the recursion `` coefficients '' are matrices , one would need _ matrix continued - fraction _ methods ( [ app : rr - cf ] ) . when the @xmath10-recurrence has an infinite coupling range ( in periodic potentials , the morse potential , etc . ) , we can convert the @xmath150-index hierarchy ( [ dcdt : generic ] ) into matrix form , but introducing recurrences in the position index @xmath131 @xmath151 that is , @xmath152 , with @xmath153 . thus , for a fixed @xmath131 one constructs the column vector with the @xmath132 for all @xmath154 , and for given @xmath155 , the matrix @xmath137 with all @xmath156 . comparing the matrix in ( [ dcdt : vectors : matrices : x ] ) with equation ( [ q : matr - elem : p : i ] ) for @xmath157 , we can construct @xmath158 diagonal by diagonal @xmath159_{\ix\ixp}\ , \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ix\ixp } \big)_{\ip,\ip-2 } & = & -\sqrt{(\ip-1)\ip}\ ; \gammamm\ , \delta_{\ix\ixp } \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ix\ixp } \big)_{\ip,\ip-1 } & = & - \sqrt{\ip}\ , \dm_{\ix\ixp } - \big [ \g_{\ip}^{0,- } \bv ' \big]_{\ix\ixp}\ , \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ix\ixp } \big)_{\ip,\ip } & = & - \gammao_{\ip}\,\delta_{\ix\ixp } \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ix\ixp } \big)_{\ip,\ip+1 } & = & - \sqrt{\ip+1}\ , \dp_{\ix\ixp } - \big [ \g_{\ip+1}^{0,+ } \bv ' \big]_{\ix\ixp}\ , \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ix\ixp } \big)_{\ip,\ip+2 } & = & -\sqrt{(\ip+1)(\ip+2)}\ ; \gammapp\ , \delta_{\ix\ixp } \\ % % % % % % % % % % % % % % % % % % \big ( \mq_{\ix\ixp } \big)_{\ip,\ip+(2\iq+1 ) } & = & - \big [ \g_{\ip+(2\iq+1)}^{\iq,+ } \bv^{(2\iq+1 ) } \big]_{\ix\ixp}\ , % % % % % % % % % % % % % % % % % % \end{array } \right . % \;.\ ] ] all other diagonals being zero @xmath160 , @xmath145 . to recognise better the structure of these matrices , we decompose them into the `` free '' and potential contributions @xmath161 . for @xmath162 , including the kinetic and irreversible terms , we have the pentadiagonal structure ( tridiagonal for @xmath107 ) @xmath163 the part due to @xmath15 , on the other hand , has the following alternate dense structure , reflecting the odd powers of @xmath67 in the wigner moyal term : @xmath164_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{3}^{1,+ } \bv^{(3)}]_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{5}^{2,+ } \bv^{(5)}]_{\ix\ixp } \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-1 [ \g_{1}^{0,- } \bv']_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{2}^{0,+ } \bv']_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{4}^{1,+ } \bv^{(3)}]_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-2 0 \!\!\!\!&\!\!\!\ ! [ \g_{2}^{0,- } \bv']_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{3}^{0,+ } \bv']_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{5}^{1,+ } \bv^{(3)}]_{\ix\ixp } \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-3 [ \g_{3}^{1,- } \bv^{(3)}]_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{3}^{0,- } \bv']_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{4}^{0,+ } \bv']_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-4 0 \!\!\!\!&\!\!\!\ ! [ \g_{4}^{1,- } \bv^{(3)}]_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{4}^{0,- } \bv']_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{5}^{0,+ } \bv']_{\ix\ixp } \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-5 [ \g_{5}^{2,- } \bv^{(5)}]_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{5}^{1,- } \bv^{(3)}]_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! [ \g_{5}^{0,- } \bv']_{\ix\ixp } \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-6 \ddots \!\!\!\!&\!\!\!\ ! \ddots \!\!\!\!&\!\!\!\ ! \ddots \!\!\!\!&\!\!\!\ ! \ddots \!\!\!\!&\!\!\!\ ! \ddots \!\!\!\!&\!\!\!\ ! \ddots \!\!\!\!&\!\ ! \ddots \end{array } \right ) % % % % % % % % % % % % % % % % % % \ ] ] with an appropriate choice of the basis @xmath115 the matrix elements @xmath118 may couple only a finite number of basis functions . then , the recurrence ( [ dcdt : vectors : matrices : x ] ) would be solvable by continued fractions when the @xmath113-recurrence is infinite and this approach is seemingly precluded . even for finite coupling in @xmath113 , the @xmath131-recurrence may be preferable in certain situations ( e.g. , at low @xmath18 ) , as indicated by risken for a cosine potential in the classical limit ( * ? ? ? 11.5.6 ) ) . in these cases , one has been able to reduce the underlying large @xmath129 problem to @xmath126 problems with associated @xmath165 matrices . once the @xmath166 or @xmath167 are solved , we can reconstruct the wigner function from its expansion coefficients @xmath132 in the double basis @xmath168 @xmath169 obtaining the _ full solution of the quantum master equation_. if preferred , one can switch to the familiar density matrix @xmath170 by the inverse of transformation ( [ wigner : def ] ) @xmath171 we can get any observable inserting the above @xmath57 in equation ( [ average ] ) for the averages . nevertheless , common observables can many times be extracted _ directly _ from the @xmath132 . let us illustrate this with transport observables , which are characterised by the averages @xmath172 . for an arbitrary function of @xmath10 only , we get from ( [ w : expansion : xp ] ) @xmath173 where we have used @xmath174 , with @xmath175 [ equations ( [ hermite ] ) and ( [ w : expansion ] ) ] . next we introduce the auxiliary integrals ( for their calculation vd.[app : auxint ] ) @xmath176 then , the moments ( case @xmath177 ) can be written in terms of the expansion coefficients as @xmath178 . taking into account that @xmath179 has the parity of @xmath113 and carrying this to @xmath180 , we get for the first two @xmath181 the first moment is the _ current _ and the second characterises _ fluctuations _ around it . the zeroth moment corresponds to the _ normalisation _ of @xmath7 , and imposses on the @xmath132 the condition @xmath182 . finally , the quantum probabilities of @xmath8 and @xmath10 , given by the marginal distributions @xmath183 and @xmath184 , can be expressed in a similar fashion @xmath185 henceforth we shall apply the method described to incorporate fluctuations and dissipation in the problem of _ quantum transport in periodic potentials_. the simplest model consists of a particle evolving in a cosine potential subjected to external forces ( tilted `` washboard '' potential ) , which is also a paradigm of _ quantum brownian motion _ others include potentials lacking inversion symmetry ( ratchets ) , which have been used to model rectification of current and _ directional motion _ in several systems @xcite . these problems are demanding because periodic potentials have ( partly ) continuous spectra @xcite , rendering inappropriate methods devised for systems with discrete levels . in this section we compute the matrix coefficients of our recurrences for arbitrary periodic @xmath15 . with them we can implement the continued - fraction method to solve the quantum master equation , which we first check by regaining the classical dissipative and quantum hamiltonian limits . in periodic potentials the @xmath10-recurrence can not be used because of its infinite coupling range ( the derivatives of @xmath15 neither vanish , nor decrease , as they are essentially the potential itself ) . however , one can guess that the @xmath8-recurrence may have a coupling range related with the number of harmonics in @xmath15 . to exploit this we introduce plane waves as basis functions and the fourier expansion of the potential derivative : @xmath186 to preserve the periodicity of @xmath7 , one extracts the external force @xmath187 from @xmath15 including it in a generalised @xmath188 ; cf . equation ( [ dmdp ] ) . then , the auxiliary potential @xmath189 is set proportional to the periodic part @xmath77 . to compute the matrix elements in @xmath158 , we only need to do the integrals ( [ matr - elem : der : v ] ) with plane waves for @xmath190 . using @xmath191 , the integral @xmath192 , and the expansion @xmath193 , we find @xmath194_{\ix\ixp}=\v_{\ix-\ixp}^{(\iq ) } % % % % % % % % % % % % % % % % % % \;.\ ] ] from these elements and @xmath195 we get @xmath196_{\ix\ixp } & = & 0 & \delta_{\ix\ixp } & - & \etapm^{2\iq+1 } \gs_{\ip}^{\iq}(-\etapro \qab^{2}\ldb^{2 } ) & \v_{\ix-\ixp } ' \\ % % % % % % % % % % % % % % % % % % \dpm_{\ix\ixp } & = & \big(\iu\,\dplmi\ix-\etapm\bf\big ) & \delta_{\ix\ixp } & - & \dplmi\epspot & \bv_{\ix-\ixp } ' % % % % % % % % % % % % % % % % % % \end{array } % % % % % % % % % % % % % % % % % % \ ] ] with @xmath197 , @xmath101 and [ cf . equation ( [ gamma : qcoeff ] ) ] @xmath198 that is , @xmath199 is the coefficient in ( [ gamma : qcoeff ] ) with @xmath200 . inserting equation ( [ dmdpvper : matr - elem ] ) into the general matrices @xmath158 of section [ sec : mqbh ] we explicitly get the matrix coefficients of our recurrences ( [ app : matrices ] ) . for the cosine potential @xmath201 the coupling range is @xmath202 , because @xmath203 for @xmath204 and we have a @xmath71-term recurrence @xmath205 . for a @xmath150-harmonic ratchet potential @xmath206 % % % % % % % % % % % % % % % % % % % \;,\ ] ] the range is @xmath150 ( @xmath207-term recurrence ) because @xmath203 for @xmath208 , and so on . as we restricted @xmath74 , we have @xmath209 , so that @xmath210 . thus , the argument of @xmath211 is positive and @xmath212 can act as a regularisation factor [ an advantage of allowing @xmath213 in ( [ w : expansion ] ) ] . this factor reduces the weight of the off - diagonal terms inside @xmath158 , enhancing the numerical stability when going into the deep quantum regime . then we typically use @xmath214@xmath215 , while for classical calculations @xmath107 performs better . concerning the auxiliary potential @xmath189 , for the problems considered below we use @xmath216 . convergence is achieved with @xmath217 hermite functions and @xmath218 plane waves [ then @xmath219 . finally , a word on the scaled quantities ( section [ sec : wkk : scaled ] ) . to get period @xmath220 , we have set @xmath221 with @xmath222 the original period . the characteristic energy @xmath31 is given by the potential amplitude @xmath223 ; then @xmath224 is scaled by @xmath225 and the other characteristic quantities are : action , @xmath226 , frequency , @xmath227 , and force , @xmath228 . ) for a particle in a tilted sinusoidal potential ( two periods are displayed along @xmath8 ) . here @xmath229 , and @xmath230 , while @xmath231 , @xmath232 , @xmath233 , and @xmath234 ( left to right , top to bottom ) . , title="fig : " ] ) for a particle in a tilted sinusoidal potential ( two periods are displayed along @xmath8 ) . here @xmath229 , and @xmath230 , while @xmath231 , @xmath232 , @xmath233 , and @xmath234 ( left to right , top to bottom ) . , title="fig : " ] ) for a particle in a tilted sinusoidal potential ( two periods are displayed along @xmath8 ) . here @xmath229 , and @xmath230 , while @xmath231 , @xmath232 , @xmath233 , and @xmath234 ( left to right , top to bottom ) . , title="fig : " ] ) for a particle in a tilted sinusoidal potential ( two periods are displayed along @xmath8 ) . here @xmath229 , and @xmath230 , while @xmath231 , @xmath232 , @xmath233 , and @xmath234 ( left to right , top to bottom ) . , title="fig : " ] before going into the quantum regime , let us check that for small enough @xmath235 we recover the classical results . recall that in the deterministic limit a particle in a periodic potential has two critical forces @xmath236 ( retrapping force ) and @xmath237 ( force at which the barrier disappears ) . for @xmath238 the attractors of the dynamical system are solutions `` locked '' around the potential minima , whereas for @xmath239 the `` running '' solution is an attractor globally stable . in the range @xmath240 the system exhibits bistability . finally , @xmath236 decreases with @xmath17 [ for the cosine potential @xmath241 , whereas @xmath236 tends to @xmath237 as the damping increases , narrowing the bistability range . thermal fluctuations are incorporated by the klein kramers equation , which for classical particles in periodic potentials was solved by continued fractions in @xcite ( using the @xmath10-recurrence ) . here we solve the caldeira leggett quantum master equation ( [ wkk : dxp ] ) using a large @xmath50 . as shown before , when @xmath242 we have a @xmath71-term recurrence which , for the static response , can be written as @xmath243 with @xmath244 , @xmath245 , and @xmath246 . solving it with the continued - fraction method we reobtain risken s impressive graphs for the classical distribution @xmath57 ( * ? ? ? some of them are displayed in figure [ fig : risken ] for weak damping and various external forces ( here @xmath247 and @xmath248 ) . at low @xmath249 the distribution , always periodic along @xmath8 , has maxima at the potential minima , while the profile in momenta is a maxwellian envelope @xmath250 . as @xmath249 is increased we see the evolution from these _ locked _ solutions ( @xmath251 ) to bistability between the locked and _ running _ solutions , and eventually purely running states @xmath252 $ ] are obtained at large forces . having checked the retrieval of the classical results , we now enter the quantum regime by decreasing @xmath253 . firstly , it will be useful to get a more quantitative feeling of the relation of our parameter @xmath50 and `` how quantal is the system '' . to this end we study the bands of the corresponding hamiltonian problem ( without force ) . inserting the _ bloch _ function @xmath254 in the schrdinger equation , we arrive at the following eigenvalue problem ( in scaled units ; section [ sec : wkk : scaled ] ) @xmath255 u(\x ) = e\ , u(\x ) % % % % % % % % % % % % % % % % % % \ ; , \quad % % % % % % % % % % % % % % % % % % -\half\leq \q\leq\half % % % % % % % % % % % % % % % % % % \;.\ ] ] expanding @xmath256 in plane waves ( [ vper ] ) and truncating @xmath257 at some large @xmath126 one gets a @xmath258 matrix which is diagonalised numerically . various bands @xmath259 pertaining to problems discussed below are displayed in figure [ fig : bands ] . we see that the rule of thumb is that @xmath260 is of the order of the number of bands laying below the barrier . finally , the quantum parameter used by kandemir @xcite is @xmath261 , while chen and lebowitz employed @xmath262 @xcite . besides , for a josephson junction problem we have @xmath263 , with @xmath264 and @xmath265 the josephson and coulomb energies . ( left ) and @xmath150 ( right ) . the `` islands '' correspond to zones of negative @xmath7 . , title="fig : " ] ( left ) and @xmath150 ( right ) . the `` islands '' correspond to zones of negative @xmath7 . , title="fig : " ] let us continue for a while in the hamiltonian case to familiarise ourselves with the structure of the schrdinger eigenstates in a phase - space representation . kandemir @xcite obtained analytical approximations for the eigenfunctions of a particle in a cosine potential in terms of mathieu functions . they were represented through the wigner function associated to the density matrix of pure states @xmath266 . ( in @xcite various dirac deltas occurring were regularised with gaussians @xmath267 and in the plots @xmath268 was used ; we attain the same setting @xmath230 since our @xmath10 is thermally rescaled . ) using a very small damping ( @xmath269 ) in the caldeira leggett master equation we reobtain kandemir s hamiltonian results for the ground state ( figure [ fig : kandemir ] ) . we see that at low @xmath50 the state is extended or delocalised , becoming more localised at the potential wells when @xmath270 is increased . according to @xcite , the regions of negative @xmath7 are associated to brillouin zone boundaries . we have represented a larger range of @xmath10 to see how these @xmath271 islands are repeated in the @xmath10 direction ( @xmath7 can be as small as @xmath272 bordering the limits of our numerical accuracy ) . to see how these structures arise , note that for small @xmath50 the potential can be treated perturbatively [ equation ( [ schrodinger : bloch ] ) ] . then , the wave functions have the form @xmath273 , with @xmath274 . the first term is a pure plane wave , and the rest first order corrections due to the sinusoidal @xmath15 . performing the wigner transform ( [ wigner : def ] ) of @xmath275 , we get for the ground state @xmath276 ( cf . @xcite ) @xmath277 \nonumber\\ % % % % % % % % % % % % % % % % % % & - 2|\lambda|^{2 } \cos(2\x)\ , \delta(\p ) + 2\iu\lambda \sin(\x ) \big [ \delta(\p-\half ) + \delta(\p+\half ) \big ] % % % % % % % % % % % % % % % % % % \;.\end{aligned}\ ] ] the first line gives the momentum localisation at @xmath278 ( unperturbed ) and at @xmath279 . the first term in the second line arises from the interference of @xmath280 and @xmath281 , and is responsible for the weak modulation of the height of the @xmath278 bell along @xmath8 . this modulation is more intense for larger @xmath282 ( i.e. , larger @xmath50 ) . the last term accounts for the interference between the plane wave centred at @xmath278 with those at @xmath279 , and can produce the negative islands . therefore , the above simple functional form captures most features of the wigner funtions of figure [ fig : kandemir ] . after checking the connection with the classical and hamiltonian quantum limits , we finally include dissipation and temperature in the quantum case . here examples of particles in cosine and ratchet potentials will be discussed . transport properties of weakly damped particles in a cosine potential were studied by chen and lebowitz @xcite . they started from the system - plus - bath density matrix , traced out the bath variables , and performed a perturbative calculation in the potential height ( followed by a resummation ) to get @xmath283 . for low forces a free - particle like behaviour @xmath284 was obtained . increasing @xmath249 , the wave - vector associated to @xmath10 approaches the first zone boundary ( figure [ fig : bands ] ) . there , while landau zener tunnelling can bring the particle to the next band , bragg scattering reduces the velocity ( @xmath285 ) . eventually , at larger forces , @xmath10 corresponding to states inside the next band become favoured , and the free - particle behaviour is progressively recovered . vs. @xmath249 for various dampings at @xmath286 and @xmath287 . right : marginal distribution of momenta @xmath288 for @xmath289 at @xmath290 , @xmath291 , @xmath292 , @xmath293 , and @xmath207 . the arrows mark the zone boundaries @xmath294 . bottom panels : two views of @xmath57 for @xmath289 and @xmath290 . , title="fig : " ] vs. @xmath249 for various dampings at @xmath286 and @xmath287 . right : marginal distribution of momenta @xmath288 for @xmath289 at @xmath290 , @xmath291 , @xmath292 , @xmath293 , and @xmath207 . the arrows mark the zone boundaries @xmath294 . bottom panels : two views of @xmath57 for @xmath289 and @xmath290 . , title="fig : " ] vs. @xmath249 for various dampings at @xmath286 and @xmath287 . right : marginal distribution of momenta @xmath288 for @xmath289 at @xmath290 , @xmath291 , @xmath292 , @xmath293 , and @xmath207 . the arrows mark the zone boundaries @xmath294 . bottom panels : two views of @xmath57 for @xmath289 and @xmath290 . , title="fig : " ] vs. @xmath249 for various dampings at @xmath286 and @xmath287 . right : marginal distribution of momenta @xmath288 for @xmath289 at @xmath290 , @xmath291 , @xmath292 , @xmath293 , and @xmath207 . the arrows mark the zone boundaries @xmath294 . bottom panels : two views of @xmath57 for @xmath289 and @xmath290 . , title="fig : " ] since high @xmath295 approximations were involved in their calculations , we set @xmath23 and @xmath296 in the master equation ( [ wkk : dxp ] ) . solving it with the continued - fraction method we reobtain the effect just described ( figure [ fig : cheleb92 ] ; here we work in the regime @xmath297 ) . note that this quantum slowing effect is reduced as @xmath17 increases , since the coupling to the bath `` broadens '' the effective energy levels . this broadening makes less relevant the presence of the band gap ( `` bridges '' it ) and brings the curves progressively closer to the free - damped - particle behaviour @xmath284 . with our method we get , in addition , the wigner function ( i.e. , the full density matrix ) . the distribution of velocities ( [ wp ] ) , @xmath184 , when the curves start to rise again ( @xmath298 ) , shows two peaks separated by a minimum corresponding to the wave vector of the first zone boundary . loosely speaking the system shows coexistence of _ two _ quantum `` running '' solutions . classical running solutions have a wiggling structure along @xmath8 , as the particle slows down near the potential maxima ( figure [ fig : risken ] ) . the full @xmath57 shows that the quantum running solutions have a nearly straight structure ( substrate insensitive ) . we also recognise the familiar crescent islands of negative @xmath7 at large @xmath10 ( there @xmath299 ) . around @xmath300 a complex structure is developed along @xmath8 ( `` locked part '' ) , with the maxima deformed and the minima becoming slightly negative ( @xmath301 ) . these features may reflect some interference between the locked and the nearby running solution . however , simply adding a plane wave @xmath302 to the @xmath303 used to analyse the wigner functions of the hamiltonial problem [ equation ( [ wigner : turkeys ] ) ] we have not been able to reproduce such structure . anyway , it is erased when integrating over @xmath8 to get @xmath304 . we now turn our attention to periodic potentials without spatial inversion symmetry . this feature combined with out - of - equilibrium conditions can give rise to directional motion with net unbiased driving ( _ ratchet effect _ ) . this directional motion has been invoked to explain the behaviour of molecular motors and , with the emergence of nanoscience , opens the way to build mesoscopic devices and engines based on it . the theoretical work has concentrated on the classical behaviour and in the high - damping regime @xcite . the few quantum studies also addressed strong system - bath coupling in two situations : semiclassical @xcite and a extreme quantum case disregarding thermal effects @xcite . an exception to the large @xmath17 studies is @xcite , but there the substrate potential ( and the force ) were treated perturbatively . here , we shall solve the quantum master equation ( [ wkk : dxp ] ) with the ( non - perturbative ) continued - fraction method for particles in ratchet potentials , taking into account finite damping ( @xmath305 non - negligible inertia ) . for the ratchet potential ( [ vrat ] ) , the fourier coefficients of @xmath306 are @xmath307 and @xmath308 ( recall the thermal scaling of @xmath309 ; section [ sec : wkk : scaled ] ) . we use @xmath310 , which smooths a small shoulder that this potential exhibits on the `` easy '' side ( figure [ fig : bands ] ) . as discussed before , the range of index coupling is @xmath150 , which gives a @xmath207-term recurrence . for the stationary response it can be written as [ cf.equation ( [ rr:3:cosine ] ) ] @xmath311 which can be folded onto a canonical @xmath71-term recurrence by introducing appropriate ( block ) vectors and matrices ( [ app : rr - cf ] ) . the potential considered is the minimal extension of the cosine potential lacking inversion symmetry , so that the average velocities can be different for positive and negative forces ( vd . figure [ fig : cl - ratchet ] ) . we excite the system with a square - wave force switching alternatively between @xmath312 , and compute the `` rectified '' current of particles @xmath313 here @xmath314 are the correponding stationary velocities since we consider adiabatic conditions ( @xmath315 ) . in what follows we briefly investigate the classical limit ( understudied for finite damping ) and then proceed to study quantum corrections . figure [ fig : cl - ratchet ] displays @xmath316 as a function of @xmath18 , showing the appearance of non - zero rectified velocities ( ratchet effect ) . the rectification is optimum at some intermediate temperature ( at too low @xmath18 there is hardly a response , and at too high @xmath18 the potential is irrelevant , so its asymmetry plays no rle in either case ) . at @xmath317 the results coindice with those obtainable from the analytical @xmath283 for overdamped classical particles ( * ? ? ? 11.3 ) ( vd . also @xcite ) @xmath318 with @xmath319/t$ ] . inertia , however , broadens the @xmath316 curves and shifts the maxima to higher @xmath18 ( for the lowest @xmath229 the maximum moves to a slightly lower @xmath18 ; vd . this broadening is accompanied by a slower decrease of @xmath316 with @xmath18 . this is consistent with the asymptotic behaviour of @xmath316 when @xmath320 @xcite . for overdamped particles it goes as @xmath321 , while for weak damping it decreases only with @xmath322 ( @xmath323 ) . we have checked that our results recover these dependences ( curves not shown due to the smallness of the asymptotic @xmath324 ) . ) with @xmath229 and @xmath230 . two forces are used @xmath325 ( left ) and @xmath326 ( right ) , to show the lack of symmetry in the response ( running part more developed to the easy side ) and the locked regions reflecting the potential profile . bottom left : rectified classical velocity vs. temperature for various values of the damping and force @xmath327 . right : @xmath328 vs. @xmath18 for the positive force . , title="fig : " ] ) with @xmath229 and @xmath230 . two forces are used @xmath325 ( left ) and @xmath326 ( right ) , to show the lack of symmetry in the response ( running part more developed to the easy side ) and the locked regions reflecting the potential profile . bottom left : rectified classical velocity vs. temperature for various values of the damping and force @xmath327 . right : @xmath328 vs. @xmath18 for the positive force . , title="fig : " ] in our case @xmath329 ( the `` retrapping '' force ; section [ sec : risken ] ) , so that at @xmath330 the attractors are locked solutions . then , the absolute @xmath328 vs. @xmath18 curves start from @xmath331 at @xmath330 , depart from zero as @xmath18 is increased and evolve towards @xmath332 for @xmath320 ( free damped particle ) . in the overdamped case this evolution is nearly a step , whereas the slope of the initial raising decreases with decreasing @xmath17 ( if @xmath238 ) . to understand this , let us assimilate @xmath283 to the escape rate @xmath333 ( modified by some mean free path ) , so that @xmath334 . kramers theory @xcite shows that @xmath335 increases monotonously with @xmath17 . therefore , the @xmath328 vs. @xmath18 curves for finite @xmath17 must lay ordered below the overdamped result . on the other hand , the maximum rectification occurs in the temperature region where @xmath328 transits from @xmath336 to @xmath249 . this intermediate region is narrower ( and the slopes steeper ) the higher the damping is ( figure [ fig : cl - ratchet ] ) , while the curves start to raise from zero at a lower @xmath18 . therefore , increasing the damping the @xmath337 curves will be narrower and the maxima shift to lower @xmath18 , as observed . eventually , this also allows to explain the anomalous curve showing the return ( @xmath229 ) . here @xmath338 , so that at @xmath330 the system starts close to the zone of bistability . then the particle can perform incursions into the running solution , increasing the initial slope of @xmath339 but now when the damping is lowered , shifting the maximum to a lower @xmath18 . after the mandatory exploration of the classical limit , we proceed to make the system quantal , by decreasing @xmath50 . figure [ fig : q - ratchet ] shows the rectified current with @xmath340 ( with @xmath341 bands below the barrier ; cf . figure [ fig : bands ] ) for various @xmath17 , together with the reference classical curves . we have set @xmath342 is all cases , to have roughly the same amount of locked and running components in the solutions . at high temperatures ( @xmath343 ) the classical and quantum curves coincide , as expected . however , decreasing @xmath18 down to @xmath344 , the ratchet effect gets reduced or enhanced , with respect to the classical result , depending on the dissipation . finally , at low temperatures ( @xmath345 ) the rectification is always reduced . with @xmath346 ( using @xmath342 ) for @xmath340 and the reference classical curve ( @xmath347 ) . the insets of @xmath348 and @xmath215 show enlargements of the peaks ( together with results for @xmath349 and @xmath350 ) . bottom right panel : kramers classical rate vs. @xmath17 at various forces ( from melnikov meshkov formula @xcite ) . the crosses mark the points with @xmath351 . , title="fig : " ] with @xmath346 ( using @xmath342 ) for @xmath340 and the reference classical curve ( @xmath347 ) . the insets of @xmath348 and @xmath215 show enlargements of the peaks ( together with results for @xmath349 and @xmath350 ) . bottom right panel : kramers classical rate vs. @xmath17 at various forces ( from melnikov meshkov formula @xcite ) . the crosses mark the points with @xmath351 . , title="fig : " ] with @xmath346 ( using @xmath342 ) for @xmath340 and the reference classical curve ( @xmath347 ) . the insets of @xmath348 and @xmath215 show enlargements of the peaks ( together with results for @xmath349 and @xmath350 ) . bottom right panel : kramers classical rate vs. @xmath17 at various forces ( from melnikov meshkov formula @xcite ) . the crosses mark the points with @xmath351 . , title="fig : " ] with @xmath346 ( using @xmath342 ) for @xmath340 and the reference classical curve ( @xmath347 ) . the insets of @xmath348 and @xmath215 show enlargements of the peaks ( together with results for @xmath349 and @xmath350 ) . bottom right panel : kramers classical rate vs. @xmath17 at various forces ( from melnikov meshkov formula @xcite ) . the crosses mark the points with @xmath351 . , title="fig : " ] to understand qualitatively the underlying physics we turn to some semiclassical results for various related problems . first , let us consider the _ reflection / transmission coefficient _ for an asymmetric saw - tooth potential @xcite . the results depend on the energy of the incident quantum particle . the reflection coefficient for energies higher than the barrier , @xmath352 , and the underbarrier transmission coefficient , @xmath353 , read @xmath354 % % % % % % % % % % % % % % % % % % % \;,\ ] ] where @xmath355 , @xmath356 are the slopes at both sides of the cusp , @xmath357 the particle energy , and @xmath358}$ ] . suppose now that we have an asymmetric potential , with an easy direction to the right , and we deform it with forces @xmath312 , computing the corresponding reflection coefficients @xmath359 and @xmath360 . for energies higher than the barrier , the above @xmath352 gives @xmath361 ( [ app : semicl ] ) . that is , _ overbarrier wave reflection _ is less intense when the slope of the potential is smaller . thus overbarrier transmission ( as thermal hopping ) is favoured along the easy direction for energies above the barrier @xcite . in contrast , using @xmath353 for energies below the barrier , we find @xmath362 , so that @xmath363 . that is , the reflection is more intense in the easy direction or , equivalently , tunnelling events are favoured in the hard direction @xcite . for a distribution of particle energies one would expect a competition between these two types of quantum effects . then , the enhancement of the rectified velocity under @xmath312 forcing for the @xmath348 curve follows from the increase of the thermal escape with @xmath17 in the intermediate - to - low damping range @xmath364 ( figure [ fig : q - ratchet ] ) . at that damping more escape events occur and the particles launched over the barrier are subjected to the discussed phenomenon of wave reflection , which favours net motion along the easy side and hence increases @xmath316 . lowering the damping , on the other hand , less escape events are produced and wave - reflection becomes progressively dominated by tunnel events ( favouring the hard direction ) . at @xmath344 one would expect the tunnel to be thermally assisted through the higher bands , which are wider . a simple estimation of channel probabilities comes from multiplying a band thermal population ( using kramers rate ) by the probability of tunnel ( given by the semiclassical transmission coefficient @xmath353 ) . here the exponential reduction of @xmath333 with @xmath357 compites with the exponential increase of @xmath353 for the upper bands ( [ app : semicl ] ) . at @xmath344 and for @xmath340 we find that tunnel is indeed favoured through the upper bands . to check that the behaviours found at the maxima are systematic , we increase @xmath50 to @xmath349 ( @xmath365 bands below the barrier ) and reduce it to @xmath366 ( @xmath291 bands ; here the temperatures around the peak are still inside the range of validity of the master equation , small @xmath26 ) . the results ( insets ) show clearly the tendency to amplify either the enhancement or the reduction when the system becomes more quantal . finally , at low temperatures ( @xmath367 ) thermal activation becomes ineficient for all dampings . then a reduction of the rectification is expected , due to dominance of below - barrier energies and hence of tunnel events ( through the lowest bands ) . note that the decrease of @xmath316 relative to the classical results is comparable in the curves shown . this suggests that we have similar crossover temperatures @xmath368 in this damping range . ( @xmath368 , below which quantum effects dominate , is useful as a measure of quantum corrections at a given @xmath18 ) . this quantity is known exactly in the problem of the escape from a `` quadratic - plus - cubic '' potential @xmath369 $ ] ( * ? ? ? 14 ) @xmath370 in our units @xmath371 and in the damping range studied the factor in brackets is close to one ( @xmath372 ) . this gives crossovers @xmath373 and hence comparable reductions of the ratchet effect , as observed . this reduction of the rectification is the precursor of the current reversals ( @xmath374 ) found in these systems when tunnelling completely dominates . although we find small @xmath374 in some cases , the results can not be fully trusted . for example , at @xmath348 , @xmath340 and @xmath375 we have @xmath376 , which is bordering the limits of validity of the quantum master equation employed here . we conclude with the study of the ( non - adiabatic ) dynamical response to oscillating forces @xmath377 . then , the coefficients of the expansion ( [ w : expansion : xp ] ) of the wigner function are periodic functions of @xmath46 , so they can be fourier expanded as follows : @xmath378 to lowest order in the probing field , the static part of the corresponding vectors [ equation ( [ dcdt : vectors : matrices : x ] ) ] obeys a recurrence relation of the type ( vd . [ app : pert : ac ] ) @xmath379 the equations for the harmonics have the form ( @xmath380 is the identity matrix ) @xmath381 % % % % % % % % % % % % % % % % % % % \;,\ ] ] with the `` forcing '' @xmath382 involving the previous order results . specifically , @xmath383 , with @xmath384 the part of @xmath385 corresponding to @xmath386 , namely @xmath387 % % % % % % % % % % % % % % % % % % line-1 \sqrt{1}\,\etam\delta f \!&\ ! 0 \!&\ ! \sqrt{2}\,\etap\delta f \!&\ ! 0 \!&\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-2 0 \!&\ ! \sqrt{2}\,\etam\delta f \!&\ ! 0 \!&\ ! \sqrt{3}\,\etap\delta f \!&\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-3 0 \!&\ ! 0 \!&\ ! \sqrt{3}\,\etam\delta f \!&\ ! 0 \!&\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-6 ( 4 ) \ddots \!&\ ! \ddots \!&\ ! \ddots \!&\ ! \ddots \!&\ ! \ddots \end{array } \right ) % % % % % % % % % % % % % % % % % % \ ] ] [ vd . [ app : matrices ] for the general @xmath158 ; the @xmath384 actually entering in @xmath382 is divided by @xmath386 , cf . equations ( [ eom : generic ] ) and ( [ coefficients : a ] ) ] . equations ( [ 3rr : statics ] ) and ( [ 3rr : harmonics ] ) can be solved sequentially with the continued - fraction method . figure [ fig : suscep - w ] shows the linear susceptibility ( @xmath388 ) computed in this way for a particle in a _ cosine _ potential as a function of @xmath389 . ( here we return to safe ground because @xmath390@xmath391 ; for related dielectric and kerr relaxation curves of a quantum rotator vd . @xcite and references therein . ) for the largest damping the results correspond to the classical limit . there the line - shape @xmath392 broadens and extends to frequencies lower than the oscillation frequency near the bottom of the potential wells ( @xmath393 in our units ) . for low damping , this classical _ spreading of oscillation frequencies _ has an important contribution from the amplitude dependence of the period of oscillation in anharmonic potentials @xcite . this non - dissipative contribution to @xmath392 is also known in classical spin problems , where the precession frequency depends on @xmath394 due to the magnetic anisotropy @xcite . and @xmath395 with @xmath289 ( @xmath305 classical ) , @xmath396 and @xmath397 ( less to more peaked curves ) . left panel : real part . right panel : imaginary part ( dashed line , @xmath398 but halving @xmath399 ) . vertical lines : loci of the transition frequencies of the associated quadratic - plus - quartic anharmonic oscillator [ equation ( [ delta - pert - fey ] ) ] . , title="fig : " ] and @xmath395 with @xmath289 ( @xmath305 classical ) , @xmath396 and @xmath397 ( less to more peaked curves ) . left panel : real part . right panel : imaginary part ( dashed line , @xmath398 but halving @xmath399 ) . vertical lines : loci of the transition frequencies of the associated quadratic - plus - quartic anharmonic oscillator [ equation ( [ delta - pert - fey ] ) ] . , title="fig : " ] the quantum regime is approached by decreasing the kondo parameter @xmath400 [ equation ( [ kondos ] ) ] . reducing @xmath17 with a fixed @xmath50 we find that the @xmath401 curves develop a multipeaked structure . these peaks could be vaguely seen as the mentioned nonlinear oscillations becoming quantised . to give content to this statement , we investigate perturbatively the effect of the anharmonic terms of a cosine potential ( those beyond @xmath402 ) on the energy levels of the harmonic oscillator part . since the main contribution should come from the quartic term , we consider the potential @xmath403 . first - order perturbation theory gives the eigenvalues ( * ? ? ? 6.3 ) @xmath404 expanding the cosine potential @xmath405 , we identify @xmath406 and @xmath407 ( @xmath408 , soft anharmonicity ) . then we can compute the level spacing @xmath409 % % % % % % % % % % % % % % % % % % % \;,\ ] ] which has acquired a dependence on the level index @xmath410 absent in the harmonic case . the associated transition frequencies @xmath411 are plotted in figure [ fig : suscep - w ] [ scaled as @xmath412 . they agree well with the location of the main peaks of the quantum @xmath392 curves , supporting their relation with the nonlinearity of the potential . furthermore , the contribution from transitions between the higher levels should be reduced when they are less populated . decreasing @xmath18 , we indeed see that the corresponding peaks ( located at lower @xmath389 ) are substantially reduced . in the suggestive language of nonlinear oscillations we would say that lowering the temperature the large amplitude oscillations ( having longer period ) become less probable . the applicability of the quantum nonlinear oscillator picture is grounded on the flattness of the lowest bands at the @xmath50 considered , which are well approximated by constant levels ( vd.figure [ fig : bands ] ) . we have computed the semiclassical width of the bands in a cosine potential in [ app : semicl ] . an upper bound for the width [ equation ( [ wb : approx : top ] ) ] reads in dimensionless units @xmath413 $ ] , with @xmath414 the band mean energy relative to the potential amplitude . for moderate @xmath50 this bound gives in fact very flat bands , specially the deepest ones ( @xmath415 ) . this also implies that the width of the @xmath392 peaks has a negligible contribution from the intrinsic level width , being dominated by the damping and thermal broadening . traditionally , the continued - fraction method has been employed to solve fokker planck equations for few - variable classical systems ( translational and rotational problems ) . when compared with numerical simulations of the langevin equation , the method has several _ shortcomings _ : ( i ) it is quite specific of the problem to be solved ( one needs to recalculate the recurrence coefficients for each potential ) , ( ii ) its convergence and stability depend on the parameters of the problem ( and can be poor in some ranges ) , and ( iii ) it does not return `` trajectories '' , which in the simulations provide helpful insight . nevertheless , when the method can be used its _ advantages _ are most valuable : ( i ) it is free from statistical errors , ( ii ) is essentially nonperturbative , ( iii ) specially apt to get stationary solutions ( static and dynamic ) , ( iv ) high efficiency ( allowing to explore parameter ranges out of the reach of the simulation ) , and ( v ) it gives the distribution @xmath7 , which certainly compensates for the lack of trajectories . for these reasons , and the lack of quantum langevin simulations , it was worth to develope continued - fraction methods for quantum dissipative systems . this had been done for problems of spins in a thermal bath and in quantum nonlinear optics . in this article we have discussed in detail the adaptation of the continued - fraction method to tackle quantum master equations in phase - space problems , taking advantage of the parallels with the classical case provided by the wigner formalism . we have seen that the quantum extension of the continued - fraction method is more problem dependent ( except for polynomial potentials ) , due to the necessity of using the @xmath8-recurrence of the quantum hierarchies and finding suitable bases . nevertheless , it inherits from the classical methods most of the abovementioned advantages , in particular , the obtaining of the wigner function , from which any observable can be computed . furthermore , the eigenvalue spectrum of the system , although helpful in understanding the physics , is not necessarily required . this is advantageous when dealing with non - bounded hamiltonians , continuous spectra , etc . finally , by changing the appropriate quantum parameter , the connection with the classical results is attained in a natural way . note that `` continued - fraction '' is a quite generic term ( like `` series expansion '' ) , which can be found in many different contexts , and in particular in works dealing with quantum dissipative systems ( see , e.g. , @xcite and references therein ) . however , in those cases one usually obtains some continued - fraction expression for a certain quantity ( e.g. , the linear susceptibility , memory functions , etc . ) , whereas we have obtained the _ complete solution of the quantum master equation _ by continued - fraction methods . the approach allows to study the interplay of quantum phenomena , nonlinearity , thermal fluctuations , and dissipation in several classes of systems . this is important , because methods optimised for one of the sides ( e.g. , nonlinearity or quantal behaviour ) tend to perform poorly for the others ( e.g. , fluctuations or dissipation ) . in addition , the visualization of @xmath57 plus the knowledge of the classical phase - space structure ( orbits , separatrices , attractors , etc . ) can provide valuable insight in difficult problems . finally , the main limitations of the approach are those that the starting master equations may have . we have considered the celebrated caldeira leggett quantum master equation , but the method can be applied to a class of equations of this type , and to possible generalisations . we have implemented the method in the canonical problem of quantum brownian motion in periodic potentials . we have considered both the cosine and ratchet potentials . for the former we have recovered and extended a number of classic results and interpreted them on the basis of the wigner formalism . results for the dynamics under oscillating forcing have also been obtained , illustrating the modification of the nonlinear effect of spreading of oscillation frequencies by quantum phenomena . for particles in ratchet potentials we have studied , under adiabatic conditions , the effects of finite damping on the rectified velocity . taking into account the competition of thermal hopping , overbarrier wave reflection and tunnel events , together with semiclassical analytical results , has allowed us to understand in great detail the physics of these systems . this work was supported by dges , project bfm2002 - 00113 , and dga , grant no . b059/2003 . we acknowledge f. falo , j. j. mazo , and a. p. piazzolla for useful discussions and valuable inspiration . here we briefly discuss the solution of recurrence relations by continued - fraction methods . we first consider @xmath71-term recurrences and then differential recurrences , vector - matrix cases , folding of larger - coupling recursions into @xmath71-term ones , and the problem of the intial conditions . ( for a brief description of the relation of continued fractions , series expansions , recursions , and orthogonal polynomials see @xcite . ) suppose we have the simplest case of a @xmath71-term recurrence relation of the form @xmath416 with @xmath417 and @xmath418 given quantities . to obtain the @xmath19 one inserts in ( [ rr ] ) the anstze @xcite ( upper and lower signs for @xmath419 and @xmath420 ) @xmath421 getting the following ladder coefficients @xmath422 and shifts @xmath423 @xmath424 if the recurrence is finite or the @xmath19 decrease quickly with @xmath425 ( e.g. , when they are coefficients of some judiciosly chosen expansion ) , we can truncate at some large @xmath426 , by setting @xmath427 and @xmath428 . next we _ iterate downwards _ in ( [ s : a ] ) getting @xmath422 and @xmath423 , down to @xmath429 . then , to get the @xmath19 from ( [ c : risken : com ] ) , we need @xmath430 , which is obtained from @xmath431 combination of equation ( [ rr ] ) at @xmath429 and @xmath432 . thus , starting from @xmath430 , we _ iterate @xmath433 upwards _ , getting the solution of the recursion . for @xmath434 ( homogeneous recurrences ) , we have @xmath435 and the solution reads @xmath436 . note that @xmath422 is given in terms of @xmath437 in the denominator , which can in turn be written as a fraction with @xmath438 in the denominator , and so on . this has the structure of a _ continued fraction _ , naming the method @xmath439 the iterative solution can be cast into explicit form as a series of products of continued fractions @xcite . the above form , however , has the virtue of being easy to implement in a computer . convergence is checked increasing @xmath440 and repeating the procedure . solving equation ( [ rr ] ) in this way requires a computational effort of order @xmath441 , instead of the order @xmath442 of an arbitrary linear algebra problem . this reduction arises from the tridiagonal structure of the matrix associated to ( [ rr ] ) . ( for analytical inversion of tridiagonal matrices see @xcite and references therein . ) we can solve analogously a differential recurrence of the type @xmath443 laplace transformation converts this equation into [ @xmath444 @xmath445 % % % % % % % % % % % % % % % % % % % \;,\ ] ] where @xmath446 has been used . then , introducing @xmath447 and @xmath448 , the above equation has the structure of the ordinary recurrence ( [ rr ] ) . the quantities involved in the above recurrences need not to be scalars , but they can be @xmath449-vectors ( @xmath19 and @xmath418 ) and the coefficients @xmath417 are @xmath450-matrices . then we speak of _ matrix continued fractions_. the only change in the above solution is that the fraction bars then mean matrix inversion ( `` from the left '' , @xmath451 ) . this is important , because a recurrence relation involving more than three coefficients can be `` folded '' into a @xmath71-term recurrence by introducing vector and matrix quantities . let us show how to do so for the following @xmath207-term recursion @xmath452 defining the @xmath150vectors @xmath453 and the @xmath454matrices @xmath455 it can be easily seen that equation ( [ rr:5 ] ) for @xmath456 and @xmath457 is equivalent to @xmath458 then , insertion of @xmath459 gives the matrix version of ( [ s : a ] ) . note that the @xmath19 and @xmath418 can already be vectors and the @xmath417 matrices . then , the above @xmath460 , @xmath461 , and @xmath462 are `` block '' vectors and matrices . finally , to iterate upwards @xmath459 , the `` initial condition '' @xmath463 is obtained from the matrix version of ( [ condini ] ) ( which we write @xmath464 ) . as in the absence of forcing ( @xmath465 ) the general solution involves an overall multiplicative constant , one adds to this system an extra equation to fix it ( e.g. , a normalisation contidion on the coefficients ; vd . section [ sec : obs ] ) . alternatively , we can fix arbitrarily one element of @xmath463 and rescale the solution at the end . in either case , one has an extra equation which added to @xmath466 yields the extended @xmath467 system @xmath468\hline\\[-2.ex ] % % % % % % % % % % % % % % % % % % \mbox{\,l.h.s.\ eq.\ , } \end{array } \right ) } \hspace{-0.5em } \begin{array}{c } \uparrow \\ { \scriptstyle ( j+1 ) } \\ \downarrow \end{array } % % % % % % % % % % % % % % % % % % \quad % % % % % % % % % % % % % % % % % % \mathbf{b}^{\ast } = \left ( \begin{array}{c } b_{1 } \\[-1.5ex ] \vdots \\[-1.5ex ] b_{j } % % % % % % % % % % % % % % % % % % \\[0.ex]\hline\\[-2.5ex ] % % % % % % % % % % % % % % % % % % \mathrm{\!\!r.h.s.\!\!\ ! } \end{array } \right ) % % % % % % % % % % % % % % % % % % \ ] ] this system can be solved by any method appropriate for cases with more equations than unknowns ( e.g. , using @xmath469- or svd - decomposition @xcite ) , getting @xmath463 . here we derive @xmath470 , the @xmath85 transform of the operator @xmath13 in the master equation with @xmath471 $ ] [ equation ( [ w : expansion ] ) ] . we split @xmath470 as @xmath86 , with @xmath87 , @xmath88 , and @xmath89 given by ( [ wkk : lir : lkin : lv ] ) . for the `` product '' of operators we use that the transformation of the product is equal to the product of the transformations : @xmath472 which follows by sandwiching `` identities '' @xmath473 between the factors . besides , the multiplication by functions of @xmath1 commutes with @xmath85 , so that only the differential terms need to be transformed . from these two properties we see that _ only @xmath474 and @xmath475 need to be calculated _ ( and then raised to some power or multiplied ) . _ calculation of @xmath474_. in order to transform @xmath121 , we apply @xmath476 , to an arbitrary function @xmath477 . the @xmath10 dependent part of @xmath73 cancels out and we obtain @xmath478 which has the form of a `` displaced '' differential operator . _ calculation of @xmath475_. applying @xmath479 to a function @xmath480 , we similarly get @xmath475 : @xmath481 this @xmath475 can be expressed in terms of the following creation and annihilation operators @xmath482 thus , using the shifted @xmath78 parameters @xmath483 [ equation ( [ etam : etap ] ) ] we arrive at @xmath484 for @xmath107 ( the choice in the classical case ) , @xmath485 and @xmath486 , so that @xmath487 . using the product property @xmath489 , we have @xmath490 . then , replacing @xmath67 by its `` bar '' form ( [ ppb ] ) and writing @xmath491 [ from ( [ bm - bp ] ) ] , we find @xmath492 here we have also used @xmath493 and the commutation rule @xmath494=1 $ ] to get a _ normally ordered _ form . if we introduce some compact notation for the coefficients @xmath495 we can finally write @xmath496 ( the choice @xmath107 gives the self - adjoint form @xmath497 @xcite . ) using equations ( [ pxb ] ) and ( [ ppb ] ) for @xmath474 and @xmath475 , as well as @xmath491 and the definition ( [ dmdp ] ) of @xmath499 we get for the kinetic part @xmath500 : @xmath501 the coefficients of @xmath90 and @xmath91 are @xmath502 and @xmath503 , also defined in ( [ dmdp ] ) , and we have @xmath504 to obtain @xmath506 [ vd . equation ( [ wkk : lir : lkin : lv ] ) ] , we use the product property ( [ bar : product ] ) to raise @xmath475 to @xmath507 , getting @xmath508 to get later the matrix elements of @xmath94 between hermite functions it is convenient to cast ( [ lvb ] ) in a normally ordered form . to this end we use pathak s theorems @xcite in the following generalised form @xmath509 ! } \ ; \frac{(\etapro/2)^{q}}{q ! } \;.\ ] ] here @xmath510 with @xmath511 $ ] ( @xmath512 is the result of moving the @xmath513 to the left as if they were scalars ) . the generalisation resides in letting @xmath514 and @xmath513 have a constant but different from one commutator . in our case @xmath515=\etapro % % % % % % % % % % % % % % % % % % % \;,\ ] ] follows from the commutator @xmath494=1 $ ] with @xmath516 . note that @xmath514 and @xmath513 are not the adjoint of each other , but the notation is symbolic . with help from equation ( [ pathak:2:gral : odd : alt : rep ] ) and @xmath517 [ equation ( [ kappa : s ] ) ] we can write ( [ lvb ] ) as follows @xmath518 where we have introduced the notations @xmath519 ! } \frac{(\etapro/2)^{q}}{q ! } \bv^{(2\iq+1 ) } % % % % % % % % % % % % % % % % % % \ ; , \qquad % % % % % % % % % % % % % % % % % % \hat{x}^{\ell } = \nol\big(\aplus+\aminus\big)^{\ell}\nor % % % % % % % % % % % % % % % % % % \;.\ ] ] for a given @xmath520 different powers of @xmath521 appear . we take a common factor @xmath522 and add coefficients , e.g. , @xmath523 with @xmath521 , @xmath524 , with @xmath525 , etc . @xmath526 for @xmath527 in @xmath528 we write @xmath529}=\px^{2q}\bv^{(2\ell+1)}$ ] and introduce the operator @xmath530 @xmath531 now the @xmath532 and @xmath533 dependences factorise , so we can sum the series of @xmath532 in @xmath534 , getting an exponential . then , recalling the definition of @xmath103 we finally obtain the normally ordered form for @xmath94 we were looking for @xmath535 the operator @xmath536 does not act on what could appear after @xmath15 , since @xmath530 was introduced as some abreviated notation for the high - order derivatives of the potential . here we calculate the matrix elements @xmath537 , with the transformed operator @xmath86 obtained in [ app : lb ] . we split in parts the calculation by introducing the corresponding decomposition : @xmath538 , @xmath539 , and @xmath540 , where we have used the custom `` bra''``ket '' notation . the calculation is simplified because we have all operators @xmath470 expressed as normally ordered forms of @xmath90 and @xmath91 . then we can take advantage of the ladder actions @xmath541 , @xmath542 , and the number property @xmath543 . from ( [ lirb ] ) for @xmath92 and the orthonormality of the @xmath545 , we find for @xmath546 @xmath547 then , we introduce the notation @xmath548 for the coefficient of the diagonal term [ cf.equation ( [ gammas ] ) ] , multiply by @xmath84 and sum over @xmath549 , getting @xmath550 using equation ( [ lkinb ] ) for @xmath93 and the `` ladder '' action of @xmath90 and @xmath91 , we have @xmath552 where we have used the kronecker delta to interchange @xmath113 and @xmath553 . then , multiplying by @xmath84 and summing over @xmath549 , one obtains @xmath554 this will take longer . first , the operator @xmath556 appearing in equation ( [ lvb : norord : el ] ) for @xmath94 can be put in explicit form by using the binomial formula @xmath557 the @xmath514 and @xmath513 are proportional to @xmath90 and @xmath91 [ equation ( [ am - ap ] ) ] , so the contribution of each summand to @xmath558 involves @xmath559 . from orthonormality only one term of the sum ( [ sum : pathak ] ) will contribute , namely @xmath560 , while @xmath561 will be odd , @xmath562 with @xmath563 ( reflecting the odd powers of @xmath67 in @xmath89 ) . then , in terms of @xmath520 , the index @xmath564 is restricted to @xmath565 [ [ relating - the - results - for - ippipmp2iq1 . ] ] relating the results for @xmath562 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the two cases can be easily related . first , the combinatorial coefficients are equal @xmath566 , because @xmath567 . second , the @xmath78 factors , coming from the proportionality of the @xmath568 , to @xmath569 , are simply obtained from each other by exchanging pluses and minuses ( @xmath516 ) : @xmath570 third and last , the matrix elements are also relatable . for @xmath101 , we have @xmath571 , while for @xmath572 @xmath573 which is formally identical to the former by replacing @xmath574 and @xmath575 . thus , from now on we consider only the case @xmath101 ( factor @xmath576 ) . the result for @xmath572 , will be readily obtained from this by replacing @xmath577 from all these considerations , we can write the matrix elements of @xmath556 as @xmath578 where we omit a kronecker delta ensuring @xmath101 . the rest of the calculation consists of multiplying this by @xmath579 and sum over all @xmath533 to get @xmath580 [ equation ( [ lvb : norord : el ] ) ] . [ [ restrictions - on - the - sum - on - ell . ] ] restrictions on the sum on @xmath533 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the contribution of the infinite sum ( [ lvb : norord : el ] ) to @xmath580 is fortunately cut - off . first , the lower index in @xmath581 should be positive and , introducing the shifted index @xmath582 , we obtain ( we rename @xmath583 at the end ) @xmath584 here we have already used that the sum is also cut - off from above at @xmath585 , since @xmath586 , for @xmath587 . on the other hand , for @xmath588 [ equation ( [ lvb : norord : el ] ) ] we have @xmath589 ! } ( -\ldb^{2}\px^{2})^{\ell } \frac{(-1)^{\iq}\ldb^{2\iq}}{(2\iq+1 ) ! } \bv^{(2\iq+1 ) } % % % % % % % % % % % % % % % % % % \;.\ ] ] combining this with @xmath590 , gathering @xmath591 with @xmath592 to form the operator @xmath530 [ equation ( [ aqlq ] ) ] , and recalling the definition ( [ kappa : s ] ) of @xmath103 , we obtain @xmath593 ! } \ ; \big\langle\ip\big| ( \bp)^{(2\iq+1)+\ell}\bd^{\ell } \big|\ipp\big\rangle \ ; \frac{(-\dlpxx)^{\ell}}{\ell ! } \ ; \bv^{(2\iq+1 ) } % % % % % % % % % % % % % % % % % % \;.\ ] ] note that the quotient of factorials equals @xmath594 , with the pochhammer symbol defined by ( * ? ? ? 13.5 ) @xmath595 [ [ computation - of - the - matrix - element . ] ] computation of the matrix element . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to conclude , we only need the matrix element @xmath596 in ( [ lvb : matele:1 ] ) . we first pass @xmath507 times the action of @xmath91 to @xmath597 by taking its adjoint @xmath90 and then we use repeteadly its downward ladder action , getting @xmath598 with @xmath101 . then , in terms of the pochhammer symbol ( [ pochhammer ] ) , we can write @xmath599 now , taking into account that the factor @xmath600 cancels that of @xmath601 and using the modified coefficient @xmath602 [ equation ( [ gamma : qcoeff ] ) ] , we arrive at @xmath603 \bv^{(2\iq+1 ) } % \;\delta_{\ip-(2\iq+1),\ipp } % % % % % % % % % % % % % % % % % % \;.\ ] ] comparing with the series definition of the confluent hypergeometric ( kummer ) function @xmath102 ( * ? ? ? 13.6 ) we get @xmath604 \bv^{(2\iq+1 ) } % % % % % % % % % % % % % % % % % % \quad % % % % % % % % % % % % % % % % % % \kum(a , c\,;z ) = \sum_{\ell=0}^{\infty } \frac{(a)_{\ell}}{(c)_{\ell } } \frac{z^{\ell}}{\ell ! } % % % % % % % % % % % % % % % % % % \;.\ ] ] the operator in front of @xmath76 is simply @xmath605 [ equation ( [ gamma : qcoeff ] ) ] , so that we can finally write @xmath606 note that @xmath607 is a polynomial because the first negative index cuts the upper pochhammer symbol . indeed the `` integer - plus - one '' second argument @xmath608 tells us that it is an associated laguerre polynomial ( * ? ? ? 13 ) @xmath609 in our case @xmath610 and @xmath611 , so that @xmath612 . [ [ final - expressions - for - sum_ippq_ipipprm - vec_ipp . ] ] final expressions for @xmath555 . + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + once we have computed @xmath613 , the matrix elements of @xmath94 , we multiply by @xmath84 and sum over @xmath549 ( i.e. , over @xmath520 ) , to get @xmath614 } \big [ \g_{\ip}^{\iq,- } \ , \bv^{(2\iq+1 ) } \big ] \ , \ec_{\ip-(2\iq+1 ) } % % % % % % % % % % % % % % % % % % \;.\ ] ] note that this @xmath520-sum is restricted because @xmath615 . this yields the upper limit @xmath616 $ ] , with @xmath617 $ ] the integer part of @xmath618 . the contribution from @xmath619 is easily obtained recalling the transformation rules ( [ rule ] ) ( here the sum is not restricted ) @xmath620 \ , \ec_{\ip+(2\iq+1 ) } % % % % % % % % % % % % % % % % % % % \;,\ ] ] with @xmath621 given by ( [ gamma : qcoeff ] ) . we have enclosed into square brackets the action of the operators @xmath622 to recall that they only act on the @xmath8-dependence of the potential . for @xmath107 , the choice in the classical case , @xmath485 , @xmath486 , and @xmath623 . then , @xmath624 and @xmath625 , so that only the @xmath626 terms survive . the chain of equations then adquires an unbalanced structure ( like in the old hierarchies @xcite ) , which may result in a poor stability when the quantum terms are important . in constrast , @xmath214 gives @xmath627 , and hence the weight of the terms at both sides of @xmath113 ( @xmath305 above and below the diagonals in the matrices @xmath158 ) is similar , improving the stability when going into the deep quantum regime . for special @xmath78 and @xmath189 , the integrals @xmath628 and @xmath180 [ equation ( [ ial : knl ] ) ] appearing in the expressions ( [ j : ial : knl ] ) for the observables , are easily done . for @xmath107 , we simply have @xmath629 , @xmath630 , and @xmath631 . if @xmath632 ( i.e. , @xmath216 ) , we get @xmath633 . then the first moments reduce to @xmath634 ( normalisation ) and to @xmath635 and @xmath636 . in the general case , one can derive recurrences relating the @xmath180 with lower order ( in @xmath533 ) ones . thus , only @xmath637 is needed , which can be found analytically . concerning the @xmath628 , we mostly use @xmath216 , and hence @xmath633 . anyway , for periodic potentials one can derive recurrence relations for them which can also be solved by continued fractions . to get the recurrence for the @xmath180 we start from their definition ( [ ial : knl ] ) with index @xmath638 , express the last @xmath10 as @xmath491 [ equation ( [ bm - bp ] ) ] , and then use @xmath639 , getting @xmath640 this recurrence can be used to get the @xmath641 from the @xmath180 . we can get analytical expressions for the first few @xmath180 with help from the tabulated integral ( * ? ? ? * equation ( 7.374 - 4 ) ) @xmath642 for @xmath643 , we use this result with @xmath644 , obtaining @xmath645 although @xmath646 and @xmath647 can now be obtained from ( [ knl : rr ] ) , they can also be done with help from equation ( [ hermite : func : int:1 ] ) , getting : @xmath648 , @xmath649 , and @xmath650 for a general periodic potential we can derive a recurrence relation for the integrals @xmath651\,u_{\ix}(\x)$ ] [ equation ( [ ial : knl ] ) ] . we integrate by parts the expression for @xmath652 and use @xmath653 with @xmath654 , getting @xmath655 for a finite number of harmonics ( @xmath656 , if @xmath657 ) , the recurrence ( [ rr : ial : vper ] ) has a finite coupling range . for instance , for the ratchet potential ( [ vrat ] ) , we have @xmath307 and @xmath308 and the above recurrence reduces to @xmath658 % % % % % % % % % % % % % % % % % % \ ; , \qquad % % % % % % % % % % % % % % % % % % \epspott=\epspot\,\vo/2\kt % % % % % % % % % % % % % % % % % % \;.\ ] ] with @xmath659 as input ( obtained numerically , e.g. , by simpson rule ( * ? ? ? * app . 2 ) ) , this recursion can be solved by continued fractions ( [ app : rr - cf ] ) . here we write expliticly the matrices @xmath158 for arbitrary periodic potentials . we shall give the results for a slightly generalised basis , obtained by shifting the origin of momenta by an amount @xmath660 [ cf.equation ( [ w : expansion : xp ] ) ] @xmath661 this momentum shift is convenient to `` catch '' solutions centred far from zero @xmath10 , where the ordinary basis would need many @xmath80 to reconstruct the distribution . ( a useful choice is , when generating a curve , to set @xmath662 of the precedent point . ) the momentum shift is handled as a change of variable @xmath663 . then , @xmath664 , but the term @xmath62 makes the force to enter in the combination ( scaled units ) @xmath665 as mentioned in section [ sec : vper ] , we extract the force @xmath187 from @xmath15 and set the auxiliary potential @xmath189 proportional to the periodic part @xmath77 . the @xmath158 are obtained by inserting the formulae for @xmath666 , and @xmath667_{\ix\ixp}$ ] [ equation ( [ dmdpvper : matr - elem ] ) ] , into the general matrices of section [ sec : mqbh ] . instead of giving a single expression , we write separately the matrices for @xmath668 and @xmath669 . the central matrix @xmath670 is determined by the terms involving @xmath671 , and reads @xmath672 % % % % % % % % % % % % % % % % % % line-1 \sqrt{1}(\iu\,\dmi\ix\!-\!\etam\bfsh ) \!\!\!\!&\!\!\!\ ! \gammao_{1 } \!\!+\!\ ! \pshift\,\iu\ix \!\!\!\!&\!\!\!\ ! \sqrt{2}(\iu\,\dpl\ix\!-\!\etap\bfsh ) \!\!\!\!&\!\!\!\ ! \sqrt{2\cdot3}\ , \gammapp \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-2 \sqrt{1\cdot2}\ , \gammamm \!\!\!\!&\!\!\!\ ! \sqrt{2}(\iu\,\dmi\ix\!-\!\etam\bfsh ) \!\!\!\!&\!\!\!\ ! \gammao_{2 } \!\!+\!\ ! \pshift\,\iu\ix \!\!\!\!&\!\!\!\ ! \sqrt{3}(\iu\,\dpl\ix\!-\!\etap\bfsh ) \!\!\!\!&\!\!\!\ ! \sqrt{3\cdot4}\ , \gammapp \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-3 0 \!\!\!\!&\!\!\!\ ! \sqrt{2\cdot3}\ , \gammamm \!\!\!\!&\!\!\!\ ! \sqrt{3}(\iu\,\dmi\ix\!-\!\etam\bfsh ) \!\!\!\!&\!\!\!\ ! \gammao_{3 } \!\!+\!\ ! \pshift\,\iu\ix \!\!\!\!&\!\!\!\ ! \sqrt{4}(\iu\,\dpl\ix\!-\!\etap\bfsh ) \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-4 0 \!\!\!\!&\!\!\!\ ! 0 \!\!\!\!&\!\!\!\ ! \sqrt{3\cdot4}\ , \gammamm \!\!\!\!&\!\!\!\ ! \sqrt{4}(\iu\,\dmi\ix\!-\!\etam\bfsh ) \!\!\!\!&\!\!\!\ ! \gammao_{4 } \!\!+\!\ ! \pshift\,\iu\ix \!\!\!\!&\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-6 ( 5 ) \ddots \!\!\!\!&\!\!\!\ ! \ddots \!\!\!\!&\!\!\!\ ! \ddots \!\!\!\!&\!\!\!\ ! \ddots \!\!\!\!&\!\!\!\ ! ! \ddots \end{array } \right ) % % % % % % % % % % % % % % % % % % \ ] ] with @xmath673 and @xmath674 given by ( [ gammas ] ) , @xmath675 , and @xmath676 . here the periodic potential does not appear ; only @xmath187 enters in the matrices @xmath670 . besides , they do depend on the index @xmath131 ( cf . below ) . the matrices @xmath677 are determined by the terms involving @xmath678 in equation ( [ dmdpvper : matr - elem ] ) and read @xmath679 % % % % % % % % % % % % % % % % % % line-1 \etam \gs_{1}^{0 } \!-\ ! \sqrt{1}\,\epspot \!\!\!&\!\!\ ! -\pshift\,\epspot \!\!\!&\!\!\ ! \etap \gs_{2}^{0 } \!-\ ! \sqrt{2}\,\epspot \!\!\!&\!\!\ ! 0 \!\!\!&\!\!\ ! \etap^{3 } \gs_{4}^{1 } \!\!\!&\!\!\ ! 0 \!\!\!&\!\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-2 0 \!\!\!&\!\!\ ! \etam \gs_{2}^{0 } \!-\ ! ! -\pshift\,\epspot \!\!\!&\!\!\ ! \etap \gs_{3}^{0 } \!-\ ! \sqrt{3}\,\epspot \!\!\!&\!\!\ ! 0 \!\!\!&\!\!\ ! \etap^{3 } \gs_{5}^{1 } \!\!\!&\!\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-3 \etam^{3 } \gs_{3}^{1 } \!\!\!&\!\!\ ! 0 \!\!\!&\!\!\ ! \etam \gs_{3}^{0 } \!-\ ! \sqrt{3}\,\epspot \!\!\!&\!\!\ ! -\pshift\,\epspot \!\!\!&\!\!\ ! \etap \gs_{4}^{0 } \!-\ ! \sqrt{4}\,\epspot \!\!\!&\!\!\ ! 0 \!\!\!&\!\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-4 0 \!\!\!&\!\!\ ! \etam^{3 } \gs_{4}^{1 } \!\!\!&\!\!\ ! 0 \!\!\!&\!\!\ ! \etam \gs_{4}^{0 } \!-\ ! \sqrt{4}\,\epspot \!\!\!&\!\!\ ! -\pshift\,\epspot \!\!\!&\!\!\ ! \etap \gs_{5}^{0 } \!-\ ! \sqrt{5}\,\epspot \!\!\!&\!\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-5 \etam^{5 } \gs_{5}^{2 } \!\!\!&\!\!\ ! 0 \!\!\!&\!\!\ ! \etam^{3 } \gs_{5}^{1 } \!\!\!&\!\!\ ! 0 \!\!\!&\!\!\ ! \etam \gs_{5}^{0 } \!-\ ! \sqrt{5}\,\epspot \!\!\!&\!\!\ ! -\pshift\,\epspot \!\!\!&\!\!\ ! \ddots \\[0.ex ] % % % % % % % % % % % % % % % % % % line-6 \ddots \!\!\!&\!\!\ ! \ddots \!\!\!&\!\!\ ! \ddots \!\!\!&\!\!\ ! \ddots \!\!\!&\!\!\ ! \ddots \!\!\!&\!\!\ ! ! \ddots \end{array } \right ) % % % % % % % % % % % % % % % % % % \ ] ] where @xmath211 is evaluated at @xmath680 , with @xmath516 . these matrices enjoy properties opposite to those of @xmath670 . the substrate potential does appear ( multiplying all elements in front of the matrices ) while the force is absent . besides , the matrices @xmath677 are independent of the index @xmath131 , which permits to write them simply as @xmath681 . the only technical complication with respect to the common choice @xmath107 is that we need to compute the kummer functions @xmath102 included in @xmath211 , which in our case are simple polynomials [ equation ( [ assoc : laguerre : poly : kummer ] ) ] . note that to compute the observables we must eventually undo the momentum shift @xmath682 . in this appendix we derive some semiclassical formulae discussed in the main text . in section [ ratqbm ] we invoked the current through a saw - tooth barrier deformed by @xmath683 ( figure [ fig : saw - width ] ) . ( b ) and by @xmath684 ( c ) . right panel : width of the energy bands @xmath685 in a cosine potential ( exact and various approximations ) . , title="fig : " ] ( b ) and by @xmath684 ( c ) . right panel : width of the energy bands @xmath685 in a cosine potential ( exact and various approximations ) . , title="fig : " ] we introduced the reflexion coefficients @xmath359 and @xmath360 for @xmath312 . one has to analyse separately the cases of incident particles with energies above or below the barrier . in the first case , the reflexion coefficient by a potential with a cusp ( with slopes @xmath355 and @xmath356 ) is given by the first equation ( [ rs ] ) @xcite . owing to the right slope in figure [ fig : saw - width ] is infinity , we make it finite and finally take the vertical limit , getting @xmath686 next we consider the case of tunnelling through the barrier . the semiclassical transmission coefficient is also written in ( [ rs ] ) . defining @xmath687 , we calculate @xmath688 and @xmath689 for positive and negative forces . after simple integrals , we find @xmath690 with @xmath357 the particle energy ( cf . reference @xcite ) . for moderate forces , @xmath691 , we get for the transmission coefficients @xmath692 $ ] @xmath693 thus , for @xmath357 below the barrier @xmath363 , while @xmath361 for @xmath357 above the barrier . in order to get the semiclassical transmision coefficient through a potential @xmath405 , we need to do the integral @xmath694 } \big| % % % % % % % % % % % % % % % % % % % \;,\ ] ] with turning points obeying @xmath695 . due to @xmath696 below the barrier , we change the sign inside the square root extracting an @xmath697 which is absorbed by the modulus . simple tranformations bring the integral into a tabulated form ( * ? ? ? * equation ( 2.576 ) ) , and we obtain @xmath698 % % % % % % % % % % % % % % % % % % \ ; , \qquad % % % % % % % % % % % % % % % % % % \varepsilon = e/\vo \ ; , \quad s=\sqrt{\half(1-\varepsilon ) } % % % % % % % % % % % % % % % % % % % \;,\ ] ] where @xmath699 while @xmath700 and @xmath701 the first and second complete elliptic integrals ( * ? ? ? insertion in @xmath692 $ ] gives the transmision coefficient through a cosine potential . this @xmath353 also gives the _ semiclassical energy bands _ @xcite @xmath702 with @xmath703 the oscillation frequency in the wells . the width of the bands is therefore @xmath704 \big\ } % % % % % % % % % % % % % % % % % % \ ; , \qquad % % % % % % % % % % % % % % % % % % \varepsilon = e/\vo % % % % % % % % % % % % % % % % % % \;.\ ] ] the parameter @xmath705 goes from @xmath202 when @xmath706 ( potential bottom ) to @xmath336 at the barrier top @xmath707 . for @xmath708 we can expand the elliptic integrals , getting @xmath709 \big\ } \nonumber\\ % % % % % % % % % % % % % % % % % % & + ( 1+\varepsilon)^{2 } \big\ { 1 + \case{1}{8 } \ln \big [ \case{1}{32 } ( 1+\varepsilon ) \big ] \big\ } \big ) \big ] % % % % % % % % % % % % % % % % % % \;.\end{aligned}\ ] ] at the potential bottom @xmath706 this gives @xmath710 $ ] . the band width increases exponentially with the energy . near the top , @xmath711 , one has @xmath712 , and we can use the series expansion of the elliptic integrals , obtaining @xmath713 % % % % % % % % % % % % % % % % % % \;.\ ] ] figure [ fig : saw - width ] shows that equations ( [ wb : approx : bottom ] ) and ( [ wb : approx : top ] ) approximate well the exact results in their respective ranges ( the later also serves as un upper bound ) . if we disregard the terms with @xmath714 in ( [ wb : approx : bottom ] ) , the result happens to fit remarkably @xmath685 in all the range . in this last appendix we shall solve the following generic differential equation @xmath715 perturbatively in the forcing @xmath716 . ( for cosine forcing @xmath717 . ) comparing with section [ sec : dynamics ] , we see that @xmath718 corresponds to the vectors @xmath719 , @xmath720 to the static part of the matrices @xmath158 , and the perturbation @xmath721 to @xmath722 , the part of @xmath670 involving @xmath386 . a periodic solution can be fourier expanded as @xmath723 with @xmath724 the unperturbed response , while @xmath725 for complex @xmath718 . separating terms oscillating with @xmath726 and @xmath727 , the left - hand side of ( [ eom : generic ] ) reads @xmath728 \e^{+\iu\omega_{\iw } t } \\ % % % % % % % % % % % % % % % % % % & + & \sum_{\iw=1}^{\infty } \dfm^{\iw}\ , \big [ -\iu(\omega_{\iw}\tau)\tilde{\ec}^{(\iw ) } + \rel\,\tilde{\ec}^{(\iw ) } \big ] \e^{-\iu\omega_{\iw } t } % % % % % % % % % % % % % % % % % % \;.\end{aligned}\ ] ] to obtain the right - hand side , we use @xmath729 , redefine the indices ( keeping the same names ) to get all the oscillating factors at @xmath730 , and introduce the definition @xmath731 , arriving at @xmath732 \e^{+\iu\omega_{\iw } t } \\ % % % % % % % % % % % % % % % % % % & + & \sum_{\iw=1}^{\infty } \dfm^{\iw } \big [ \vpert\,\tilde{\ec}^{(\iw-1 ) } + \left(\dfp\dfm\right ) \vpert\,\tilde{\ec}^{(\iw+1 ) } \big ] \e^{-\iu\omega_{\iw } t } % % % % % % % % % % % % % % % % % % \;.\end{aligned}\ ] ] equating terms with the same oscillating factors at both hand sides ( _ uniqueness _ of the fourier expansion ) , we obtain @xmath733 \\[1.ex ] % % % % % % % % % % % % % % % % % % \iu(\omega_{\iw}\tau)\ec^{(\iw ) } + \rel\,\ec^{(\iw ) } & = & \vpert\ , \big[\ ; \ec^{(\iw-1 ) } & + & \left(\dfp\dfm\right ) \ec^{(\iw+1 ) } \big ] % % % % % % % % % % % % % % % % % % \end{array } % \;,\ ] ] the @xmath734 are included as @xmath735 with @xmath736 . the contribution @xmath737 comes from products of rotating " terms @xmath738 . for instance , the oscillating factor @xmath739 in @xmath740 when multiplied by the part @xmath741 of the field raises the harmonic from @xmath742 to @xmath743 . the term @xmath744 comes from products of counter - rotating " terms @xmath745 . for example , the product of the oscillating factor @xmath746 and the @xmath747 part of the field lowers the order from @xmath748 to @xmath743 . this contribution ( the contamination " from higher harmonics ) has an order higher in @xmath716 than the contribution of @xmath740 , which is therefore the leading one . indeed , expanding the @xmath749 in powers of @xmath716 we get the leading terms at each harmonic @xmath750 the @xmath751 equation gives the static response [ corresponding to ( [ 3rr : statics ] ) ] , @xmath388 the linear dynamical response , etc . [ corresponding to equation ( [ 3rr : harmonics ] ) ] .

one of the many achievements of quantum field theories like the standard model is the confirmed prediction of a scale dependence of the physical couplings @xmath2 . in order to know in which way the couplings @xmath3 of a given quantum field theory depend on the energy scale @xmath4 one usually has to regularize and renormalize the theory . however , when it comes to gravity a consistent and predictive renormalization and therefore a predictive quantum field theoretical description is still to be found . independently of how this theory of quantum gravity will look like , in most approaches it is expected to introduce a non - trivial running to the couplings of classical gravity which are newtons `` constant '' @xmath5 and the cosmological constant @xmath6 . in an effective description those couplings are expected to be present in an improved action and the corresponding solutions . since on the one hand black holes are key objects in every classical or quantum gravitational theory and their understanding is crucial for the understanding of the whole model and on the other hand the scale dependence of gravitational couplings has been extensively studied in the context of exact renormalization group ( erg ) equations @xcite , it is natural to study black holes in the context of erg results . this has been done mostly by improving the classical solutions @xcite . this procedure has however two weaknesses which basically motivated this study : the first problem is that in order to study black holes in the erg context one has to relate the radial scale of a black hole solution @xmath7 to the energy scale @xmath4 of the erg calculation , this procedure is however not uniquely defined . the second problem is that the improved solution does not ( at least not at all scales @xcite ) resolve the improved equations of motion nor does it minimize the erg improved action . as complementary contribution to this program we will follow a philosophy that is somewhat inverse to the existing studies on erg improved black hole solutions . as starting point we will take the improved equations of motions which contain scale depending couplings @xmath0 and @xmath1 which are a priory undetermined . the working hypothesis will then be to ask for which functional form of those couplings it is possible to solve those equations of motion with the most symmetrically possible metrical ansatz . the paper is organized as follows . in section [ secsol ] it is shown how solving the resulting system of equations determines then the resulting black hole metric and the functional form of @xmath0 and @xmath1 up to the existence of four integration constants . general properties of this solution such as differences to the studies in the literature , singularities , and the existence of classical - like parameter choices are then discussed . since both couplings of the present solution are functions of the radial scale @xmath7 , the corresponding adimensional couplings @xmath8 and @xmath9 can be combined in a coupling flow . in section [ secrun ] this induced flow with an ultra violet fixed point for @xmath8 and @xmath9 is derived , the flow is compared to the flow for @xmath10 and @xmath11 known from erg calculations . finally , the anomalous dimensions of the induced couplings and the product of couplings are discussed and and compared to the findings in the erg approach . after summarizing remarks in section [ secconcl ] we give complementary discussions in the appendix [ secappend ] . up to now black holes where studied in the context of erg equations by taking the classical solutions ( for constant @xmath12 and @xmath13 ) and then replacing @xmath5 and @xmath14 in those solutions . the physical interpretation of those erg improved solutions depends on how the erg scale @xmath4 is related to the physical scale @xmath7 . this procedure however was only partially successful , since the improved solutions were actually no more solutions of any form of einstein or modified einstein equations . yet , the improved solution with a cosmological term seems at least to solve the improved equations of motion asymptotically in the uv @xcite . in order to avoid the danger which is involved with the choice of @xmath15 we will consider directly the fact that all scales are in the end functions of the physical scale @xmath16 and @xmath17 . when doing so we pretend to find a solution which is still an exact solution of the equations motion . the corresponding improved equation of motion is @xcite ( see also appendix [ appendeom ] ) @xmath18 with t_=g(r)(g_- _ _ ) . at first instance we are interested in spherically symmetric solutions in regions where the classical matter contribution @xmath19 . as ansatz for the metric tensor we use [ ansatz ] ds^2=-f(r ) dt^2 + 1/f(r)dr^2 + r^2 d^2 + r^2 ( ) d^2 where [ fvonr ] f(r)=(1 - 2-r^2 ) . with this ansatz the einstein equations ( [ eom2 ] ) reduce to three independent differential equations for the a priory unknown functions @xmath20 , and @xmath21 . please note that the constant @xmath22 would only be the mass of the black hole , if @xmath12 and @xmath23 would be constants . in the context of variable constants , however , @xmath22 is only an arbitrary constant with units of mass , which can take arbitrary ( even negative ) values . it was found that , apart from the well known solutions , which imply constant couplings , there exists a solution with a non trivial @xmath7 dependence [ gvonr ] g(r)&=&- , + [ lvonr ] ( r)&=&\{-2 c_1 ( c_1 ^ 2 ( 12 c_1 ^ 2 + 384 c_2+c_3)+24 r^3 c_1 ^ 3 c_4 + .. + & & .3 r^2 ( 384 c_2+c_3 - 24 c_1 ^ 4 c_4)+6 r c_1 ( -c_1 ^ 2 - 384 c_2-c_3 + 8 c_1 ^ 4 c_4 ) ) + & & .+3 r ( r^2 - 3 r c_1 + 2 c_1 ^ 2 ) ( 384 c_2+c_3 ) -3 r ( r^2 - 3 r c_1 + 2 c_1 ^ 2 ) ( 384 c_2+c_3 ) } , + [ llvonr ] l(r)&= & c_4+\{++ . + -3 ( 384 c_2+c_3 ) + 3 ( 384 c_2+c_3 ) } . those solutions contain four constants of integration @xmath24 . the role and the interpretation of those constants in terms of different physical perspectives will be discussed throughout this paper . one of the first questions one might ask for this new kind of solution is whether it has the same singularity problem at the origin as the standard solution or whether there exist parameter configurations for which the singularity can be avoided . this can be checked by calculating the invariant tensor density which gives to lowest order in @xmath25 r _ r^ = + ( r^-5 ) . the divergence to this order can be avoided by choosing @xmath26 with c_2 = - . this choice removes a number of divergent terms such that the remaining next to next to next to leading singularity reads r _ r^|_c_2 = + ( r^-1 ) . if one further tries to remove this singularity , one is forced to take @xmath27 , where the value for @xmath28 has to be chosen , before one takes the limit in @xmath29 . this leads to a finite tensor density r _ r^|_c_2 , c_1 = c_4 ^ 2 , but this choice of the integration constants is actually almost trivial because corresponds to flat ( anti ) de sitter space f(r)|_c_2 , c_1 = 1-r^2 . since the equation of motion ( [ eom2 ] ) can be interpreted as a special case of an @xmath30 theory @xcite ( see @xcite for a review ) one has to check , whether the above solution has already been discussed in the context of the more general @xmath30 theories . a class of solutions with constant curvature @xmath31 and perturbative expansions around the `` classical '' solutions has been discussed in @xcite and special solutions with conformal anomaly have been discussed in @xcite . other perturbative solutions ( assuming @xmath32 ) to @xmath30 theories in their transition limit to general relativity can be found in @xcite . further variations of the classical solution have also been studied in the context of kerr black holes @xcite . for an exact solution that assumes a finite @xmath33 and constant curvature @xmath34 , which generalizes to charged , rotating black holes see @xcite . none of the above solutions contains ( [ gvonr]-[llvonr ] ) since here , @xmath35 , @xmath36 is not constant , and the solution is not a perturbation of a classical solution . thus , to our current knowledge ( [ gvonr]-[llvonr ] ) has not been discussed in the context of @xmath30 theories . an other very similar approach is brans dicke theory @xcite where one allows for a varying gravitational constant @xmath37 . in difference to this approach , the brans dicke action contains a kinetic term @xmath38 . it the limit @xmath39 this theory corresponds to constant coupling and standard general relativity . solutions of bd with the limit @xmath40 correspond to our erg inspired approach . black hole solutions in pure brans dicke theory have been discussed . first black hole solutions for this theory have been found by brans @xcite . in @xcite it has been shown that all bd black holes that have constant @xmath41 in the outside region and that are static solutions correspond to the known schwarzschild , kerr or kerr - newman solutions . if one relaxes the condition of constant fields further non - trivial solutions have been found @xcite . this solution does not apply here , since it has @xmath42 . in @xcite bd black holes have been studied in the lifshitz context . bd black hole solutions with a specific cosmological term @xmath43 have been found and discussed in @xcite . based on those studies @xcite discussed further generalizations of bd -ads bhs . by using dilaton black hole solutions with a cosmological term @xmath44 they construct via a conformal transformation @xmath45 the bd solution with the cosmological term @xmath43 . this method , however , works only for larger space - time dimensions @xmath46 since this conformal transformation is ill defined for @xmath47 . this reflects also in the fact that the metric coefficients of those bd solutions diverge for four space - time dimensions . it is also instructive to do a comparison with an approximated black holes solution of brans dicke theory given by weinberg @xcite ( page 183 and page 247 ) for the case that gives the schwarzschild solution plus corrections in @xmath25 [ epand1 ] ds^2=(1 - 2++ )dt^2 -(1++ ) dr^2 -r^2 d^2 -r^2 ^2d^2 . comparing this approximation with @xmath48 to the series expansion in @xmath25 of the exact solution one finds that one can at best fix the constants @xmath29 , and @xmath49 in combination with a time rescaling ( @xmath50 ) , such that [ epand2 ] ds^2=(1 - 2++ )dt^2 -(2 + 2 + ) dr^2 -r^2 d^2 -r^2 ^2d^2 . both expansions ( [ epand1 ] and [ epand2 ] ) do not agree in the constant coefficient of the @xmath51 component . if one wants to demand a regime where the exact solution has a constant factor @xmath52 for both @xmath53 and @xmath54 one has fix the constants for an expansion around @xmath55 , as it is discussed in the following subsection . this also can not be fixed by just varying the initial metric ansatz by @xmath56 . in terms of physical viability this is a severe problem , because the above expansion for @xmath48 basically predicts that this new solution would contradict all gravitational lensing effects with relativistic trajectories . in order to address this concern we will first ask , whether there exist special parameter choices ( with @xmath57 ) for which the new solution approximates to the ( for wide ranges of @xmath7 ) well confirmed classical solution . given the problems for the @xmath25 expansion for @xmath48 we will now study the case @xmath57 and by trying to approximate the new solution to the standard solution where [ fstandard ] f_s(r)=1 - 2 -r^2 . as it can bee seen from equation ( [ epand2 ] ) , not even reproducing the standard result without cosmological constant is trivially possible . thus , one has to ask the question , whether there is any configuration of the parameters ( @xmath58 ) which is in agreement with current experimental results , which basically confirm ( [ fstandard ] ) . by constructing an approximation to the standard metric ( [ fstandard ] ) it will now be shown that such parameter choices exist . an important property of the standard metric in the de sitter case is that it approximates to a plane flat space for large @xmath7 , before it runs into an other horizon at even larger distances . thus , we try to fix the constant @xmath59 such that there exists a @xmath60 with f(r_m)=1 f|_r_m=0 . an analytic solution of those two combined conditions is possible , it gives r_m&= & , + [ c4new ] c_4,s&= & , where @xmath61 . the next step is to choose the constants @xmath29 and @xmath62 such that the two horizons of @xmath63 approximate to the horizons of @xmath64 r_02 g_0 m_0 r_1 , where we are interested in the cases @xmath65 . a numerical optimization to those conditions gives [ c1new ] c_1,s&= & , + [ c3new ] c_3,s&=&. by demanding that @xmath66 for @xmath67 one can also fix the missing constant [ c2new ] c_2,s=. using the definitions ( [ c4new ] , [ c1new ] , [ c3new ] , [ c2new ] ) , the solution ( [ fvonr ] , [ gvonr ] , [ lvonr ] ) can be written as f(r)&= & ( 96 ^ 2/3 r ( ) ^1/3.-1446 ^ 1/3 g_0 ^ 2 m_0 ^ 2 r^2 _ 0 - 186 ^ 2/3 ( ) ^4/3 _ 0 + & & [ fvonrnew ] + 48 ( g_0 m_0)^5/3 r^3 _ 0 ^ 4/3 + 36 r^2 ( g_0 ^ 2 m_0 ^ 2 _ 0)^1/3 - 108 r ( g_0 ^ 5 m_0 ^ 5 _ 0)^1/3 - 86 ^ 1/6 r^3 ( g_0 m_0 _ 0 ^ 2)^1/3 + & & + 12 6 ^ 1/3 g_0 m_0 r^3 _ 0 ( [ 3]+1)-48 6 ^ 2/3 ( g_0 m_0)^7/3 r^3 _ 0 ^ 5/3 [ 3 ] + & & + 126 ^ 1/3 g_0 m_0 r^3 ( 1 - 4 6 ^ 1/3 ( g_0 m_0)^4/3 _ 0 ^ 2/3 ) _ 0 + & & . + ( 96 6 ^ 2/3 ( g_0 m_0)^7/3 r^3 _ 0 ^ 5/3 -246 ^ 1/3 g_0 m_0 r^3 _ 0 ) ) , [ gvonrnew ] g(r)= , [ lvonrnew ] ( r)&= & \ { 36g_0 ^ 2 m_0 ^ 2 r _ 0 ^ 2/3.-6 ( g_0 m_0)^2/3 r + 126 ^ 5/6 g_0 m_0 r^2 _ 0 ^ 2/3 + & & + 864 ( g_0 m_0)^8/3 r _ 0 + 36 ( g_0 m_0)^5/3 r^2 _ 0 - 5766 ^ 1/3 g_0 ^ 3 m_0 ^ 3 r^2 _ 0 ^ 5/3 + 192 ( g_0 m_0)^8/3 r^3 _ 0 ^ 2 + & & -4326 ^ 1/6 g_0 m_0 r^2 ( g_0 m_0 _ 0)^4/3 - 32 6 ^ 1/6 r^3 ( g_0 m_0 _ 0)^4/3 - 246 ^ 2/3 r ( g_0 ^ 4 m_0 ^ 4 _ 0)^1/3 - 186 ^ 2/3 ( g_0 ^ 7 m_0 ^ 7 _ 0)^1/3 + & & -216 ( g_0 m_0)^8/3 r _ 0 [ 3]-108 ( g_0 m_0)^5/3 r^2 _ 0 [ 3]+4326 ^ 1/3 g_0 ^ 3 m_0 ^ 3 r^2 _ 0 ^ 5/3 [ 3]+486 ^ 1/3 g_0 ^ 2 m_0 ^ 2 r^3 _ 0 ^ 5/3 ( [ 3]+1 ) + & & -1926 ^ 2/3 ( g_0 m_0)^10/3 r^3 _ 0 ^ 7/3 [ 3]+9 6 ^ 2/3 g_0 m_0 r ( g_0 m_0 _ 0)^1/3 [ 3 ] + & & -6 r ( -36 ( g_0 m_0)^5/3 r _ 0 + 8 g_0 ^ 2 m_0 ^ 2 ( -9 ( g_0 m_0)^2/3 + 26 ^ 1/3 r^2 _ 0 ^ 2/3 ) _ 0 + 36 ^ 2/3 ( g_0 ^ 4 m_0 ^ 4 _ 0)^1/3 . + & & .-166 ^ 1/3 g_0 ^ 3 m_0 ^ 3 r _ 0 ^ 5/3 ( -9 + 46 ^ 1/3 r ( g_0 m_0 _ 0 ^ 2)^1/3 ) ) [ r ] + & & -3 g_0 m_0 r ( 72 ( g_0 m_0)^5/3 _ 0 + 36 ( g_0 m_0)^2/3 r _ 0 - 1446 ^ 1/3 g_0 ^ 2 m_0 ^ 2 r _ 0 ^ 5/3 - 166 ^ 1/3 g_0 m_0 r^2 _ 0 ^ 5/3 . + & & .-36 ^ 2/3 ( g_0 m_0 _ 0)^1/3 + 646 ^ 2/3 r^2 ( g_0 m_0 _ 0 ) ^7/3 ) + & & + ( 432 ( g_0 m_0)^8/3 r _ 0 -18 6 ^ 2/3 g_0 m_0 r ( g_0 m_0 _ 0)^1/3 216 ( g_0 m_0)^5/3 r^2 _ 0 + 3846 ^ 2/3 ( g_0 m_0)^10/3 r^3 _ 0 ^ 7/3 . + & & . -.8646 ^ 1/3 g_0 ^ 3 m_0 ^ 3 r^2 _ 0 ^ 5/3 -966 ^ 1/3 g_0 ^ 2 m_0 ^ 2 r^3 _ 0 ^ 5/3 ) } . one can see that now there appears only the physical mass @xmath68 and not the parameter @xmath22 from the initial ansatz in this set of solutions . the physical meaning of @xmath68 can be verified by expanding ( [ fvonrnew ] ) for small @xmath69 f(r)=1 - 2+(_0 ^ 2/3 ) . this special choice of the parameters ( @xmath70 ) is not exactly the most compact but it shows that the classical result @xmath64 with the parameters ( @xmath71 ) can be approximated very well by the exact solution @xmath63 with variable @xmath0 and @xmath1 . as it can be seen , from figure [ figfr ] , @xmath63 and @xmath64 are practically indistinguishable for very small values of @xmath69 . comparison of the classical metric coefficient ( [ fstandard ] ) with the solution ( [ fvonrnew ] ) as dashed curves . the numerical values were chosen to be @xmath72 and from left to right @xmath73.,width=377 ] since the classical form of ( [ fstandard ] ) has been reproduced by various experiments at different scales , and since the special choice of @xmath59 allows to approximate very well to this result , it can be concluded that the solution ( [ fvonr ] ) allows also for parameter choices which are also in agreement with current experimental limits and observations . the question to which extend observational data such as dark matter effects can be incorporated in the current solution ( in the spirit of @xcite ) will be subject of future studies @xcite . the above choice of parameters is most likely not the only one that achieves this goal , but the point is that it shows by construction that physically viable parameter choices are perfectly possible . when discussing the flow of scale dependent couplings like ( @xmath74 ) one usually does so for the dimensionless couplings ( @xmath75 ) . therefore , the free parameters that describe this flow , should also be adimensional . hence one defines a set of four dimensionless parameters parameters @xmath76 instead of the dimensionfull parameters @xmath77 [ replfp ] . ccc c_1&=&- + c_2&=&- + c_3&= & + c_4&=&- } \ { ccc ^*_u&= & - + l_i&=&- + g^*_u & = & + g_i & = & -16 c_2 ^3 } . with this reparametrization the metric solution reads [ fvonrnew2 ] f(r)&= & \{g_i ( -6 g^*3_u ^2 r^2 + 4 g_i^3 ^*_u -6 g_i^2 g^*_u r ^*_u + g_i g^*2_u r ( 6+^2 r^2 l_i+12 r ^*_u ) ) . + & & . + 6 g^*3_u ^3 r^3 ( g^*_u -2 g_i ^*_u ) } . one observes that the remaining dimensionfull parameters are @xmath7 and @xmath22 , which only appear in dimensionless pairs @xmath78 in ( [ fvonrnew ] ) . now that convenient parameters have been defined one can define the dimensionless couplings . in order to make he couplings dimensionless one has to multiply them with dimension - full quantities that describe the physical system . in our case the two dimension - full quantities that can vary and that describe the physical system are @xmath22 and @xmath7 . in principle , every adequate power of those two quantities can do the job , thus one can write [ adimg ] g(r)&= & g(r ) + [ adiml ] ( r)&=&-(r ) , where @xmath79 and @xmath80 are numbers that determine the respective impact of @xmath7 and @xmath22 adimensionalization ! " and the minus sign in ( [ adiml ] ) is pure convention . the constants @xmath79 and @xmath80 are crucial for the expected fixed point behavior of the adimensional coupling flow . only for certain values there exists a well behaved , non - trivial fixed point . motivated from the results in the erg approach , which will be introduced in the following section , one demands the existence of a non - trivial uv fixed point . this uv fixed point exists for both couplings for the values @xmath81 and @xmath82 [ uvfp ] g_u(r)&= & g(r ) ^2 + _ u(r)&= & -(r ) . the values of the uv fixed points are [ uvfplim ] g_u(r0)&=&g_u^ * + _ u(r0)&=&_u^ * . the limits in ( [ uvfplim ] ) show that two of the new dimensionless parameters , are such that they represent the numerical value of the corresponding uv fixed points . therefore , part of the possible physical results are already encoded in the numerical value of this fixed point . the choice ( [ uvfp ] ) is further interesting in the sense that in the adimensional couplings only appear the adimensional quantities @xmath83 , and @xmath84 . with those parameters the adimensional gravitational couplings read @xmath85 and _ u(r)&= & \ { g_i ( g_i^3 ( r l_i+2 ^*_u)-12g^*3_u ^2 r^2 + 3 g_i^2g^*_u r ( r l_i+8 ^*_u ) ) . g_i^2g^*2_u r ( 2 ^2 r^2 l_i-11 + 24 r ^*_u ) + 6 g^*_u r ( g_i^2 + 3 g_ig^*_u r+2g^*2_u ^2 r^2 ) ( g^*_u-2 g_i ^*_u ) } . it is straight forward to depict the flow of this uv fixed point scenario @xmath81 and @xmath86 . the corresponding flow is shown in figure [ figflowuv ] , where the numerical values of @xmath87 and @xmath88 where taken from @xcite . schematic flow of the scale dependent couplings @xmath89 and @xmath90 for @xmath91 , @xmath92 , @xmath93 , and @xmath94 . the different curves correspond to @xmath95 which are depicted from left to right in @xmath96blue , green , black , orange , red@xmath97 . , width=377 ] in the context of erg induced coupling flows the dimensionless couplings are defined by the use of the energy scale @xmath4 [ gvong ] g(k)=k^2 g(k ) , ( k)=. those running couplings and the corresponding fixed points have been repeatedly calculated numerically @xcite . in order to obtain a tractable analytic solution for the running couplings ( [ gvong ] ) we will use a similar approximation procedure as it was used in @xcite . according to @xcite the flow equations are given by : @xmath98g(k ) \nonumber \\ \partial_{t}\lambda(k)&= & \beta_{\lambda } ( \lambda_{k},g_{k}),\end{aligned}\ ] ] where @xmath99 and @xmath100 is the anomalous dimension and the beta functions are for @xmath101 _ t g=_g= , and for @xmath102 _ t = _ = . with p_1&=&d ( 2+d ) g(k ) ( -3+d-16 g(k)+8 d g(k))+4 ( -1 + 10 d g(k)+d^2 g(k)-d^3 g(k ) ) + & & ( k)+4 ( 4 - 10 d g(k)-3 d^2 g(k)+d^3 g(k ) ) ( k)^2 - 16 ( k)^3 , and p_2=2 + 8 ( -d g(k)-(k)+(k)^2 ) . expanding those beta functions for small values of the couplings @xmath103 and for four space - time dimensions one gets [ betagapprox ] _ g = g(k)(2 -24 g(k ) ) and [ betalapprox ] _ = 12 g(k)-2 ( k ) . the approximated beta function ( [ betagapprox ] ) can be integrated to [ gerg ] g_erg(k)=. using ( [ gerg ] ) one can also integrate ( [ betalapprox ] ) giving [ lambdalit ] ( k)_erg&= & ^*_u+ _ 0 - . note that the fixed points for this flow equations are the gaussian fixed point ( @xmath104 ) and in this approximation the uv non gaussian fixed point with @xmath105 and @xmath106 . the values of the non gaussian fixed point were replaced in the solution by their symbols @xmath107 and @xmath108 , which will subsequently be treated as free parameters . the relation between @xmath101 and @xmath109 ( [ gerg ] ) can be inverted in order to express equation ( [ lambdalit ] ) in terms of @xmath110 , giving [ lambdaergg ] _ erg(g)= ^*_u + ( _ 0 g_0 ( 1-g /g^*_u ) -^*_u g^*_u ( 1- g /g^*_u ) ) this result can be plotted and compared to the coupling flow from figure [ figflowuv ] , the resulting graphical comparison of the two flows is shown in figure [ figflowuv2 ] . flow of the scale dependent couplings due to the erg result from equation ( [ lambdaergg ] ) which are depicted as dashed lines , and by using the parameters @xmath91 , @xmath92 , and @xmath111 with @xmath112 ( @xmath96blue , green , black , orange , red@xmath97 ) . this flow is compared to the black hole induced flow @xmath89 and @xmath90 as solid lines for the same parameters with the identifications @xmath113 and @xmath114 with the additional choice of @xmath93 and @xmath115 . , width=377 ] since the graphical similarity of the black hole induced flow and of the analytic erg flow in figure [ figflowuv2 ] is quite striking we will now proceed with an analytical comparison . for the adimensional gravitational constant one finds that the black hole induced @xmath116 ( [ adimgu ] ) and the erg result @xmath110 ( [ gerg ] ) are exactly identical if one uses the scale setting [ setscaler ] r . this scale setting result is interesting since intuitively one might have expected something like @xmath117 . please note that this scale setting definition still leaves @xmath118 arbitrary . the next step is to compare the couplings @xmath119 ( [ lambdalit ] ) and @xmath120 ( [ adiml ] ) . by using the scale setting ( [ setscaler ] ) one finds [ lambdabh ] _ u(k)&= & + & & \{g_0 ^ 2 g_ik^4 ( + 2 k^2 _ u^*)-12 g_u^*3 k^2 + 3g_0 g_i g_u^ * k^2 ( + 8 k^2 _ u^*)+g_u^*2 ( -11g_0 k^4 + 2 g_i(+12 k^2 _ u^ * ) ) . + & & .+6 ( 2 g_u^*2 + 3g_0 g_u^ * k^2 + g_0 ^ 2 k^4 ) ( g_u^*-2 g_i_u^ * ) } the limit @xmath121 suggests the identification @xmath122 . apparently , ( [ lambdabh ] ) is not identical to ( [ lambdalit ] ) , but the question is whether and to which extent both are similar . since equation ( [ lambdalit ] ) is an analytic approximation which is assumed to be best close to a small valued fixed point it is instructive to compare ( [ lambdalit ] ) and ( [ lambdabh ] ) in the uv regime for large values of @xmath109 . for the comparison we separate ( [ lambdabh ] ) in a logarithmic and a non - logarithmic part and perform a taylor expansion of the coefficients to lowest order in @xmath123 , @xmath108 , and @xmath107 , which is analogous to the expansion that was used when deriving @xmath124 . this gives [ lambdabhexp ] _ u(k)|_uv&= & ^*_u+ + & & - . one observes that the approximated black hole induced function ( [ lambdabhexp ] ) has the same functional structure as the approximated erg function ( [ lambdalit ] ) . even more , by choosing the remaining free constants to be @xmath125 and @xmath126 , the matching is exact . the approximated bh induced cosmological constant ( [ lambdabhexp ] ) is then identical to the ( approximated ) erg function ( [ lambdalit ] ) . please note that due to the uv approximation the infrared limit @xmath127 of ( [ lambdabhexp ] ) is a factor of two different from the infrared limit of the complete expression ( [ lambdabh ] ) . one relevant point in the discussion of the running parameter is the behavior of the anomalous dimensions in the uv or ir region . the anomalous dimension @xmath100 is connected to the previously defined beta function @xmath128 @xmath98g(k)\end{aligned}\ ] ] where @xmath99 . by using ( [ gvong ] ) @xmath100 can be written as @xmath129 the conditions for the existence of non trivial fixed points in the uv limit , non gaussian fixed points , is that : @xmath130 and @xmath131 . with those conditions the uv behavior of the anomalous dimension is given by @xmath132 ; @xmath133 . in order to study the behavior of the anomalous dimension in the ir region we select one of the trajectories which connects the uv non gaussian fixed point and run to the ir region as it is done in figure [ anom123 ] one can see from the figure that the trajectories have a classical behavior in the ir where the anomalous dimension goes to zero , while the anomalous dimension smoothly goes to the fixed point value in the uv : * in the ir limit @xmath134 ; @xmath135 * in the uv limit @xmath132 ; @xmath136 since in the previous discussion it has been shown that @xmath137 and @xmath138 are exactly equivalent due to the scale setting condition ( [ setscaler ] ) it is sufficient to work with one of the two , for example @xmath90 ( [ adimgu ] ) . considering this result we find that anomalous dimension has the form : @xmath139 one can see nicely that @xmath7 and @xmath118 only appear in pairs @xmath140 which explains that the different values of @xmath118 in figure [ anom123 ] actually correspond to a rescaling of @xmath7 in the same function @xmath141 . anomalous dimension as a function of @xmath7 , for @xmath142 , @xmath91 , with @xmath143 in red , @xmath144 in black , and @xmath145 in blue . , width=377 ] in the variety of erg calculations it turned out that even though the values of the fixed points @xmath146 and @xmath147 are scheme dependent , the product @xmath148 is rather robust throughout the different calculations . therefore , figure ( [ figltimesg ] ) shows the product @xmath149 compared to the product @xmath150 as a function of @xmath4 . product of the scale dependent couplings due to the erg result from equation ( [ lambdaergg ] ) which are depicted as dashed lines , and by using the parameters @xmath91 , @xmath92 , and @xmath111 with @xmath112 ( @xmath96blue , green , black , orange , red@xmath97 ) . this is compared to the black hole induced product @xmath151 as solid lines for the same parameters with the identifications @xmath113 and @xmath114 with the additional choice of @xmath93 and @xmath115 . , width=377 ] one can see from figure [ figltimesg ] that also the erg result and the black hole induced results for @xmath152 are in good agreement . in this paper we have studied the possibility of a scale dependent gravitational coupling @xmath0 and cosmological coupling @xmath1 . such scale dependent couplings have to be studied in the context of improved equations of motion ( [ eom2 ] ) . we asked the question of how the scale dependent couplings would have to look like in order to permit the most simple spherical symmetric metric solution with @xmath153 . solving the equations of motion lead to a non - trivial metric @xmath154 with non - trivial functions for @xmath0 , and @xmath1 . this solution contains four constants of integration . since a naive expansion with with one of the constants zero lead to unphysical predictions ( [ epand2 ] ) for large radii , we showed the existence of parameter choices where this problem does not exist ( see section [ subsecclaslike ] ) . from the functional form of the dimension - full coupling constants @xmath155 , we defined the dimensionless coupling constants @xmath156 with a fixed point in the regime of small radii . the flow of those dimensionless couplings was then compared to the ( approximated ) functional form of the running couplings of the erg approach @xmath157 . after a proper scale setting @xmath158 ( [ setscaler ] ) it was found that there exists an exact equivalence between @xmath159 and a structural correspondence between @xmath160 . by approximating the black hole induced result @xmath161 to the same order as the erg result @xmath124 one even finds an exact agreement between @xmath162 and @xmath124 . finally , the behavior of the anomalous dimension and of the product of the two couplings was discussed . given the good qualitative and quantitative agreement between the actual erg result and the black hole induced results , one is tempted to believe that the solution ( [ fvonr ] , [ gvonr ] , and [ lvonr ] ) is actually a selfconsistent and good approximation to a still unknown complete solution of the erg improved equations of motion . the solution is defined for all scales , but its correspondence to the erg black hole is expected to be best for small values of @xmath107 and @xmath108 and for large energy scales @xmath4 . this similarity ( and approximate correspondence ) was taken as an unexpected surprise and is the main result of this study . the work of b.k . was supported proj . fondecyt 1120360 and anillo atlas andino 10201 ; the work of c.c . was supported proj . fondecyt 1120360 and dgip grant 11.11.05 . coupling the einstein - hilbert action to matter with scale dependent couplings @xmath163 and @xmath164 gives the improved action @xmath165&=&\int_{m } d^4 x \sqrt{-g}\left(\frac{r-2\lambda _ { k}}{16\pi g _ { k}}+\mathcal{l}_m\right ) -\frac{1}{8 \pi}\int_{\partial m } d^3x \sqrt{-h}\frac{k}{g(k ) } \quad . \label{action}\end{aligned}\ ] ] the gibbons - hawking boundary term with the trace of the extrinsic curvature @xmath166 , can become relevant for the consistency of solutions that are not asymptotically flat such as ( a)ds . the equations of motion for the metric field in ( [ action ] ) are @xmath167 where the possible coordinate dependence of @xmath168 induces an additional contribution to the stress energy tensor @xcite t_=g _ k(g_- _ _ ) . demanding a self consistency of those equations of motion and a conserved stress energy tensor for matter _ t_=0 the following condition is found @xcite [ condi ] r_()- 2_()=0 . given a certain form of the scale dependent couplings @xmath168 and @xmath163 , for example from the erg approach , the above relation allows to relate this scale @xmath4 to the scalar curvature @xmath36 of a supposed solution . it it interesting to look for further solutions where @xmath169 . just adding an other unknown function to ( [ ansatz ] ) would however leave more functions than independent equations . thus , it is straight forward to study scenarios with @xmath170 and [ ansatz2 ] ds^2=-f(r ) dt^2 + 1/h(r)dr^2 + r^2 d^2 + r^2 ( ) d^2 . with this ansatz we did not succeed to find general solutions with variable @xmath0 . a special solution was however found [ solution2 ] f(r)&=&1 + h(r)&=&1 + + g(r)&=&c_2 it is quite interesting that for this solution the adimensional coupling is the square root of the adimensional coupling of the previous solution . given the schematic behavior of the induced coupling flow in figure [ figflowuv ] , one wonders how the classical - like choice of parameters fits into this picture . the relations ( [ c4new ] , [ c1new ] , [ c3new ] , and [ c2new ] ) in combination with ( [ replfp ] ) allow to express the dimensionless parameters @xmath171 , @xmath120 , @xmath172 , and @xmath116 in terms of the physical parameters @xmath173 , @xmath69 , @xmath68 and the undetermined scale @xmath22 that was introduced in order to define the dimensionless parameters . assuming a small physical cosmological constant @xmath69 adimensional constants that correspond to the solution ( [ fvonrnew]-[lvonrnew ] ) read approximately [ liclas ] l_i , s&&-8 + ( _ 0 ^ 4/3 ) + [ luclas ] _ u , s^*&=&-4 ( 4/3)^1/3 + [ giclas ] g_i , s&=&-(9/2)^1/3 + [ guclas ] g_u , s^*&= & g_0 ^2 those adimensional constants can be combined in various ways in order to study the fixed point behavior when varying @xmath22 and @xmath68 . but it allows also to form combinations which only depend on the general physical parameters @xmath69 and @xmath173 , like @xmath174 which establishes a global relation between the uv fixed point of @xmath175 and the infrared parameter @xmath176 . the positivity of the gravitational coupling , further suggests a negative value for the infrared parameter @xmath171 for positive @xmath69 . one also can ask whether the classical - like choice of parameters can be made compatible with the values of the uv fixed points @xmath92 , @xmath91 known from the erg approach @xcite . imposing those fixed point values on the classical - like parameters ( [ luclas]-[guclas ] ) allows to fix the parameters @xmath22 and @xmath69 and leaves @xmath173 and @xmath68 as only free physical parameters . in figure [ figflowuv3 ] it is shown how the corresponding flow would look like for the classical - like scenario . flow of the scale dependent couplings @xmath89 and @xmath90 for @xmath91 . the lower curves are plotted for @xmath93 , and @xmath94 whith @xmath95 which are depicted from left to right in @xmath96blue , green , black , orange , red@xmath97 . the black upper curves are the real values of the classical - like scenario for @xmath111 and from left to right @xmath177 . , width=377 ] please note that ( [ liclas]-[guclas ] ) is not the only possible choice for the parameters of this solution . it only means that this choice of parameters reproduces the standard form of the black hole metric ( [ fstandard ] ) despite of the fact that @xmath0 and @xmath1 are not constants . this result is new and unexpected , because it means that from solely observing the classical de sitter black hole metric to high precision in the range between the horizons , one can not conclude that @xmath12 and @xmath13 are actually constants . the horizons for this solution are given by the conditions @xmath179 and the black hole temperature is obtained by using the radial derivative at this point . this analysis reproduces only in very special cases such as the classical - like scenario all the features that are known for the classical solution ( [ fstandard ] ) . depending on the choice of the four parameters the horizon structure can be largely different . in most other cases the finding of the horizons boils down to solving non - analytic equations which only can be done numerically . this numerical study of the possible horizons and corresponding thermodynamical behavior is postponed to a future study . c. wetterich , phys . b * 301 * , 90 ( 1993 ) . d. dou and r. percacci , class . grav . * 15 * , 3449 ( 1998 ) [ arxiv : hep - th/9707239 ] . w. souma , prog . phys . * 102 * , 181 ( 1999 ) [ arxiv : hep - th/9907027 ] . m. reuter and f. saueressig , phys . d * 65 * , 065016 ( 2002 ) [ arxiv : hep - th/0110054 ] . d. f. litim , phys . lett . * 92 * , 201301 ( 2004 ) [ arxiv : hep - th/0312114 ] . p. fischer and d. f. litim , phys . b * 638 * , 497 ( 2006 ) [ arxiv : hep - th/0602203 ] . r. percacci , arxiv:0709.3851 [ hep - th ] . d. f. litim , arxiv:0810.3675 [ hep - th ] . g. narain and c. rahmede , class . * 27 * , 075002 ( 2010 ) [ arxiv:0911.0394 [ hep - th ] ] . k. groh and f. saueressig , j. phys . a * 43 * , 365403 ( 2010 ) [ arxiv:1001.5032 [ hep - th ] ] . a. bonanno and m. reuter , phys . b * 527 * , 9 ( 2002 ) [ arxiv : astro - ph/0106468 ] ; a. bonanno and m. reuter , phys . d * 65 * , 043508 ( 2002 ) [ arxiv : hep - th/0106133 ] . m. reuter and f. saueressig , new j. phys . * 14 * , 055022 ( 2012 ) , arxiv:1202.2274 . m. reuter and f. saueressig , arxiv:1205.5431 . d. benedetti , arxiv:1301.4422 [ hep - th ] . a. bonanno and m. reuter , phys . d * 60 * , 084011 ( 1999 ) [ arxiv : gr - qc/9811026 ] . a. bonanno and m. reuter , phys . d * 62 * , 043008 ( 2000 ) [ arxiv : hep - th/0002196 ] . m. reuter and h. weyer , phys . d * 69 * , 104022 ( 2004 ) [ arxiv : hep - th/0311196 ] . k. falls , d. f. litim and a. raghuraman , arxiv:1002.0260 [ hep - th ] . m. reuter and e. tuiran , arxiv:1009.3528 [ hep - th ] . f. girelli , s. liberati , r. percacci and c. rahmede , class . * 24 * , 3995 ( 2007 ) [ arxiv : gr - qc/0607030 ] . a. codello , r. percacci and c. rahmede , int . j. mod . a * 23 * , 143 ( 2008 ) [ arxiv:0705.1769 [ hep - th ] ] . b. koch , phys . b * 663/4 * , 334 ( 2008 ) [ arxiv:0707.4644 [ hep - ph ] ] . b. koch , phys . b * 663 * , 334 ( 2008 ) . t. burschil and b. koch , zh . * 92 * , 219 ( 2010 ) [ arxiv:0912.4517 [ hep - ph ] ] . r. casadio , s. d. h. hsu and b. mirza , arxiv:1008.2768 [ gr - qc ] . b. koch , f. saueressig , renormalization group improved ( anti ) de sitter black holes , in preparation . s. i . nojiri and s. d. odintsov , econf c * 0602061 * , 06 ( 2006 ) [ int . j. geom . phys . * 4 * , 115 ( 2007 ) ] [ hep - th/0601213 ] . t. p. sotiriou and v. faraoni , rev . phys . * 82 * , 451 ( 2010 ) [ arxiv:0805.1726 [ gr - qc ] ] . s. i . nojiri and s. d. odintsov , phys . * 505 * , 59 ( 2011 ) [ arxiv:1011.0544 [ gr - qc ] ] . v. faraoni , arxiv:0810.2602 [ gr - qc ] . a. de la cruz - dombriz , a. dobado and a. l. maroto , phys . d * 80 * , 124011 ( 2009 ) [ erratum - ibid . d * 83 * , 029903 ( 2011 ) ] [ arxiv:0907.3872 [ gr - qc ] ] . g. cognola , e. elizalde , s. i . nojiri , s. d. odintsov and s. zerbini , jcap * 0502 * , 010 ( 2005 ) [ hep - th/0501096 ] . s. h. hendi and d. momeni , eur . j. c * 71 * , 1823 ( 2011 ) [ arxiv:1201.0061 [ gr - qc ] ] . g. j. olmo , phys . d * 75 * , 023511 ( 2007 ) [ gr - qc/0612047 ] . y. s. myung , phys . d * 84 * , 024048 ( 2011 ) [ arxiv:1104.3180 [ gr - qc ] ] . d. psaltis , d. perrodin , k. r. dienes and i. mocioiu , phys . * 100 * , 091101 ( 2008 ) [ phys . * 100 * , 119902 ( 2008 ) ] [ arxiv:0710.4564 [ astro - ph ] ] . e. barausse and t. p. sotiriou , phys . * 101 * , 099001 ( 2008 ) [ arxiv:0803.3433 [ gr - qc ] ] . a. larranaga , arxiv:1108.6325 [ gr - qc ] . c. brans and r. h. dicke , phys . rev . * 124 * , 925 ( 1961 ) . c. h. brans , phys . rev . * 125 * , 2194 ( 1962 ) . s. w. hawking , commun . phys . * 25 * , 167 ( 1972 ) . h. kim , phys . d * 60 * , 024001 ( 1999 ) [ gr - qc/9811012 ] . h. s. kim , h. maeda and g. giribet , jhep * 1111 * , 015 ( 2011 ) [ arxiv:1105.1331 [ gr - qc ] ] . c. j. gao and s. n. zhang , phys . d * 70 * , 124019 ( 2004 ) [ hep - th/0411104 ] . c. j. gao and s. n. zhang , phys . b * 605 * , 185 ( 2005 ) [ hep - th/0411105 ] . m. reuter and h. weyer , jcap * 0412 * , 001 ( 2004 ) [ arxiv : hep - th/0410119 ] . m. reuter and h. weyer , phys . d * 70 * , 124028 ( 2004 ) [ arxiv : hep - th/0410117 ] . d. c. rodrigues , p. s. letelier and i. l. shapiro , jcap * 1004 * , 020 ( 2010 ) [ arxiv:0911.4967 [ astro-ph.co ] ] . s. h. tye and j. xu , arxiv:1008.4787 [ hep - th ] . a. bonanno , a. contillo and r. percacci , arxiv:1006.0192 [ gr - qc ] . o. piatella , d. rodriguez , i. shapiro , and b. koch , work in preparation .

because of their simple crystal structure , the perovskite oxides present a special opportunity for the development of a detailed theoretical understanding of the ferroelectric phase transition . within this family of materials , one finds transitions to a wide variety of low - symmetry phases , including ferroelectric and antiferroelectric transitions . both first- and second - order transitions are observed , with a full spectrum of transition behavior ranging from displacive to order - disorder . the properties of batio@xmath0 , a much - studied prototypical example of this class of compounds,@xcite exemplify this rich behavior . batio@xmath0 undergoes a succession of first - order phase transitions , from the high - temperature high - symmetry cubic perovskite phase ( fig . [ stru ] ) to slightly distorted ferroelectric structures with tetragonal , orthorhombic and rhombohedral symmetry . there is increasing evidence that the cubic - to - tetragonal transition , at first thought to be of the simple displacive kind , may instead be better described as of the order - disorder type . the variety exhibited by the perovskite oxides shows that the phase transformation behavior depends on details of the chemistry and structural energetics of each particular compound . therefore , it is of the first importance to develop a microscopic theory of the materials properties which determine the ordering of the phases , the character and thermodynamic order of the transitions , and the transition temperatures . the value of a microscopic approach has long been appreciated , but its realization was hindered by the difficulty of determining microscopic parameters for individual compounds . the forms of phenomenological model hamiltonians @xcite were limited by the available experimental data , leading to oversimplification and ambiguities in interpretation . for the perovskite oxides , empirical @xcite and non - empirical pair potential methods @xcite did not offer the high accuracy needed for the construction of realistic models . first - principles density functional calculations offer an attractive approach for enhancing our microscopic understanding of perovskites and other ferroelectrics . the all - electron full - potential linearized - augmented - plane - wave ( flapw ) method has been used by several groups to study ferroelectricity in perovskites within the local density approximation ( lda).@xcite recently , king - smith and vanderbilt performed a systematic study of structural and dynamical properties and energy surfaces for eight common perovskites,@xcite using the first - principles ultra - soft pseudopotential method and the lda . these calculations demonstrate that ferroelectricity in the perovskites reflects a delicate balance between long - range electrostatic forces which favor the ferroelectric state , and short - range repulsions which favor the cubic phase . while constrained to calculations of zero - temperature properties , these calculations yield correct predictions of ground state structures and occurrence of ferroelectric phases for certain materials . they show that high - quality lda calculations can provide considerable insight into the nature of the total - energy surface in the perovskites . for further insight into the energetics of ferroelectric compounds , the polarization generated by various distortions can be studied directly , using a recent first - principles method by king - smith and vanderbilt.@xcite this approach has been applied to the investigation of the zone - center phonons in the common perovskite oxides.@xcite the application of these first - principles methods can clearly form a foundation for the realistic study of the finite - temperature phase transitions . while an _ ab - initio _ molecular - dynamics simulation of the structural phase transition is not computationally feasible at present , we purse an alternative first - principles approach to study ferroelectric phase transitions and demonstrate its application to batio@xmath0 . in particular , we ( i ) construct an effective hamiltonian to describe the important degrees of freedom of the system,@xcite ( ii ) determine all the parameters of this effective hamiltonian from high - accuracy _ ab - initio _ lda calculations,@xcite and ( iii ) carry out monte carlo ( mc ) simulations to determine the phase transformation behavior of the resulting system . an abbreviated presentation of this work has already appeared in ref . . the remainder of this paper is organized as follows . in sec . ii , we go through the detailed procedure for the construction of the effective hamiltonian , and give the explicit formula . in sec . iii , we describe our first - principles calculations and the determination of the expansion parameters in the hamiltonian . the technical details of the monte carlo simulation are presented in sec . iv . in sec . v , we report our calculated transition temperatures , order parameters , and phase diagram , as well as thermodynamic order and nature of the phase transitions . the role of different interactions in determining the phase transition behavior is also discussed . vi concludes the paper . the central quantity for studying the equilibrium properties of a system at finite temperature is its partition function . this can be determined from the energy surface , i.e. , the total potential energy as a functional of the atomic coordinates . since the contribution to the partition function decreases exponentially with increasing energy , it is possible to obtain an accurate partition function for low - temperature applications from a simplified energy surface including only low - energy configurations . our goal is to construct a parameterized hamiltonian which ( i ) is _ ab initio _ , involving no empirical or semi - empirical input ; ( ii ) results in an accurate partition function for the temperature range of interest ; ( iii ) is fully specified by a few _ ab - initio _ total energy calculations ; and ( iv ) involves only approximations that are systematically improvable and removable . our first fundamental approximation is to use an energy surface represented by a low - order taylor expansion . both experiments and first - principles total energy calculations suggest that the ferroelectric ( fe ) phase transition involves only very small atomic displacements and strain deformations from the equilibrium cubic structure . it is reasonable to assume that all the atomic configurations with significant contribution to the partition function would be close to this cubic structure in the temperature range of interest . thus , it is natural to represent the energy surface by a taylor series in the displacements from the cubic structure . we include up to fourth - order terms in our expansion ; this is clearly a minimum , since ferroelectricity is intrinsically an anharmonic phenomenon . by including higher - order terms , this approximation could later be systematically improved . it is convenient to describe the small distortions from the cubic structure in terms of the three acoustic and twelve optical normal - mode coordinates per k - point . while this could be regarded as only a change of basis , it motivates our second fundamental approximation , which is to restrict the expansion to include only low - energy distortions . to achieve this separation , we note that both experimentally - measured and lda - calculated phonon dispersion relations@xcite show that only the lowest to modes ( soft modes ) and long - wavelength acoustic phonons ( strain variables ) make significant contributions to the phonon density of states at low energy . experimental studies also suggest that the fe phase transitions are accompanied by a softening of the lowest to mode and the appearance of a strain . all other phonons are hardly affected by the transitions . it is then our second approximation to express the energy surface only as a function of the soft - mode amplitudes and strain . this approximation reduces the number of degrees of freedom per cell from fifteen to six , and greatly reduces the number of interaction parameters needed . if necessary , this approximation could later be relaxed by including additional modes . it is convenient to describe the soft mode over the whole brillouin zone ( bz ) in terms of a collective motion of `` local modes , '' just as one describes an acoustic phonon in terms of a collective displacement of individual atoms . however , there is more than one choice of local mode which will generate the same soft mode throughout the bz ; an intelligent choice can simplify the hamiltonian and reduce the number of calculations needed.@xcite first , the local mode should be as symmetric as possible , so as to minimize the number of expansion parameters needed . second , the interactions between local modes at different sites are more difficult to treat than their on - site energy , so the local mode should be chosen so as to minimize intersite interactions . for perovskite abo@xmath0 compounds , the highest symmetry is achieved by centering the local mode on either atom a or b. in the case of batio@xmath0 , the ti o bond is much stronger than the ba o bond and the motion of the ti is more important in the fe transition , so we choose the local mode which is centered on the ti atom . the soft zone - center ( @xmath1=0 ) fe mode in batio@xmath0 is a @xmath2 mode which can be characterized by the four parameters @xmath3 , @xmath4 , @xmath5 , and @xmath6 ( for a mode polarized along the @xmath7th cartesian direction , these refer to the displacements of the a atom , the b atom , the o atom that forms a b o bond along direction @xmath7 , and the other two o atoms , respectively ) . we take the corresponding local mode to consist of a motion of the central a atom by amount @xmath3 , the eight neighboring b atoms by amounts @xmath8 , and the six neighboring o atoms by amounts @xmath9 or @xmath10 , along the @xmath7th cartesian direction . this mode is illustrated in fig . ( [ stru ] ) ; its amplitude is denoted @xmath11 . an arbitrary @xmath12 soft mode can then be realized as a linear superposition of these local modes having identical amplitudes @xmath13 in every cell . the harmonic interactions between the local - mode amplitudes @xmath14 connecting neighboring cells @xmath15 must be chosen to reproduce the harmonic behavior of the soft - mode branch throughout the bz . long - range coulomb forces are known to play an important part in these interactions ; they are characterized in terms of the calculated born ( or `` transverse '' ) effective charges.@xcite thus , the harmonic intersite interactions are represented by a sum of two contributions : an infinite - range piece that is precisely the interaction of point dipoles whose magnitude is given by the born effective charge , and corrections which we take to be of covalent origin and therefore local . to be completely general , anharmonic intercell interactions between neighboring @xmath16 would likewise have to be included . instead , we include only _ on - site _ anharmonic interactions , which are chosen in such a way that the anharmonic couplings for @xmath12 modes of the real system are correctly reproduced . this `` local anharmonicity approximation '' is an important feature which helps make our scheme tractable and efficient . to go beyond this approximation , one could carry out a careful series of frozen - phonon lda calculations on supercells to determine anharmonic couplings at other points in the bz . however , past experience has shown that calculations of this kind are very cumbersome because of the large number of parameters which have to be determined.@xcite with these approximations , our hamiltonian consists of five parts : a local - mode self energy , a long - range dipole - dipole interaction , a short - range interaction between soft modes , an elastic energy , and an interaction between the local modes and local strain . symbolically , @xmath17 where @xmath18 is the local soft - mode amplitude vector , and @xmath19 is the six - component local strain tensor in voigt notation ( @xmath20 , @xmath21 ) . in the following subsections , we present the explicit formulae for these five contributions . the first term is @xmath22 where @xmath23 is the energy of an isolated local mode at cell @xmath24 with amplitude @xmath16 , relative to that of the perfect cubic structure . to describe the fe phase , @xmath23 must contain anharmonic as well as harmonic contributions . since the reference structure is cubic , only even - order terms can enter ; we choose to truncate at fourth order . symmetry considerations then require that it take the form @xmath25 where @xmath26 , and @xmath27 , @xmath28 , and @xmath29 are expansion parameters to be determined from first - principles calculations . the second term in the effective hamiltonian represents long - range interactions between local modes . only dipole - dipole interactions are considered , since higher - order terms tend to be of short range and their effect will be included in the short - range contribution @xmath30 . the dipole moment associated with the local mode in cell @xmath15 is @xmath31 . here , @xmath32 is the born effective charge for the soft mode , which can be obtained as @xmath33 from the eigenvector of the soft mode , once the born effective charges for the ions are known.@xcite in atomic units ( energy in hartree ) , we have @xmath34 here , @xmath35 is the optical dielectric constant of the material , @xmath36 , @xmath37 , and @xmath38 . in practice , ( [ dipolesum ] ) is not directly useful for three - dimensional systems with periodic boundary conditions ; instead , we use an ewald construction to evaluate @xmath39 . we effectively terminate the sum in such a way that the @xmath12 modes of the supercell will represent physical to(@xmath40 ) modes . for a to mode , the induced depolarization electric field is zero ; from the point of view of the dipole sum , it is as though the material were surrounded by a layer of metal . in the ewald construction , this is equivalent to setting the surface terms to zero.@xcite under these conditions , and choosing the decay @xmath41 of the gaussian charge packets to be small enough so that the real - space summation can be entirely neglected , we have @xmath42 \ ; , \ ] ] where @xmath43 is the cell volume and * g * is the reciprocal lattice vector . because of its long - range nature , the calculation of @xmath39 is the most time - consuming part of our monte carlo simulations . it is thus worth some special treatment to reduce the computational load . in principle , the term @xmath44 appearing in the denominator of eq . ( [ dipolesum ] ) should be strain - dependent . however , as we have chosen to expand the intersite interactions between local modes only up to harmonic order , it is consistent to ignore this effect , since strain - induced changes of the dipole - dipole interaction will enter only at higher order . this is equivalent to fixing the reciprocal lattice vectors @xmath45 and all the atomic position vectors @xmath24 . the dipole energy can then be written as @xmath46 with @xmath47 \ ; .\ ] ] here , @xmath28 and @xmath48 denote cartesian components . the matrix @xmath49 is thus treated as a constant ; it is calculated once and for all , and stored for later calculation of the dipole energy . this strategy increases the computational efficiency by at least one order of magnitude . @xmath50 is the energy contribution due to the short - range interactions between neighboring local modes , with dipole - dipole interaction excluded . this contribution stems from differences of short - range repulsion and electronic hybridization between two adjacent local modes and two isolated local modes . together with the dipole - dipole interaction , this interaction determines the soft - mode energy away from the zone center . expanded up to the second order , it can be written as , @xmath51 the coupling matrix @xmath52 is a function of @xmath53 and should decay very fast with increasing @xmath54 . here , we consider the short - range interaction up to third nearest neighbor ( nn ) , whose local mode shares atoms with the local mode on the origin . local modes between further neighbors involve displacements of atoms at least two hops away ( in tight - binding language ) and their core - core repulsion or hybridization should be much less important than the dipole - dipole interaction which is taken care of in @xmath39 . the interaction matrix @xmath52 can be greatly simplified by symmetry . for a cubic lattice , we have , @xmath55 where @xmath56 is the @xmath28 component of @xmath57 . so we have only seven interaction parameters for a cubic lattice . the coefficients @xmath58 , @xmath59 , ... , @xmath60 in the above equations have physical meanings that are sketched schematically in fig . for example , @xmath58 represents the interaction strength of `` @xmath61 '' like interactions between first neighbors . we will describe the state of elastic deformation of the batio@xmath0 crystal using local strain variables @xmath62 , where the voigt convention is used ( @xmath63 ) and @xmath24 labels a cell center ( ti ) site . in fact , the six variables per unit cell @xmath64 are not independent , but are obtained from three independent displacement variables per unit cell . in our analysis , these are taken as the dimensionless displacements @xmath65 ( in units of the lattice constant @xmath66 ) defined at the unit cell corner ( ba ) positions @xmath67 . in terms of these , the _ inhomogeneous _ strain variables @xmath68 are defined in the next subsection . because of our use of a periodic supercell in the monte carlo simulations , however , homogeneous strain deformations are not included in the configuration space @xmath69 . therefore , we introduce six additional _ homogeneous _ strain components @xmath70 to allow the simulation cell to vary in shape . the total elastic energy is expanded to quadratic order as @xmath71 where the homogeneous strain energy is simply given by @xmath72 here @xmath73 , @xmath74 , and @xmath75 are the elastic constants expressed in energy units ( @xmath76 , etc . ) , and @xmath77 is the number of primitive cells in the supercell . rather than use an expression like ( [ eqhom ] ) for the inhomogeneous strain energy , we have found it preferable to express this part directly in terms of the @xmath65.@xcite this approach keeps the acoustic phonon frequencies well behaved throughout the brillouin zone . to satisfy requirements of invariance under translations and rotations of the crystal as a whole , the energy is expanded in scalar products of differences between the @xmath65 . the cubic crystal symmetry leads to a great reduction of the independent parameters in the expansion . the energies of the long - wavelength strain deformations can be reproduced by an expansion of the form @xmath78 ^ 2 \nonumber\\ \phantom{\big| } & + & \gamma_{12 } [ v_x({\bf r}_i ) - v_x({\bf r}_i \pm { \bf x } ) ] [ v_y({\bf r}_i ) - v_y({\bf r}_i \pm { \bf y } ) ] \nonumber\\ \phantom{\big| } & + & \gamma_{44 } [ v_x({\bf r}_i ) - v_x({\bf r}_i \pm { \bf y } ) + v_y({\bf r}_i ) - v_y({\bf r}_i \pm { \bf x } ) ] ^2 \nonumber\\ & + & { \rm \;cyclic\;permutations\ ; } \big\ } \label{eqinhom}\end{aligned}\ ] ] corresponding to bond stretching , bond correlation , and bond bending , respectively . here , @xmath79 , @xmath80 , @xmath81 , and @xmath82 indicates multiple terms to be summed . the @xmath29 coefficients are related to the elastic constants by @xmath83 , @xmath84 , and @xmath85 . to describe the coupling between the elastic deformations and the local modes , we use the on - site interaction @xmath86 as a result of cubic symmetry , there are only three independent coupling constants @xmath87 : @xmath88 the strain contains both homogeneous and inhomogeneous parts . the latter are expressed in terms of the local displacement vectors @xmath90 as follows . we first define the six average differential displacements associated with site @xmath24 as @xmath91 \;,\nonumber\\ \delta v_{xy } & = & \sum_{{\bf d } = 0,{\bf y},{\bf z},{\bf y}+{\bf z } } \left [ v_y ( { \bf r}_i - { \bf d } - { \bf x } ) - v_y ( { \bf r}_i - { \bf d } ) \right ] \;,\nonumber\end{aligned}\ ] ] and their cyclic permutations , recalling that @xmath65 is associated with position @xmath67 . then @xmath92 , @xmath93 , etc . we have shown that , with the two approximations we made , the total energy functional of the perovskite system is fully specified by a set of parameters . these parameters can be obtained from first - principles calculations . we use density - functional theory within the local density approximation ( lda ) . the technical details and convergence tests of the calculations can be found in refs . . the most important feature of the calculations is the vanderbilt ultra - soft pseudopotential,@xcite which allows a low energy cutoff to be used for first - row elements . this makes high - accuracy large - scale calculations of materials involving oxygen and @xmath94 transition metal elements affordable . the ultra - soft pseudopotential also allows for exceptionally transferable pseudopotentials . it ensures that all - electron and pseudo - atom scattering properties agree over a very large energy range and preserves the chemical hardness of the atom . a generalized kohn - sham functional is directly minimized using a preconditioned conjugate gradient method.@xcite we use a ( 6,6,6 ) monkhorst - pack k - point mesh@xcite for single - cell calculations , i.e. , 216 k - points in the full brillouin zone ( fbz ) . for supercell calculations , the k mesh is kept the same to minimize errors due to the k - point sampling . therefore , a smaller number of k - points is used because of the smaller fbz . we start with the construction of the local mode vectors . all the eigenvalues and eigenvectors of the force - constant matrix at @xmath12 for the cubic batio@xmath0 structure are calculated from frozen - phonon calculations , as in ref . . the mode with imaginary frequency is identified as the soft mode . the soft - mode eigenvector has been reported previously@xcite as @xmath95 , @xmath96 , @xmath97 , and @xmath98 . the local mode is then constructed from it using the scheme proposed in ii.1 ( fig.[stru ] ) . determination of many of the parameters in the effective hamiltonian involves only calculations for zone - center distortions . these parameters have been reported previously.@xcite they include the fourth - order terms of on - site energy @xmath28 and @xmath29 ; the elastic constants @xmath99 ; and the on - site elastic - mode interaction parameters @xmath100 . the mode effective charge @xmath32 of eq . ( [ effcharge ] ) is calculated from the values @xmath101=2.75 , @xmath102=7.16 @xmath103=@xmath1045.69 , and @xmath105=@xmath1042.11 published in ref . . ( the resulting value @xmath32=9.96 is slightly different from the one given in ref . ; there , the eigenvector of the _ dynamical _ matrix , not the _ force - constant _ matrix , was used . ) we use the experimental value @xmath106 of the optical dielectric constant , since for this quantity , the lda seems not to be a well - justified approximation , while exact density - functional theory results are not accessible . we find , however , that the effect of a small inaccuracy in the dielectric constant affects thermodynamic properties such as transition temperatures only slightly . the second - order energy parameter @xmath107 for zone - center distortions is a linear combination of the local mode self energy parameter @xmath27 , the intersite interactions @xmath108 , and the dipole - dipole interaction . the calculations of intersite interaction parameters involve determination of the energy for distortions at the zone - boundary k - points x=(@xmath109,0,0 ) , m=(@xmath109,@xmath109,0 ) , and r=(@xmath109,@xmath109,@xmath109 ) , where @xmath66 is the lattice constant . five frozen - phonon calculations on doubled unit cells are sufficient to extract all the information available from these k - points . the arrangement of the local mode vectors for each case , as well as for the zone - center distortion at @xmath40=(0,0,0 ) , are shown in fig . the actual ionic configurations are constructed by superpositions of displacements associated with adjacent local modes . for example , letting @xmath110 be the amplitude of the ti - centered local mode defined in sec . [ sec : approx ] , the displacement of the ti atoms is just @xmath111 in fig . [ cell](a ) and @xmath112 in fig . [ cell](b ) , while ba is affected by eight neighboring local modes so that its displacement is @xmath113 in fig . [ cell](a ) and @xmath114 in fig . [ cell](b ) . the above five doubled - cell calculations can be used to determine the parameters @xmath58 , @xmath59 , @xmath115 , @xmath116 , and @xmath117 . the determination of @xmath118 requires a four - cell calculation involving 20 atoms with low symmetry [ fig . [ cell](g ) ] . table [ table1 ] lists the energy expressions for all the configurations calculated in terms of the quadratic expansion parameters . a further decomposition of @xmath119 and @xmath60 would require an expensive eight - cell calculation . furthermore , the interaction parameter @xmath60 is the third nearest - neighbor interaction and is thus presumably not very important . this argument is justified by our monte carlo simulations which show that the calculated transition temperature is insensitive to different decompositions of @xmath119 and @xmath60 . this prompts us to make an approximate decomposition based on a simple physical argument : we expect the interaction to be smallest for two interacting local modes oriented such that reversing the relative sign of the vectors produces the least change of bond lengths . applied to third nearest neighbors , this argument implies @xmath120 , thus fixing the value of @xmath60 . the resulting interaction parameters are shown in table [ table2 ] , together with other parameters published previously . it may be surprising to see that the on - site @xmath27 is positive , while the cubic structure is known to be unstable against @xmath12 distortion . the cubic structure is thus stable against forming an isolated local mode ; instability actually comes from the intersite interactions between local modes . to be more precise , we find that it is the long - range coulomb ( dipole - dipole ) interaction which makes the ferroelectric state favorable . if we turn off the dipole - dipole interaction by setting the effective charge @xmath121 , we find that the ferroelectric instability disappears . this is consistent with the previous point of view that long - range coulomb forces favor the ferroelectric state , while short - range repulsions favor the nonpolar cubic state . from table [ table2 ] , we see that the intersite interaction parameters decay very fast with increasing distance , indicating the short - range nature of the intersite interactions after the long - range coulomb interactions have been separated out . the ratio of the magnitudes of the strongest first- , second- , and third - neighbor interactions turns out to be approximately @xmath122 . this decays even faster than the dipole - dipole interactions , for which the corresponding ratio ( @xmath123 ) is @xmath124 . these results help justify our approximation of including only up to third nearest neighbors for the short - range interactions . for the quantitative study of non - universal finite - temperature behavior of a given model , monte carlo simulation@xcite has emerged as the most reliable and powerful technique . it is especially appropriate for a model such as ours , with two continuous vector degrees of freedom per unit cell and both short and long range interactions , for which analytical approaches such as renormalization group or high - temperature expansions would be cumbersome and involve additional approximations . in comparison , the monte carlo approach requires only the ability to compute changes in total energy as the configuration is changed . furthermore , with suitable analysis of statistical errors and finite - size effects , the results of monte carlo simulation can be made arbitrarily accurate . finally , with little additional effort , a number of physical quantities can be computed to aid in characterization of the transition . we solve the effective hamiltonian [ eqs . ( [ eqself ] ) , ( [ eqdpl ] ) , ( [ eqshort ] ) , ( [ eqstrain ] ) , and ( [ eqelint ] ) ] using monte carlo simulations with the metropolis algorithm@xcite on an @xmath125 cubic lattice with periodic boundary conditions . since most energy contributions ( except @xmath39 ) are local , we choose the single - flip algorithm . that is , a trial move consists of an attempted update of a single variable , after which the total energy change is calculated to determine whether to accept the move . the step sizes are adjusted to ensure an acceptance ratio of approximately 0.2 . in one monte carlo sweep ( mcs ) , we first make a trial move on each @xmath16 in sequence , then each @xmath126 in sequence , then iterate several times ( typically @xmath127 times ) on the homogeneous strain variables . for @xmath128 , each mcs takes about one second on an hp 735 workstation . the typical correlation time for the total energy is found to be several hundred mcs close to the phase transition ; this long correlation time makes certain new mc techniques using energy distribution functions@xcite unfavorable . the correlation times for the local mode amplitudes are one order of magnitude shorter , and thus 10,000 mcs are usually enough to equilibrate and to obtain averages of local mode variables with statistical error @xmath129% . in our simulation , we concentrate on identifying the succession of low - temperature phases , determining the phase transition temperatures , and extracting qualitative features of the transitions . this analysis will allow us to identify the features of the effective hamiltonian which most strongly affect the transition properties . for these purposes , it is most convenient to monitor directly the behavior of the homogeneous strain and the vector order parameter . in the case of the ferroelectric phase transition , the latter is just the average local - mode vector @xmath130 , which is proportional to the polarization . here , @xmath16 is the local mode vector at site @xmath15 and @xmath77 is the total number of sites . as a reference , the average local mode amplitude @xmath131 is also monitored . to avoid effects of symmetry - equivalent rotations of the order parameter and to identify the different phases clearly , we accumulate the absolute values of the largest , middle , and smallest components of the averaged local - mode vector for each step , denoted by @xmath132 , @xmath133 , and @xmath134 , respectively ( @xmath135 ) . the cubic ( c ) , tetragonal ( t ) , orthorhombic ( o ) , and rhombohedral ( r ) phases are then characterized by zero , one , two , and three non - zero order - parameter components , respectively . the effect of symmetry - equivalent rotations on the homogeneous strain is handled analogously , with the largest , the medium and the smallest linear strain components denoted by @xmath136 , @xmath137 and @xmath138 , respectively , and shear strain components by @xmath139 , @xmath140 and @xmath141 . the ferroelectric phase transition is very sensitive to hydrostatic pressure , or equivalently , to lattice constant . the lda - calculated lattice constants are typically 1% too small , and even this small error can lead to large errors in the zero - pressure transition temperatures . one approach which largely compensates for the effect of this systematic error is to exert a negative pressure that expands the lattice constant to the experimental value . we determine the value of the pressure by calculating volumes for four different phases and comparing with experimental measurements.@xcite we find @xmath142 gpa gives the best overall agreement ( although the application of pressure does lead to a slight change in the low temperature structure ) . except for the simulations for the construction of the temperature - pressure phase diagram , the following simulations and analysis are for this pressure . in this section , the finite - temperature behavior of the model is presented and analyzed . first , we examine the order parameters as a function of temperature in a typical simulation to obtain a measure of the transition temperatures . from the results of simulations for a range of pressures , we construct the temperature - pressure phase diagram . for the system at ambient pressure , more detailed simulations are performed . the order of the transitions , the nature of the paraelectric phase and the properties of the low - temperature phases are investigated and compared with experimental observations . finally , the roles played by different terms in the effective hamiltonian and the sensitivity of the results to various approximations are examined . we start the simulations at a high temperature ( @xmath143k ) and equilibrate for 10,000 mcs . the temperature is then reduced in small steps , typically 10k . after each step , the system is allowed to equilibrate for 10,000 mcs . the order parameter averages are then accumulated over a period of 7,000 mcs , yielding a typical standard deviation of less than 10% . the temperature step size is reduced and the number of mcs is doubled for temperatures close to the phase transition . [ u - t ] shows the averaged local - mode vectors @xmath132 , @xmath133 , @xmath134 and averaged local mode amplitude @xmath110 as functions of temperature in a typical simulation for an @xmath128 lattice at @xmath142 gpa . at high temperatures , @xmath132 , @xmath133 , and @xmath134 are all very close to zero . as the system is cooled down below 295k , @xmath132 increases and becomes significantly larger than @xmath133 or @xmath134 . this indicates the transition to the tetragonal phase . two additional phase transitions occur as the temperature is reduced further . the t o transition ( sudden increase of @xmath133 ) occurs at 230k and the o r transition ( sudden increase of @xmath134 ) occurs at 190k . the sequence of transitions exhibited by the simulation is the same as that observed experimentally . the averaged homogeneous strain variables obtained from the above simulation are shown in fig . [ e - t ] . these strains are measured relative to the lda calculated equilibrium cubic structure , so the linear strains are significantly non - zero at higher temperatures due to the negative pressure applied . as expected , the simulation cell changes shape at the same temperatures at which the jumps of order parameter components are observed . at high temperatures , we have approximately @xmath144 and @xmath145 , corresponding to the cubic structure . as the system is cooled down , the shape of the simulation cell changes to t , o , and r phases . the orthorhombic ( o ) structure has a non - zero shear strain , in agreement with the centered orthorhombic structure observed experimentally . quantitatively , our calculated distortions are also in good agreement with the experiment , with the calculated distortions tending to be slightly smaller . for example , @xmath146 for the tetragonal phase is 1.1% as measured from experiment@xcite and 0.9% from our calculation . the simulations are repeated for a range of applied pressures to obtain the temperatures at which the order parameter components and homogeneous strain jump on cooling down . the resulting temperature - pressure phase diagram is shown in fig . [ p - t ] . ( this measure of the transition temperature is actually a lower bound , due to hysteresis around @xmath147 for t o and o r transitions and negligible for the c t transition , to be discussed further below . ) all three transition temperatures decrease almost linearly with increasing pressure . at the experimental lattice constant , the values for @xmath148 are found to be @xmath10428 , @xmath10422 , and @xmath10415 k / gpa for the c t , t o , and o r transitions respectively . the experimental values for the c t transition range from @xmath10440 k / gpa@xcite to @xmath10466 k / gpa.@xcite for the t o transition the measured value is @xmath10428 k / gpa,@xcite and for the o r transition it is @xmath10410 to @xmath10415 k / gpa.@xcite at pressures as high as @xmath149gpa , the sequence of phases c t o r is still observed in the simulation . when the pressure is increased further , the phase boundary in the simulation becomes unclear due to fluctuations . our calculated critical pressure ( beyond which the cubic structure is stable at t=0 k ) is @xmath150gpa . taking into account the pressure correction for the lda volume underestimate , this corresponds to a predicted physical @xmath151 gpa . we are not aware of any experimental value for @xmath152 with which to compare this value . however , we find that the magnitude of our @xmath148 is significantly smaller than experimental value , at least for the c t and t o transitions . this may partly be due to the neglect of higher - order strain coupling terms in the effective hamiltonian . we have tested the effect of including a volume dependence for the short - range interaction parameters @xmath108 . this correction does not change the sequence of phases , and it only increases the magnitudes of @xmath148 slightly . therefore , our results are reported without this correction . the accuracy of the phase diagram may be further improved by including higher - order terms in the elastic energy , or the coupling of @xmath108 to anisotropic strain . for the investigation of the order of the transitions , the nature of the paraelectric phase and the properties of the low - temperature phases , we performed more detailed simulations at @xmath142gpa for the three system sizes @xmath153 , @xmath154 and @xmath155 . in the cooling - down simulations , the length of each simulation was increased from 10,000 to up to 35,000 mcs at temperatures close ( @xmath8210k ) to the phase transition to include a longer equilibration . the size of the temperature step was decreased to 5k or less in the vicinity of the transition . in addition , a heating simulation was performed , starting from the lower - temperature phase , to detect any possible hysteresis . the calculated transition temperatures , obtained as the average of the cooling and heating transition temperatures , are given in table [ table3 ] . the error estimates in the table are determined by the width of the hysteresis , which persists even for the longest simulation lengths considered . t transition temperature for @xmath153 is difficult to identify because of large fluctuations between phases . ) the calculated transition temperatures are well converged with respect to system size , and are in reasonable agreement with experiment . table [ table3 ] also gives the saturated spontaneous polarization p@xmath156 at @xmath157 in the r phase , and just above the o r and t o transitions in the o and t phases respectively . these are calculated from the average local mode vector @xmath18 and the local mode @xmath32 . we find almost no finite - size effect for this quantity , and the agreement with experiment is very good for the o and t phases . the disagreement for the r phase may result in part from twinning effects in the experimental sample.@xcite from the jumps in structural parameters and the observed hysteresis in heating and cooling , we conclude that the phase transitions are first order . an accurate determination of the latent heats would require considerable effort;@xcite here , we only try to provide estimates sufficiently accurate for meaningful comparison with experiment . we approach each transition from both high - temperature and low - temperature sides until the point is reached where both phases appear equally stable . ( that is , the typical time for the system to fluctuate into the opposite phase is roughly independent of which phase the simulation is started in . ) the difference of the average total energy is then the latent heat.@xcite this approach is practical as long as some hysteresis is present . the calculated latent heats ( table [ table3 ] ) show non - negligible finite - size dependence . taking this into account , we find that the latent heats for all three transitions are significantly non - zero and in rough agreement with the rather scattered experimental data . for the t o and o r transitions , the first - order character of the transition is predicted by landau theory , since in these two cases the symmetry group of the low temperature structure is not a subgroup of that of the high temperature structure . for the c t transition , the first - order character is not a consequence of symmetry , but rather of the values of the effective hamiltonian parameters . although it has the largest latent heat of the three transitions , the c t transition also exhibits large finite - size effects in the latent heat and in the smearing of order parameter components and strain discontinuities in the simulation ( figs . [ u - t],[e - t ] ) . this suggests the presence of long - wavelength fluctuations characteristic of second - order phase transitions and critical phenomena , and the classification of the c t transition as a weak first - order transition . using the microscopic information available in the simulations , we are able to investigate the extent to which the cubic - tetragonal transition can be characterized as order - disorder or displacive . these possibilities can be distinguished by inspecting the distribution of the real - space local - mode vector @xmath16 in the cubic phase just above the transition . a displacive ( microscopically nonpolar ) or order - disorder ( microscopically polar ) transition should be characterized by a single - peaked or double - peaked structure , respectively . the distribution of @xmath158 at @xmath159k is shown in fig . it exhibits a rather weak tendency to a double - peaked structure , indicating a transition which has some degree of order - disorder character . we also see indications of this in the @xmath110@xmath160 relation in fig . [ u - t ] . even in the cubic phase , the average of the local - mode magnitude @xmath110 is significantly non - zero and close to that of the rhombohedral phase , while the magnitudes of the average local mode components change dramatically during the phase transitions . in reciprocal space , a system close to a displacive transition should show large and strongly temperature - dependent fluctuations of certain modes associated with a small portion of the brillouin zone ( bz ) ( for a ferroelectric transition , near @xmath40 ) . for an extreme order - disorder transition , on the other hand , the fluctuations are expected to be distributed over the whole bz . for batio@xmath0 , we calculated the average fourier modulus @xmath161 for eigenmodes at several high - symmetry k - points ( along @xmath40x , @xmath40m and @xmath40r ) for a range of temperatures above the c t transition . these eigenmodes are identified by their symmetry properties as one longitudinal optical ( lo ) branch and two transverse optical ( to ) branches at each point . for a purely harmonic system , @xmath162 can be shown to be a temperature - independent constant proportional to the square of the eigenfrequency @xmath163 of the corresponding eigenmode . a strong decrease of @xmath164 as @xmath165 from high temperature can be interpreted as mode softening due to anharmonicity . the results at the k - point ( @xmath166 , @xmath166 , @xmath114 ) illustrate the main features of this analysis . from symmetry , three nondegenerate eigenmodes lo , to1 , to2 are identified . the polarization of lo , to1 , and to2 are in the direction of @xmath167 , @xmath168 , and @xmath169 , respectively . for each eigenmode , the temperature dependence of the calculated @xmath170 is shown in fig . the almost linear behavior of @xmath171 vs. @xmath160 ( the curie - weiss form ) is observed for the other k - points as well . both the lo and to1 branches are almost temperature independent . the to2 branch is strongly temperature dependent and is thus a `` soft '' mode . according to the soft - mode theory of structural phase transitions , @xmath172 is the lowest temperature at which all @xmath173 . linear extrapolation indicates that the to2 mode frequency goes to zero at @xmath174k , which is a lower bound for @xmath172 , consistent with the value obtained in monte carlo simulations . a similar calculation of @xmath170 for the to modes at @xmath40=(0,0,0 ) extrapolates to zero at the higher temperature @xmath175k , in excellent agreement with the monte carlo value of @xmath176k . within this formalism , the microscopic character of the paraelectric phase is determined by the extent of the soft mode in the bz . we define a quantity @xmath177 to indicate the hardness of the modes . in fig . [ u2g2 ] , @xmath178 is shown for the various k - points along some special directions in the bz . if @xmath179 , the corresponding eigenfrequency extrapolates to zero at some positive temperature , and the mode is regarded as soft . if @xmath163 is independent of temperature , @xmath180 , corresponding to the hardest mode . for all the k - points considered , all the lo modes are almost temperature independent [ @xmath180 ] and are not included in the figure . along the @xmath40x direction , the doubly - degenerate to modes are soft all the way to the zone boundary . in contrast , along the @xmath40r direction , both to modes become hard immediately after leaving the @xmath40 point . along the @xmath40m direction , the to1 mode becomes hard quickly , while the to2 branch remains soft at least halfway to the zone boundary . this behavior , especially along @xmath40x , does not conform completely to the displacive limit . this supports the interpretation of the c t transition as intermediate between displacive and order - disorder , with a slightly stronger order - disorder character . thus , from the example of batio@xmath0 , it seems that a positive on - site quadratic coefficient does not automatically imply a displacive character for the transition . rather , the relevant criterion is the extent to which the unstable modes extend throughout the bz . our theoretical approach allows us to investigate the roles played by different types of interaction in the phase transition . first , we study the effect of strain . recall that the strain degrees of freedom were separated into local and homogeneous parts , representing finite- and infinite - wavelength acoustic modes , respectively . both parts were included in the simulations . if we eliminate the local strain ( while still allowing homogeneous strain ) , we find almost no change in the transition temperatures . this indicates that the effect of the short - wavelength acoustic modes may not be important for the ferroelectric phase transition . if the homogeneous strain is frozen at zero , however , we find a direct cubic rhombohedral phase transition , instead of the correct series of three transitions . this demonstrates the important role of homogeneous strain . second , we studied the significance of the long - range coulomb interaction in the simulation . to do this , we changed the effective charge of the local mode ( and thus the dipole - dipole interaction strength ) , while modifying the second - order self - energy parameter @xmath27 , so that the frequencies of the zone - center and zone - boundary modes remain in agreement with the lda values . we found only a slight change ( 10% ) of the transition temperatures when the dipole - dipole interaction strength was doubled . however , elimination of the dipole - dipole interaction altogether changed the results dramatically ; the ground state becomes a complex antiferroelectric structure similar to the room - temperature structure of pbzro@xmath0 . this result shows that it is essential to include the long - range interaction , although small inaccuracies in the calculated values of the effective charges or dielectric constants may not be very critical . third , we investigated the sensitivity of our results to variations of the short - range interaction parameters . we find the accuracy of the first - neighbor interaction parameters ( @xmath58,@xmath59 ) is very important , and a mere 10% deviation can change the calculated transition temperatures dramatically , and can sometimes even change the ground state structure . second nearest - neighbor interactions are less important , and for the third - neighbor interactions , even a 100% change does not seem to have a strong effect on the values of t@xmath182 . this result is to be expected , and partly justifies our choice of including only up to third neighbors for the short range interactions . we have also tested the effect of our assumption @xmath183 . we find that any reasonable choice leads to a barely noticeable change in @xmath172 . in short , highly accurate lda calculations do appear to be a prerequisite for an accurate determination of the transition temperatures , but as long as certain features of the energy surface are correctly described , other approximations can be made without significantly affecting the results . we have developed a first - principles approach to the study of structural phase transitions and finite - temperature properties in perovskite compounds . we construct an effective hamiltonian based on taylor expansion of the energy surface around the cubic structure , including soft optical modes and strain components as the possible distortions . the expansion parameters are determined by first - principles density - functional calculations using vanderbilt s ultra - soft pseudopotential . we have applied this scheme to batio@xmath0 and calculated the pressure - temperature phase diagram . we have obtained the sequence of low - temperature phases , the transition temperatures , and the spontaneous polarizations , and found them to be in good agreement with experiment . we find that long - wavelength acoustic modes and long - range dipolar interactions both play an important role in the phase transition , while short - wavelength acoustic modes are not as significant . accurate lda calculations are required for accurate determination of the transition temperatures . the c t phase transition is not found to be well described as a simple displacive transition ; on the contrary , if anything it has more order - disorder character . with slight modifications , our approach should be applicable to other perovskite compounds , as long as their structure is close to cubic and there are some low energy distortions responsible for the phase transitions . it can be easily applied to ferroelectric materials like pbtio@xmath0 ( under study by another group@xcite ) and knbo@xmath0 . it can also be applied to antiferroelectric materials like pbzro@xmath0 . the application to antiferrodistortive materials like srtio@xmath0 is slightly more involved , though also successful.@xcite we would like to thank r.d . king - smith , u. v. waghmare , r. resta , z. cai , and a.m. ferrenberg for useful discussions . this work was supported by the office of naval research under contract number n00014 - 91-j-1184 and n00014 - 91-j-1247 . k. binder and d. w. heermann , _ monte carlo simulation in statistical physics _ ( springer - verlag , berlin , 1988 ) ; k. binder , ed . _ the monte carlo method in condensed matter physics _ , ( springer - verlag , berlin , 1992 ) ; k. binder ed . _ application of the monte carlo method in statistical physics _ , ( springer - verlag , berlin , 1984 ) . under a _ physical _ applied pressure , the latent heat would be the change of enthalpy , not energy . if this were applied to our case of a _ fictitious _ pressure , the calculated latent heats would be approximately twice as large .

cellular structures@xcite appear in a wide range of natural phenomena , and have puzzled and fascinated scientists for decades@xcite . they can be generally described as packings of space - filling cells of roughly polygonal shape , separated by thin interfaces to which a surface energy is associated . they arise as a consequence of competition between domains under the constraint of space - filling , and there is often some mechanism , such as the migration of a conserved quantity across the interface , which makes them evolve in time , i.e. coarsen . + foams such as those obtained by shaking soapy water are the examples of cellular structures closest to our everyday experience . two - dimensional foams may be obtained by confining a soap froth between two closely spaced glass plates . in spite of their apparent simplicity , foams display much of the phenomenology appearing in coarsening cellular structures . foams have been the subject of interest since their relevance in the problem of grain growth was pointed out by smith@xcite . despite the attention they have received , the understanding of their dynamical properties proved to be a tricky problem . even some of the most basic issues , such as their asymptotic scaling properties , has been a matter of debate until recently ( see references above ) . numerically exact simulation procedures@xcite , as well as analytical@xcite and numerical@xcite approximations , have been useful in understanding the dynamics of ideal foams , but the system sizes accessible to available simulation procedures are strongly limited ( for a complete account see refs.@xcite ) . + a satisfactory theoretical framework for the description of foams has not yet been achieved . the main result of this work consists in establishing a _ rigorous connection _ between foams and partitions(vp ) . this connection provides a set of fundamental degrees of freedom for the foam ( the source s locations in space ) and therefore constitutes a step towards the above mentioned theoretical understanding . on the other hand , an efficient method for the numerical simulation of ideal foams would also be highly desirable , and we propose that such a method could be obtained by exploiting the equivalence between foams and vp reported in this work . more precisely , we establish the following _ correspondence _ between equilibrated two - dimensional foams ( ef ) and a generalization of vps , the sectional multiplicative partition ( smvp ) : + _ given an arbitrary ef , it is always possible to find sources in space such that a smvp with respect to them exactly gives this foam . _ + in section [ sec : tessellations ] the vp and its generalizations are reviewed . the simplest generalizations of the vp concept are the sectional partition and the multiplicative partition . these correspond to adding a constant to the ( square ) distance and to multiplying it by a constant respectively . at the end of this section a combination of both is introduced , the sectional multiplicative partition . we will use this partition in order to describe foams . + the demonstration of the above mentioned correspondence between foams and smvps is divided in two parts ( sections [ sec : smvprecognition ] and [ sec : equilibrium ] ) for clarity . in section [ sec : smvprecognition ] , the _ recognition problem _ for smvps is solved . the recognition problem consists in giving the sufficient geometrical conditions that a circular partition has to satisfy in order to be a smvp . we will see that if a circular partition has all of its vertices _ aligned _ and admits an _ oriented reciprocal figure _ , then it is a smvp . in other words , it is always possible to find sources in three - dimensional space that give this circular partition as a smvp . in this section we also describe the procedure to _ find _ the sources when these conditions are met . some material that is needed in this section is described in the appendix . + section [ sec : equilibrium ] deals with the equilibrium conditions for foams and its geometrical consequences . we start by writing the force and pressure equilibrium conditions in compact form in section [ sec : eq_equations ] . in sections [ sec : eq_align ] and [ sec : eq_reciprocal ] it is shown that an equilibrated foam has all of its vertices _ aligned _ and admits an _ oriented reciprocal figure _ respectively . therefore all equilibrated foams satisfy the conditions required in section [ sec : smvprecognition ] , and can be described as smvps . + finally in section [ sec : conclusions ] the implications of this result are discussed , and some perspectives for future work are advanced . given a set of @xmath2 point - like _ sources _ @xmath3 in @xmath4-space , the _ partition _ ( vp ) or _ tessellation _ of space with respect to @xmath3 is a classification of space into _ cells _ @xmath5 defined such that @xmath6 if @xmath7 is closer to @xmath8 than to any other source @xcite . @xmath9 where @xmath10 denotes the distance between @xmath7 and source @xmath8 . this construction is also known under the names : wigner - seitz cells , dirichlet tessellations , thiessen polygons . in two dimensions , the dual lattice of a partition is called delaunay triangulation . cell @xmath5 can be seen as the `` region of influence '' of @xmath8 , were the sources competing for some spatially distributed resource . it is common to use these constructions , and their generalizations , as _ approximate _ models for cellular structures occurring in nature@xcite . + two neighboring cells @xmath5 and @xmath11 are delimited by an _ interface _ @xmath12 of points @xmath7 equidistant from @xmath8 and @xmath13 . @xmath14 these interfaces are @xmath15-dimensional hyperplanes normal to @xmath16 . in two dimensions @xmath5 are convex polygons and the interfaces @xmath12 are straight lines ( fig . [ fig : voronoi_partition ] ) . three interfaces @xmath12 , @xmath17 and @xmath18 meet at a _ @xmath19 , which is equidistant from @xmath8 , @xmath13 and @xmath20 . this means that @xmath19 is the center of a circle through @xmath8 , @xmath13 and @xmath20 . vertices of higher multiplicity are possible for particular locations of the sources . for example a fourfold vertex would exist if four sources lay on the same circle . we will ignore the existence of these particular configurations , that is we will assume _ generic _ source locations . under this assumption , all vertices are of multiplicity three . + the concept defining partitions is _ equidistance_. a simple way to generalize them is changing the way in which distances to the sources are measured . one lets each source measure the distance to points @xmath7 according to its own `` rule '' . the interface @xmath12 is then the set of points @xmath7 for which two neighbors `` claim '' the same distance . two simple ways to do this are : adding an arbitrary constant to the square of the distance ( sectional partition @xcite ) , and , multiplying the distance by a constant ( multiplicative partition @xcite ) . a _ sectional partition _ ( svp ) is defined @xcite as a @xmath21-dimensional cut of a higher dimensional partition . the sources @xmath22 of the sectional partition are defined as the projections of the original sources @xmath3 onto this lower dimensional @xmath21-space . cell @xmath5 associated to source @xmath0 is the intersection of cell @xmath23 ( associated to @xmath8 ) with this @xmath21-dimensional hyperplane . for example , take sources @xmath3 in three dimensions ( fig . [ fig : svpview ] ) and construct a vp with three - dimensional cells @xmath24 and plane interfaces @xmath25 . now project the sources @xmath26 onto the @xmath27 plane ( which we call the @xmath28 plane ) to obtain the projected sources @xmath29 , which have associated `` heights '' @xmath30 . assign to each source @xmath0 the intersection @xmath5 of @xmath23 with @xmath28 . this defines the a sectional partition of @xmath28 with respect to sources @xmath22 with heights @xmath30 . cells @xmath5 are thus defined as @xmath31 where @xmath10 is the distance , on the @xmath28 plane , between @xmath0 and point @xmath7 . + in the same way , interfaces @xmath12 are defined as @xmath32 properties of these partitions in two dimensions are : * interfaces are straight lines and ( generically ) meet at triple vertices . ( in dimension @xmath33 , vertices have multiplicity @xmath34 ) . * interface @xmath12 is normal to @xmath1 , but not in general equidistant from @xmath0 and @xmath35 . * the partition is unchanged if all @xmath30 are changed according to : @xmath36 for @xmath37 an arbitrary real . an important difference with the partition , is that in this case there may be sources to which no cell is associated . this happens if @xmath23 is not cut by @xmath28 . it is easy to see that svps are _ equivalent _ to laguerre partitions @xcite . + appendix [ sec : rprecognition ] discusses the _ recognition problem _ for this kind of partitions , that is giving the sufficient conditions for an arbitrary rectilinear partition to be a svp . we will use similar concepts in section [ sec : smvprecognition ] in order to solve the recognition problem for sectional multiplicative partitions . + as an example of the application of sectional partitions , consider the case of sources @xmath8 whose location in space changes in time , giving rise to a time - dependent partition of @xmath28 . if for example source @xmath38 moves away from @xmath28 , the two - dimensional cell @xmath39 will shrink and finally disappear when @xmath40 no longer cuts @xmath28 . therefore the number of cells can vary without changing the number of sources . thus sectional partitions can be used as dynamical models for crystal growth @xcite , and other processes in which some cells disappear or are created . the multiplicative partition ( mvp ) @xcite is defined by assigning to each source @xmath8 a positive _ intensity _ @xmath41 , and defining the _ multiplicative distance _ @xmath42 . the cell @xmath5 associated to source @xmath8 is then the set of points @xmath7 closer to @xmath8 ( in terms of this multiplicative distance ) than to any other source @xmath43 interfaces @xmath12 are hyperspherical surfaces ( circle arcs in two dimensions ) satisfying @xmath44 in two dimensions , the circular interface of a mvp with two sources is one of the apollonius circles ( see for example @xcite ) of those two points . in arbitrary dimensions let @xmath45 be the distance between two sources @xmath8 and @xmath13 , and @xmath46 . without loss of generality we take @xmath47 . this means that @xmath13 has the smallest intensity and therefore @xmath11 will be the interior of a hypersphere , while @xmath5 will be its exterior . the following properties are satisfied in any dimension . * @xmath8 is contained in @xmath5 . * @xmath12 has radius , and its center @xmath48 is on the straight line going through @xmath8 and @xmath13 . * @xmath48 is located at a distance @xmath49 from @xmath8 , that is , it never lays between @xmath8 and @xmath13 . three sources give rise to a tessellation like the one shown in fig . [ fig : mvp_vertex_2d ] . the exterior of the two `` bubbles '' is the infinite cell associated to the source with the largest intensity , @xmath38 in this example . for some choices of @xmath50 the interfaces will not intersect . in this case the tessellation is simply a pair of circles , each containing one of the sources with smaller intensities , while the exterior of these two circles is the cell of the source with larger intensity . mvps can be interpreted again as `` areas of influence '' of sources @xmath8 , but now each source has a different strength . + consider a mvp of two dimensional space with respect to a set of sources @xmath3 . if three interfaces @xmath12 , @xmath17 and @xmath18 meet at a vertex ( again , vertices of higher multiplicity are only possible for particular configurations , which we ignore here ) , then centers @xmath48 , @xmath51 and @xmath52 lay on a line . the reason why this is so is simple . if two interfaces @xmath12 and @xmath17 have a common point @xmath19 , then their _ continuations _ must meet again at another point @xmath53 , the _ conjugate vertex_. but then the third interface @xmath54 must also go through this point since @xmath55 and @xmath56 implies @xmath57 . therefore two points exist ( @xmath58 and @xmath59 ) at which all three circles intersect . then the centers of these circles must be on a line ( fig . [ fig : mvp_vertex_2d ] ) . + let us define a _ circular partition_(cp ) of two - dimensional space to be a classification of space in cells delimited by arbitrary circle arcs that meet at triple points called vertices . no three centers of these arbitrarily defined interfaces will in general be on a line . we will say that a vertex of a cp is _ aligned _ if the centers of the three interfaces defining it are on a line . as we saw , all vertices of a mvp are aligned in 2d . therefore for each triple vertex of a mvp in 2d , the three sources and the three centers form the configuration @xcite @xmath60 of projective geometry . since @xmath48 can not be between @xmath0 and @xmath35 , the three centers can only be on one of the two external segments of the configuration . + in a three - dimensional mvp , four cells meet at each vertex @xmath61 , giving rise to six spherical interfaces @xmath62 . the centers of these interfaces are aligned in triplets so that the six centers also form the configuration @xmath60 , this time in three - dimensional space . this has the implication that the six centers are necessarily on the same plane . + the mvp was first introduced by boots @xcite to describe areas of influence in geography . ash and bolker @xcite also studied the recognition problem , that is , under which conditions a cp is a mvp . the visual similarity between this kind of partitions and two - dimensional foams is striking , and suggests the idea to find a connection between them . obviously the centers of the films of a foam must be aligned for each triple vertex , if the foam is to be described as a mvp , since mvps are aligned . one finds @xcite that this alignment condition is indeed satisfied by all vertices of arbitrary equilibrated foams . despite this ( which is a necessary but not sufficient condition for a foam to be a mvp ) , two - dimensional mvps can not describe all possible equilibrated foams in two dimensions . this work shows the reason of this limitation : in order to describe arbitrary equilibrated two - dimensional foams we must introduce the sectional variant of a multiplicative partition . in other words , instead of confining the sources to the plane , we let them exist in a three - dimensional space and obtain the foam as _ a two - dimensional cut of a three - dimensional mvp_. we will restrict the description to the case of a plane cut of a three dimensional mvp . the generalization to a @xmath21-dimensional cut of an @xmath4-dimensional mvp is straightforward . + a plane cut , with a plane @xmath63 , of a three - dimensional mvp defines a sectional multiplicative partition ( smvp ) of @xmath28 ( see figs . [ fig : smvp_vertex ] and [ fig : smvp_upper ] ) . the sources @xmath0 of this smvp are the normal projections onto @xmath28 of the original sources @xmath8 , and may be seen to have as attributes both an intensity @xmath41 and a height @xmath30 . cells @xmath64 associated to sources @xmath0 are defined as @xmath65 in the same way , interfaces @xmath12 are circle arcs satisfying @xmath66 an example of a smvp with three sources is shown in figs . [ fig : smvp_vertex ] and [ fig : smvp_upper ] . we see that there are two vertices on @xmath28 at which the three interfaces meet . in a general case ( for example in a partition with respect to many sources like in fig . [ fig : mvp2d - many ] ) , if vertex @xmath19 exists , then the continuations of interfaces @xmath12 , @xmath17 and @xmath18 meet at a _ conjugate _ vertex @xmath53 also . this implies the alignment of centers @xmath48 , @xmath51 and @xmath52 , which could also be deduced from the alignment of @xmath67 , @xmath68 and @xmath69 in three - dimensions ( fig . [ fig : smvp_vertex ] ) . + we notice that the smvp is equivalent to a multiplicative laguerre partition , since a svp is equivalent to a laguerre partition @xcite . as is usual in sectional partitions , in the smvp there may be sources @xmath0 with no associated cell , those for which the corresponding three - dimensional `` bubble '' @xmath23 does not cut @xmath28 . + it was seen in section [ sec : mvp ] that the spherical interface @xmath25 in three dimensions is cut normally by the segment @xmath16 . as a projective consequence of this , the straight line containing @xmath0 and @xmath35 on @xmath28 is normal to the circular interface @xmath12 . in other words , @xmath0 , @xmath35 and @xmath48 are on the same line . furthermore and since @xmath67 never lays between @xmath8 and @xmath13 in 3d , we notice that @xmath48 is always outside the segment @xmath1 . in this section the sufficient conditions for a cp to be a smvp are given . we first introduce the notion of oriented reciprocal figure for cps , and then proceed to demonstrate that an aligned cp admitting such a reciprocal figure is a smvp . we will now generalize the concept of reciprocal figure as appropriate for circular partitions . we say that a graph @xmath70 made of points @xmath0 connected by edges @xmath71 forms a reciprocal figure for a cp @xmath72 if , * points @xmath0 are in correspondence with the cells @xmath5 of @xmath73 . * edges @xmath71 are in correspondence with the interfaces @xmath12 of @xmath73 . * for each @xmath12 in @xmath73 , points @xmath0 , @xmath35 and @xmath48 lay on the same line . consider two regions @xmath5 and @xmath11 separated by a circular interface @xmath12 . a reciprocal figure @xmath70 will be said to respect _ orientation _ if for each @xmath12 in @xmath73 : * @xmath48 is not between @xmath0 and @xmath35 . * starting from @xmath48 and traveling along @xmath74 to infinity , points @xmath0 and @xmath35 are found in the same order as regions @xmath5 and @xmath11 are . we saw already that all smvps are aligned ( all vertices satisfy the alignment condition introduced in sec . [ sec : mvp ] ) . on the other hand , it is clear that the sources @xmath0 of a smvp form a reciprocal figure that satisfies orientation . therefore @xmath75 figure [ fig : convexity ] shows four possible partitions that share the same centers @xmath76 , @xmath77 and @xmath78 . all four admit reciprocal figures . but only in cases a ) and c ) a reciprocal figure that satisfies orientation can be constructed . therefore neither b ) nor c ) can be a smvp . the reason is that the orientation condition can not be satisfied if any of the vertices is not _ convex_. a vertex is _ convex _ if all the internal angles formed by the interfaces are less than @xmath79 . convexity of all vertices is clearly a necessary condition for a circular partition to be a smvp . + we are now ready to give the _ sufficient _ conditions for a circular partition to be a smvp . we will show in section [ sec : find_sources ] that if a circular partition @xmath73 is aligned and admits a reciprocal figure that satisfies orientation , then it is a smvp . @xmath80 given a circular partition @xmath73 of a region of two dimensional space , all whose ( triple ) vertices are aligned , we show here that , if @xmath73 admits a reciprocal figure that satisfies orientation , then @xmath73 is a smvp with respect to sources @xmath8 located somewhere above points @xmath0 . our demonstration is constructive , that is we explicitly show how the sources and intensities are determined . we will use for this purpose the inversion transformation ( appendix [ sec : inversion ] ) to _ straighten _ each vertex in turn , that is , transform a circular vertex into a rectilinear one . this rectilinear vertex is one of a of a svp . sources are located in this _ straight _ representation ( appendix [ sec : rprecognition ] ) , and then transformed back to the original system . the intensities @xmath41 of the corresponding smvp are obtained in this back - transformation , since the smvp is invariant under inversion . + let @xmath58 be an aligned vertex on which three interfaces @xmath81 meet , and @xmath82 the three points of the reciprocal figure associated to the three bubbles sharing the vertex , as in fig . [ fig : reciproca1 ] . the conjugate vertex @xmath59 is obtained as the intersection point of the _ continuations _ of the interfaces , which happens at a unique point because of alignment . let furthermore @xmath83 be the straight line through @xmath59 and @xmath0 . we start by showing that [ th : recircles ] a monoparametric family of triplets of circles @xmath84 exists , which has the following properties : * @xmath85 , @xmath86 and @xmath87 intercept each other on @xmath59 . * @xmath88 . * the intersection point @xmath89 between @xmath90 and @xmath91 lays on @xmath83 . the non - trivial content of the theorem is the fact that the three intersection points @xmath89 of these normal circles @xmath90 lay on the lines @xmath83 . for example take an arbitrary point @xmath92 on @xmath93 and draw a circle @xmath85 through @xmath92 , @xmath59 and normal to @xmath94 . let @xmath95 be the intersection of @xmath85 with @xmath96 . draw now a circle @xmath86 trough @xmath95 , @xmath59 and normal to @xmath97 . let @xmath98 be its intersection with @xmath99 . then the circle @xmath87 through @xmath59 , @xmath92 and @xmath98 is normal to @xmath100 . + * proof . * the demonstration is done by first performing an inversion with center @xmath59 and arbitrary radius , whereupon the interfaces @xmath12 are transformed into straight lines @xmath101 meeting at @xmath102 ( fig . [ fig : reciproca2 ] ) . this transformed system will be referred to as the `` straight '' representation of that vertex . primed names are used in the straight representation , with the exception of @xmath59 , whose original location is kept in fig . [ fig : reciproca2 ] ( @xmath59 itself is mapped to @xmath103 by the inversion ) . lines originally not going through @xmath59 are transformed into circles through @xmath59 , as is the case of @xmath104 , which form the reciprocal figure in the original system . circles @xmath90 are transformed into straight lines @xmath105 going through @xmath106 and @xmath107 , respectively on @xmath83 and @xmath108 , which are invariant . + the inversion transformation preserves angles , therefore circle @xmath109 is normal to the now straight interface @xmath101 . this means that its center @xmath110 must lay on @xmath101 . consider the figure formed by @xmath102 , @xmath111 , @xmath112 and @xmath113 , and the six lines joining them . maxwell @xcite showed that a figure made of four points joined by six lines _ always _ has a reciprocal figure ( to see this just consider the centers of the four circles going through three of these points ) . we will now identify it on fig . [ fig : reciproca2 ] as the figure formed by @xmath59 , @xmath114 , @xmath115 and @xmath116 . first we notice that circles @xmath109 and @xmath117 intercept each other at two points @xmath59 and @xmath118 , both on @xmath108 . therefore lines @xmath119 are normal to @xmath108 , and we have identified the first three lines of the reciprocal figure . the other three lines @xmath120 of the reciprocal figure have to be normal to the segments @xmath121 , so they are normal to the interfaces @xmath101 . since the inversion preserves angles , this means that @xmath88 in the original system ( fig . [ fig : reciproca1 ] ) . @xmath122 + 0.2 cm we see that the existence of triplets of circles @xmath123 with the above mentioned properties is a consequence of the existence , in the straight representation , of a figure @xmath124 , which is the reciprocal of @xmath125 . in other words , theorem [ th : recircles ] means that in the straight representation , triangle @xmath126 forms a reciprocal figure for the rectilinear vertex . clearly this reciprocal figure satisfies orientation if the original figure satisfies orientation in the circular system . therefore as discussed in appendix [ sec : rprecognition ] , it is always possible to find three sources @xmath127 , located at heights @xmath128 above points @xmath106 , such that this rectilinear vertex is a svp with respect to them . once these sources @xmath129 are known in the straight representation , a second inversion around the same point @xmath59 takes us back to the original system and provides the original locations @xmath8 of the sources . + if @xmath130 is the first vertex to be considered in the system , we have one degree of freedom in the determination of the circles @xmath90 , or equivalently of points @xmath89 . once in the straight representation , there is one more degree of freedom : the determination of _ one _ of the heights @xmath128 . we will now show that , if this @xmath128 is chosen appropriately , the three back - transformed sources @xmath8 are located above their respective points @xmath0 . + consider a plane @xmath131 that is normal to @xmath28 and contains @xmath83 , as shown in fig . [ fig : above ] . draw a circle @xmath132 through @xmath89 , @xmath59 and normal to @xmath28 . let @xmath133 be the intersection of this circle with the normal @xmath134 to @xmath28 through @xmath0 . now under an inversion with center @xmath59 , point @xmath135 is above @xmath106 , since circle @xmath132 is transformed in a straight line @xmath136 , normal to @xmath28 . straight line @xmath134 normal to @xmath28 through @xmath0 is now a circle @xmath137 through @xmath59 , @xmath135 and normal to @xmath28 . its intersection with @xmath28 determines the image @xmath138 . now we now that source @xmath129 is located somewhere on @xmath136 . the sources can be displaced vertically ( simultaneously ) according according to what we saw in appendix [ sec : rprecognition ] , but not independently . fixing the height @xmath128 of one of them determines the other two uniquely . what we want to demonstrate is that if one of the sources @xmath129 coincides with its point @xmath135 , then the other two also do . in order to do this , take for example the sphere @xmath139 through @xmath59 , @xmath140 , @xmath141 and normal to @xmath28 . in the straight representation , interface @xmath142 of the partition with respect to sources @xmath140 and @xmath141 is a plane , equidistant from @xmath140 and @xmath141 , and normal to @xmath143 . this means that @xmath142 is normal to @xmath139 . as a consequence , the intersection @xmath144(a circle ) of @xmath139 with @xmath28 is normal to the intersection @xmath145(a straight line ) of @xmath146 with @xmath28 . now assume that @xmath147 . this means that @xmath148 since @xmath149 is in this case contained in @xmath139 . but then the circle @xmath144 is the circle @xmath150 ( fig . [ fig : reciproca2 ] ) that goes through @xmath151 , @xmath152 and @xmath59 and is normal to @xmath145 . this implies that @xmath152 is also in @xmath139 , which in turn implies @xmath153 . the same reasoning can be repeated for the pair of sources @xmath154 and @xmath155 . + we have thus shown that all sources @xmath8 are , in the back - transformed , original system , located above the points @xmath0 of the reciprocal figure . in other words , that the given reciprocal figure is the projection of the sources @xmath8 on the @xmath28 plane . + now we have to determine the heights @xmath30 and intensities @xmath41 . the heights @xmath30 of the sources can be found found by back - transforming the sources @xmath129 from the straight system . but in practice this is not necessary . by looking at fig . [ fig : above ] we notice that triangle is rectangle . therefore @xmath156 which suffices to locate the sources @xmath8 with the sole knowledge of points @xmath0 and @xmath89 in the original system . knowing now the spatial locations of the sources @xmath8 in three dimensions , the intensities @xmath41 follow immediately since the vertex @xmath59 is equidistant @xcite from the three sources . this means @xmath157 here @xmath158 is an arbitrary positive constant . alternatively we can get the same result by using the transformation properties ( eq . [ eq : dist_transf ] ) of the @xmath41 s and the fact that the intensities are all equal in the straight system . + we have used the inversion transformation to demonstrate that the given vertex is a smvp with respect to sources @xmath8 located above @xmath0 , and showed how the sources @xmath8 can be located without the need to explicitly perform an inversion for each vertex . all we need is to know the location of points @xmath89 , which we are free to choose for the first vertex under consideration . this first choice determines all subsequent @xmath89 points , and therefore all sources . it is clear from fig . [ fig : above ] that @xmath159 has to be satisfied . therefore one has to choose the triplet of circles @xmath160 such that this condition is verified for the three points @xmath89 . if one of the @xmath89 coincides with @xmath0 , this means that the corresponding height @xmath30 is zero , i.e. the source @xmath8 lays on the @xmath28 plane . a point @xmath89 closer to @xmath59 than the corresponding @xmath0 is not acceptable . + the general procedure to determine the sources can now be described . assume we are given a circular tessellation which satisfies the required conditions of alignment and convexity , and that we are able to find , or are given , a reciprocal figure @xmath0 for it . we start from one of the cells of the partition , as in fig . [ fig : bubblesources ] . take an arbitrary vertex to start with , for example vertex @xmath161 , and determine tentative locations for point @xmath162 on @xmath163 . this fixes @xmath92 and @xmath95 as discussed in section [ sec : find_sources ] . after fixing @xmath162 the rest of the construction is uniquely determined . the locations @xmath164 determine @xmath165 and @xmath166 , and the corresponding intensities @xmath167 and @xmath168 , through equations ( [ eq : zi ] ) and ( [ eq : ai ] ) . now proceed to the neighboring vertex @xmath169 . two of the sources , @xmath170 and @xmath171 are already known , only the height @xmath172 of @xmath173 has to be found . first one has to find the _ locations of @xmath162 and @xmath95 when defined from vertex @xmath174 . in order to do this , lines @xmath175 from vertex @xmath174 are drawn , and @xmath162 and @xmath95 are located using the fact that the heights @xmath176 and @xmath166 are already known . equation ( [ eq : zi ] ) implies @xmath177 once the new @xmath162 and @xmath95 are known , the next step is to draw circles @xmath178 and @xmath86 through them and normal to the respective interfaces . their intersection gives the location of @xmath98 ( this intersection always occurs on @xmath179 , as theorem [ th : recircles ] shows ) . this determines @xmath172 and @xmath180 using equations ( [ eq : zi ] ) and ( [ eq : ai ] ) . this procedure of triangulation is repeated until all sources are located . eventually it may happen that a point @xmath89 is found to be closer to the conjugate vertex @xmath59 than the corresponding @xmath0 . this is not acceptable and means that the _ tentative _ starting position of @xmath162 has to be changed . it must be shifted away from @xmath181 and the whole procedure repeated . it is easy to see that taking a large enough value of @xmath162 always solves this problem . + one could ask whether this construction can be closed self - consistently . for example after determining @xmath182 from @xmath170 and @xmath183 in our example of fig . [ fig : bubblesources ] , one can go on with the procedure as if @xmath38 were unknown . would the position of @xmath38 determined by @xmath170 and @xmath182 be the same one as found initially ? alternatively : if we used @xmath170 and @xmath38 to determine @xmath182 instead of going around the bubble in the opposite sense , would its position be the same as found after going around the bubble ? the answer is yes because of unicity . as discussed in appendix [ sec : rprecognition ] , the height of one of the sources of a vertex determines the other two uniquely . this means that @xmath184 are all uniquely determined by @xmath170 . in this section we show that a two dimensional foam in equilibrium satisfies all conditions required for it to be a smvp . in order to do this let us first write down the equilibrium conditions for an arbitrary vertex of the foam , in compact form . we will consider the case of foams with arbitrary surface tensions , and also allow forces to act on the foam s vertices . let @xmath185 be the location of a vertex @xmath58 at which three interfaces @xmath186 , @xmath187 and @xmath188 meet , as shown in fig . [ fig : eq_sign_convention ] . each interface @xmath189 is a circle arc with radius @xmath190 and center @xmath191 . it produces on @xmath58 , due to its surface tension @xmath192 , a force @xmath193 of modulus @xmath194 in the direction of the tangent to the film at @xmath185 . these forces @xmath195 can therefore be written as @xmath196 where @xmath197 and @xmath198 . the asterisk stands here for a counterclockwise rotation in @xmath199 , so that @xmath200 . + let us more generally assume that an external force @xmath201 acts on @xmath58 . equilibrium of all forces acting on the vertex implies @xmath202 there is still one more condition , which is related to pressure equilibrium around the vertex . the pressure drop across an interface can be written as @xmath203 we will adopt the convention for @xmath204 to be positive if the pressure decreases when crossing the interface in the counterclockwise sense of rotation around @xmath58 . this is of course related to a convention for the signs of the @xmath190 . in fig . [ fig : eq_sign_convention ] @xmath205 and @xmath206 are negative according to this convention . notice that pressure jumps and radii have different signs when considered from the two opposite vertices joined by a film . + the fact that the total accumulated pressure drop around a vertex has to be zero implies then @xmath207 this is the pressure equilibrium condition for the vertex . equations ( [ forceeq ] ) and ( [ pressureeq ] ) are satisfied if the vertex is equilibrated , and are together equivalent to @xmath208 where @xmath209 is an _ arbitrary _ point . this is what we will call the equilibrium condition for the vertex , and encloses both force and pressure equilibrium . the alignment of the centers of a two - dimensional foam was already known long time ago by plateau @xcite for equal surface tensions , and in the case of small self - standing clusters of 2 and 3 bubbles . for the case of a cluster of two bubbles and zero load it is a trivial consequence of symmetry @xcite . the demonstration for the case of clusters of three bubbles is referred to by boys as being``rather long and difficult '' @xcite . + it is not difficult to see that the alignment of the centers is in no way a property of clusters , and also not restricted to vertices with equal surface tensions and zero loads , but a general consequence of equilibrium . we will find that under very general conditions such as arbitrary surface tensions and external loads applied on the vertices , if a vertex is equilibrated then the centers of the three arcs converging to it lay on a line . + consider equation ( [ eq : aligned ] ) , and assume for a moment that the external load is zero . then @xmath210 and @xmath211 lay on the same line , as can be seen by taking @xmath212 for example . thus all vertices in equilibrium are aligned if no external force is applied on it . this alignment property is even true under non - zero load conditions , if the force is perpendicular to the line of centers . a load satisfying such condition will be called a _ normal load_. the alignment condition has the geometrical consequence that the interfaces @xmath189 , when continued , meet each other again at a unique point @xmath59 which we called the _ conjugate vertex _ @xcite of @xmath58 . this also means that the vertex @xmath58 could be physically realized as a self - standing cluster of two bubbles ( by continuing its interfaces ) , and is a kind of a `` separability '' condition for the static equilibrium conditions , in the sense that each vertex of a foam might as well be that of an isolated cluster of two bubbles . thus _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ a vertex in equilibrium under a normal load is aligned . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ or , equivalently _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ a vertex in equilibrium under a normal load has a conjugate vertex . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ next we would like to consider the existence of a reciprocal figure , since this is also a condition that has to be satisfied in order for a foam to be a smvp . this condition must be separately considered . the reader may easily build examples of bubbles with @xmath4 neighbors , all of whose @xmath4 vertices are aligned , but yet do not admit a reciprocal figure . the reason is that the attempted reciprocal figure will not in general `` close '' around that bubble , the same case as described in appendix [ sec : rprecognition ] . we will now show that if all vertices of a foam are equilibrated , then an oriented reciprocal figure exists for it . we start by considering a bubble and show that the reciprocal figure can be found for it . the figure for the whole foam can then be formed by patching together those of neighboring bubbles . consider a bubble with @xmath4 neighbors @xmath213 , as shown in figure [ fig : bubblea ] . at each vertex @xmath214 of this bubble , three films @xmath215 , @xmath216 and @xmath217 meet . interface @xmath215 separates the central bubble from its neighbor @xmath218 , while interface @xmath217 is the limit between neighbors @xmath218 and @xmath219 . we will assume vertices @xmath214 to be in equilibrium under arbitrary _ loads @xmath220 . these loads we can generally write as @xmath221 that is , we have decomposed the external load of vertex @xmath214 in the _ dependent _ basis formed by the three centers of the films meeting at the vertex . because these centers are aligned , and the load is normal to the line of centers , the following condition is always satisfied by the coefficients . @xmath222 the equilibrium condition ( [ eq : aligned ] ) now reads @xmath223 and holds for _ arbitrary _ @xmath224 s . the sign of @xmath225 is determined by the sign convention at vertex @xmath214 , therefore @xmath226 must appear with a minus sign in the equilibrium equation for vertex @xmath218 . + since we use a dependent basis ( equation . ( [ eq : loadgeneral ] ) ) the coefficients @xmath227 ( the `` representation '' of the load ) are not uniquely determined for a given load . there is instead a monoparametric family of coefficients @xmath227 , all satisfying both ( [ eq : loadgeneral ] ) and ( [ eq : nullsum ] ) , for each load @xmath228 . we will use these degrees of freedom to choose a representation @xmath229 that satisfies @xmath230 notice that this condition relates the coefficients of two consecutive loads . we now show that such a representation always exists . we start by making the degree of freedom in the representation of the loads explicit . for arbitrary @xmath231 , we add the null vector ( [ eq : geneq ] ) multiplied by @xmath231 to the load ( [ eq : loadgeneral ] ) and get @xmath232 where @xmath233 condition ( [ eq : ring ] ) then implies @xmath234 which has always a solution in the generic case . + using this representation of the loads we can rewrite the equilibrium condition ( [ eq : geneq ] ) as @xmath235 where @xmath236 this amounts to hiding the loads in a redefinition of the surface tensions and pressures : @xmath237 . as we have shown , this can always be done for aligned loads . we notice that the coefficients @xmath238 satisfy @xmath239 as can be verified by subtracting eq . ( [ eq : geneq1 ] ) with two different values of @xmath240 , and adding up the result for @xmath241 . + more generally the fact that all vertices of the bubble are in equilibrium has the consequence that @xmath242 for @xmath243 arbitrary . condition ( [ eq : vectornullaround ] ) can be generalized to any closed path in the foam . the sum is in that case over all films cut by the closed path . the smallest such path , enclosing a vertex , gives the vertex equilibrium condition ( [ eq : aligned ] ) . now that we have written the bubble equilibrium condition in the convenient form ( [ eq : geneq1 ] ) , we will write an algebraic condition for the existence of a reciprocal figure . + the reciprocal figure was defined to be a set of points @xmath244 , each associated to a bubble ( but not necessarily contained in it ) , such that the straight line passing through @xmath245 and @xmath246 is normal to the interface @xmath12 between bubbles @xmath247 and @xmath248 . this means that @xmath249 , @xmath245 and @xmath246 are on the same line , as in fig . [ fig : bubblec ] . we will require that @xmath250 and @xmath251 be arbitrary ( with the only condition that @xmath250 , @xmath251 are @xmath252 are on a line ) . this will allow us to patch together the reciprocal figures of neighboring bubbles to form that of the whole foam . this is equivalent to the translation and dilatation degrees of freedom existent in the definition of a reciprocal figure for a svp . for arbitrary @xmath250 , take @xmath251 anywhere on the line @xmath253 . all other points are now uniquely determined . @xmath254 is located in the intersection of @xmath255 and @xmath256 , next @xmath257 is found as the intersection of @xmath258 and @xmath259 , etc . when the figure is closed with the last point @xmath260 , there is an extra condition , since it has to be the intersection of _ three _ lines : @xmath261 , @xmath262 , @xmath263 . the construction of the reciprocal figure is thus an _ overdetermined _ problem , and would not in general be possible if the centers @xmath48 were arbitrarily located . let us now write the conditions for this reciprocal figure to close , in an algebraic form . the points @xmath0 of the reciprocal figure are determined by a set of coefficients @xmath264 satisfying the following conditions . @xmath265 substituting [ rfcond1 ] into [ rfcond2 ] we get @xmath266 we will now show that the equilibrium conditions ( [ eq : geneq1 ] ) ensure that ( [ eq : rfcond3 ] ) always has a solution , and therefore that a reciprocal figure exists . comparison with equation ( [ eq : geneq1 ] ) lets us conclude that a solution will exist for @xmath250 arbitrary , if we are able to find a set of coefficients @xmath264 that satisfy @xmath267 this is equivalent to @xmath268 it is not difficult to see that these equations are satisfied if the @xmath269 are related by @xmath270 starting from an _ arbitrary _ @xmath271 , this recursion relation gives us the following values of @xmath272 . once all are known , equation ( [ eq : rel1 ] ) provides the values of the @xmath273 , which in turn determine @xmath274 using ( [ rfcond1 ] ) . the condition that the figure can be closed is @xmath275 , and is equivalent to @xmath276 . this condition is satisfied because ( [ eq : nullaround ] ) holds , and therefore is a consequence of equilibrium . + if we were given an arbitrary circular partition , it would not in general be possible to find a reciprocal figure for it . the fact that this cp is an equilibrated foam imposes geometrical constraints on it ensuring , for example , that it admits a reciprocal figure . + we have thus shown that , for an arbitrary equilibrated bubble , it is always possible to find a reciprocal figure . we can arbitrarily fix @xmath250 since @xmath224 in ( [ eq : geneq1 ] ) can be arbitrarily chosen , and we can also choose the `` scale '' @xmath277 of the drawing at will , since the starting value @xmath271 , that fixes this scale , is arbitrary . therefore the reciprocal figures of neighboring bubbles can be patched together to form a reciprocal figure for the whole foam . each selection of a starting point @xmath250 and a `` scale '' @xmath277 determines the other points uniquely , therefore there are three degrees of freedom in the determination of the reciprocal figure . once @xmath250 and @xmath251 are chosen , all other points are found as intersections of two lines passing through the centers @xmath48 and one already existing point @xmath0 . now we have to show that it is always possible to find , among all possible reciprocal figures , at least one that satisfies orientation as defined in section [ sec : smvp ] . in the first place , if all the surface tensions are positive then all films will be under traction and therefore the vertices will be convex . if there are non - convex vertices in the foam ( which would happen if some of the films are compressed instead of stretched ) then we know that it is not possible to satisfy orientation ( fig . [ fig : convexity ] ) . positive surface tensions is thus a necessary condition @xcite for the foam to be a smvp , although their modulus can be arbitrary for each film . + the orientation condition could fail in the first place because a center @xmath48 lays in between two points @xmath0 and @xmath35 . it is always possible to avoid this by choosing @xmath251 close enough to @xmath250 . in this way all following points @xmath0 are confined within a ( arbitrarily chosen ) small region of space where no center is located . this ensures that no center @xmath48 lays between @xmath0 and @xmath35 . + now regarding the second part of the orientation condition ( sec . [ sec : smvp ] ) , consider a vertex @xmath130 , which is convex and aligned . two orientations of the triangle @xmath278 are possible , as shown in figures [ fig : convexity]a and [ fig : convexity]c . notice that both constitute reciprocal figures for both vertices , but only one of them satisfies orientation in each case . in our construction of the reciprocal figure for the whole foam , we can decide the orientation of the initial triangle , choosing the one that respects orientation . the question is now if the correct orientation of this starting triangle ensures that of all subsequent ones , whose locations are determined by @xmath250 and @xmath251 . to demonstrate that this is indeed the case , we notice that the triangle of figure [ fig : convexity]a , if considered as a reciprocal figure for the vertex of [ fig : convexity]c , has all three segments @xmath1 wrongly oriented . the point we want to make is that , if the vertex is convex , there are only two possibilities : either all pairs @xmath279 satisfy orientation , or all are wrongly oriented . then if one of the pairs of a triangle forming part of a reciprocal figure is correctly oriented , the other two must necessarily also be . this demonstrates that if the starting pair @xmath280 is chosen with the correct orientation , then all subsequent triangles must be correctly oriented since they share at least a pair of sources with one preexisting triangle . therefore in order to ensure correct orientation of the whole figure it is enough to correctly choose the orientation of the first pair . a dissection of space into cells separated by circular interfaces that meet at triple vertices is called a circular partition ( cp ) . a two - dimensional foam therefore defines a circular partition of two - dimensional space . the equilibrium conditions for the foam impose geometrical constraints on this cp . we have here shown that the cp defined by an equilibrated foam is _ aligned _ and admits an _ oriented reciprocal figure_. this result is valid in general for heterogeneous foams , each of whose films may have an arbitrary ( positive ) surface tension , and even if loads are applied on the vertices , with the sole requirement of equilibrium . we have seen in sec . [ sec : find_sources ] , that any cp satisfying the conditions of alignement and existence of oriented reciprocal figure is a sectional multiplicative partition ( smvp ) . a smvp is a plane cut of a multiplicative partition , thus two - dimensional foams are plane cuts of three - dimensional foams , these last being a multiplicative partition with respect to sources in three dimensions . therefore given an arbitrary equilibrated two - dimensional foam , it is always possible to find sources @xmath3 in three - dimensional space , and amplitudes @xmath41 such that the given foam is a smvp with respect to those sources . + a first implication of this correspondence is the identification of a new set of degrees of freedom for the foam : the intensities @xmath41 and locations of the sources @xmath8 in three dimensions . this allows a more natural description of a foam , than the one that is done in terms of films and vertices . when a foam is interpreted as a tessellation of space with respect to some sources , we see the foam s films and vertices are secondary constructions , and their evolution is a consequence of that of the sources . the dynamical description is conceptually simpler using the smvp interpretation . for example the processes of neighbor switching ( @xmath281 ) and cell disappearance ( @xmath282 ) are described in a unified manner ( fig . [ fig : process ] ) . both are due to the fact that a fourfold vertex in 3d crosses the projection plane @xmath28 . depending on the spatial orientation of the vertex with respect to @xmath28 , this is seen as a @xmath281 or @xmath282 process . + an evolving foam can be seen as a particular instance of a dynamical random lattice@xcite , in which the evolution of a cellular structure is fixed by assigning a given dynamics to the sources of a mathematically defined tessellation . in the case of foams the dynamics is usually fixed by gas diffusion across the membranes . alternatively other dynamical evolution rules may also be interesting , but in any case one has to make the translation to obtain the source s dynamics . the next step is then to find , for a given proposed dynamic evolution for the foam , the corresponding dynamics@xcite for the sources and intensities . + foams are usually studied inside a bounded region or cage , which imposes the only constraint that the films be normal to its boundary . boundaries of this kind do not affect the fact that the foam is a smvp . we have shown that the result holds with the sole requirement of vertex equilibrium . but it is obvious that smvps are always closed on themselves forming a self standing cluster , that is , there are no open film ends . this implies that even bounded foams _ must _ be a region of a larger self - standing cluster of bubbles that is closed on itself , and everywhere equilibrated . the point is not trivial in that it ensures that all films ending at the boundary can be continued , eventually forming new ( phantom ) vertices , and that the resulting foam will have all its vertices in equilibrium . we see then that there is no fundamental difference between bounded and self - standing foams , since all foams are regions of a self - standing cluster . this does not mean that the boundaries have no effect , which would of course be false . if the foam s dynamic inside the cage is for example produced by gas diffusion across the films , the evolution of the `` phantom '' bubbles existing as continuations of the physical foam outside the cage , will _ not _ follow this dynamic , but a different one , which is determined by the constraint that the films be normal to the cage s boundaries . + in the field of joint - bar structures , an old result due to maxwell@xcite states that if a lattice accepts a reciprocal figure then it can support a self - stress , and conversely . more recently ash and bolker@xcite have shown that the existence of a reciprocal figure is sufficient condition for the lattice to be a sectional partition . in this case there is the additional requirement that all vertices are convex , therefore all stresses in the lattice must be of the same sign , and the lattice can be an equilibrated spider web . a chain of results that span a century allow us to see equilibrated spider webs as svps , and conversely . the alert reader may have noticed that equilibrated foams can be seen as a kind of `` generalized '' spider webs , in which the pressure difference between cells is the new ingredient , and is equilibrated by the curvature of the interfaces . it therefore turns out as no surprise that these generalized spider webs ( foams ) are equivalent to an appropriate generalization of svps , namely smvps , which include a multiplicative constant that gives rise to curved interfaces . + i have benefited from discussions with k. lauritsen , h. herrmann , h. flyvbjerg , d. le caer and d. weaire . i thank d. weaire for reference @xcite , and d. le caer for making me aware of plateau s results and sending me references @xcite and @xcite . i am also grateful to h. flyvbjerg and d. stauffer for their efforts to help me make this work more readable . + a classification of space into cells @xmath5 separated by rectilinear interfaces @xmath12 is called a _ rectilinear partition _ of space . given a rectilinear partition @xmath73 of the plane , a _ reciprocal figure _ @xmath283is a planar graph composed of sites @xmath247 joined by edges @xmath71 satisfying @xcite : * sites @xmath247 of @xmath70 are in one - to - one correspondence with cells @xmath5 of @xmath73 . * edges @xmath71 of @xmath70 are in one - to - one correspondence with interfaces @xmath12 of @xmath73 . * edges @xmath71 of @xmath70 are * normal * to interfaces @xmath12 of @xmath73 . clearly a reciprocal figure is defined up to arbitrary global dilatations and translations , since angles are not changed by them . therefore , if @xmath73 admits a reciprocal figure , there are three degrees of freedom in its determination @xcite . an arbitrary partition @xmath73 will _ not _ in general admit a reciprocal figure . we can see this with a simple example . draw an arbitrary polygonal cell @xmath284 with @xmath4 faces , and take arbitrary rectilinear interfaces between its @xmath4 neighbors(fig . [ fig : closing ] ) . now take an arbitrary point @xmath285 on the plane to start with , an assign it to the central cell . this starting point is arbitrary since a reciprocal figure is defined up to arbitrary translations . the other @xmath4 points @xmath286 associated to the external cells must be somewhere on the @xmath4 rays @xmath190 stemming from @xmath285 and normal to the faces @xmath287 of @xmath284 . the global length scale is also arbitrary so that say @xmath288 can be freely chosen . then we choose @xmath289 somewhere on @xmath205 . now point @xmath290 is determined as the intersection of ray @xmath291 with the normal to face @xmath94 going through @xmath289 . this can be repeated to obtain all @xmath4 external points , but in general the figure will not _ close _ , that is , segment @xmath292 will not be normal to interface @xmath293 , whose orientation is arbitrary . + the planar graph formed by joining the sources @xmath0 of a svp with edges @xmath71 , one for each nonempty interface @xmath12 , constitutes a reciprocal figure for the svp . therefore every svp has a reciprocal figure . @xmath294 recently ash and bolker have shown that the existence of a reciprocal figure satisfying _ orientation _ @xcite is also a _ sufficient _ condition for @xmath73 to be a svp . a reciprocal figure @xmath70 satisfies orientation if for each bond @xmath295 , the sites @xmath247 and @xmath248 are oriented in the same way as cells @xmath5 and @xmath11 are with respect to @xmath12 . @xmath296 this result solves the recognition problem for svps . the orientation condition can only be satisfied if all vertices of @xmath73 are _ convex_. we will say that a vertex is convex if the internal angles formed by the interfaces are all smaller than @xmath79 . figure [ fig : orientation ] shows two partitions with three cells . one of them is convex , the other is not . in the second case it is not possible to find a reciprocal figure that satisfies orientation , and therefore it can not be a svp . + given a rectilinear partition @xmath297 , and an oriented reciprocal figure for it ( as in fig . [ fig : recfig ] ) , it is always possible@xcite to find sources @xmath3 in three dimensions , located at heights @xmath30 above the points @xmath0 , such that @xmath297 is the section with @xmath28 of a three - dimensional vp with respect to @xmath3 . the procedure to determine the heights @xmath30 can be easily described . vertices @xmath19 of @xmath297 will be equidistant from sources @xmath8 , @xmath13 , @xmath20 ( see fig . [ fig : svplocate_sources ] ) . start from an arbitrary vertex , say @xmath19 , and draw a spherical surface of _ arbitrary _ radius @xmath298 with center at that vertex . now define sources @xmath8 , @xmath13 and @xmath20 as the intersections of this surface with the verticals ( normals to @xmath28 ) through @xmath0 , @xmath35 and @xmath299 respectively . next go to a neighboring vertex , which shares two sources with this one . let us call it @xmath300 . for this vertex , only source @xmath301 has to be located since @xmath8 and @xmath13 are known . draw a spherical surface with center @xmath300 and containing @xmath8 and @xmath13 . both will be simultaneously contained , since interface @xmath12 , on which @xmath300 is located , is equidistant from @xmath8 and @xmath13 . the radius @xmath302 of this surface is determined by the locations of @xmath8 and @xmath13 , which in turn is fixed by @xmath303 . its intersection with the vertical @xmath304 through @xmath305 determines @xmath301 . if this spherical surface does not intersect @xmath304 , just choose a larger value of @xmath303 and start all over again ( from the initial vertex ) . the construction proceeds in this manner until all sources have been determined . as mentioned , the initial value of @xmath303 is tentative , in the sense that it may have to be modified ( increased ) if at some point during the procedure , a normal line is not cut by the corresponding spherical surface from the vertex . it is easy to see that increasing the value of the starting radius @xmath303 is always enough to solve this problem . + there is thus one degree of freedom in this construction ( @xmath303 ) . we conclude that , given a two - dimensional partition @xmath73 that admits a reciprocal figure , there is a four - parametric family of source locations such that @xmath73 is a svp with respect to them . three degrees of freedom come from the determination of the reciprocal figure itself ( since a dilatation and/or translation of a reciprocal figure is again a reciprocal figure ) and the last one from @xmath303 . this last degree of freedom results from the fact that a svp is invariant if all heights are changed according to @xmath306 with @xmath218 arbitrary ( see equations ( [ eq : svpcell ] ) and ( [ eq : svpinterface ] ) ) . + reciprocal figures were first studied by maxwell @xcite in relation with the rigidity of bar - joint frameworks in the plane . maxwell pointed out that frameworks that have a reciprocal figure are able to support a self - stress , and conversely . the reason is that the edges of the reciprocal figure can be taken to represent forces transmitted by the edges of the original framework ( rotated by @xmath307 ) . since these edges form closed polygons , the existence of a reciprocal figure implies the existence of an equilibrated set of internal stresses in the absence of external load . the addition of the orientation condition ( a condition not required by maxwell s definition of a reciprocal figure ) has the statical consequence that all signs of the stresses are equal , for example all traction or all compression . it is clear that no equilibrium is possible in the case of fig . [ fig : orientation]a if all three stresses are to have the same sign . fig . [ fig : orientation]b on the other hand , can be in equilibrium under compression or traction on all three interfaces . + the conclusion is that every svp is an equilibrium configuration for a spider web @xcite , and conversely , each such equilibrium configuration is a svp . + the existence of a reciprocal figure has also projective consequences , which have been studied by crapo @xcite and whiteley @xcite . we briefly describe here a geometric transformation called _ inversion _ @xcite . we will find it extremely useful for our purpose of discussing circular partitions . an inversion with radius @xmath21 around a point @xmath308 located at @xmath309 transforms a point @xmath310 located at @xmath311 into a point @xmath312 at @xmath313 satisfying @xmath314 where @xmath315 . the sphere of radius @xmath21 and centered at @xmath308 is invariant under this transformation , while the inside and outside of this sphere are interchanged . obviously this transformation is self - inverse . let us now describe some of the properties of this transformation in two dimensions @xcite . most of them apply trivially in higher dimensions . * circles not through @xmath308 are transformed in circles not trough @xmath308 . * circles through @xmath308 are transformed into straight lines not through @xmath308 . * straight lines not through @xmath308 are transformed in circles through @xmath308 . * straight lines through @xmath308 are invariant . * angles are preserved ( in modulus ) by the inversion . given two points @xmath251 and @xmath254 at distances @xmath205 and @xmath291 from the inversion center , the distance @xmath316 between them transforms as @xmath317 using this result we can easily see that the inversion is a `` symmetry '' of a mvp in any dimension if the intensities are also appropriately transformed @xcite . more precisely , a mvp of @xmath318 with respect to sources @xmath22 with intensities @xmath50 is transformed by an inversion into a mvp with respect to @xmath319 with intensities @xmath320 , where the new intensities satisfy @xmath321 here @xmath322 is an arbitrary prefactor , the same for all @xmath41 s , and @xmath190 is the distance between source @xmath247 and the inversion center @xmath308 . to see this it suffices to demonstrate that if @xmath323 then after an inversion , @xmath324 , which is easily done using ( [ eq : mvpinterface ] ) , ( [ eq : dist_transf ] ) and ( [ eq : transforma ] ) . the inversion transformation is of course also a symmetry of the smvp ( section [ sec : smvp ] ) , if the inversion center @xmath308 is on the cutting plane @xmath28 , since in this case the inversion leaves this plane invariant . + if the inversion center @xmath308 happens to be located on an interface @xmath12 of a mvp ( initially a spherical surface ) , the transformed interface @xmath101 is a plane not through @xmath308 . the resulting interface thus corresponds to a partition with respect to sources @xmath247 and @xmath248 in their new locations . therefore the transformed intensities @xmath325 and @xmath326 have to be equal after the inversion , which is verified using ( [ eq : transforma ] ) @xmath327 in the same way a smvp with respect to two sources @xmath247 and @xmath248 is transformed into a svp if @xmath328 . the intensities @xmath41 are transformed according to ( [ eq : transforma ] ) , where @xmath190 is the distance between @xmath308 and source @xmath8 in three - dimensional space . the way in which heights @xmath30 transform is easily found using ( [ eq : inversion ] ) . notice that if the inversion center coincides with the location of a conjugate vertex @xmath59 , then the transformed partition has a rectilinear vertex since three interfaces are simultaneously transformed into straight lines . we will use this property of the inversion transformation when solving the recognition problem for smvps in section [ sec : find_sources ] . n. rivier , _ on the structure of random tissues or froths , and their evolution . _ , philosophical magazine b * 47 * ( 1983 ) , l45-l49 . d. weaire and n. rivier , _ soap , cells and statistics random patterns in two dimensions _ , contemp . * 25 * ( 1984 ) , 59 - 99 . j. glazier and d. weaire , _ the kinetics of cellular patterns _ , j.phys.:condens . matter * 4 * , ( 1992 ) , 1867 - 1894 . j. stavans , _ the evolution of cellular patterns _ , rep . prog . phys . * 56 * ( 1993 ) , 733 - 789 . c. s. smith , _ the shape of things _ , sci . am . * 190 * ( 1954 ) , 58 . c. s. smith , _ grain shapes and other metallurgical applications of topology _ , in : metal interfaces , am . metals , cleveland , 1952 . d. weaire and j. kermode , _ computer simulation of a two - dimensional soap froth i. method and motivation _ , philosophical magazine b * 48 * ( 1983 ) , 245 - 259 . + j. kermode and d. weaire , _ 2d - froth : a program for the investigation of 2-dimensional froths _ , comput . phys . * 60 * ( 1990 ) , 75 - 109 . + t. herdtle and h. aref , _ numerical experiments on a two - dimensional foam _ , j. fluid mech . * 241 * ( 1992 ) , 233 - 260 . h. flyvbjerg and c. jeppesen , _ a solvable model for coarsening soap froths and other domain boundary networks in two dimensions _ , physica scripta * t38 * , ( 1991 ) , 49 - 54 . + h. flyvbjerg , _ dynamics of soap froth _ , physica a * 194*,(1993 ) , 298 . + h. flyvbjerg , _ model for coarsening froths and foams _ , phys . rev . e * 47 * , ( 1993 ) , 4037 . k. kawazaki , t. nagai and k. nakashima , phil . mag . b * 60 * , ( 1989 ) , 399 . + k. kawasaki , physica a * 163 * , ( 1990 ) , 59 . , a. okabe , b. boots and k. sugihara , wiley 1992 . h. honda , _ description of cellular patterns by dirichlet domains : the two - dimensional case _ , j. theor . * 72 * , ( 1978 ) , 523 . v. icke and r. van de weygaert , fragmenting the universe . i. statistics of two - dimensional voronoi foams , astron . astrophys . * 184 * , ( 1987 ) , 16 - 32 . hiroshi imai , masao iri and kazuo murota , _ diagram in the laguerre geometry and its applications . _ , siam j. comput . * 14 * , ( 1985 ) , 93 . b. n. boots , _ weighting thiessen polygons _ , _ economic geography _ ( 1979 ) , 248 - 259 . , c. vernon boys , ny : dover 1959 , pp . 120 - 127 . + reprinted in : _ the world of mathematics _ , tempus books of microsoft press 1988 vol . ii , pp . 883 - 886 . , darcy wentworth thompson , cambridge university press 1961 , page 96 . maxwell , j. c. _ on reciprocal figures and diagrams of forces _ , _ phil . mag . series _ * 4 * ( 1864),250 - 261 . maxwell , j. c. _ on reciprocal figures , frames , and diagrams of forces _ , _ trans . royal soc . edinburgh _ * 26 * ( 1869 - 72),1 - 40 . peter f. ash and ethan d. bolker , _ generalized dirichlet tessellations _ , _ geometriae dedicata _ * 20 * ( 1986),209 - 243 . henry crapo , _ structural rigidity _ , _ structural topology _ * 1 * ( 1979 ) , 26 - 45 . walter whiteley , _ realizability of polyhedra _ , structural topology , * 1 * ( 1979 ) , 46 - 58 . h. telley , _ modelisation et simulation bidimensionelle de la croissance des polycrystaux _ , unpublished , thesis report n. 780 . ecole polytechnique federale de lausanne , ( 1989 ) . this is called `` phantom vertex '' in @xcite . , h. s. m. coxeter , second edition , wiley and sons 1989 . a configuration @xmath329 is formed by @xmath330 points and @xmath331 lines , such that each line is adjacent to @xmath79 points , and each point is adjacent to @xmath332 lines . see for example , _ geometry and the imagination _ , d. hilbert and s. cohn - vossen , chelsea publishing co. , new york 1952 , chapter iii . c. moukarzel,_voronoi foams _ , physica * a 199 * ( 1993),19 - 30 . k. lauritsen , c. moukarzel and h. j. herrmann , _ statistical physics and mechanics of random lattices _ , _ j. physique i _ ( france ) * 3 * ( 1993),1941 . c. moukarzel , to be published . within our generalized notion of distance , that is , including the multiplicative constants . although it is statically unstable , a foam all of whose films are under compression would also be geometrically possible .

charge - parity ( cp ) violation is one of the least understood phenomena in high energy physics . in the standard model ( sm ) , cp violation is parametrized by one arbitrary phase in the quark mixing matrix and its value fixed by experimental observations . models beyond the sm , especially those including supersymmetry , predict several new cp - odd phases . the effects of some of these phases , particularly the ones coming from the soft supersymmetry breaking scalar masses , have been studied extensively @xcite . in the minimal supersymmetric standard model ( mssm ) , these phases need to either be unreasonably small , or the scalar fermion masses unusually large @xcite . a solution to this so called `` susy cp - problem '' has been suggested by the authors in @xcite . by enlarging the supersymmetric gauge sector to be left - right symmetric , parity forces yukawa couplings to be hermitean , and mass terms in the superpotential to be real . thus the cp problem is solved at the right - handed ( @xmath1 , or seesaw ) scale @xmath2 . if @xmath2 is of the order of the electroweak scale , then no other phase can be generated and the problem is solved . if however @xmath2 is a higher scale , the scenario favored by neutrino masses and grand - unification schemes with left - right models as intermediate steps , this is no longer the case . in some variants of the model , the phases appear only at two loop level and are thus naturally small @xcite , in others the electroweak phase is generated form the renormalization group equations evolution from the seesaw scale , and the electron dipole moment ( edm ) problems persist @xcite . the left - right supersymmetric model ( lrsusy ) enlarges the minimal supersymmetric standard model ( mssm ) gauge symmetry @xmath3 to @xmath4 . this symmetry allows for the seesaw mechanism within supersymmetry and predicts neutrino masses and mixing naturally . in the supersymmetric sector of the theory , it introduces right - handed gauginos in addition to several higgsinos . we expect this change to be most pronounced in the electroweak chargino sector , where previously forbidden right - right interactions are now allowed to proceed with a strength comparable to left - left interactions . this could have significant effects on the cp - phases in the sector , where more phases are allowed , which could distinguish the model from the mssm . by contrast , we expect the effect to be less pronounced in the neutralino sector , where right - handed gaugino interactions are permitted in mssm . in this work , we investigate the effects of such phases on the masses , production , and decay rates of charginos . our emphasis is on an analysis of a general unconstrained model , assuming arbitrary phases , and finding the combination of angles that would affect the phenomenology , as well as presenting experimentally testable signals from the chargino sector . we assume in what follows that the phases can be taken arbitrarily large , and that cp - violation restrictions in the quark and lepton sectors can be satisfied either by allowing scalar masses to be large , or through cancellations of various contributions @xcite . we perform a detailed and complete analysis of the chargino sector including masses , production rates , forward - backward asymmetry in production and decays . for decays , we chose to analyze the channel most sensitive to left - right symmetry , the decay into a sneutrino and a lepton . we analyze the dependence on production and subsequent decay from a polarized @xmath0 beam and the t - odd asymmetry associated with this . our paper is organized as follows : in section [ sec : model ] we present the general framework of the left - right supersymmetric model with special emphasis on the chargino sector . we then discuss the chargino masses , with particular attention paid to the effect of the cp - odd angles in section [ sec : masses ] . in section [ sec : production ] , we investigate the cross section for chargino production in an unpolarized @xmath0 collider and its dependence on the phases in the theory . we discuss the variation of the decay rates of charginos with cp - phases as well as the t - odd asymmetry in polarized production and decay in section [ sec : decays ] . section [ sec : conclusion ] is devoted to our summary and conclusions . the minimal supersymmetric left - right model is based on the gauge group @xmath4 @xcite . the matter fields of this model consist of three families of both left and right handed quark @xmath5 and lepton @xmath6 fields . the higgs sector consists of two bidoublet ( @xmath7 ) and four triplet higgs ( @xmath8 and @xmath9 ) superfields . the bidoublet higgs superfields implement the @xmath10 symmetry breaking and generate a cabibbo - kobayashi - maskawa ( ckm ) mixing matrix . the triplet higgs @xmath11 bosons allow for the seesaw mechanism , while the additional triplet fields @xmath12 are needed to cancel triangle gauge anomalies in the fermionic sector . the most general superpotential involving these superfields is : @xmath13 where @xmath14 and @xmath15 are the yukawa couplings for the quarks and leptons , respectively and @xmath16 is the coupling for the triplet higgs bosons . the parameters @xmath17 and @xmath18 are the higgs mass parameters . left right symmetry requires all @xmath19-matrices to be hermitean in the generation space and @xmath16 matrix to be symmetric . here @xmath20 denotes ( possible ) non - renormalizable terms arising from higher scale physics or planck scale effects @xcite . the presence of these terms insures that , when the susy breaking scale is above @xmath21 , the ground state is r - parity conserving . left - right supersymmetric models have one additional problem : two of the physical higgs bosons can contribute flavor - violating neutral currents ( fcnc ) by violating flavor by two units and by generating too large a contribution to the @xmath22 mass splitting . this problem can be circumvented by raising the right - handed mass scale to @xmath23 tev @xcite . in addition , the lagrangian also includes soft supersymmetry breaking terms as well as @xmath24- and , @xmath25-terms . for a more detailed description of the model , see @xcite . in the supersymmetric sector of the model there are six singly - charged charginos , corresponding to @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , and @xmath31 . the model also has eleven neutralinos , corresponding to @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , and @xmath42 . the doubly charged higgs and higgsinos do not affect quark phenomenology , but the neutral and singly charged components do , through mixings in the chargino and neutralino mass matrices . the terms relevant to the masses of charginos in the lagrangian are @xmath43 where @xmath44 and @xmath45 , and @xmath46 where we have taken , for simplification , @xmath47 . here @xmath48 and @xmath49 are the bidoublet higgs bosons vacuum expectation values ( vevs ) , @xmath50 and @xmath51 are the triplet higgs bosons vevs , and @xmath52 the @xmath53 and @xmath1 gaugino masses , respectively . the chargino mass eigenstates @xmath54 are obtained by @xmath55 with @xmath56 and @xmath57 unitary matrices satisfying @xmath58 the diagonalizing matrices @xmath59 and @xmath56 are obtained by computing the eigenvectors corresponding to the eigenvalues of @xmath60 and @xmath61 , respectively . from the above general mass matrices for charginos , we can obtain analytic expressions for the eigenvalues . first , note that the vevs of the triplet higgs bosons , @xmath50 and @xmath51 , are responsible for giving masses to @xmath62 bosons , as well as implementing the seesaw mechanism . as such , @xmath63 several tev , and thus we can safely decouple @xmath64 and @xmath65 from the lighter chargino states . we now incorporate the most general set of cp violating phases in the mass matrix . that is , we do not restrict the phases according to any particular scenario , and allow the most general set of phases . these can easily be generated from the rge evolution running down from @xmath2 to the electroweak scale . since @xmath52 and @xmath66 , the higgsino coupling parameter are assumed complex at the electroweak scale , with nontrivial phases @xmath67 and @xmath68 , respectively , then they can include phases : @xmath69 and , in addition we allow the higgs vacuum expectation values to be complex , with phases @xmath70 : @xmath71 and @xmath72 . the chargino mass matrix @xmath73 with the most general allowed set of @xmath74 violating phases becomes @xmath75 where the phase angles ( i.e. @xmath76 ) are in the interval @xmath77 $ ] . the results in the @xmath78 conserving limit , maybe obtained by simple taking the values @xmath79 or @xmath80 , in the matrix . using the transformation @xmath81 where @xmath82 is a unitary matrix . then the matrix @xmath83 can be written as @xmath84 where we have defined @xmath85 and @xmath86 . these phases allow for both explicit and/or spontaneous cp violation ; in the latter case @xmath87 and @xmath88.the dependence with respect to only two angles is now clear . the phases @xmath89 and @xmath90 are the new sources of @xmath74 violation , which can vary in the range @xmath91 , and @xmath92 , respectively . to diagonalize the matrix @xmath83 we use the following transformation @xmath93v^{-1 } = { \rm diag}(\mathrm{\tilde{m}}_{\chi^{\pm}_{1}},\mathrm{\tilde{m}}_{\chi^{\pm}_{2}},\mathrm{\tilde{m}}_{\chi^{\pm}_{3}},\mathrm{\tilde{m}}_{\chi^{\pm}_{4}})\equiv m_{d}^{2},\end{aligned}\ ] ] where we assume @xmath94 . notice that the transformation matrix @xmath56 is now a @xmath95 matrix and a function only of @xmath89 and @xmath96 . the chargino masses can be determined by solving the eigenvalue equation analytically . we use the method outlined in @xcite . we get the exact analytic expressions for the chargino masses as functions of the cp angles , i.e. , @xmath97 . the analytic formulas for the chargino masses are given by @xmath98^{\frac{1}{2}},\\ \mathrm{\tilde{m}}_{\chi^{\pm}_{2 } } & = & { \cal r}e\left[\frac{a}{4}-\frac{\alpha}{2}+\frac{1}{2}\,\sqrt{\zeta-\varpi-\frac{\lambda}{4\alpha } } \right]^{\frac{1}{2}},\\ \mathrm{\tilde{m}}_{\chi^{\pm}_{3 } } & = & { \cal r}e\left[\frac{a}{4}+\frac{\alpha}{2}-\frac{1}{2}\,\sqrt{\zeta-\varpi+\frac{\lambda}{4\alpha } } \right]^{\frac{1}{2}},\\ \mathrm{\tilde{m}}_{\chi^{\pm}_{4 } } & = & { \cal r}e\left[\frac{a}{4}+\frac{\alpha}{2}+\frac{1}{2}\,\sqrt{\zeta-\varpi+\frac{\lambda}{4\alpha}}\right]^{\frac{1}{2}},\end{aligned}\ ] ] where @xmath99 represents the real value of the function and @xmath100,\\ \zeta & \equiv & \nonumber \left[\frac{a^{2}}{2}-\frac{4b}{3}-\nu\right],\,\,\;\nu \equiv\nonumber\frac{(2^{\frac{1}{3}}\,\gamma)}{3\,\epsilon},\\ \varpi & \equiv & \nonumber\frac{\epsilon}{32^{\frac{1}{3}}},\,\,\,\epsilon \equiv \nonumber(\delta+\sqrt{\eta})^{\frac{1}{3}},\:\quad\qquad\eta \equiv \nonumber ( -4\,\gamma^{3}+\delta^{2}),\\ \gamma & \equiv & \nonumber ( b^{2}+3\,a\,c+12\,d),\;\delta \equiv\nonumber(2\,b^{3}+9\,a\,b\,c+27\,c^{2}+27\,a^{2}\,d-72\,b\,d),\\ \lambda & \equiv&\nonumber ( a^{3}-4\,a\,b-8\,c),\end{aligned}\ ] ] with parameters @xmath101 given in terms of the chargino mass - squared matrix elements @xmath102 as : @xmath103 , \nonumber \\ c & \equiv & \frac16\left [ ( { \rm tr}m_{ij})^3- 3 { \rm tr}m_{ij}{\rm tr}m^2_{ij}+2 { \rm tr}m^3_{ij}\right ] , ~ d \equiv{\det } m_{ij}.\end{aligned}\ ] ] similarly , we can obtain analytic expressions for the diagonalizing matrices @xmath104 . while expressions for these exist in the literature @xcite , here we will fix the masses and @xmath105 and use the numerical values for the elements which include the phases . we now investigate the phase dependence of the masses and take as an example the light chargino mass scenario . in numerical analyses we assume the following relations among the parameters of the model : @xmath106 , and we set @xmath107 gev , @xmath108 gev , and @xmath109 . as mentioned before , we further assume the vev s @xmath50 and @xmath51 to be big enough for @xmath110 to decouple from the lighter charginos . in fig . [ fig : masses ] , the cp - odd phase dependences of the lightest and the heaviest chargino masses are shown in a two - dimensional contour plot . the masses of the four charginos in the cp - conserving case are @xmath111 gev . among four , the second heaviest is independent of phases and determined completely by the higgsino mass ( @xmath66 ) . this is clear from the chargino mass matrix given in eq . ( [ eq : charmassmatrix ] ) where the phase dependence of the @xmath66 entry disappears when we consider @xmath112 . from analyses , we observe that the second lightest mass is practically independent of cp - odd phases so that we only include the figures for the lightest and the heaviest masses . 3.5 in 3.5 in -0.2 in in the left graph of fig . [ fig : masses ] , we give the phase dependence of the heaviest chargino mass . this chargino is important for the analysis in section [ sec : decays ] . it can be seen that the dependence is not strong and the mass varies in 272 - 280 gev range and the cp - odd phases can enhance its value by at most @xmath113 , which occurs in the parameter regions @xmath114 , @xmath115 , or @xmath116 . these regions are denoted with dashed - dotted line contours in the graph . the ones with the solid line are the regions where the mass takes its lowest value . from the graph , these are approximately @xmath117 , @xmath118 , and @xmath119 . the second graph in fig . [ fig : masses ] shows the same phase dependence for the lightest chargino mass @xmath120 . unlike the case of the heaviest chargino , there is a strong dependence on the cp - odd phases @xmath89 and @xmath96 . when we span the whole region in the @xmath121 space , the mass runs in the interval @xmath122 gev . this means that up to @xmath123 destructive effects in the lightest chargino mass could be generated from cp - odd phases . if one divides the @xmath96 interval into two equal regions and the interval for @xmath89 into 4 equal parts , the mass is at its lowest value for either very small @xmath96 values ( @xmath124 ) with @xmath125 or maximum @xmath96 value ( @xmath126 ) with two possible values for @xmath89 , either @xmath127 or @xmath128 . note that the regions bounded by the contours corresponding to 104 gev are not experimentally allowed . at the future linear collider one will be able to determine the masses of charginos and the pair production cross section to high accuracies . in this part , we evaluate and illustrate the chargino pair production in unpolarized @xmath0 annihilation with non - trivial phases . the cross sections without cp - violation phases are analyzed in @xcite . for completeness , we include the expressions for the cross sections here . we introduce the following variables @xmath129 where the momenta of the incoming particles are represented by @xmath130 and the momenta of the outgoing particles by @xmath131 . the chargino production occurs at tree - level through @xmath132 in the s - channel , and @xmath133 in the t - channel . note that the channels through @xmath134 and @xmath135 are new channels specific to lrsusy . the corresponding production cross section and effects of cp - phases have been analyzed in mssm @xcite . 5.8 in -0.6 cm the lagrangian responsible for the interaction is : @xmath136{\chi}^{+}_{j } \nonumber \\&+&\frac { g_{r}\sqrt { \cos 2 \theta_{w}}}{\cos\theta_{w } } z^{\mu}_{l } \bar{\chi}^{+}_{i}\gamma^{\mu}\biggl[o^{\prime l}_{ij}p_{l}+o^{\prime r}_{ij}p_{r}\biggr ] { \chi}^{+}_{j } \nonumber \\ & -&g \sum_i { \bar \chi}_i^{+ c } \left \ { \left [ v_{i1}^{\ast}p_l-\left ( \frac{m_{\nu_k}u_{i3}}{\sqrt{2}m_w \sin \beta } + \frac{m_{e_k}u_{i4}}{\sqrt{2}m_w \cos \beta } \right)p_r\right ] e_k { \tilde \nu}_{lk}^{\ast } \right . \nonumber \\ & + & \left . \left [ u_{i2}p_r- \left ( \frac{m_{\nu_k}v_{i3}^{\ast}}{\sqrt{2}m_w \sin \beta } + \frac{m_{e_k}v_{i4}^{\ast}}{\sqrt{2}m_w \cos \beta}\right ) p_l\right ] e_k { \tilde \nu}_{rk}^{\ast } \right\},\end{aligned}\ ] ] where @xmath137 is the left(right ) chirality projection operator and we define the following matrices @xmath138 we perform the calculation in the center of mass frame , where the expression for the cross section is given by @xmath139 where we neglect the lepton masses . the triangle function @xmath140 is defined as @xmath141 and @xmath142 is the invariant amplitude of the collision process . the @xmath143 vertex is defined as @xmath144 where @xmath145 , and @xmath146 the total cross section is obtained by integrating over the angle @xmath147 . it yields @xmath148 where the terms in the cross section are as follows @xmath149\lambda^{1/2}(1,y , z ) , \\ \sigma_{\tilde{\nu}_{l } } & & \hspace*{-0.2cm}= \frac{g^{4}}{128\pi s}|v_{i1}|^{2}|v_{j1}|^{2}\biggl\{\biggl[\frac{1}{(a^{2}-b^{2})}-\frac{(y - z)^{2}}{(a^{2}-b^{2})}-4\lambda^{-1/2}(1,y , z)\nonumber\\ & & \times \biggl(\ln(\frac{a+b}{a - b})-\frac{2ab}{a^{2}-b^{2}}\biggr)+4\biggl(1-\frac{a}{b}\ln(\frac{a+b}{a - b})+\frac{a^{2}}{a^{2}-b^{2}}\biggr)\biggr]\biggr\}\lambda^{1/2}(1,y , z ) , \\ \sigma_{\tilde{\nu}_{r } } & & \hspace*{-0.2cm}= \frac{g^{4}}{128\pi s}|u_{i2}|^{2}|u_{j2}|^{2}\biggl\{\biggl[\frac{1}{(a^{\prime 2}-b^{2})}-\frac{(y - z)^{2}}{(a^{\prime 2}-b^{2})}-4\lambda^{-1/2}(1,y , z)\nonumber\\ & & \times \biggl(\ln(\frac{a^{\prime}+b}{a^{\prime}-b})-\frac{2a^{\prime}b}{a^{\prime 2}-b^{2}}\biggr)+4\biggl(1-\frac{a^{\prime}}{b}\ln(\frac{a^{\prime}+b}{a^{\prime}-b})+\frac{a^{\prime 2}}{a^{\prime 2}-b^{2}}\biggr)\biggr]\biggr\}\lambda^{1/2}(1,y , z ) , \\ \sigma_{\gamma z_{l } } & & \hspace*{-0.2cm}= \frac{e^{2}g^{2 } } { 32 \pi\cos^{2}\theta_{w}}\delta_{ij}(c_{l}+c_{r}){\cal r}e[d_{z_l}(s)(o^{\prime l}_{ij}+o^{\prime r}_{ij})]\nonumber\\ & & \times [ 1-(y - z)^{2}+\frac{1}{3}\lambda(1,y , z)+4x]\lambda^{1/2}(1,y , z ) , \\ \sigma_{\gamma z_{r } } & & \hspace*{-0.2cm}= \frac{e^{2}g^{2 } } { 32\pi\cos^{2}\theta_{w}}\cos2\theta_{w}\delta_{ij}(c_{l}^{\prime}+c_{r}^{\prime}){\cal r}e[d_{z_r}(s)(o^{l}_{ij}+o^{r}_{ij } ) ] \nonumber\\ & & \times [ 1-(y - z)^{2}+\frac{1}{3}\lambda(1,y , z)+4x]\lambda^{1/2}(1,y , z ) , \\ \sigma_{\gamma \tilde{\nu}_{l } } & & \hspace*{-0.2cm}= \frac{-e^{2}g^{2}}{64 \pi s}|v_{i1}|^{2}\delta_{ij}\biggl[[1-(y - z)^{2}+4x ] \frac{1}{b}\ln(\frac{a+b}{a - b})-4(1+a)\nonumber\\ & & \times \biggl(2-\frac{a}{b}\ln(\frac{a+b}{a - b})\biggr)\biggr]\lambda^{1/2}(1,y , z ) , \\ \sigma_{\gamma \tilde{\nu}_{r } } & & \hspace*{-0.2cm}= \frac{-e^{2}g^{2}}{64 \pi s}|u_{i2}|^{2}\delta_{ij}\biggl[[1-(y - z)^{2}+4x ] \frac{1}{b}\ln(\frac{a^{\prime}+b}{a^{\prime}-b}-4(1+a^{\prime})\nonumber\\ & & \times \biggl(2-\frac{a^{\prime}}{b}\ln(\frac{a^{\prime}+b}{a^{\prime}-b})\biggr)\biggr]\lambda^{1/2}(1,y , z ) , \\ \sigma_{z_{l } \tilde{\nu}_{l } } & & \hspace*{-0.2cm}= \frac{-g^{4}}{64 \pi\cos^{2}\theta_{w}}c_l\biggl\{{\cal r}e[d_{z_l}(s)v_{i1}^{\ast}v_{j1}o^{\prime l}_{ij } ] \biggl([1-(y - z)^{2}]\frac{1}{b}\ln(\frac{a+b}{a - b})-4(1+a)\nonumber\\ & & \times [ 2-\frac{a}{b}\ln(\frac{a+b}{a - b})]\biggr)+{\cal r}e[d_{z_l}(s)v^{\ast}_{i1}v_{j1}o^{\prime r}_{ij}]\frac{4x}{b}\ln(\frac{a+b}{a - b})\biggr\}\lambda^{1/2}(1,y , z ) , \\ \sigma_{z_{l } \tilde{\nu}_{r } } & & \hspace*{-0.2cm}= \frac{-g^{4}}{64 \pi\cos^{2}\theta_{w}}c_r\biggl\{{\cal r}e[d_{z_l}(s)u_{i2}u^{\ast}_{j2}o^{\prime l}_{ij } ] \biggl([1-(y - z)^{2}]\frac{1}{b}\ln(\frac{a^{\prime}+b}{a^{\prime}-b } ) - 4(1+a^{\prime})\nonumber\\ & & \times [ 2-\frac{a^{\prime}}{b}\ln(\frac{a^{\prime}+b}{a^{\prime}-b})]\biggr)+{\cal r}e[d_{z_l}(s)u_{i2}u^{\ast}_{j2}o^{\prime r}_{ij}]\frac{4x}{b}\ln(\frac{a^{\prime}+b}{a^{\prime}-b})\biggr\}\lambda^{1/2}(1,y , z ) , \\ \sigma_{z_{r } \tilde{\nu}_{l } } & & \hspace*{-0.2cm}= \frac{-g^{4}}{64 \pi\cos^{2}\theta_{w}}\cos2\theta_{w } c_l^{\prime}\biggl\{{\cal r}e[d_{z_r}(s)v_{i1}^{\ast } v_{j1}o^{l}_{ij } ] \biggl([1-(y - z)^{2 } ] \frac{1}{b}\ln(\frac{a+b}{a - b})\nonumber\\ & & -4(1+a)[2-\frac{a}{b}\ln(\frac{a+b}{a - b})]\biggr)+{\cal r}e[d_{z_r}(s)v^{\ast}_{i1}v_{j1}o^{r}_{ij}]\frac{4x}{b } \ln(\frac{a+b}{a - b})\biggr\}\nonumber \\ & & \times \lambda^{1/2}(1,y , z ) , \\ \sigma_{z_{r } \tilde{\nu}_{r } } & & \hspace*{-0.2cm}= \frac{-g^{4}}{64 \pi\cos^{2}\theta_{w}}\cos2\theta_{w } c_r^{\prime}\biggl\{{\cal r}e[d_{z_r}(s)u_{i2}u^{\ast}_{j2}o^{l}_{ij } ] \biggl([1-(y - z)^{2 } ] \frac{1}{b}\ln(\frac{a^{\prime}+b}{a^{\prime}-b})\nonumber\\ & & -4(1+a^{\prime})[2-\frac{a^{\prime}}{b}\ln(\frac{a^{\prime}+b}{a^{\prime}-b})]\biggr)+{\cal r}e[d_{z_r}(s)u_{i2}u^{\ast}_{j2}o^{r}_{ij}]\frac{4x}{b } \ln(\frac{a^{\prime}+b}{a^{\prime}-b})\biggr\}\nonumber \\&&\times \lambda^{1/2}(1,y , z ) , \\ \sigma_{z_{l , r } } & & \hspace*{-0.2cm}= \frac{g^{4}s}{64 \pi\cos^{4}\theta_{w}}\cos2\theta_{w } ( c_{l}c_l^{\prime}+ c_{r}c_r^{\prime } ) \biggl [ { \cal r}e[d_{z_l}(s)d_{z^{\ast}_r } ( s ) ( o^{\prime l}_{ij}o^{l \ast } _ { ij}+o^{\prime r}_{ij}o^{r \ast } _ { ij } ) ] \nonumber \\ & & \times [ 1-(y - z)^{2}+\frac{1}{3}\lambda(1,y , z)]+ 4x{\cal r}e[d_{z_l}(s)d_{z^{\ast}_r}(s)(o^{\prime l}_{ij}o^{r \ast}_{ij}+o^{l \ast}_{ij}o^{\prime r}_{ij } ) ] \biggr ] \nonumber \\ & & \times ] \lambda^{1/2}(1,y , z ) , \\ \sigma_{z_l } & & \hspace*{-0.2cm}= \frac{g^{4}s}{64 \pi\cos^{4}\theta_{w}}|d_{z_l}(s)|^{2}(c_{l}^{2}+c_{r}^{2 } ) \biggl\{(|o^{\prime l}_{ij}|^{2}+|o^{\prime r}_{ij}|^{2 } ) \nonumber \\ & & \times [ 1-(y - z)^{2}+\frac{1}{3}\lambda(1,y , z)]+4x{\cal r}e[{o}^{\prime l}_{ij}\,{o}^{\prime r \ast}_{ij } ] \biggr\ } \lambda^{1/2}(1,y , z ) , \\ \sigma_{z_r } & & \hspace*{-0.2cm}= \frac{g^{4}s \cos^{2}2\theta_{w}}{64 \pi\cos^{2}\theta_{w}}|d_{z_r}(s)|^{2}(c_{l}^{\prime 2}+c_{r}^{\prime 2})\biggl\{(|o^{l}_{ij}|^{2}+|o^{r}_{ij}|^{2 } ) \nonumber \\ & & \times [ 1-(y - z)^{2}+\frac{1}{3}\lambda(1,y , z)]+4x{\cal r}e [ o^{l}_{ij}\,o^{r \ast}_{ij } ] \biggr\}\lambda^{1/2}(1,y , z).\end{aligned}\ ] ] here , we have taken @xmath150 , @xmath151 and @xmath152 ; @xmath153 , and the variables @xmath154 are defined as @xmath155 in fig . [ fig : cs11 ] , the cross section for the process @xmath156 in the ( @xmath157)-plane is depicted at two different center of mass energies , 500 and 1000 gev . the value of the cross section without cp - odd phases is around 50 pb for @xmath158 gev and 0.35 pb for @xmath159 gev . cp - odd phases can lead to either constructive or destructive interference effects , about a maximum @xmath160 for constructive and @xmath161 for destructive effects at @xmath158 gev . these occur at @xmath162 and @xmath163 in the @xmath164-plane , respectively . for @xmath159 gev , there exists a similar situation with about @xmath165 constructive ( destructive ) effects . in this case , the maximum occurs at @xmath163 . the case with final states @xmath166 is shown in fig . [ fig : cs22 ] . in this case , constructive ( destructive ) effect can reach up to @xmath167 at @xmath158 gev and @xmath168 at @xmath159 gev . 3.5 in 3.5 in -0.2 in 3.5 in 3.5 in -0.2 in 3.5 in 3.5 in -0.2 in figs . [ fig : cs12 ] and [ fig : cs14 ] represent the cp - odd phase dependence of the cross sections for @xmath169 and @xmath170 , respectively . the extrema for the cross section occur in more than one region in the @xmath164 plane . for the @xmath171 case at @xmath158 gev the maximum cross section is obtained for @xmath172 and @xmath173 $ ] , and the minimum at around @xmath174 and @xmath175 . the cp - odd phase effects are slightly smaller for constructive interference : around @xmath176 for the constructive ( destructive ) case . at @xmath159 gev , there is almost no destructive interference in the entire plane . the cross section can get enhanced at most by @xmath123 in three different intervals shown in fig . [ fig : cs12 ] . the @xmath177 case is similar to the @xmath178 or @xmath166 and the @xmath89 values for getting a maximum ( minimum ) value for the cross section are the same ( around @xmath179 for constructive and @xmath180 for destructive ) each at different center of mass energies , 500 and 1000 gev . the enhancement can be more than @xmath181 for both energies . the destructive effects can not exceed @xmath123 at @xmath159 gev , but become more than @xmath182 at @xmath158 gev . 3.5 in 3.5 in -0.2 in 3.5 in 3.5 in -0.2 in we also consider the angular distribution @xmath183 where @xmath184 is angle between the chargino and the electron beam , as well as the forward backward asymmetry ( @xmath185 ) defined as : @xmath186 3.5 in 3.5 in -0.2 in 3.5 in 3.5 in -0.2 in in fig . [ fig : afb11 ] , the forward - backward asymmetry ( @xmath185 ) is shown in the @xmath164-plane for @xmath187 at both @xmath158 gev ( on the left ) and @xmath159 gev ( on the right ) . the asymmetry without cp - odd phases is around @xmath188 at @xmath158 gev and @xmath189 at @xmath159 gev . even though the asymmetry is a lot bigger at @xmath159 gev than at @xmath158 gev , it is more sensitive to the cp - odd phases for latter case . the change in @xmath185 can be as much as @xmath181 and can even change sign ( there is an almost symmetric pattern with respect to @xmath190 line ) . however , the increase in @xmath185 can not exceed @xmath191 at @xmath159 gev . for @xmath192 and @xmath193 , cp - odd phases effects cancel each other . maximum constructive effects occur around @xmath179 for both angles . as depicted in fig . [ fig : afb22 ] , the maximum asymmetry for @xmath194 is obtained at around @xmath195 at both @xmath158 and @xmath196 gev ( around @xmath197 and @xmath123 increase , respectively ) . maximum destructive interference effects are slightly smaller and take place for @xmath198 . the final set of graphs for @xmath185 is shown in fig . [ fig : afb12 ] for the process @xmath169 . at @xmath158 gev , cp - odd phases have constructive effects in almost the entire range and the asymmetry can reach a maximum @xmath199 , for which @xmath185 is about @xmath200 bigger than the case without cp phases . for values of the phases around @xmath179 , destructive effects get maximized at @xmath159 gev and suppress the @xmath185 to be as small as @xmath201 for the phases around @xmath179 . the suppression in percentage is almost equal to the enhancement at @xmath158 gev . in this section we consider chargino production by a polarized @xmath0 beam , followed by two body chargino decay . we calculate the associated production and decay cross section and the observable t - odd asymmetry . two - body decays of charginos have been studied in the context of the mssm without cp violating phases @xcite . there , the rates of interest are for decays to sm particles : @xmath202 , if kinematically allowed . though these processes are also allowed in lrsusy , their decay ratios do not differ significantly from the ones in mssm . we choose to concentrate here on those decays which could show a difference in signals between mssm and lrsusy . these are the decays to leptons / sleptons , which are sensitive to the right - handed gauginos . these decay widths are : @xmath203 \right ] , \nonumber \\ \gamma_d ( \chi^+_i \rightarrow l_j^+{\tilde \nu}_{rj})&= & \frac{g^2{\tilde m}_{\chi_i^+ } \lambda^{1/2}(1 , y^2_l , y^2_{{\tilde \nu}_r})}{64 \pi y_w^2 \cos^2 \beta } \left [ \left ( 1+y_l^2-y_{{\tilde \nu}_r}^2 \right ) \left ( y_l^2 |v_{i4}|^2 + 2y_w^2 \cos^2\beta|u_{i2}|^2 \right ) \right . \nonumber \\ & -&\left . 4\sqrt{2 } y_w y_l \cos \beta \;{\cal r}e[u_{i2}v_{i4 } ] \right ] , \\ \gamma_d ( \chi^+_i \rightarrow { \tilde l}_{lj}^+ { \nu}_{j})&= & \frac{g^2{\tilde m}_{\chi_i^+ } \lambda^{1/2}(1 , y^2_{{\tilde l}_l } , 0)}{64 \pi y_w^2 \cos^2 \beta } \left [ \left ( 1+y_l^2-y_{{\tilde l}_l}^2 \right ) \left ( y_l^2 |v_{i4}|^2 + 2y_w^2 \cos^2\beta|u_{i1}|^2 \right ) \right ] , \nonumber \\ \gamma_d ( \chi^+_i \rightarrow { \tilde l}_{rj}^+ { \nu}_{j})&= & \frac{g^2{\tilde m}_{\chi_i^+ } \lambda^{1/2}(1 , y^2_{{\tilde l}_r } , 0)}{64 \pi y_w^2 \cos^2 \beta } \left [ \left ( 1+y_l^2-y_{{\tilde l}_r}^2 \right ) \left ( y_l^2 |u_{i4}|^2 + 2y_w^2 \cos^2\beta|v_{i2}|^2 \right ) \right ] , \nonumber\end{aligned}\ ] ] where @xmath204 . similarly , charginos could decay to quarks / squarks , if kinematically allowed , though if sleptons are lighter , we expect the decay rates to ( s)leptons to dominate over the ones to ( s)quarks . in principle the decay processes are similar , and we have , as in the mssm @xcite : @xmath205 if kinematically allowed . these types of decays would be less important phenomenologically than the decays into neutralinos if the lightest neutralino is the lightest supersymmetric particle ( lsp ) . the first two would be dominant if the sneutrino is the lsp and the last two would be possible for a heavier chargino if @xmath206 , and if the neutralino is the lsp . in the case the neutralino is the lps , the scalar leptons and scalar neutrinos would decay further , @xmath207 . we concentrate in this section on the first two decay channels , since we expect these to show the strongest deviations from mssm ( due to the availability of the right - handed sneutrino ) . we study the chargino production @xmath208 with longitudinally polarized beams , and the subsequent decay of one of the charginos into a sneutrino and anti - lepton @xmath209 . we define the triple product : @xmath210 and the t - odd asymmetry : @xmath211 where here the cross section @xmath212 is the product of the chargino pair production cross section and the branching ratio for the decay into the sneutrino . this asymmetry is sensitive both to the phases in the chargino mass matrix , and to absorptive contributions , which do not contribute to cp violation . to eliminate the absorptive effects , we study the cp asymmetry : @xmath213 with @xmath214 the asymmetry for the cp conjugated process @xmath215 . in order to calculate the cross section for the combined process of chargino production and decay of @xmath216 , we use the spin density matrix formalism for the production and decays of charginos as in @xcite . the amplitude for the whole process is : @xmath217 where @xmath218 is the helicity amplitude for the production , @xmath219 the helicity amplitude for the decay , and @xmath220 is the chargino propagator . in the above formulas @xmath221 is the helicity of the decaying chargino , and @xmath222 its four - momentum squared . the total differential cross section is defined as : @xmath223 where the index @xmath224 runs over the charginos , sneutrino and emitted lepton momenta and @xmath225 is the energy of the incoming beam . in the definition of the asymmetry : @xmath226 the only cp sensitive contribution comes from the production and decay amplitudes corresponding to @xmath227 polarization perpendicular to the production plane . in the notation of @xcite , the production amplitudes are : @xmath228 with @xmath229 \nonumber \\ & & \times e_b^2m_{\chi_i^- } q \sin \theta \ ; { \cal i}m \left ( o^{\prime l}_{ij } o^{\prime \ast r}_{ij}\right ) , \nonumber \\ \sigma^2_p(z_lz_r)&= & \frac{g^4\cos 2 \theta_w}{\cos^4 \theta_w } \left [ c_rc_r^{\prime}(1+p_{e^-})(1-p_{e^+})-c_lc_l^{\prime}(1-p_{e^-})(1+p_{e^+ } ) \right ] \nonumber \\ & & \times e_b^2m_{\chi_i^- } q \sin \theta \ ; { \cal i}m \left ( d_{z_l}d_{z_r}^{\ast}o^{\prime l}_{ij } o^ { \ast r}_{ij}+ d^{\ast}_{z_l}d_{z_r}o^ { l}_{ij } o^ { \prime \ast r}_{ij}\right ) , \nonumber \\ \sigma^2_p(z_rz_r)&= & 2 \frac{g^4\cos^2 2 \theta_w}{\cos^4 \theta_w } |d_{z_r}|^2 \left [ c_r^{\prime 2}(1+p_{e^-})(1-p_{e^+})-c_l^{\prime 2}(1-p_{e^-})(1+p_{e^+ } ) \right ] \nonumber \\ & & \times e_b^2m_{\chi_i^- } q \sin \theta \ ; { \cal i}m \left ( o^ { l}_{ij } o^ { \ast r}_{ij}\right ) , \nonumber \\ \sigma^2_p({\tilde \nu}_lz_l)&= & \frac{g^4}{\cos^2 \theta_w } c_l^2(1-p_{e^-})(1+p_{e^+ } ) e_b^2m_{\chi_i^- } q \sin \theta \ ; { \cal i}m \left ( v_{i1}^{\ast}v_{j1 } o^{\prime r}_{ij } d_{z_l } d_{{\tilde \nu}_l}^{\ast } \right ) , \nonumber \\ \sigma^2_p({\tilde \nu}_r z_l)&= & \frac{g^4}{\cos^2 \theta_w } c_r^2(1+p_{e^-})(1-p_{e^+ } ) e_b^2m_{\chi_i^- } q \sin \theta \ ; { \cal i}m \left ( u_{i2}u_{j2}^{\ast } o^{\prime l}_{ij } d_{z_l } d_{{\tilde \nu}_r}^{\ast } \right ) , \nonumber \\ \sigma^2_p({\tilde \nu}_l z_r)&= & \frac{g^4 \cos 2 \theta_w}{\cos^2 \theta_w } c_l^{\prime 2}(1-p_{e^-})(1+p_{e^+ } ) e_b^2m_{\chi_i^- } q \sin \theta \ ; { \cal i}m \left ( v_{i1}^{\ast}v_{j1 } o^ { r}_{ij } d_{z_r } d_{{\tilde \nu}_l}^{\ast } \right ) , \nonumber \\ \sigma^2_p({\tilde \nu}_rz_r)&= & \frac{g^4 \cos 2 \theta_w}{\cos^2 \theta_w } c_r^{\prime 2}(1+p_{e^-})(1-p_{e^+ } ) e_b^2m_{\chi_i^- } q \sin \theta \ ; { \cal i}m \left ( u_{i2}u_{j2}^{\ast } o^ { l}_{ij } d_{z_r } d_{{\tilde \nu}_r}^{\ast } \right ) . \nonumber \\\end{aligned}\ ] ] in these expressions the angle @xmath230 is the scattering angle between the incoming electron beam and the @xmath231 chargino in the laboratory frame , @xmath232 is the beam energy , @xmath233 is the positron ( electron ) polarization , and @xmath234 . similarly , the chargino decay amplitudes are : @xmath235 and the conjugated decay process @xmath236 amplitudes have an overall @xmath237 sign . here @xmath238 is the spin polarization component perpendicular to the production plane and @xmath239 is the momentum of the antilepton . see @xcite for details . the production contribution independent of the chargino polarization vectors can be obtained from the expressions for the cross sections in section [ sec : production ] , in the same way as the polarization - independent part of the chargino decay can be obtained from the partial decay widths , eq . ( [ decays ] ) . for example : @xmath240{\tilde m}_{\chi_i^+ } \right ] , \nonumber \\ d(l^+ { \tilde \nu}_r)\hspace*{-0.2cm}&&=\frac{g^2}{2}\left [ \left ( |u_{i2}|^2-\frac { m_l^2}{2m_w^2 \cos^2\beta}|v_{i4}|^2\right ) ( { \tilde m}_{\chi_i^+}^2-m_{{\tilde \nu}_r}^2-m_l^2 ) \right . \nonumber \\ & & \left . - 2\sqrt{2 } \frac { m_l^2}{m_w \cos\beta } { \cal r}e[u_{i2}^{\ast } v_{i4}]{\tilde m}_{\chi_i^+ } \right ] . \end{aligned}\ ] ] with these expressions for the density matrices , the asymmetry becomes : @xmath241 where @xmath242 . using the narrow width approximation for the chargino propagator : @xmath243 the total cross section appearing in the denominator is obtained as @xmath244 and is polarization - independent . an interesting sign of lrsusy would be given by the case in which at least one of charginos could decay to an anti - lepton and a right - handed sneutrino . this might be an unlikely scenario , since @xmath245 is expected to be large . the splitting between the left and right sneutrinos is proportional to the dirac neutrino mass , and it is conceivable that this is in the mev region . if so , the right handed neutrino could be kinematically accessible for the decay of a heavier chargino . to achieve this , we consider a different mass scenario than in the previous section . we take the right - handed gaugino mass @xmath246 gev , larger than the left - handed one ( @xmath107 gev ) . under these circumstances , with the notation @xmath247 , the fourth chargino @xmath248 is mainly a right - handed gaugino and it is heavy enough to decay to an anti - lepton and a right - handed sneutrino ( taken to have mass of 500 gev ) . the third lightest state is now mainly higgsino and we will consider @xmath249 in conjunction with its production process from @xmath0 scattering . since @xmath250 vanishes for the production of same charginos , we will only consider @xmath251 and @xmath252 processes , even though some other channels ( with @xmath253 chargino final states ) are allowed for cross section consideration . 3.5 in 3.5 in -0.2 in 3.5 in 3.5 in -0.2 in in fig . [ fig : sigatoddl ] , we present , on the left , the total production - decay cross section @xmath254 ( @xmath255 or @xmath66 ) at @xmath256 gev in the cp - odd phases @xmath164-plane with electron - positron beam polarization @xmath257 , and , on the right , the corresponding asymmetry @xmath250 for the same chain with the same parameter set , defined as in the caption of the figure . the total cross section including the particular decay channel is @xmath258 fb without cp - odd phases and depends only weakly on the phases . the maximum enhancement is around @xmath191 . the cross section is approximately constant on the contours @xmath259 , and maximum enhancement occurs along such contours . there is a similar pattern in the asymmetry @xmath250 which varies in ( -3.5%,3.5% ) interval . cp - odd phases can give large asymmetry in the entire @xmath164 range . @xmath250 takes values of @xmath260 along @xmath261 or @xmath262 , and @xmath263 along @xmath264 . these values for both the cross section and the asymmetry are pretty much consistent with the mssm results , as expected . however , the process @xmath252 which is shown in fig . [ fig : sigatoddr ] , is peculiar to the left right supersymmetric model and measuring the t - odd asymmetry associated with it could be used to distinguish it from mssm . even though the total cross section is comparatively small with respect to the previous case , the asymmetry can reach @xmath265 . since the asymmetry is around @xmath266 without the phases , the experimental availability of an asymmetry measurement would allow us to bound the cp - odd phases in the chargino sector of the model . we have analysed the influence of the cp - violating phases in the chargino sector of the left - right supersymmetric model . our emphasis has been mostly on the presence of a right - handed wino , and less so on additional higgsinos , which arise from ( model - dependent ) ways to break the symmetry . to achieve this , we decoupled the additional higgsinos by retaining only those originating from the bidoublet higgs bosons , which are common to all left - right models . the added advantage of this method is that we are able to solve analytically for the chargino mass eigenvalues . from the most general set of phases in the mass matrix , two cp - odd phases arise ( as opposed to just one in mssm ) and their influence is the strongest on the lightest chargino ( mostly a mixture of left and right winos ) mass , where effects of up to 20% can be detected . the phases influence weakly the heaviest ( corresponding to the mssm higgsino ) mass , and not much the other two states . for the chargino production from @xmath0 pairs , a large cp - odd phase effect in the cross section is obtained for either same or different states production , at both @xmath267 and @xmath196 gev . the variation of the cross section has distinct features : it is enhanced or depressed for a single region of combined parameter space for same charginos , while for different charginos the enhancement can occur for several parameter regions . the forward - backward asymmetry @xmath185 ( 1.2% at @xmath267 gev and 12% @xmath268 gev without cp phases ) is a very sensitive probe of cp - odd angles , being enhanced to 50% for @xmath267 gev for @xmath269 production . this is even more pronounced for @xmath270 where the maximum forward - backward asymmetry can be almost four times its value without cp phases . a distinguishing sign of a left - right gaugino sector would be the decay of charginos into left and right , if kinematic restrictions allow it , scalar neutrino . we study these decays seperately , in conjunction with the chargino production from a polarized electron - positron beam , and investigate the dependence of the cross section for chargino production and decay , as well as the cp asymmetry in this decay , @xmath250 , as a function of the two cp - odd phases . though for this decay to occur , the right - handed gaugino mass must be larger than the right sneutrino mass , and neither should decouple from the energy spectrum of the lower lying charginos , we find this possibility very interesting for detecting a signal for left - right supersymmetry . while the cross section for production and decay is enhanced , but not very sensitive to the cp - phases , the t - odd asymmetry is very sensitive to the phases . if the decay into the right - handed sneutrino is open , the enhancement of the asymmetry could be four times the one obtained in mssm and would serve as a distinguishing signal for this model . in summary , additional cp - odd phases arise in lrsusy . their effect is strongest in forward - backward asymmetry of the production , and in the cp - asymmetry in polarized production , followed by the decay into a neutrino and slepton . both of these asymmetries show large enhancements with respect to their values in mssm . 99 j. r. ellis , s. ferrara and d. v. nanopoulos , phys . b * 114 * , 231 ( 1982 ) ; j. polchinski and m. b. wise , phys . b * 125 * , 393 ( 1983 ) ; m. dugan , b. grinstein and l. j. hall , nucl . b * 255 * , 413 ( 1985 ) . a. i. sanda , phys . d * 32 * , 2992 ( 1985 ) ; y. kizukuri and n. oshimo , phys . rev . d * 46 * , 3025 ( 1992 ) ; y. kizukuri and n. oshimo , phys . rev . d * 45 * , 1806 ( 1992 ) . r. n. mohapatra and a. rasin , phys . * 76 * , 3490 ( 1996 ) [ arxiv : hep - ph/9511391 ] ; r. n. mohapatra and a. rasin , phys . d * 54 * , 5835 ( 1996 ) [ arxiv : hep - ph/9604445 ] ; r. kuchimanchi , phys . * 76 * , 3486 ( 1996 ) [ arxiv : hep - ph/9511376 ] . m. frank , phys . d * 59 * , 055006 ( 1999 ) ; m. frank , j. phys . g * 25 * , 1813 ( 1999 ) . t. ibrahim and p. nath , phys . d * 61 * , 093004 ( 2000 ) [ arxiv : hep - ph/9910553 ] ; t. ibrahim and p. nath , phys . d * 57 * , 478 ( 1998 ) [ erratum - ibid . d * 58 * , 019901 ( 1998 ) , errat , d * 60 * , 079903 ( 1999 ) , errat , d * 60 * , 119901 ( 1999 ) ] [ arxiv : hep - ph/9708456 ] . r. m. francis , m. frank and c. s. kalman , phys . d * 43 * , 2369 ( 1991 ) . z. chacko and r. n. mohapatra , phys . d * 58 * , 015003 ( 1998 ) [ arxiv : hep - ph/9712359 ] ; b. dutta and r. n. mohapatra , phys . d * 59 * , 015018 ( 1999 ) [ arxiv : hep - ph/9804277 ] . m. e. pospelov , phys . d * 56 * , 259 ( 1997 ) [ arxiv : hep - ph/9611422 ] ; g. couture , m. frank , h. konig and m. pospelov , eur . j. c * 7 * , 135 ( 1999 ) [ arxiv : hep - ph/9701299 ] . m. frank and i. turan , phys . d * 72 * , 035008 ( 2005 ) [ arxiv : hep - ph/0506197 ] . m. frank , phys . d * 62 * , 053004 ( 2000 ) ; g. barenboim , k. huitu , j. maalampi and m. raidal , phys . b * 394 * , 132 ( 1997 ) [ arxiv : hep - ph/9611362 ] ; k. huitu , j. maalampi , a. pietila and m. raidal , nucl . b * 487 * , 27 ( 1997 ) [ arxiv : hep - ph/9606311 ] . p. ball , j. m. frere and j. matias , nucl . b * 572 * , 3 ( 2000 ) [ arxiv : hep - ph/9910211 ] ; p. ball and r. fleischer , phys . b * 475 * , 111 ( 2000 ) [ arxiv : hep - ph/9912319 ] . g. j. gounaris , c. le mouel and p. i. porfyriadis , phys . d * 65 * , 035002 ( 2002 ) [ arxiv : hep - ph/0107249 ] ; g. j. gounaris and c. le mouel , phys . rev . d * 66 * , 055007 ( 2002 ) [ arxiv : hep - ph/0204152 ] . m. frank and h. n. saif , z. phys . c * 69 * , 673 ( 1996 ) ; m. frank , c. s. kalman and h. n. saif , z. phys . c * 59 * , 655 ( 1993 ) . a. bartl , h. fraas , o. kittel and w. majerotto , phys . b * 598 * , 76 ( 2004 ) [ arxiv : hep - ph/0406309 ] ; o. kittel , a. bartl , h. fraas and w. majerotto , phys . d * 70 * , 115005 ( 2004 ) [ arxiv : hep - ph/0410054 ] . s. y. choi , j. s. shim , h. s. song and w. y. song , phys . b * 449 * , 207 ( 1999 ) ; s. y. choi , a. djouadi , h. s. song and p. m. zerwas , eur . phys . j. c * 8 * , 669 ( 1999 ) [ arxiv : hep - ph/9812236 ] . s. y. choi , j. kalinowski , g. moortgat - pick and p. m. zerwas , eur . j. c * 22 * , 563 ( 2001 ) [ addendum - ibid . c * 23 * , 769 ( 2002 ) ] [ arxiv : hep - ph/0108117 ] . j. f. gunion and h. e. haber , phys . d * 37 * , 2515 ( 1988 ) . o. kittel , a. bartl , h. fraas and w. majerotto , phys . d * 70 * , 115005 ( 2004 ) [ arxiv : hep - ph/0410054 ] ; a. bartl , h. fraas , o. kittel and w. majerotto , phys . b * 598 * , 76 ( 2004 ) [ arxiv : hep - ph/0406309 ] ; a. bartl , k. hohenwarter - sodek , t. kernreiter and h. rud , eur . j. c * 36 * , 515 ( 2004 ) [ arxiv : hep - ph/0403265 ] ; g. moortgat - pick , h. fraas , a. bartl and w. majerotto , eur . j. c * 7 * , 113 ( 1999 ) [ arxiv : hep - ph/9804306 ] .

with the discovery of the higgs boson at the lhc , an important step towards a complete identification of the structure of the mechanism of mass generation in the electroweak theory has been taken . however , it is clear that one needs suitable extensions of the current version of the standard model ( sm ) in order to solve some important theoretical and phenomenological issues which do not find an answer within the same model . these include , for instance , the absence of a mechanism which could account for the masses of the three generations of neutrinos , or the origin of its chiral nature as well as of the gauge charge assignments of the flavor sector . at theoretical level , the latter are constrained only by the gauge symmetry together with the mechanism of cancelation of the gauge and of the gravitational anomalies , which provide significant restrictions on the possible interactions . + analogously , the issue concerning the instability of the scalar potential of the sm under the perturbative renormalization group ( rg ) evolution up to the planck scale , which is another critical aspect of the same model , noticed long ago , has been exploited in the past @xcite to constrain the higgs mass , and it has been instrumental for later higgs searches , but it has also been viewed as one of its unappealing features . in this case , the basic motivation hinged on the belief that any approach which tries to connect the sm directly with the planck scale is probably untenable , as epitomised by the gauge - hierarchy problem . + with the recent lhc activity and the significant exclusion limits on the parameter space of possible supersymmetric models which have emerged from it , recurrent attempts to connect the sm with larger physical scales , particularly the planck scale , have resurfaced . in particular , the assumption of a _ big desert _ scenario , without any intermediate new physics , have forced many to reconsider the issue of its vacuum stability . in this framework , finer inspections of the perturbative rg evolution of the sm scalar potential have become quite popular @xcite . it has been pointed out that although the sm scalar potential develops a new deeper minimum , the lifetime of the unstable electroweak vacuum is found to be much larger than the age of the universe . this feature , called metastability , taken at face value , suggests that the extrapolation of the sm up to the planck scale does not necessarily require the introduction of new physics . these issues have been and are widely debated also in cosmology @xcite , for a potential both at zero and at finite temperature . the role of these analyses has been to clarify , in view of the measured value of the higgs mass , if the sm stands solidly in a quantum field theory and cosmological context and , specifically , whether additional contributions to the running would resolve the issue related to its behavior . the study of the shape of the higgs effective potential @xmath2 , for large values of the higgs field , larger than the higgs vev @xmath3 @xmath4 , is critically important for coming to reasonable conclusions concerning the stability of the sm . if another minimum is found at @xmath5 , then the requirement of stability imposes that @xmath6 . in the opposite situation , the potential could be either completely unstable or metastable . + as mentioned above , if @xmath7 , a metastable condition could be accepted as far as the tunnelling time @xmath8 for the transition from the false to the true vacuum of the higgs field is larger than the age of our universe @xmath9 . in the opposite case , with @xmath10 , one encounters a complete instability , which must be excluded . the rg analyses points to a metastable behavior which runs quite close to the region of instability in the parameter space @xcite . + one of the unappealing issues of the sm potential lays in its critical dependence on the top quark @xmath11 and higgs @xmath12 masses . in fact , the effect of the fermions , and of the yukawa of the top quark in particular , on the rg evolution , is to drive the higgs quartic couplings @xmath13 towards negative values , upsetting the vacuum stability already at @xmath14 gev , but allowing for a long - lived vacuum . for this reason , in the pure sm scenario , precise measurements of the value of the top mass and the resolution of the ambiguities in the use of different renormalization schemes for its evaluation are important in order to lift the controversy @xcite @xcite . + this leaves open the possibility that new physics , around the electroweak scale and above , but below the planck scale @xmath15 , will drastically change this scenario , reducing such sensitivity or eliminating it all together @xcite . in these analyses , the magnitude and the sign of the different contributions to the beta function @xmath16 of the higgs quartic coupling is crucial for a correct prediction . + obviously , even simple modifications of the gauge structure of the sm , with the inclusion of one extra @xmath17 symmetry in the neutral currents sector allow to envision completely different scenarios where the issue of stability or of metastability finds alternative solutions . also in this case , though , the large yukawas of other heavy fermions play a similar role as the top quark , hinting that a complete solution of this issue involves the entire heavy flavor sector @xcite . + the goal of our analysis , in this work , is to illustrate how the requirement of stability under the rg evolution of the parameters of the model is linked to the flavor sector of the same theory , as soon as we enlarge the scalar potential even in a minimal way . we will investigate this point rather thoroughly , by addressing the case of an extra @xmath0 symmetry , where @xmath18 and @xmath19 are the baryon and lepton numbers respectively , extending our previous one - loop analysis @xcite to two - loop level , and detailing many features which we have not discussed before . the model invokes a single extra higgs scalar ( the heavier higgs ) in the breaking of the extra abelian symmetry , beside the ordinary higgs doublet of the sm . several previous phenomenological analyses have addressed some of the salient features of this model , which is a front - runner in the hunting for a possible extra @xmath20 at the lhc @xcite . b - l is an anomaly - free and conserved symmetry of the sm , which plays an important role in grand unified theory ( gut s ) . + the current study is also characterized by the obvious inclusion in the rg running of the one - loop matching conditions at the electroweak scale , formerly considered by us only at tree - level . as we are going to show , these are essential in order to determine correctly the regions of stability of the potential . our equations also keep into account the effects due to the kinetic mixing of @xmath0 with the hypercharge . we have assumed for simplicity that the coupling @xmath21 which parameterizes the mixing is vanishing at the electroweak scale , and gets only generated radiatively in the upward running . for this reason , our analysis is here focused on a _ pure _ b - l extra gauge structure at the electroweak scale . however , using our results , one can easily address different scenarios , under the assumption of a diagonal b - l and hypercharge ( @xmath22 ) symmetries at the gut scale , with a kinetic mixing induced by a reverse evolution from the gut scale towards the electroweak one . + moreover , our work invokes a suitable mechanism of mass generation of the light neutrinos of the sm . as in a previous work @xcite , we consider a scenario based on a type - i seesaw realized with three heavy right - handed ( rh ) neutrinos which can be easily generalized to other mechanisms , such as to a type - ii , -iii , or to an inverse seesaw . + concerning the specific charge assignments of the rh neutrinos , we rely on the general conditions of anomaly cancellation . the solutions of the corresponding equations depend significantly on the number of heavy rh neutrinos ( @xmath23 ) selected in the model , which in our case is one per generation , and we have opted for a symmetric charge assignment of the @xmath24 under @xmath0 . obviously , other solutions can be considered as well . + in any case , the pattern that emerges from this analysis is quite insensitive to the gauge charges , while remaining quite sensitive to all the other parameters which characterize the potential and the fermion sector . as we are going to point out , these can be traded for the mass of the extra @xmath20 , @xmath25 , and its gauge coupling @xmath26 , the mixing angle @xmath27 of the two higgses , the mass of the heavy higgs @xmath28 and the masses of the rh neutrinos . the gauge symmetric extension that we consider has the same content of the sm augmented by a single extra @xmath29 factor , i.e. @xmath30 , with @xmath29 later identified with @xmath0 . the higgs sector includes one extra complex scalar @xmath31 , beside the higgs , @xmath32 , which will be discussed thoroughly in the next sections . + in this section we briefly review the conditions imposed by the cancellation of the anomalies , which constrain significantly the @xmath29 charge assignments of the model . as already discussed in previous analysis @xcite , these conditions can be combined with extra requirements on the mass generation in the fermion sector , i.e. on the type of operators chosen in order to provide masses for the neutrinos . in general , these extra requirements may involve operators of mass dimension larger than four . in this context , the choice , for instance , of a type - i seesaw for the neutrino masses with three sm singlet right - handed neutrinos is , for sure , quite economical and carries the advantage of preserving the renormalizability of the model . + we assign generic charges @xmath33 for the left - handed ( lh ) quark and lepton doublets @xmath34 , and the charges @xmath35 for the right - handed ( rh ) @xmath36 , @xmath37 and @xmath38 quarks and leptons . the charges of the rh neutrinos @xmath39 are denoted as @xmath40 . notice that we will omit the indices from the definition of the charges since the conditions for anomaly cancellation will be universal , the same for each fermion generation . finally , the charges of the two scalars @xmath32 and @xmath31 are denoted as @xmath41 and @xmath42 respectively . + we have the following cancellation conditions for the non - abelian @xmath43 and @xmath44 anomalies u(1) su(2 ) su(2 ) : & & 3 z_q + z_l=0 , + u(1) su(3 ) su(3 ) : & & 2 z_q - z_u - z_d=0 , [ y33 ] which fix @xmath45 and @xmath46 in terms of @xmath47 and @xmath48 . two other conditions are u(1) u(1)_y u(1)_y : & & z_q -8 z_u -2 z_d + 3 z_l - 6 z_e=0 , + u(1) u(1) u(1)_y : & & z_q^2 - 2 z_u^2 + z_d^2 - z_l^2 + z_e^2=0 , [ zyy ] for the mixed @xmath17 anomalies . from the first of the two requirements in eq . ( [ zyy ] ) one can immediately extract the relation @xmath49 , while the second condition of the same equation is automatically satisfied using the solutions found from eq . ( [ y33 ] ) . we have summarized in tab.[table1 ] the spectrum of the theory that we will be using in our phenomenological study . + the constraints on the charges @xmath40 of the sm singlet fermions are obtained from the conditions of cancellation of the @xmath29 cubic anomalies , together with those from the gravitational anomaly . the latter involve the @xmath29 current and two gravitons @xmath50 ( i.e. two insertions of the stress - energy tensor of the sm ) . in the general case with @xmath51 singlet fermions one has , respectively u(1) u(1) u(1) : & & _ k=1^n z_k^3 = 3= 3 ( z_u - 4 z_q)^3 , for the cubic anomaly and u(1) g g : & & _ k=1^n z_k = 3= 3 ( z_u - 4 z_q ) , [ zgg ] for the gravitational anomaly , where we have used the constraints extracted from eqs . ( [ y33 ] ) and ( [ zyy ] ) . finally , combining together the two conditions in eq . ( [ zgg ] ) , one obtains the cubic relation ( _ k=1^n z_k)^3 = 9 _ k=1^n z_k^3 . [ cubic ] for instance , the constraints in eq . ( [ zgg ] ) imply , for @xmath52 and @xmath53 , the condition @xmath54 , which brings either to the trivial solution or to a solution which is @xmath22-sequential . in this latter case the @xmath29 charge assignment is proportional to that of @xmath55 . in the @xmath56 case one obtains , instead , @xmath57 . + an interesting solution which is non - sequential in @xmath22 is found for @xmath58 . in this case each of the @xmath39 is assigned to a separate generation . for example , the choice @xmath59 allows to find the simple solution @xmath60 . the cancellation instead becomes inter - generational by choosing , for instance , the less restrictive condition @xmath61 . in this case ( [ zgg ] ) gives @xmath62 , with @xmath63 , as one can easily verify . + the cancellation of the gravitational anomalies , can be imposed , in general , at inter - generational level . in the present analysis we will opt , however , for a completely symmetric ( family independent ) assignment of the rh neutrinos charges @xmath64 , which allow to reduce the corresponding parameter space . with this choice , the @xmath65 constraint from the gravitational anomaly reduces to a single equation for just one charge . on the other hand , the cancellation of the analogue gravitational anomalies in the sm , obtained from the @xmath66 sector , is a natural consequence of the hypercharge assignments of the same model , and does not generate any additional constraint . .charge assignment of fermions and scalars in the @xmath29 sm extension . [ table1 ] [ cols="^,^,^,^,^",options="header " , ] ) , = , whose expressions , determined by the minimization conditions , are v^2 = , v^2 = . [ vevs ] after spontaneous symmetry breaking , the mixing between the two scalar fields can be removed by an orthogonal transformation which rotates @xmath32 and @xmath31 into the two mass eigenstates @xmath67 and @xmath68 ( c h_1 + h_2 ) = ( cc & - + & ) ( c h + ) with @xmath69 . the masses of the scalar eigenstates are m_h_1,2 ^ 2 = _ 1 v^2 + _ 2 v^2 , and one can easily derive the relations [ lambdas ] _ 1 & = & ( 1 + 2 ) + ( 1 - 2 ) , + _ 2 & = & ( 1 - 2 ) + ( 1 + 2 ) , + _ 3 & = & 2 ( ) , which can be used to set the initial conditions on the scalar couplings through the physical masses @xmath70 , the two vevs @xmath71 and the mixing angle @xmath27 . from eq . ( [ lambdas ] ) one can immediately derive the relation 2 = , [ theta ] which allows to express the mixing of the two scalars in terms of the three independent couplings of the potential and the two vevs @xmath71 . notice that while the couplings @xmath72 in eq . ( [ lambdas ] ) can be taken as the independent parameters of the potential after imposing the vacuum conditions eq . ( [ vevs ] ) , eq . ( [ theta ] ) is a trivial identity , derived from the former . after spontaneous symmetry breaking , one obtains for the @xmath73 mass the usual sm expression , @xmath74 , while the masses of the neutral gauge bosons are obtained from the lagrangian ( g_2 w^3_- g b_- |g b_)^2 + ( g z_b_)^2 , where we have assumed a nonzero kinetic mixing of the two @xmath75 gauge fields , as expressed in eq . ( [ covder ] ) . we have also defined the new coupling @xmath76 , which involves the kinetic mixing @xmath21 and the @xmath29 gauge coupling @xmath26 . @xmath41 and @xmath42 are the @xmath29 charges of the two scalar fields . in the case of a b - l symmetry , since @xmath77 , then the new coupling @xmath78 coincides with the coupling of the kinetic mixing , @xmath79 . + the diagonalization of the mass matrix extracted from the previous expression gives the relations between mass and interaction eigenstates ( c b^ + w_3^ + b^ ) = ( ccc _ w & - _ w & _ w + _ w & _ w & - _ w + 0 & & ) ( c a^ + z^ + z^ ) where @xmath80 is the usual weinberg angle of the sm , and @xmath81 is a new mixing angle , with @xmath82 defined as [ thetaprime ] 2 = . the relation above can be re - expressed in terms of the masses of the two gauge bosons . the masses of the @xmath83 and @xmath20 gauge bosons are given by [ zzpmasses ] m_z , z = ^ . a bound on the mixing angle @xmath81 has been obtained from the lep experiment @xcite which constrains @xmath81 to small values , namely @xmath84 . in this case the expressions for the gauge boson masses simplify considerably to [ zzpmassessmallth ] m_z , m_z , and [ thetaprimemasses ] |g can be used in order to quickly grasp the dependence of the mixing angle @xmath81 in terms of @xmath85 . + notice that the kinetic mixing between the two @xmath17 gauge fields can be taken to vanish , without loss of generality , at any particular scale , but it can be reintroduced by the rg evolution @xcite . as already mentioned above , in the present analysis we have chosen a vanishing @xmath86 at the electroweak scale , which is one of the boundary conditions of our rg study . also notice that , for @xmath87 , the tree - level mixing between the @xmath20 and the sm gauge bosons is absent only for @xmath88 , which occurs when the @xmath29 charges are those of @xmath89 . we conclude this section with few comments about the physical parameters of the model , which we will be using in our numerical study in the sections below . the original 5 parameters of the potential @xmath90 , after imposing the vacuum conditions eq . ( [ vevs ] ) , can be replaced with the two physical masses of the higgses , the two vevs @xmath71 , and the scalar mixing angle @xmath27 . after eliminating @xmath3 and @xmath91 ( respectively identified with the sm vev and higgs mass , @xmath92 gev , @xmath93 gev ) using the current lhc data , we are left with @xmath94 , the mass of the extra higgs , @xmath27 and @xmath95 as extra parameters . we can re - express @xmath95 in terms of the physical mass of the @xmath20 using eq . ( [ zzpmassessmallth ] ) , which involves as relevant unknown parameter the gauge coupling @xmath26 and the kinetic mixing @xmath21 . henceforth , the potential can be conveniently investigated in terms of the physical masses @xmath25 and @xmath94 , and of the mixing angle @xmath27 . for a given mass of the extra @xmath20 , one can use the two couplings @xmath96 as additional parameters , whose values are taken as boundary conditions at the electroweak scale . a typical study of the stability of the potential would then involve @xmath97 , together with @xmath96 . the inclusion of the rh neutrinos renders this study more complex , since the evolution will couple the previous parameters to the mass of the rh neutrino @xmath98 . as we are going to see , the requirement of stability of the potential through the running of its rg equations , will in general select specific regions of the @xmath99 plane , for assigned values of the couplings , @xmath25 and @xmath27 . + as we will discuss in the sections below , the variables @xmath100 , @xmath27 and @xmath94 will have to satisfy some bounds from lep ( eqs . ( [ stu1 ] ) and ( [ stubound ] ) respectively ) while @xmath96 will be constrained by the requirements of perturbativity along the entire rg flow ( eq . ( [ pertconds ] ) ) . + one final comment concerns the mass of the heavy rh neutrino , which is also directly related to the vev @xmath95 as for @xmath25 , as shown in eqs . ( [ yuk],[typeiseesawmasses1 ] ) . therefore , an independent variation of @xmath98 respect to @xmath25 , causes @xmath101 , the yukawa of the rh neutrino , to vary . for a given mass of the extra @xmath20 , with fixed couplings @xmath96 at the electroweak scale , the variation of @xmath98 is entirely accounted for by a variation of @xmath101 . we are now in the position to start a renormalization group analysis using a basis of diagonal kinetic terms and with eq . ( [ nondiagcovderivative ] ) as the minimal form of the covariant derivative . consequently , the contribution coming from the abelian sector to the couplings include @xmath103 , @xmath26 and @xmath21 as given by eq . ( [ orthogonal ] ) , the latter accounting for the mixing . we will simplify the structure of the sm yukawa sector by retaining only the contribution coming from the top quark ( @xmath104 ) , thereby omitting the yukawas of the 3 light neutrinos . on the other hand , we have taken the yukawas of the heavy neutrinos to be universal ( generation independent ) @xmath105 . + we present , in this section , the one - loop @xmath102-functions for the dimensionless parameters for a generic charge assignment . the corresponding two - loop corrections can be found in appendix [ betafuncs ] . + we start from the rotated gauge coupling sector which , as already said , contain no trace of @xmath106 , if the @xmath102-functions undergo the same rotation given in eq . ( [ explicitrot ] ) . at one - loop order , the evolution of the gauge couplings is defined by the @xmath102 functions [ abelianbeta ] ^(1)_g & = & , ^(1)_g_2 = -,^(1)_g_3 = -7 g_3 ^ 3 , + ^(1)_g & = & g^2 ( + ) + + g^3 ( -44 z_q z_u+134 z_q^2 + 18 z_u^2 ) , coupled with the @xmath102-function accounting for the mixing between the two @xmath17 factors [ mixbeta ] ^(1 ) _ & = & ( -44 g^2 z_q z_u+134 g^2 z_q^2 + 18 g^2 z_u^2)+^2 ( + ) + & + & + g^2 g z_q+ g^2 g z_u . it is clear from eq . ( [ mixbeta ] ) that even imposing a zero starting value for @xmath21 , an abelian mixing can be radiatively generated by the evolution . + the yukawa sector is described by the evolution of just two real terms ^(1)_y_t & = & y_t ( -g z_q-4 g z_u----8 g_3 ^ 2 - 3 g^2 z_q^2 - 3 g^2 z_u^2)+ + ^(1)_y_n & = & y_n ( 48 g^2 z_q z_u-96 g^2 z_q^2 - 6 g^2 z_u^2)+10 y_n^3 , and together with the evolution of the scalar quartic parameters [ betal1 ] ^(1)__1 & = & _ 1 ( 12 g z_q-12 g z_u-3 ^2 - 3 g^2 - 9 g_2 ^ 2 + 24 g^2 z_q z_u-12 g^2 z_q^2 - 12 g^2 z_u^2 + 12 y_t^2 ) + & & -3 g^2 g z_q+3 g^2 g z_u+ g^2 ^2 - 36 g^3 z_q z_u^2 + 36 g^3 z_q^2 z_u-12 g^3 z_q^3 + 12 g^3 z_u^3 + & & -18 g^2 ^2 z_q z_u+9 g^2 ^2 z_q^2 + 9 g^2 ^2 z_u^2 - 3 g ^3 z_q-3 g_2 ^ 2 g z_q+3 g ^3 z_u+3 g_2 ^ 2 g z_u + & & + + g_2 ^ 2 ^2 + -6 g^2 g^2 z_q z_u+3 g^2 g^2 z_q^2 + 3 g^2 g^2 z_u^2 + g_2 ^ 2 g^2 - 6 g_2 ^ 2 g^2 z_q z_u + & & + 3 g_2 ^ 2 g^2 z_q^2 + 3 g_2 ^ 2 g^2 z_u^2 + -24 g^4 z_q z_u^3 + 36 g^4 z_q^2 z_u^2 - 24 g^4 z_q^3 z_u+6 g^4 z_q^4 + & & + 6 g^4 z_u^4 + 24 _ 1 ^ 2+_3 ^ 2 - 6 y_t^4 , [ betal2 ] ^(1)__2 & & = -1536 g^4 z_q z_u^3 + 9216 g^4 z_q^2 z_u^2 - 24576 g^4 z_q^3 z_u+24576 g^4 z_q^4 + 96 g^4 z_u^4 + & & + _ 2 ( 384 g^2 z_q z_u-768 g^2 z_q^2 - 48 g^2 z_u^2 + 24 y_n^2)+20 _ 2 ^ 2 + 2 _ 3 ^ 2 - 48 y_n^4 , ^(1)__3 & & = _ 3 ( 6 g z_q-6 g z_u---+204 g^2 z_q z_u-390 g^2 z_q^2 - 30 g^2 z_u^2 . + 12 _ 1 + 8 _ 2 + 12 y_n^2 + 6 y_t^2)-432 g^3 z_q z_u^2 + 1152 g^3 z_q^2 z_u-768 g^3 z_q^3 + 48 g^3 z_u^3 + & & -96 g^2 ^2 z_q z_u+192 g^2 ^2 z_q^2 + 12 g^2 ^2 z_u^2 - 480 g^4 z_q z_u^3 + 1584 g^4 z_q^2 z_u^2 + & & -1920 g^4 z_q^3 z_u+768 g^4 z_q^4 + 48 g^4 z_u^4 + 4 _ 3 ^ 2 , provide a closed system of differential equations to evolve the scalar potential towards any final scale . before coming to a discussion of the main phenomenological implications of our analysis , we pause for a moment , discussing a point which is essential in order to secure the consistency of the evolution , which concerns the determination of the initial ( matching ) conditions . + we recall that the rge s employed in this work have been computed in the @xmath107 renormalization framework at two - loop order . these equations must be supplemented with suitable one - loop boundary conditions defined in the same scheme . these consist of @xmath107 renormalized couplings and masses evaluated at a given energy scale which correspond to the starting scale of the rg running . + in general , the initial conditions can be unknown free parameters introduced by the specific model which is under investigation , and are directly associated to some measured observables . in order to determine the latter , the @xmath107 parameters at the starting scale must be related to these physical observables . this task can be accomplished in two different ways : 1 ) one can adopt the @xmath107 renormalization from the very beginning and obtain the needed @xmath107 parameters directly from a set of measured observables or 2 ) use a scheme , as the on - shell ( @xmath108 ) one largely used in the electroweak theory , in which the renormalized parameters are expressed in terms of the physical quantities , the pole masses and the fermi constant , and then translate the on - shell parameters to the corresponding @xmath107 expressions through appropriate matching conditions . in this work we adopt the second strategy which is quite common in the literature on the perturbative corrections in the sm . + the matching conditions are easily extracted from the obvious relation [ match ] _ 0 = _ + _ = _ ( ) + _ , where @xmath109 , @xmath110 and @xmath111 denote , respectively , the bare , the on - shell and the @xmath107 expressions of a generic parameter @xmath112 . from eq . ( [ match ] ) one can extract a @xmath107 parameter in terms of its on - shell version obtaining , at one - loop order , [ ostoms ] _ = _ + _ - _ = _ + _ |_finite where the last expression is simply a consequence of the definition of the @xmath107 renormalization scheme , in which the counterterms only subtract the uv singular parts . it is clear from eq . ( [ ostoms ] ) that the matching conditions between the os and the @xmath107 schemes are defined from the finite part of the os counterterm . notice that , at tree level , the @xmath107 parameters coincide with their os version . + in our analysis we have adopted a mixed renormalization procedure in which the known sm parameters are renormalized in the on - shell scheme , while the @xmath107 is used for the additional couplings and masses introduced by the extended gauge and neutrino sectors and for the vacuum expectation value @xmath95 of the extra scalar . on the other hand , for the sake of simplicity , all the remaining parameters in the scalar sector , containing both the higgs - like quartic coupling and the two new quartic interactions , are renormalized in the on - shell scheme . such a mixed scheme is not unusual in qfts as it is already employed , for instance , in the sm , and in particular for the computation of the strong corrections to electroweak observables . + as shown in @xcite , one of the interesting features of this hybrid renormalization scheme , besides its simplicity , is that the appelquist - carrazzone decoupling theorem is explicit manifest . indeed , the sm limit of measurable quantities is straightforwardly obtained for @xmath113 , with @xmath95 defined in the @xmath107 scheme . this has been the main motivation for our renormalization setup . the sm - like parameters which enter into this rg study are the quartic coupling @xmath114 of the @xmath43 doublet @xmath32 , the yukawa of the top quark @xmath115 and the gauge coupling constants @xmath103 , @xmath116 and @xmath117 . these are computed in terms of the pole masses of the higgs @xmath91 , of the top @xmath118 , of the weak gauge bosons @xmath119 and @xmath120 , and of the fermi constant @xmath121 . all these quantities are then translated in the @xmath107 scheme using eq . ( [ ostoms ] ) . notice that for the @xmath44 strong coupling constant @xmath117 there is no need to introduce matching conditions because it is directly extracted in the @xmath107 renormalization framework as @xmath122 . + on the other hand , the other unknown dimensionless parameters introduced by the @xmath29 extension , with the only exception of the those of the scalar potential , are directly employed in the @xmath107 scheme from the very beginning ( or , equivalently , are matched to their counterparts at tree level ) . these are the abelian gauge coupling constants , @xmath26 and @xmath86 , and the yukawa of the right - handed neutrinos @xmath101 . instead , the quartic couplings @xmath123 and @xmath124 are matched at one - loop from the on - shell physical mass of the heavy scalar @xmath94 and from the the mixing angle @xmath27 . + one of the most important sm parameters needed in the determination of the initial conditions of our rg study is the fermi constant @xmath121 . using its definition in the effective fermi theory = ( 1 + r_0 ) = ( 1 + r_0 ) , we obtain the counterterm of the vev @xmath3 in the on - shell scheme @xmath125 . we recall that @xmath121 is extracted from the muon lifetime , computed in the fermi theory augmented by qed corrections . as a consequence , the computation of the @xmath126 electroweak corrections to the @xmath127 decay requires the subtraction of the pure qed contributions . at one - loop order @xmath126 can be decomposed as r_0 = v - + b + e , where @xmath128 and @xmath18 denote vertex and box corrections , @xmath129 is the @xmath130 boson self - energy evaluated at zero momentum and @xmath131 corresponds to the wave - functions contributions . all of them are computed at zero external momenta and are affected by sm - like corrections as well as new - physics effects . notice also that we have chosen a renormalization prescription in which the tadpoles are included in the perturbative expansion . this property has the advantage to provide a gauge - independent definition of the mass counterterms and of @xmath126 . nevertheless , the dimensionless parameters appearing in the lagrangian , which are the interesting ones for our analysis , are not affected by this choice and the results are independent of the tadpole corrections . in fig.[fig.pertgf ] we show some of the one - loop perturbative contributions to the @xmath127 decay which are needed in the computation of the matching condition of @xmath121 . the corrections proportional to the neutrino mixing angle @xmath132 , as the ones depicted in fig.[fig.pertgf2 ] , have been neglected due to the smallness of the ratio between the dirac and the majorana masses . + + the counterterms of the top yukawa and of the @xmath43 gauge couplings , needed in the matching procedure , are obtained exploiting the relations m_t = y_t , m_w^2 = g_2 ^ 2 , which preserve their sm - like form , and read as y_t = y_t ( - ) , g_2 = g_2 ( - ) . the determination of counterterms of the abelian gauge coupling @xmath103 deserves a special attention because its defining relation gets modified in the @xmath29 extension with respect to the sm case as one can see from eq . ( [ zzpmasses ] ) . nevertheless , the departure from the sm expression , parametrized by the mixing angle @xmath81 , is very small and eq . ( [ zzpmassessmallth ] ) can be safely employed leading to g _ = g _ ( - ) . in the previous equation @xmath133 is obtained from @xmath126 as explained above , while the top , the @xmath130 and the @xmath83 boson mass counterterms in the on - shell renormalization scheme are computed from the corresponding self - energies m_t = re _ t(p /= m_t ) , m_w^2 = re _ ww(p^2 = m_w^2 ) , m_z^2 = re _ zz(p^2 = m_z^2 ) . the remaining two abelian coupling constants , @xmath26 and @xmath86 , could be related , in principle , to the pole mass of the @xmath20 and to the os expression of the mixing angle @xmath81 through eq . ( [ thetaprime ] ) and ( [ zzpmasses ] ) . nevertheless , being these two free parameters of the @xmath29 extension , the use of the matching conditions in this case is not mandatory and we can work directly with their @xmath107 expression . therefore , in all of the following analyses , @xmath26 and @xmath86 should always be understood as computed within the @xmath107 . + the major differences in the matching relations , with respect to the sm case , are found in the scalar sector due to the presence of new quartic interactions . using the defining equations in eq . ( [ lambdas ] ) we obtain _ 1 & = & ( 1 + 2 _ ) ( - 2 - 2 _ _ ) + & + & ( 1- 2 _ ) ( - 2 + 2 _ _ ) , + _ 2 & = & ( 1- 2 _ ) ( + 2 _ _ ) + & + & ( 1 + 2 _ ) ( - 2 _ _ ) , + _ 3 & = & _ 3 ( - + 2 _ 2 _ ) , where m_h_i^2 = re _ h_i(p^2 = m_h_i^2 ) , _ = . notice that , in the last of the previous equations , the mixed scalar self - energy @xmath134 has been evaluated at @xmath135 . this choice enforces the absence of mixing between the two tree - level mass eigenstates @xmath136 also at one - loop level and at a particular scale , given by @xmath91 . the allowed parameter space of the model can be strongly restricted if we require the vacuum stability and the perturbativity up to a given scale @xmath137 . this may be chosen to coincide with the gut , planck or any other scale where new physics scenarios are expected to appear . obviously , these conditions are often not sufficient to select a bounded region of the same space , and new assumptions must be added , for instance in the form of suitable benchmark points where to sharpen the physics predictions . + significant bounds are derived mainly from previous lep - ii analysis @xcite , which limit the gauge mixing angle @xmath84 , and force us to consider only the sector for which the condition m_z/g 7 , [ stu1 ] holds @xcite . besides , the addition of a new heavy scalar in the spectrum modifies the coupling of the light higgs to the sm particles by a factor of @xmath138 . therefore the electroweak precision measurements , through the @xmath139 parameters @xcite , may be used to constrain the scalar mixing angle together with the heavy higgs mass leading to the bound [ stubound ] 0.44 , m_h_2 500 for @xmath140 gev . is shown for comparison . [ fig.ynyt ] ] as we have discussed in the previous sections , the requirement of cancellation of the gauge and of the gravitational anomalies has allowed us to reduce the @xmath29 charges to just two parameters @xmath47 and @xmath48 . together with the constraint from the lep - ii results , reported in eq . ( [ stu1 ] ) , we also require the perturbativity of the evolution of the coupling of the new abelian sector which gives the conditions [ pertconds ] g(q ) < , ( q ) < q q , up to the given final evolution scale @xmath141 . here , the free parameters are the values @xmath142 and @xmath143 of the couplings at the electroweak scale , identified with the top quark mass . in addition , to process any allowed point in parameter space and investigate the perturbativity of the couplings and the stability of the scalar potential under the two - loop rg evolution we need to implement the appropriate matching conditions as illustrated in section [ matcondsection ] . we show in fig . [ fig.pertcoupgeneric ] the region of parameter space in which the perturbativity is maintained up to @xmath144 gev ( blue region ) , @xmath145 gev ( green region ) , @xmath146 gev ( yellow region ) and @xmath147 gev ( red region ) , as a function of the free charges @xmath47 and @xmath48 , or @xmath148 and @xmath149 . the shadowed areas in pink and green are the excluded regions according to eq . ( [ stu1 ] ) . they correspond to two different choices of @xmath25 equal to @xmath150 and @xmath151 tev respectively . the plots shows that the validity of perturbation theory up to @xmath15 requires the charge values @xmath152 and @xmath153 , and that this bound becomes more stringent as @xmath26 grows . the thick point in each of the three plots corresponds to the charge assignment of a @xmath0 abelian symmetry . we have set to zero the parameter of the kinetic mixing ( @xmath21 ) and varied only @xmath26 , the coupling of the extra @xmath20 , set to 0.1 and 0.2 in panel ( a ) and ( b ) respectively . clearly , in both cases , the @xmath0 charge assignment guarantees a perturbative evolution up to the planck scale . in panel ( c ) we have plotted the same regions of perturbativity and the corresponding exclusion areas of ( a ) using the @xmath154 parameterization for a generic extra @xmath29 symmetry . as we have already discussed in the previous sections , the cancellation of gauge and gravitational anomalies naturally requires the introduction of sm singlet fermions which can dynamically acquire a majorana mass through their yukawa coupling to the extra scalar @xmath31 controlled by @xmath101 in a type - i seesaw . in this scenario the mass of the heavy neutrinos can be of the same order of magnitude of the vev of the extra higgs , which can be easily chosen in the tev range , with a yukawa coupling @xmath155 . values of @xmath156 require a detailed rg analysis in order to identify the perturbativity and stability regions in the parameter space of these models . as shown below , the yukawa coupling @xmath101 , and , equivalently , the mass of the heavy neutrinos , has a fundamental role in this study . [ fig.ynyt ] we show the evolution of @xmath101 for different initial values at the electroweak scale . differently from the yukawa of the top quark , @xmath101 grows at a higher rate as we increase the starting value of @xmath101 . the analysis shows that for @xmath157 and @xmath158 tev , a value of @xmath159 at the electroweak scale spoils the perturbativity of the model at @xmath160 . the decreasing behavior of @xmath115 is due to the negative contributions of the qcd corrections to the corresponding @xmath102 function which are obviously absent in the evolution of the yukawas of the rh neutrinos . the study of the perturbativity of the abelian sector , investigated by the evolution of the coupling of the extra @xmath20 and the size of the kinetic mixing @xmath21 , provides significant information on the possible scenarios available in the b - l case . this analysis is presented in fig . [ fig.pertcoup ] where we show five plots related to the evolution of the abelian couplings . in particular , we try to identify the regions in the parameter space of this model where the evolution is perturbative along the rg running up to a given final scale @xmath141 . we show in panel ( a ) the maximum value of such evolution scale ( @xmath141 ) allowed for a given initial value of the coupling @xmath26 , with @xmath161 ( pure b - l case ) and @xmath162 tev , @xmath163 gev . for instance , for @xmath164 at the electroweak scale , the rg evolution remain perturbative only up to the gut scale ( @xmath146 gev ) , but it violates the limits from lep given in eq . ( [ stu1 ] ) . for smaller values of the same coupling , such as @xmath165 , the model is weakly coupled for any value of the final scale @xmath141 up to @xmath15 and satisfies eq . ( [ stu1 ] ) . we have varied the mass of the heavy neutrinos @xmath98 continuously from 500 up to 1500 gev , thereby identifying a colored area bounded by the two black and red curves . the continuous boundary curves refer to the choices of @xmath166 gev and 1.5 tev respectively . the two horizontal black lines in the same plot identify the values of @xmath141 corresponding to the gut ( @xmath146 gev ) and planck scales ( @xmath147 gev ) . + the analysis is repeated in the case of @xmath167 tev ( panel ( b ) ) , where now the experimentally allowed initial values of @xmath26 are for @xmath168 , with the shadowed region being therefore excluded by lep data . notice that in this case the two boundary curves describing the variation with the heavy neutrino mass appear to be superimposed and undistinguishable . as it can be inferred from eq . ( [ abelianbeta ] ) , the yukawa of the heavy neutrino affects the evolution of the abelian coupling only at two - loop level , thus explaining such mild dependence on it . moreover , for a heavier @xmath20 the effect of the rh neutrino mass is more suppressed , since for a given @xmath98 , a larger @xmath25 ( thus a bigger @xmath95 ) corresponds to a smaller @xmath101 . this explains the different bandwidths in the two panels for the same span of @xmath98 . + notice that in both panels the bounds from the requirement of perturbativity are implicitly taken into account in the identification of the maximum stability scale , while the lep bounds are shown an shadowed regions . incidentally , both for @xmath158 and 5 tev and for the chosen initial parameters , we find that the gauge coupling has to satisfy the condition @xmath165 at the electroweak scale in order to maintain perturbativity up to planck scale and to fulfill the lep bound eq . ( [ stu1 ] ) . in panel ( a ) the lep and the perturbativity bounds tend to coincide , while in ( b ) the latter is far more significant , with the shadowed region excluding values of the gauge coupling @xmath169 . + in panels ( c ) , ( d ) and ( e ) we have investigated the evolution of the @xmath26 and of the @xmath21 couplings , the latter describing the impact of the kinetic mixing , which are responsible for the behaviors discussed above . in panels ( c ) and ( d ) , we plot the running of @xmath26 as a function of the evolution scale @xmath141 , for a varying heavy neutrino mass . the shadowed areas are the regions excluded by the bound eq . ( [ stu1 ] ) . notice that the growth of @xmath26 is very similar in both panels , and in the case of a heavier extra @xmath20 it is quite insensitive to the change of the heavy neutrino mass , for the same reasons illustrated for the previous plots . the growth of this coupling is quite moderate for initial conditions @xmath170 but its rg trajectory undergoes a rapid steepening for larger initial values . + a similar analysis ( panel ( e ) ) is presented for the coupling which parameterizes the kinetic mixing of @xmath55 and @xmath0 . the evolution of the kinetic mixing @xmath21 is strongly correlated with the size of the initial condition on @xmath26 . the rg trajectories , in this case , steepen very fast as @xmath26 grows . we have chosen four values of @xmath26 at the electroweak scale equally spaced in the interval ( 0.1 - 0.4 ) . we start with a vanishing kinetic mixing at the electroweak scale and let the coupling be radiatively generated by the evolution . by selecting initial values of @xmath171 the evolution of the mixing satisfies the perturbativity bound up to the planck scale . for larger initial values , such as @xmath172 , both couplings @xmath96 will violate the perturbativity below the planck scale . in this plot , the shadowed regions are bounded by the two trajectories generated by masses of the heavy rh neutrinos @xmath166 gev ( black curve ) and 1.5 tev ( red curve ) which appear to be separated only for @xmath173 and @xmath174 gev . . two - loop evolution with the one - loop matching ( continuous line ) ; two - loop evolution without matching ( dotted line ) ; one - loop evolution without matching ( dashed line ) . ] the information about the vacuum stability is enclosed in the structure of the potential for large fields value , that is , in the quartic scalar interaction terms , stated in the three positivity conditions of eq . ( [ stabilitycond ] ) . in the sm , where only one term of this kind is present ( @xmath13 ) , the stability condition coincides with the positivity of this coupling . as we are going to show in this section , the stability bound is significantly affected also by the order of the evolution , which is crucial in order to reach definitive conclusions regarding the dynamics of the @xmath0 model . the result of this analysis is shown in fig . [ fig.coupev2 ] . in this study we want to emphasize the role of the two - loop corrections on the rg running and of the one - loop matchings on the initial conditions . for the sake of simplicity we choose fixed values for the mass of the heavy higgs and of the right - handed neutrinos both of 500 gev , and @xmath175 . the mass of the @xmath20 is reported in each plot . + we show in panel ( a ) the rg running of @xmath114 at one- and two - loop levels . notice that at one - loop level , for the chosen initial conditions , the maximum stability scale @xmath141 in the rg trajectory of this coupling is below the gut scale , at @xmath176 gev . two loop effects are clearly essential in order to verify its positivity up to the planck scale . + a similar investigation is carried out for the quartic coupling of the heavy higgs field @xmath123 . in panel ( b ) it is shown that both the one - loop and the two - loop evolutions indicate that this coupling evolves rather slowly and stays positive up to the planck scale for the chosen @xmath94 and @xmath98 . we have selected two different initial conditions at the electroweak scale corresponding to two different values of the mass of the @xmath20 , @xmath158 tev ( black curve ) , @xmath177 tev ( red curve ) . in this case the two - loop effects are quite small , independently of the chosen initial conditions . + in panel ( c ) we extend our analysis to the mixed @xmath178 quartic coupling , which enters into the stability condition by the quantity @xmath179 , the latter plotted in panel ( d ) . notice that , in this case , the flip in sign of @xmath124 , as shown in ( c ) , is not problematic for the stability bound if the two - loop rg equations are employed , since @xmath180 stays anyhow positive over the entire range of the evolution , as also shown in ( d ) . also in this case , it is clear that two - loop effects are essential in order to revert the trend of the evolution , which otherwise causes @xmath180 to turn negative below the gut scale , thereby generating a vacuum instability . in ( d ) we show the trajectories of @xmath180 for two values of @xmath25 reported in the same figure . in both cases an instability is generated at @xmath176 gev by the one - loop running , which is completely eliminated by the two loop effects . notice that two loop effects , in general , tend to change the concavity of the @xmath180 evolution . this interesting behavior can be understood from a cursory look at all the panel ( a ) , ( b ) and ( c ) , due to the different slopes of the corresponding curves over the entire evolution region . we conclude that the inclusion of two - loop effects in the rg evolution is mandatory for a consistent analysis of the stability of the scalar potential and for the identification of the available parameter space of the model . in order to show the impact of the matching conditions we study the behavior of @xmath114 in fig . ( [ c3 ] ) where the matched and the unmatched evolutions are shown . we have indicated with a continuous line the two - loop rg running of the same coupling with one - loop matching and with a dotted line the unmatched case . the dashed line refers to the one - loop result , with the tree - level matching . it is clear that , as for the two - loop @xmath102 functions , the one - loop matching relations on the initial condition improve the stability of the potential and are necessary for the consistency of the perturbative approach . moreover , we have explicitly verified that these contributions become more sizeable as soon as we switch on the scalar mixing parameters @xmath27 of the two higgses . having shown the importance of the two - loop corrections in the analysis of the stability of @xmath0 , we are now going to investigate this issue more closely by focusing our attention on its parametric dependence on the scalar mixing angle @xmath27 , on the mass of the heavy higgs and of the rh neutrinos . + the bounds from the electroweak precision measurements reported in eq . ( [ stubound ] ) can then be used to reduce the space of parameters in the @xmath182 plane . in fig . [ mtheta ] we plot in different colors the regions of stability and perturbativity for different final evolution scales @xmath141 . the @xmath182 region shown in all the panels is the one allowed by the bounds in eq . ( [ stubound ] ) , with the areas marked in blue , green , yellow and red identifying stable evolutions with final scales @xmath141 equal to @xmath144 , @xmath145 , @xmath146 and @xmath147 gev respectively . we have selected @xmath162 tev and a mass of the rh neutrinos @xmath98= 500 gev . the choices of the values of the couplings , @xmath157 in panel ( a ) and @xmath183 in panel ( b ) show the sensitivity of these regions on such small changes of @xmath26 . notice that in panel ( b ) the requirement of having a stable evolution up to the planck scale requires that @xmath184 tev . in general , as clear from the plots , such a requirement reduces significantly the bound given in eq . ( [ stubound ] ) , suggesting that the smallness of the scalar mixing can be inferred by the condition of rg stability . in particular , @xmath27 has to be smaller that @xmath185 in order to have a stable evolution of the scalar potential at least up to the gut scale . a study of the dependence of the stability region on the mass of the rh neutrino is presented in fig . [ shadows ] , where ( panel ( a ) ) the curves in black and red refer to values of the mass of the extra @xmath20 of 2.5 and 5 tev respectively , each with a more stable ( @xmath186 0.1 ) and a less stable ( @xmath186 0.3 ) coupling assignment . notice that both branches are characterized by a drastic change of the @xmath141 value of maximum stability , around @xmath187 2.2 and 4.3 tev respectively . as the mass of the heavy neutrino grows , the effects of its yukawa coupling @xmath101 dominate on the evolution of @xmath123 driving it towards negative values and , therefore , destabilizing the vacuum of the scalar potential . this behavior is caused by the large and negative @xmath188 contribution to the @xmath102 function of @xmath123 , as one can see already at one - loop level from eq . ( [ betal2 ] ) @xcite . clearly , the role of this coupling is similar to that of @xmath115 , the yukawa of the top , which is responsible for driving the sm vacuum towards an instability . once more we want to stress that the destabilizing role of a large @xmath101 is twofold : it can spoil both the stability of the scalar potential and the perturbativity , as shown in fig . [ fig.ynyt ] . moreover , it is clear from the same figure that @xmath101 grows significantly for large final evolution scales , thus increasing its destabilizing effect on vacuum . + more insight into this behavior of the potential can be deduced also from panel ( b ) and ( c ) of the same figure . here we show a plot of the @xmath123 coupling versus the evolution scale @xmath141 . the region of parameter space explored in panel ( b ) corresponds to the case @xmath25=2.5 tev and @xmath189 , with the mass of the rh neutrinos between 2.1 and 2.4 tev . as shown an instability emerges for @xmath98 around 2.2 tev turning @xmath123 negative and destabilizing the potential . a similar analysis is illustrated in panel ( c ) where @xmath190 has been chosen . it is clear , from this analysis , that a stable evolution up to a significantly large scale , such as the gut scale or larger , is favored by values of @xmath191 2 tev and is worsened by an increase in the size of the coupling @xmath26 . in fig . [ c4c ] we show the stability regions up to @xmath144 gev ( blue region ) , @xmath145 gev ( green region ) , @xmath146 gev ( yellow region ) and @xmath147 gev ( red region ) in the @xmath99 plane . in panel ( a ) we compare their shapes in the case of a two - loop versus a one - loop implementation of the running for a vanishing scalar mixing @xmath27 . the two - loop rg result is pictured in the larger panel , while the smaller subpanel on the bottom right reproduces the same plot but uniformly rescaled and at one - loop level . as we have already observed , the two plots show clearly that the inclusion of the two - loop corrections is essential in order to improve the stability of the potential up to the planck scale . notice that in the small subpanel , the red and the yellow regions , which in the larger plot characterize a stable potential up to the planck and gut scale respectively , are replaced by a green area , corresponding to a stability which can reach only @xmath192 gev . + in panel ( b ) we have varied the scalar mixing angle to the value of @xmath193 . also in this case the effects of a nonzero @xmath27 to the shape of the stability regions are quite significant . in general , the upper bound on @xmath94 in the @xmath194 plane is far stronger compared to panel ( a ) , showing that , for a rh neutrino of 2 tev , no values of the heavy higgs mass can accomodate the requirement of stability above @xmath144 gev . notice also that the upper bound on the heavy higgs mass @xmath94 is barely influenced by the heavy neutrino mass but it is strongly controlled by its vacuum expectation value @xmath95 and the scalar mixing angle @xmath27 , see fig . [ mtheta ] . + in panels ( c ) and ( d ) the region of planck scale stability is investigated in terms of its dependence on the initial conditions at the electroweak scale of the gauge coupling @xmath26 ( panel ( c ) ) and of @xmath27 ( panel ( d ) ) . we find that the region of stability grows larger as the gauge coupling and the scalar mixing decrease . the identification of possible approaches which could set significant exclusion limits on the parameter space of new physics models at the lhc is , for sure , of direct theoretical and experimental significance . among these , the requirement of stability of the scalar potential up to a very large scale , such as the gut scale or even the planck scale , has a clear physical appealing , although metastable vacuum states , as in the case of the sm , could also be acceptable . the case of simple abelian extensions of the gauge structure , which require a minimally enlarged higgs sector , with just one extra scalar , define a minimal framework in which to explore the requirements of stability . + our results clearly indicate that the sensitivity of the evolution on the large yukawa coupling of the top quark , present in the sm , is probably part of a more general scenario which includes other similar couplings sharing a similar behaviour . this picture emerges as soon as we allow a more complete description of the light neutrino masses . with these motivations , we have investigated in some detail the case of a b - l extension of the sm , accompanied by a type - i seesaw for the generation of the neutrino masses , realized by rh neutrinos . in such @xmath29 extensions , the gauge charges of the rh neutrinos are naturally constrained by the condition of cancellation of the gauge and gravitational anomalies . in our analysis we have specifically chosen a family universal charge assignment for these neutrinos , one per generation . the choice of a type - i seesaw , obviously , is model dependent , but the approach can be extended to other mechanisms as well . the phenomenological advantage of a mechanism of this type , besides its simplicity , is to involve a vev of the extra higgs which can be easily accommodated in the tev range , at the reach , in principle , of lhc searches . in this regard the model has been largely investigated starting from early lep searches @xcite , to the most recent analyses of its lhc visibility . such minimal yet rich extension of the sm has found a lot of attention along the years , covering the entire spectrum of the model , in search of specific experimental signatures at the lhc . from the extra scalar sector @xcite to its gauge @xcite and neutrino sector @xcite , the tev b - l scenario has shown to provide a promising window of experimental visibility , if realized in nature , for the present and forthcoming experiments at the lhc . moreover the aforementioned analysis have been carried out in the minimal @xmath0 extension in which the mixing between the @xmath83 and the @xmath20 gauge bosons has been neglected . a more complete analysis of the kinetic mixing of this model , although quite constrained by electroweak precision data @xcite , and the possibility of exploring different charge assignments for the new abelian gauge symmetry are left to future studies . + the rg analysis has been performed employing two - loop @xmath102 functions and one - loop matching relations on the initial conditions at the electroweak scale . we have shown that such higher - order corrections affect considerably the running of the couplings and , probably , should always be kept into account in any significant phenomenological analysis . they are essential for a consistent identification of the allowed regions in the parameter space of a given model , under the requirements of perturbativity and of vacuum stability . + following this reasoning , one finds that a charge assignment of b - l type , as shown in figs . [ fig.pertcoupgeneric ] and [ fig.pertcoup ] , can be characterized by a perturbative behaviour up to the planck scale . + we have seen that the requirements of perturbativity and stability of the scalar potential constrain the masses of the heavy higgs and of the rh neutrinos quite significantly . the boundaries of the allowed region , for a given mass of the @xmath20 , depend on the values of the scalar mixing angle @xmath27 and of the abelian coupling @xmath26 . we have also verified that the transition from a region of stability to an unstable one can be quite abrupt , and may depend quite significantly on the mass of the rh neutrino , as illustrated in fig . [ shadows ] . this feature can be ascribed to the evolution of the scalar coupling @xmath123 , and , in particular , to the yukawa ( @xmath101 ) of the rh neutrinos which affects their masses . as already mentioned above , the roles of @xmath101 and of @xmath115 , from this perspective , appear to be quite similar . finally , we have also investigated the impact of the kinetic mixing on the evolution of the @xmath195 symmetries . + there are other aspects , not touched in this analysis , and that we leave to future studies , which concern the inclusion of this model into a specific unified theory . it is well know that multiple @xmath17 factors are often encountered in the breaking of a gut scenario , or in the construction of effective theories from string models @xcite so that it is natural to ask what a vacuum stability analysis can add in predictivity using the corresponding rge s . from our point of view its use in this context can be dual . without any prior knowledge of a particular breaking pattern , the study of the instabilities of the scalar potential gives an indication of the validity of the @xmath196 gauge symmetry as a residual one , valid at low energy , before calling for a uv completion ( either of gut type or else ) . more significant theoretical constraints on the parameter space can instead be drawn when the model is embedded in a particular gut scenario . in this case the stability of the vacuum between the threshold scales of the intermediate breakings is a necessary condition for its consistency . from this perspective , our study clarifies that a heavy neutrino mass may challenge specific gut models . indeed , we have shown that the running of the corresponding yukawa couplings could spoil the perturbativity and the stability of the vacuum if a new threshold is not sufficiently near the tev scale . + another important point concerns the role of a tev b - l scenario respect to the constraints that could be imposed by requiring a successful leptogenesis . analysis in this direction have been presented in @xcite . we have verified that also in the model that we have considered , the value of the baryon abundance @xmath197 , which depends on product of the cp - asymmetry of the heavy neutrinos decays ( @xmath198 ) and on the efficiency factor @xmath199 , can account for the current estimated value of @xmath200 . @xmath201 is temperature dependent and is calculated at the sphaleron freeze - out temperature @xmath199 . it is responsible for the amount of washout of the baryon asymmetry , which can be dangerously low when the condition @xmath202 is realized @xcite . always in a context of a very strong near - degeneracy of the heavy neutrinos , a pattern which can be completely encompassed by our choice of a scalar @xmath101 , we have found regions , in the parameter space that we have analysed , which are compatible with leptogenesis . for instance , for a mass of the extra @xmath20 of 2.5 tev and a coupling @xmath157 , we have verified that the requirement of a maximal value of the cp - asymmetry parameter @xmath198 starts to allow leptogenesis ( @xmath203 ) if @xmath204 gev . meanwhile , less stringent values of @xmath198 , @xmath205 , are feasible if @xmath206 gev and @xmath207 tev respectively . we have found , as expected , that an increment of the strength of the @xmath20 coupling @xmath26 strongly disfavours the leptogenesis scenario , since it intensifies the washout of the asymmetry , for a 2.5 tev gauge boson . + * acknowledgements * we thank alon faraggi and pasquale di bari for discussions . l.d.r . thanks kostas skenderis for hospitality at the university of southampton , and nora brambilla and antonio vairo for their hospitality during a recent visit at the technical university of munich . l.d.r . is grateful to f. lyonnet for pointing out a typo in the @xmath102 functions and for accurate cross - checks . the work of c.c . is supported by a _ the leverhulme trust visiting professorship _ at the university of southampton in the stag research centre and mathematical sciences department , whose members he thanks for the kind hospitality . we give in this appendix the complete set of the two - loop @xmath102-functions , for the charge assignment @xmath208 and @xmath209 , required for the stability analysis . ^(2)_g_2 = g_2 ^ 3 ( 4 g + + + 12 g_3 ^ 2 + 4 g^2-)+ ^(2)_g_3 = g_3 ^ 3 ( + + + + -2 y_t^2 ) - 26 g_3 ^ 5 ^(2)_g = g^3 ( + + + + -)+ & ^(2)_&= ^2 ( + 12 g_2 ^ 2 g++- ) + & & + ^3 ( + + + - ) + & & + ( + - g^2 y_t^2 + 9 g_2 ^ 2 g^2 + g_3 ^ 2 g^2 + 12 g_2 ^ 2 g^2 . + & & . + g_3 ^ 2 g^2 + -12 g^2 y_n^2-^2 y_t^2 ) + & & + + + + + & & - g^2 g y_t^2 + 12 g^2 g_2 ^ 2 g+ g^2 g_3 ^ 2 g & ^(2)_g & = g^3 ( + + 12 g_2 ^ 2 + -12 y_n^2- ) + & & + g^2 ( - y_t^2++12 g_2 ^ 2 + g_3 ^ 2 ) + & & + g ( g^2 ^2- ^2 y_t^2++ g_2 ^ 2 ^2 + g_3 ^ 2 ^2 ) + & & + + & ^(2)_y_t & = y_t^3 ( + + + + 36 g_3 ^ 2 + 3 g^2 - 12 _ 1 ) + & & + y_t ( g^2 g + g^2 ^2++^2 ^2 + . + & & . + g_2 ^ 2 g - g_3 ^ 2 g + - g_2 ^ 2 ^2 + g_3 ^ 2 ^2 + . + & & .+ - g_2 ^ 2 g^2 + g_3 ^ 2 g^2 + g_2 ^ 2 g^2- g_3 ^ 2 g^2- . + & & .-108 g_3 ^ 4 + 9 g_2 ^ 2 g_3 ^ 2 + + 6 _ 1 ^ 2+)-12 y_t^5 & ^(2)_y_n & = y_n ( --^2 ^2 - 127 g^4 + 4 _ 2 ^ 2 + _ 3 ^ 2 ) + & & + ( 206 g^2 - 32 _ 2 ) y_n^3 - 44 y_n^5 & ^(2)__1 & = - g^4 g - g^4 ^2 - 13 g^2 g^2 ^2 - 10 g^2 g y_t^2- g^2 g ^3 + & & - g_2 ^ 2 g^2 g + _ 1 ^ 2 ( 36 ^2 + 36 g^2 + 108 g_2 ^ 2 - 144 y_t^2 ) + & & - g^2 ^2 y_t^2- g^2 ^4- g_2 ^ 2 g^2 ^2+_1 ( g^2 g + g^2 ^2 . + 34 g^2 ^2 + y_t^2++ ^2 y_t^2++ g_2 ^ 2 ^2 . + & & .++ g^2 y_t^2 + g_2 ^ 2 g^2 + g_2 ^ 2 y_t^2 + 80 g_3 ^ 2 y_t^2- . + & & .+^2 y_t^2 - 10 _ 3 ^ 2 - 3 y_t^4)+20 g^2 _ 3 ^2 - 4 g^2 ^2 y_t^2 - 13 g^2 ^4 + & & -13 g_2 ^ 2 g^2 ^2- y_t^4 - 10 g ^3 y_t^2 + 6 g_2 ^ 2 g y_t^2- + & & - g_2 ^ 2 g ^3- ^2 y_t^4- ^4 y_t^2 + g_2 ^ 2 ^2 y_t^2 - - g_2 ^ 4 ^2 + & & - g_2 ^ 2 ^4 - - g^4 y_t^2- g_2 ^ 2 g^4- g^2 y_t^4 + g_2 ^ 2 g^2 y_t^2 + & & - g_2 ^ 4 g^2 - 32 g_3 ^ 2 y_t^4- g_2 ^ 4 y_t^2++32 g^2 _ 3 ^ 2-^2 y_t^4 + & & -312 _ 1 ^ 3 - 4 _ 3 ^ 3 - 12 _ 3 ^ 2 y_n^2 + 30 y_t^6 & ^(2)__2 & = --^4 ^2 + 40 g^2 _ 3 ^2 + + & & _ 2 ( + ^2 ^2 + 2112 g^4 + 120 g^2 y_n^2 - 20 _ 3 ^ 2 + 48 y_n^4 ) + & & + 4 _ 3 ^ 2 ^2 + 4 g^2 _ 3 ^ 2 + 12 g_2 ^ 2 _ 3 ^ 2 - 7168 g^6 + 768 g^4 y_n^2 + _ 2 ^ 2 ( 448 g^2 - 240 y_n^2 ) + & & + 192 g^2 y_n^4 - 240 _ 2 ^ 3 - 8 _ 3 ^ 3 + 768 y_n^6 - 12 _ 3 ^ 2 y_t^2 & ^(2)__3 & = - g^2 g^3 + _ 3 ^ 2 ( ^2+g^2 + 3 g_2 ^ 2 + 16 g^2 - 72 _ 1 - 48 _ 2 - 24 y_n^2 - 12 y_t^2 ) + & & - g^2 g^2 ^2+_3 ( g^2 g + g^2 ^2++^2 ^2 . + & & .+ y_t^2++24 _ 1 ^2 + ^2 y_t^2++ g_2 ^ 2 ^2 + . + & & .+24 g^2 _ 1 + g^2 y_t^2 + g_2 ^ 2 g^2 + 72 g_2 ^ 2 _ 1 + g_2 ^ 2 y_t^2 + 40 g_3 ^ 2 y_t^2- . + & & .+672 g^4 + 256 g^2 _ 2 + 60 g^2 y_n^2+^2 y_t^2 - 60 _ 1 ^ 2 - 40 _ 2 ^ 2 - 96 _ 2 y_n^2 . + & & .-72 y_n^4 - 72 _ 1 y_t^2-)-656 g^4 ^2 - 160 g^3 y_t^2-^3 ^3 + & & + 120 g^2 _ 1 ^2 + 80 g^2 _ 2 ^2 + 48 g^2 ^2 y_n^2 - 76 g^2 ^2 y_t^2-^2 ^4 + & & -45 g_2 ^ 2 g^2 ^2 - 64 g^4 y_t^2 - 11 _ 3 ^ 3 to investigate the nature of the neutrino mass eigenstates we introduce a six dimensional vector in flavor space @xmath210 so that eq . ( [ typeiseesaweffective ] ) becomes [ typeiseesaweffective2 ] - l_y^= ^t m + h.c . , with a @xmath211 mass matrix @xmath212 of the form [ typeiseesawmassmatrix ] m = ( ) . for the sake of simplicity , all the flavor indices have been omitted . the mass matrix obtained in eq . ( [ typeiseesawmassmatrix ] ) can be formally set in a block - diagonal form with the help of a unitary matrix @xmath213 , implicitly defined as [ diagonalization ] ( ) , where @xmath214 and @xmath215 are ( generally non diagonal ) @xmath216 blocks . when expressed in terms of the rotated states @xmath217 the following matrix structure arises - l_y^&= & m__l^ij _ l^i _ l^j + m__h^ij ^i_h ^j_h + h.c . , notice that the dirac mass matrix @xmath218 has entries at most of the order of the electroweak scale , meanwhile the majorana mass @xmath219 for the sterile neutrinos is usually allowed to rise up to the gut scale or even up to planck scale . + such hierarchy is necessary for the implementation of the seesaw mechanism and translates into a hierarchy between @xmath214 and @xmath215 . the same separation in scales justifies the use of a particular ansatz @xcite for the unitary matrix @xmath213 employed to block - diagonalize the mass matrix in eq . ( [ typeiseesawmassmatrix ] ) [ uexp ] u = ( ) , where @xmath220 is a non - diagonal @xmath221 matrix with small elements . inserting the matrix expansion given in eq . ( [ uexp ] ) into eq . ( [ diagonalization ] ) one obtains in the following constraints [ constraints ] 0 & & m_d - m _ mv^ - v^t v^ * m_d - v^t m_d^t v^ - m_d v v^ , + m__l & & -v^ * m_d - m_d^t v^ + v^ * m_m v^ , + m__h & & m_m + v^t m_d^t + m_d v - v^t v^ * m_m - m_m v^ v . the first equality in eq . ( [ constraints ] ) , evaluated at leading order in @xmath220 , is used to fix @xmath222 . from the remaining conditions we are now able to recover the usual type - i seesaw relations for the light and the heavy mass matrices given in eq . ( [ typeiseesawmasses1 ] ) . + in order to completely resolve the neutrino spectrum , the mass matrices @xmath223 and @xmath98 must be separately diagonalized , leading to [ diag2 ] l^t m__l l^-1 & = & ( m__l1 , m__l2 , m__l3 ) , + h^t m__h h^-1 & = & ( m__h1 , m__h2 , m__h3 ) , for the light and the heavy neutrino mass eigenstates . this further step is of fundamental importance to investigate the ( non - unitary ) nature of neutrino mixing in a type - i seesaw scenario , given its role in defining the corresponding pontecorvo - maki - nakagawa - sakata ( pmns ) matrix @xcite as @xmath224 , where @xmath225 diagonalizes the mass matrix for the charged leptons and @xmath226 . + such fine - grained analysis , however , is an unnecessary complication in our study and , for the sake of simplicity , we consider diagonal yukawa couplings . d. buttazzo , g. degrassi , p. p. giardino , g. f. giudice , f. sala , _ et . _ , _ investigating the near - criticality of the higgs boson _ , _ jhep _ * 1312 * ( 2013 ) 089 , [ http://xxx.lanl.gov/abs/1307.3536[arxiv:1307.3536 ] ] . v. branchina , e. messina , and m. sher , _ lifetime of the electroweak vacuum and sensitivity to planck scale physics _ , _ phys . _ * d91 * ( 2015 ) 013003 , [ http://xxx.lanl.gov/abs/1408.5302[arxiv:1408.5302 ] ] . m. herranen , t. markkanen , s. nurmi , and a. rajantie , _ spacetime curvature and the higgs stability during inflation _ , _ phys . * 113 * ( 2014 ) , no . 21 211102 , [ http://xxx.lanl.gov/abs/1407.3141 [ arxiv:1407.3141 ] ] . j. r. espinosa , g. f. giudice , e. morgante , a. riotto , l. senatore , a. strumia , and n. tetradis , _ the cosmological higgstory of the vacuum instability _ , http://xxx.lanl.gov/abs/1505.04825 [ arxiv:1505.04825 ] . a. hook , j. kearney , b. shakya and k. m. zurek , _ probable or improbable universe ? correlating electroweak vacuum instability with the scale of inflation _ , _ jhep _ * 1501 * ( 2015 ) 061 , [ http://xxx.lanl.gov/abs/1404.5953[arxiv:1404.5953 ] ] . s. alekhin , a. djouadi , and s. moch , _ the top quark and higgs boson masses and the stability of the electroweak vacuum _ , _ phys . * b716 * ( 2012 ) 214219 , [ http://xxx.lanl.gov/abs/1207.0980 [ arxiv:1207.0980 ] ] . l. basso , s. moretti , and g. m. pruna , _ phenomenology of the minimal b - l extension of the standard model : the higgs sector _ , _ _ * d83 * ( 2011 ) 055014 , [ http://xxx.lanl.gov/abs/1011.2612 [ arxiv:1011.2612 ] ] . l. basso , s. moretti , and g. m. pruna , _ a renormalisation group equation study of the scalar sector of the minimal b - l extension of the standard model _ , _ phys.rev . _ * d82 * ( 2010 ) 055018 , [ http://xxx.lanl.gov/abs/1004.3039[arxiv:1004.3039 ] ] . a. datta , a. elsayed , s. khalil , and a. moursy , _ higgs vacuum stability in the b - l extended standard model _ , _ phys.rev . _ * d88 * ( 2013 ) , no . 5 053011 , [ http://xxx.lanl.gov/abs/1308.0816[arxiv:1308.0816 ] ] . j. chakrabortty , p. konar , and t. mondal , _ constraining a class of b - l extended models from vacuum stability and perturbativity _ , _ phys . rev . _ * d89 * ( 2014 ) , no . 5 056014 , [ http://xxx.lanl.gov/abs/1308.1291[arxiv:1308.1291 ] ] . l. a. anchordoqui , v. barger , h. goldberg , x. huang , d. marfatia , l. h. m. da silva , and t. j. weiler , _ majorana dark matter through the higgs portal under the vacuum stability lamppost _ , _ phys . * d92 * ( 2015 ) , no . 6 063504 , [ http://xxx.lanl.gov/abs/1506.0470 [ arxiv:1506.0470 ] ] . s. oda , n. okada , and d .- s . takahashi , _ classically conformal u(1 ) extended standard model and higgs vacuum stability _ , * d92 * ( 2015 ) , no . 1 015026 , [ http://xxx.lanl.gov/abs/1504.0629 [ arxiv:1504.0629 ] ] . n. haba and y. yamaguchi , _ vacuum stability in the @xmath228 extended model with vanishing scalar potential at the planck scale _ , _ ptep _ * 2015 * ( 2015 ) , no . 9 093b05 , [ http://xxx.lanl.gov/abs/1504.05669 [ arxiv:1504.05669 ] ] . c. corian , l. delle rose , and c. marzo , _ vacuum stability in u(1)-prime extensions of the standard model with tev scale right handed neutrinos _ , _ phys . * b738 * ( 2014 ) 1319 , [ http://xxx.lanl.gov/abs/1407.8539[arxiv:1407.8539 ] ] . l. basso , a. belyaev , s. moretti , and c. h. shepherd - themistocleous , _ phenomenology of the minimal b - l extension of the standard model : z and neutrinos _ , _ phys . * d80 * ( 2009 ) 055030 , [ http://xxx.lanl.gov/abs/0812.4313[arxiv:0812.4313 ] ] . l. basso , a. belyaev , s. moretti , g. m. pruna , and c. h. shepherd - themistocleous , _ @xmath20 discovery potential at the lhc in the minimal b - l extension of the standard model _ , _ eur . phys . j. _ * c71 * ( 2011 ) 1613 , [ http://xxx.lanl.gov/abs/1002.3586 [ arxiv:1002.3586 ] ] . l. basso , k. mimasu , and s. moretti , _ non - exotic @xmath20 signals in @xmath230 , @xmath231 and @xmath232 final states at the lhc _ , _ jhep _ * 11 * ( 2012 ) 060 , [ http://xxx.lanl.gov/abs/1208.0019 [ arxiv:1208.0019 ] ] . e. accomando , a. belyaev , j. fiaschi , k. mimasu , s. moretti , and c. shepherd - themistocleous , _ forward - backward asymmetry as a discovery tool for z bosons at the lhc _ , http://xxx.lanl.gov/abs/1503.02672[arxiv:1503.02672 ] . e. accomando , a. belyaev , j. fiaschi , k. mimasu , s. moretti , and c. shepherd - themistocleous , _ @xmath233 as a discovery tool for @xmath20 bosons at the lhc _ , http://xxx.lanl.gov/abs/1504.03168 [ arxiv:1504.03168 ] . f. del aguila , m. masip , and m. perez - victoria , _ physical parameters and renormalization of @xmath235 models _ , _ nucl.phys . _ * b456 * ( 1995 ) 531549 , [ http://xxx.lanl.gov/abs/hep-ph/9507455 [ hep - ph/9507455 ] ] . g. cacciapaglia , c. csaki , g. marandella , and a. strumia , _ the minimal set of electroweak precision parameters _ , _ phys.rev . _ * d74 * ( 2006 ) 033011 , [ http://xxx.lanl.gov/abs/hep-ph/0604111 [ hep - ph/0604111 ] ] . f. lyonnet , i. schienbein , f. staub , and a. wingerter , _ pyr@te : renormalization group equations for general gauge theories _ , _ comput . * 185 * ( 2014 ) 11301152 , [ http://xxx.lanl.gov/abs/1309.7030[arxiv:1309.7030 ] ] . a. ibarra , e. molinaro , and s. petcov , _ tev scale see - saw mechanisms of neutrino mass generation , the majorana nature of the heavy singlet neutrinos and @xmath236-decay _ , _ jhep _ * 1009 * ( 2010 ) 108 , [ http://xxx.lanl.gov/abs/1007.2378[arxiv:1007.2378 ] ] . e. salvioni , a. strumia , g. villadoro and f. zwirner , _ non - universal minimal @xmath20 models : present bounds and early lhc reach _ , _ jhep _ * 1003 * ( 2010 ) 010 , [ http://arxiv.org/abs/arxiv:0911.1450[arxiv:0911.1450 ] ]

typical proton - conducting polymer electrolyte membranes ( pem ) for fuel cell applications consist of a perfluorinated polymeric backbone and side chains that are terminated by strongly acidic sulfonic head groups . they provide the transport medium for the protons generated during the anodic oxidation reaction of the fuel . pem fuel cells offer numerous advantages as power generators for emission - free vehicles , portable applications , and individual homes @xcite . when such ionomers are exposed to water or humid air , the acidic groups terminating the side chains dissociate into so@xmath1 groups on the chains , and protons in the aqueous sub - phase . the sub - phase extends through the membrane , creating a bi - continuous nano - phase separated network of aqueous pores and polymer . the percolation structure and its ability to support proton diffusion have to meet strict requirements for a pem to be considered as an ` ideal ' membrane for fuel cells : a ) it must form a chemically , mechanically , thermally and morphologically stable structure over a wide range of water content , b ) show steady performance and tolerance of elevated temperatures up to 150@xmath2c at high proton conductance , and c ) be impermeable to gases and contaminating ions . development of new pem membranes that could meet all these demands is a primary task for many research groups involved in fuel - cell studies . one extensively used pem material is nafion@xmath0 , a perfluorinated polymer from dupont . the integrity and structural stability of nafion membranes is provided by the polytetrafluoroethylene ( ptfe ) backbones , but it is the hydrophilic clusters of side chain material that facilitate the transport of ions in the membrane . the microstructure of these clusters consists of an inter - facial region of solvated perfluoroether side chains separating the polymer matrix from the more bulk - like water found in the pores . in the presence of water , one finds a nano - separation of hydrophobic and hydrophilic domains . according to the early work of gierke _ et al . _ @xcite these pores are inverted micellar spheres with diameters around 4 nm , and are interconnected by channels with a diameter and length of about 1 nm . recent experimental investigations suggest that the spherical shape and uniform spacing of these pores are serious oversimplifications @xcite . new small - angle x - ray scattering ( saxs ) and small - angle neutron scattering ( sans ) experiments by gebel _ @xcite for nafion membranes in different swelling states have given a different picture for the morphology . their results suggest that elongated bundles of hydrophobic backbone material with diameters of about 4 nm and lengths greater than 100 nm are surrounded by sulfonate groups and water molecules . in other words , hydrophobic material is surrounded by hydrophilic domains . the fact that controversy continues to surround the nature of the hydrated morphology and the shape of the domains suggests that there is still room for theoretical investigation and computer simulations to address this issue . _ ab initio _ electronic structure calculations of polymeric fragments in contact with water @xcite and quantum molecular dynamics studies on model pem systems @xcite have provided a basis for understanding the energetically favorable conformations of nafion chains and the molecular ingredients of the conduction process @xcite . for systems large enough to contain several examples of hydrophilic inclusions and their interconnecting networks , however , treatment of the polymer in an _ ab initio _ manner , i.e. a full electronic treatment with molecular orbital theory , is not computationally feasible . the objective of this paper is to develop a deeper understanding of the driving mechanism behind the nanophase aggregation of sulfonate head groups , and to test the extent to which coarse - graining methods can be applied . we attempt this by using a simple dipolar head group model for ionomer side chains . we show that the dipole - dipole correlations strongly depend on the molar fraction of acidic groups and dielectric permittivity of the membrane . globular micelles formed in the case of zero dipole moment evolve to percolated cluster structures for a non - zero dipole moment of head groups when an artificial attraction between the end groups of side chains is introduced . accounting for the dipole moment of sulfonate head groups thus changes the conformational structure of clusters from a gierke model of spherical aggregations to a gebel model of elongated aggregations . the determination of the parameters necessary for success in modeling nanophase aggregation will help us understand the essential role played by the electrostatic interactions of the head group in determining microstructure formation in pem materials . the rest of this paper is organized as follows . after a brief review of segregation models in section ii , we present our dipolar model for side chains in section iii . section iv is devoted to our simulation results for dry pem ionomers . we conclude with a discussion of results and planned future work in section v. significant experimental and theoretical efforts have been put into the characterization of the microstructure of dry and swollen nafion membranes . although the fact of microphase segregation and its influence on the thermodynamic and transport properties of nafion membranes are well established , some uncertainty regarding the morphology of the subphases in swollen nafion remains . it has been generally assumed that the hydrophilic phase is continuous , in accord with the experimental evidence of high conductivity and water permeability . the conventional model of microphase segregation in nafion membranes , put forward by gierke _ @xcite , where the protons and water diffuse between globular clusters through cylindrical narrow channels , was further modified by mauritz and rogers @xcite and eikerling _ et al . also , ioselevich _ et al . _ @xcite developed a lattice based micelle - channel model to describe a connecting structure of a water bridges between the cages of the micelles . however , there is as yet no direct experimental evidence for the existence of the channels connecting the clusters in gierke s model . other models explaining the properties of hydrated nafion exploit bilayer , lamellar , and sandwich - like structures for nanophase separation in pem membranes . gebel @xcite describes the nafion membrane as an aggregation of polymeric chains forming elongated objects ( simplified as cylinders ) , embedded in a continuous ionic medium . at larger scales , those aggregates form bundles with definite orientational order . this new picture of multi - scale structure can explain membrane swelling as a continuous process from the dry state to a colloidal suspension . scanning calorimetry experiments on cast nafion films @xcite led to the conclusion that a continuous water phase exists only when the relative humidity of the membrane exceeds 90% . this conclusion contradicts the simulation results of khalatur _ et al . _ @xcite , where formation of continuous channels at very low hydration , and even in almost dry membranes , is reported . this is certainly consistent with the conductivity measurements in nafion films with very little water present @xcite . in contrast to all the above - mentioned network models , vishnyakov _ et al . _ @xcite showed that the water clusters do not form a continuous hydrophilic subphase . instead , the cluster - size distribution evolves in time as a consequence of the formation and rupture of temporary bridges between clusters . molecular dynamics ( md ) studies , using both classical force fields and an _ ab initio _ approach , appear to support the idea of irregularly shaped ionic clusters of sulfonic groups connected by smaller channels of similar composition embedded in the hydrophobic covalent matrix . however , the characteristic length scales of the morphological conformations suggested by experimental data are currently inaccessible to atomistic modeling . to probe the morphology at the nanometer scale , a mesoscale model is required . in this work we are concentrating on the effects of electrostatic interactions among the elements of the side chains , and so in order to avoid the excessively slow processes resulting from chain entanglement we consider a segmented system in which the side chains are detached from the backbone . the later introduction of very strong interactions between the ends of the side chains and the backbone , which have the effect of knitting together the segments into a single macromolecule , do not seem to affect our results significantly . a generic acidic side chain of nafion is commonly denoted as cf@xmath3ocf@xmath4cf(cf@xmath3)o(cf@xmath4)@xmath4so@xmath3h . in order to expedite these simulations , we adopt a simplified model for the side chain by using a united - atom presentation of the cf@xmath4 and cf@xmath3 groups @xcite and the so@xmath1 sulfonate head groups @xcite . these groups are modeled as lennard - jones ( lj ) monomers . we employ the same united - atom approach for backbone chain segments . in the framework of our dipole model all the partial charges on the ether oxygen , carbon and fluorine atoms of the side chain are set to zero . the fluorocarbon groups of the backbone skeleton are also assigned zero partial charges . the electrostatic charge is located entirely on the sulfonate head group , so@xmath1 , and on the hydrogen ion h@xmath5 permanently bound to it . the dipole moment produced when the proton and the sulfonate group , modeled as lj monomers , carry the full formal charges of @xmath6 and @xmath7 respectively , was reduced by a factor @xmath8 in some of our simulations . though our model is not applicable for the measurement of proton diffusion in pem membranes , it is a good starting point for a step - by - step exploration of the nanophase morphology in dry pem materials . the overall inter - monomer interaction in our model system is a combination of electrostatic coulomb interactions between each pair of charged particles , 12 - 6 lennard - jones ( lj ) potentials between neighboring monomers , and intrachain stretch and bend energies . the lennard - jones interaction parameters for the side chain are chosen to agree in most instances with the nafion model of paddison @xcite . we make a further simplification for the lj parameters of the side chain by assigning a universal value of @xmath9 , chosen to be 3.5 for all monomers . the total potential energy of the system is then given as @xmath10 where @xmath11 and @xmath12 are the lennard - jones , electrostatic , bond - stretching ( two - body term ) , and angle - bending ( three - body term ) components of the total energy , and are given by @xmath13 @xmath14 @xmath15 @xmath16 we omit consideration of the dihedral potential term ( four - body term ) , as this would be expected to make only a small contribution to the energetics of the single chain conformations in these dense polymeric systems . ( this assumption was tested in sample simulations in which this term was restored , and found to have little effect on our results . ) the following parameters for the polymer chains were used : an equilibrium bending angle @xmath17 , an equilibrium bond length @xmath18 ( 0.44@xmath9 ) , a bending force constant @xmath19 kcal / mol deg@xmath20 , and a stretching force constant @xmath21 kcal / mol @xmath22 . the last two constants correspond to the thermal energy @xmath23 at room temperature for a bond - length displacement of 0.1 and bond - angle displacement of @xmath24 from the equilibrium position . the bond length between the so@xmath1 head - group monomer and the hydrogen ion was chosen to be @xmath9 , and hence 0.35 nm . the lj interaction coefficient @xmath25 was chosen to be 3@xmath23 for hydrophobic - hydrophobic ( hh ) interactions , and 1@xmath23 for both hydrophobic - hydrophilic ( hp ) and hydrophilic - hydrophilic ( pp ) interactions . in the latter case the lj potential was modified to be purely repulsive by truncating at its minimum , where @xmath26 , and raising it by the addition of an amount @xmath25 . the reason for this choice is that the interaction between protons is not well represented by van der waals forces , which arise from virtual dipole - dipole interactions , and also that such attractive forces are small in comparison to the coulomb interaction . the normalized charge @xmath27 ( @xmath28 ) in eq.(3 ) is + 1 for protons and -1 for sulfonate groups . we used this coarse - grained model to perform molecular dynamics simulations of ensembles of ionomer segments . because chain entanglement greatly slows the rate of equilibration in polymers , we adopted the strategy of temporarily cutting the system into more mobile segments , which were later reassembled . in order to increase the diffusional rate of our model ionomer we detached the side chains from the backbone skeleton , an approach previously adopted by vishnyakov @xcite , and also cut the backbone skeleton into segments @xcite . in total the system then consists of @xmath29 sidechain segments and @xmath30 backbone segments . taking into account that the sulfonic acid groups are hydrophilic while the ethers and fluorocarbon groups are hydrophobic @xcite , we adopted the following general representation for a coarse - grained architecture of side chains and backbone segments . each side chain contains @xmath31 monomers of which @xmath32 are hydrophilic and @xmath33 are hydrophobic . its architecture is written as : @xmath33h+@xmath32p . the backbone segments are fully hydrophobic with an architecture @xmath34h . in the majority of systems studied in this paper , the side chain architecture was 8h+2p and the backbone segment architecture was 10h . the simulations were then performed in several stages . the initial configuration was first prepared by using monte - carlo techniques to grow a mixture of polymer chain segments with fixed bending angle @xmath35 and bond length @xmath36 . after an initial equilibration the constraints on @xmath35 and @xmath36 were removed . subsequently , md runs were performed for time periods between 50ps and 500ps at constant volume @xmath37 and constant temperature @xmath38 until the system of polymer segments was fully equilibrated . the system temperature was controlled by coupling the system to a langevin thermostat with a friction coefficient @xmath39 and gaussian white - noise force of strength @xmath40 . the equations of motion were integrated using the velocity verlet algorithm with a time step of 0.5 fs . we imposed standard periodic boundary conditions to our system , filling space with translational replications of a fundamental cell . long - range electrostatic interactions were calculated using the standard lekner summation algorithm @xcite . to check whether equilibrium had been reached we examined the proton pair distribution function @xmath41 in every picosecond time interval , and continued the run until no significant further change was observed . to avoid mistaking the trapping of the simulated system into a metastable glassy state for true equilibrium , we repeated each run with several different initial configurations . for each computer experiment we typically gathered system statistics over time intervals of a few nanoseconds . we believe this to be sufficient to allow for the self - organization of the side chains and backbones into clustered structures . in the next stage of the simulations the segments were reassembled into the branched chain characterizing the original nafion . this was achieved by the simultaneous introduction of a specific mutually attractive force between the ends of each backbone segment , which united them into a single chain , and a similar force acting between the tail monomer of the detached side chains and the fifth monomer of every backbone segment . this was applied in the form of a lj potential of strength @xmath42 . to avoid the formation of star - like branched polymers , only a single occupancy of the backbone attachment sites was permitted . the simulation was then resumed , and run until a new equilibrium was attained . it was found that the bonds on the ends of all the backbone segments were occupied , and that a large majority ( typically around 95% ) of the side chains were reattached to a backbone segment . this fraction was a function of the ratio @xmath43 of the number @xmath29 of side chains ( and hence of sulfonate groups ) to the number @xmath44 of backbone monomers . the quantity @xmath43 thus measures the concentration of sulfonate groups relative to the concentration of backbone monomers . the few side - chains that did not bond to the special sites on the backbone attached themselves more loosely to the backbone through the weaker regular hydrophobic - hydrophobic lj interaction potential . we also introduce the absolute molar concentration of sulfonate groups , @xmath45 , with @xmath46 being avogadro s number and @xmath47 the volume @xmath48 of the simulation cell , as an alternative parameter with which to describe the sulfonate concentration , independently of the amount of backbone polymer . while it is true that in real dry nafion , no significant free volume exists , and so the total number of monomers per unit volume is roughly fixed at the time of synthesis , in a simulation , however , it is possible to change the density of the material . it is of interest in understanding the mobility of ions in nafion to know what structures would form in the presence of imposed free volume , were such a reduction in density of the material possible . in order to observe the effects of the dipolar nature of the sulfonate - proton group on the structural characteristics of side chain aggregates , we calculated the normalized pair distribution function @xmath41 of head group protons ( which are the end group monomers of the side chains ) . the function @xmath41 indicates the relative probability of finding two protons at a separation distance @xmath49 averaged over the equilibrium trajectory of the simulated system , so that @xmath50 here @xmath51 is the number of protons located in a shell of thickness @xmath52 at the distance @xmath49 from a fixed proton , and @xmath53 is the total number of protons in the system of volume @xmath54 . in addition to this spherically averaged quantity , it is also useful to note some indications of the local anisotropy in order to distinguish between the gierke and gebel models . to this end we also evaluated the partial structure factors @xmath55^{2}+\left[\sum_{i=1}^{n_{+}}\sin q_{j}r_{i}\right]^{2}\right\rangle .\ ] ] here @xmath56 denotes the @xmath57 , @xmath58 and @xmath59 components of the vector @xmath60 , with @xmath61 , and the sums can proceed over the entire sample or over smaller volumes . .list of key variables . [ cols="<,^ " , ] the goal of these extensive coarse - grained molecular - dynamics simulations was to investigate the relative importance of the various model parameters on the aggregation of the sulfonate ( acidic ) head groups . to achieve this we performed a series of simulation runs in which the key parameters were varied as summarized in table ii . these parameters included the absolute concentration and the relative concentrations of side chains and backbone polymer , as well as the strength of the electrostatic interactions . this last quantity is characterized by a coulomb coupling parameter defined as @xmath62 . here @xmath63 , which is known as the bjerrum length , is about 56 nm for a dielectric constant @xmath64 and @xmath65 k. the coupling parameter @xmath66 was varied between 2 and 160 by varying the system dielectric permeability @xmath67 between 80 and 1 , with higher @xmath66 representing stronger electrostatic correlations in the system . in a typical simulation , a few thousand side chains are confined in a cubic box whose side is of length @xmath68 , with @xmath69 nm being the monomer diameter . each simulation run was started from a randomly generated assembly of side chains and backbone segments within the system volume . however , in order to avoid packing problems , and to make this process effective for high monomer densities , a preliminary simulation of the same number of monomers was first performed in a larger box of side @xmath70 , and this was then followed by a gradual reduction of the system volume until the desired molar concentration was achieved . in this first simulation the constraints imposed by the attachment of the side chains to the backbone were removed in order to concentrate on the effects of the electrostatic interactions . we then analyzed the proton - proton pair distribution function @xmath41 calculated for an assembly of unattached side chains for three different dielectric constants @xmath67 of the ionomer medium , and for two different molar concentrations . the higher concentration , for which @xmath71 mol / l , is characteristic of typical dry nafion samples . figure [ fig4 ] shows the proton - proton pair distribution function for the parameters described in table ii as run 1 and run 7 . the first observation is that , as expected , small @xmath67 , or , equivalently , high @xmath66 , facilitates clustering of the head groups , as shown in the red dot - dashed lines in figure [ fig4 ] . the graph of @xmath41 shows well - defined maxima at @xmath72,2.0 , and 3.4 . a detailed investigation shows that this is a consequence of the formation of three simple multiplet structures : compact and linear multiplets , each created by two head groups , and a branched multiplet composed of three head groups , as shown in fig . [ clusters ] . for @xmath73 there is no correlation between head groups on different side chains except for steric hindrances , as seen in the continuous blue lines in figure [ fig4 ] . a more surprising result is the effective repulsion between acidic head groups at intermediate dielectric permittivities , as seen in the green dashed lines in fig . it appears that , in the absence of strong correlations , each head - group dipole prefers to stay in the bulk of the sample in order to increase its polarization energy . this is similar to the well known salt - depletion effect in colloidal physics , when salt pairs are effectively repelled from the surface of a neutral colloid @xcite . the other unexpected result is the suppression of the correlations between the end groups at high molar sulfonate concentrations @xmath74 , as seen in the dot - dashed line and symbols in figure [ fig4](b ) . this can be traced back either to the reduction of the debye screening length , which effectively shields the dipoles from each other , or to the enhancement of the hydrophobic - hydrophobic attraction in the system . the latter emerges from the hydrophobic tails of the side chain segments . to decide which of these two possible mechanisms was responsible , we added into the system described in fig . [ fig4](b ) a small number of backbone segments in order to boost the effective hydrophobicity ( or association energy ) in our model system . one backbone segment of 10 monomers was added for each side chain unit , making @xmath75 . after equilibration , the side chains were then loosely attached to the backbone segments and the backbone segments were mostly united in the manner previously described , whereupon the system was further equilibrated . the results , shown as run 2 in figure [ fig6 ] , clearly indicate that the backbone segments have a minor impact on the head group distribution function . the system has almost the same microphase structure as observed for the original side chain system . a fivefold increase in the backbone concentration in figure [ fig6](b ) has only a small effect on @xmath41 . therefore , it is the dipole - dipole interaction that tends to destroy the correlations when the molar fraction of sulfonates is high . this effect takes place when the screening length becomes less than the mean separation distance between the dipoles . this has some relevance to the search for ionomer membranes with superior transport properties : whereas for better proton conductance one needs a higher sulfonate molar fraction , the latter hinders sulfonate aggregation , which is vital for water - sulfur channel formation in wet ionomers . typical snapshots of the simulation box , shown in figure [ fig5 ] , make visible the process of microphase separation in ionomer systems . the tips of the side chains , which correspond to end group protons , are shown as balls , whereas the rest of the side chains are plotted as lines . the backbone segments , drawn also as lines , mostly cluster together to create large aggregates . the side chains either anchor onto the surface of the backbone aggregate , or create their own small clusters . figure [ fig3d - density ] presents a 3-dimensional density distribution of monomers for run 5 . the areas of low ( high ) density colored in blue ( red ) show the channel structure of the aggregates . the polymer matrix is composed of hydrophobic tails of side chains and backbone segments . the hydrophilic head groups form the walls of these channels . though this 3d density picture differs from the long - bundle structural model of nafion @xcite , it supports the model of a dynamic system of non - spherical clusters with temporary inter - cluster bridges @xcite . in the previous section , in which we varied the strength of the coulomb interactions , we observed that in the systems with the higher sulfonate concentration , the coulomb correlations were suppressed . this suggested that a study of the morphology of aggregates over a wider range of concentrations would be helpful to understand the physics behind sulfonate cluster formation . simulation results for @xmath41 for systems with sulfonate concentrations @xmath74 varying between 1.6 and 2.5 mol / l are shown in fig . [ fig9 ] for runs 8 and 10 - 12 . results for the much wider range of 0.16 to 4.8 mol / l are shown in fig . [ fig9a ] , but for these it was necessary to delete the backbone segments entirely in order to accommodate the very large concentration of side chains at the upper part of the range constituting run 13 . the dipole - dipole correlations are seen to depend strongly on the molar concentration of acidic groups . for a typical dilute case , run 10 in fig . [ fig9 ] , the first peak of @xmath41 reaches a value @xmath76 , indicating a strong condensation of nearly all the side chains . a possible explanation for this effect can be drawn from the clustering features of pure dipolar liquids . in a dilute system , where the average interchain distance is larger than the gyration radius @xmath77 of the side chain , the lj interaction between the neighboring chains is negligible . the main contribution to the energy in this case comes from the electrostatic dipole - dipole interaction of head groups . although this interaction decays with a @xmath78 distance dependence , it is not weak over distances comparable with the simulation box length @xmath79 , and is capable of forcing the chains to aggregate . however , for a dense system , where the lj contribution to the energy becomes comparable to the dipolar term , the side chains start to lose their rotational entropy , and thus can not reorient themselves in order to minimize the electrostatic energy of the aggregate . while the heights of the maxima of @xmath41 gradually decrease as the system becomes more tightly packed , their positions do not change and remain fixed at @xmath80 and@xmath81 . the line with circles in fig . [ fig9a ] corresponds to the case where the concentration of backbone segments was double the case for run 8 , and corresponds to run 9 . it is seen that the proton pair distribution function is hardly affected by the existence of the extra hydrophobic segments . however this conclusion is not necessarily true for a system with the more attractive lj parameters for hh and hp interactions that will be considered in section d. both the snapshot picture in fig . [ fig15 ] and the 3d monomer density in fig . [ fig8 ] show microphase separation . another way to examine the head group clustering is to calculate the partial structure factors @xmath82 ( @xmath83 , @xmath58 , @xmath59 ) of the protons . because these did not exhibit any marked anisotropy , an averaged @xmath84 is plotted in fig . [ figsk ] . the position of the peak at @xmath85 is associated with a correlation distance , @xmath86 in the sample structure via the bragg relationship @xmath87 . a sharp growth at small @xmath88 of the values of @xmath84 in fig . [ figsk ] results from microscopic aggregation of hydrophilic and hydrophobic monomers in the system . to summarize the tendencies of dipole - dipole correlations presented in this and the previous section , we conclude that the hydrophilic head group aggregation takes place preferentially at low dielectric permittivity @xmath67 and low densities of sulfonate groups . although the bulk of this work describes the conformational structures of side chains of length @xmath90=10 , it is worth while investigating systems with shorter side chains , since recent experiments @xcite show a high proton transfer rate in ionomer membranes with shorter pendant segments . this section contains simulation results for different side chain lengths , @xmath91 . the length of the hydrophilic part ( the number of hydrophilic monomers ) is fixed at @xmath92 and the length of the hydrophobic tail ( the number of hydrophobic monomers ) @xmath33 is varied between 0 and 8 ( which corresponds to the variation of total side chain length @xmath90 between 2 and 10 ) in order to reveal how the side chains hydrophobic tail affects the pair distribution function of protons . no backbone material was included in these systems . simulation results for run 14 are shown in fig . [ fig12 ] . note that the case @xmath93 corresponds to a system of dipolar dumbbells . even a short hydrophobic tail with @xmath94 strongly suppresses the aggregation of head groups . thus the hydrophobicity of side chains , or the length of hydrophobic tail , is a strong obstacle to head group clustering . in other words , the dipole - dipole attraction between sulfonates of longer side chains is not sufficient to produce a nanophase separation . this conclusion is in contrast to the suggestion of eisenberg @xcite that each ionic group in the vicinity of a cluster of sulfonate groups is drawn in by dipolar attraction to form a depleted zone free of charge . one intriguing question in fuel cell technology is how the morphology of nanophase separation in pem membranes depends on the polymeric structure of nafion . in other words , what is the optimal comb - like structure for the ionomer to be a good material for a fuel cell ? using our segmental model for nafion we can mimic different comb - like nafion molecular conformations by changing the parameter @xmath43 , which denotes the ratio of the number of sulfonate groups to the number of backbone monomers . the impact of @xmath43 on the distribution function @xmath41 was shown to be negligible for @xmath95 for hh and @xmath96 for both hp and pp interactions in fig . [ fig6 ] for runs 1 - 6 . in order to explore more generally the effect of changing @xmath43 , we exaggerate the hydrophilic and hydrophobic interactions by tripling the interaction constants @xmath25 . the results for runs 1 - 3 presented in fig . [ fig16 ] for @xmath97 for hh , and @xmath95 both for hp and pp interactions , now show that the addition of backbone chain segments brings more conformational disorder into the head group correlations . this is also evident from the system snapshots for run 3 in fig . [ fig18 ] for the two different strengths of lennard - jones interaction parameters . in other words , a strong interaction between hydrophobic and hydrophilic monomers can disrupt the side chain aggregation , causing all the side chains to be evenly distributed on the surface of backbone aggregations , as in fig . [ fig18](b ) . this distribution will have a globular , lamellar , or cylindrical - rod form . to complete this study of the effects of interactions on morphology , we look at the effect of both decreasing and increasing the interaction between head groups . we have already seen in section iiia some of the consequences of reducing the coulomb interaction by increasing the dielectric constant @xmath67 . in the absence of interactions other than the dipole - dipole interaction , this is equivalent to a reduction in the strength of the dipole moments . in fig . [ fig1 ] we show the effect on the structure factor @xmath84 of the protons of a reduction in the strength of the dipole moments by various factors @xmath8 ( run 8) . when the parameter @xmath8 is small , there is no phase separation in the ionomers , and the function @xmath84 has a liquid - like structure with a single maximum . the behavior of @xmath84 is changed for an unreduced dipole strength @xmath8=1 in a manner similar to the behaviour of @xmath84 in the less dense system shown in fig . [ figsk ] : there is a strong peak at @xmath98=4.2 , and a sharp growth at small @xmath98 . the latter , again , is an indication of a strong microscopic separation of hydrophilic and hydrophobic monomers in the system . it is clear that the effective attraction between the acidic groups of side chains can not be accounted for by the sulfonate - hydrogen dipole moment alone . the sulfonate end group has its own internal dipole moment due to the charge distribution between the sulfur atom and the three oxygen atoms bound to it . partial charges on the other side chain monomers can also contribute to the effective attraction between head groups . to take all these contributions into account at a phenomenological level we introduce an artificial attractive interaction between the end groups of the side chains , in an approach similar to that employed by shirvanyanz @xcite . for this purpose we use a yukawa - type attractive potential , @xmath99 here @xmath100 is an effective yukawa charge for the head group sulfonates . for the sake of simplicity , we restrict ourselves to the case of constant debye screening length @xmath101 . in fact , the parameters @xmath100 and @xmath102 determine the stability , size and shape of the aggregates forming in the system . the simple attractive yukawa term can serve as a convenient basis for interpreting all the important features that can be observed in real systems or predicted by statistical mechanical theories . in the absence of electrostatic interactions in the system , i.e. when the head group dipole moment vanishes and @xmath103 , the ionomer system becomes spatially inhomogeneous . the pair distribution functions @xmath41 for runs 7 - 9 , plotted in fig . [ fig19 ] for @xmath104 ( case ( a ) ) and @xmath105 ( case ( b ) ) show strong structuring . the aggregation is substantially stronger for a system missing backbone polymer ( run 7 ) than for a system with a full complement of backbone monomers ( run 9 ) , as seen from a comparison of the respective lines in fig . [ fig19 ] . there is a clear tendency to micellar arrangement , which is characterized by some short - range order . snapshot pictures , given in fig . [ fig23 ] , show formation of a space - filling web - like network of cluster - like aggregates surrounded by neutral polymer sections . the size of these clusters grows as the density of side chains increases . a comparison of figs . [ fig23](a ) and [ fig23](b ) reveals that the effective yukawa charge @xmath100 , or in other words , the strength of the the screened coulomb attraction , has a strong impact on the shape of the clusters . the structure of the system for a moderate yukawa attraction of @xmath106 between the end groups is mobile : the aggregates can diffuse , decay and reappear again . they can also create temporary bridge - like connections . therefore the nanophase separation of the head groups , shown in fig . [ fig23](a ) can be viewed as an example for a gierke - like sulfonate aggregation . however , a strong attraction between the end groups of the side chains , the case @xmath105 in fig . [ fig23](b ) , results in a frozen network of globular clusters of diameter roughly 1.5 nm in the polymer matrix . this cluster picture resembles the ion segregation phenomenon proposed by eisenberg @xcite . when the dipole moment of the sulfonate head groups is restored , in this case with a value of @xmath107 , the globular micelles form super - clusters of diameter 3 - 4 nm , as seen is fig . [ fig24](b ) for @xmath105 . on the other hand , this nonzero dipole moment facilitates the formation of percolated cluster structures for @xmath104 , as in fig . [ fig24](a ) . the inclusion of the dipole moment of the sulfonate head groups thus changes the conformational structure of clusters from a gierke model of spherical aggregations to a gebel model of elongated aggregations . a snapshot picture of the same system but with no artificial attraction between the head groups is given in the fig . [ fig20 ] . it is clear that the introduction of a screened coulomb attraction results in the formation of compact and interconnected clusters of sulfonate head groups . in this study we have explored the issue of nanophase separation in dry nafion@xmath0 films by means of simulations within a simple dipolar - head - group model for the side chains . our procedure involved detaching the side chains from the backbone and cutting the backbone into segments to speed equilibration , and then reassembling the macromolecule by means of a strong imposed attractive force between the cut ends of the backbone , and between the non - ionic ends of the side chains and the midpoints of the backbone segments . as we vary the interaction parameters and dielectric permittivity of the medium we find different structures to be formed by the head groups . small permittivities facilitate the clustering of head groups with a well defined maximum in the proton - proton pair distribution function @xmath41 that indicates three simple multiplet structures : compact and linear multiplets each created by two head groups , and a branched cluster composed of three head groups . at higher permittivity no correlations between head groups are obtained . an effective repulsion between acidic head groups at intermediate dielectric permittivities was revealed from our data . we assume that , in the absence of strong correlations , each head - group dipole prefers to stay in the bulk of the sample in order to increase its polarization energy . for the ionomer interaction parameters considered in this work the microphase structure of the head groups was found to have little dependence on the concentration of backbone segments . since the molar fraction of sulfonates in a fuel cell membrane is known to be a crucial parameter for its performance , this factor was also studied . we find that for a dilute case a strong condensation of nearly all side chains takes place . the main contribution to the hamiltonian in this case comes from a long - range dipole - dipole interaction of head groups which forces the chains to aggregate . however , for a dense system , where the hydrophobic lj contribution to the hamiltonian becomes comparable with the dipolar term , the side chains can not reorient themselves in order to minimize the electrostatic energy of the aggregate . while on the one hand , for high proton conductance one needs higher sulfonate molar fractions @xmath74 , on the other hand we find that a high value of this parameter @xmath74 hinders the sulfonate aggregation that is vital for water - sulfonate channel formation in ionomers . overall , our simulations show that the dipole - dipole interaction is the main driving force for the aggregation of sulfonates . when the dipole moment of sulfonates is small , so that @xmath108 , there is no phase separation in the ionomers . in contrast , for @xmath109 a microscopic aggregation of hydrophilic and hydrophobic monomers takes place in the system . the most significant feature learned from the simulations is the fact that the dipolar - head - group approximation is not sufficient to drive the compact side chain cluster formation suggested by experimental measurements . an additional artificial attraction between the sulfonates , introduced as a screened yukawa interaction potential , boosts formation of nanometer - sized clusters dispersed throughout the system volume . the final size of these clusters , and the geometry of the cluster network they create , depend on the dipole moment @xmath8 of the sulfonate head groups for fixed yukawa parameters . as the strength of @xmath8 increases from zero to one , the conformational structure of clusters changes from a gierke model of spherical aggregations to a gebel model of elongated aggregations . more work is required to develop a deeper understanding of the mechanisms behind the aggregation picture observed experimentally . because the shape and size distribution of the aqueous pores in wet membranes is not completely known , it will be useful to extend our preliminary results for the dry nafion considered in this paper to systems with water present and with a more detailed chain structure . in future work , we plan to extend the model presented here to more complex models . in particular we will employ partial charges and lennard - jones parameters @xcite for side chain and backbone monomers of nafion - like oligomers . the issues of how the partial charges of ionomer molecules , in fixed and free proton models for side chains , affect the nanophase structure of the pem materials will be addressed . a complete understanding of the role of head group charges on the cluster conformations of side chains can provide helpful guidelines for understanding existing membranes and designing new membrane materials for fuel cells with improved characteristics . this work was supported by the us department of energy under grant de - fg02 - 05er46244 , and was made possible by use of facilities at the case its high performance computing cluster and the ohio supercomputing center . 98 m. i. perry , t. f. fuller , j. electrochemical society * 149 * , s59 ( 2002 ) . t. d. gierke , g. e. munn , f. c. wilson , j. polym . phys . ed . * 19 * , 1687 - 1704 ( 1981 ) . w. y. hsu , t. d. gierke , macromolecules * 15 * , 101 - 105 ( 1982 ) ; _ ibid _ j. memb . * 13 * , 307 - 326 ( 1982 ) . for ( a ) run 1 ( low density ) , and ( b ) run 7 ( high density ) for @xmath67=1 ( dot - dashed line ) , 15 ( dashed line ) and 80 ( full line ) . simulation results for @xmath67=1 are scaled down by a factor of four to fit the figure dimensions . the @xmath110 ( symbols ) in ( b ) , given for comparison , corresponds to the @xmath64 line in ( a ) . [ fig4],title="fig:",scaledwidth=60.0% ] for ( a ) run 1 ( low density ) , and ( b ) run 7 ( high density ) for @xmath67=1 ( dot - dashed line ) , 15 ( dashed line ) and 80 ( full line ) . simulation results for @xmath67=1 are scaled down by a factor of four to fit the figure dimensions . the @xmath110 ( symbols ) in ( b ) , given for comparison , corresponds to the @xmath64 line in ( a ) . [ fig4],title="fig:",scaledwidth=60.0% ] for run 2 and different dielectric permittivities . @xmath67=1 ( dot - dashed line ) , 15 ( dashed line ) and 80 ( full line ) . results for run 1 , not shown here , are very close to the results for @xmath67=1 . simulation results for @xmath67=1 are scaled down by a factor of four to fit the figure dimensions . ( b ) proton - proton pair distribution function @xmath41 for different amounts of backbone polymer ( runs 2 and 6 ) . results for run 3 - 5 ( not shown here ) that correspond to intermediate amount of backbone polymer , lay between these two curves . [ fig6 ] , title="fig:",scaledwidth=60.0% ] for run 2 and different dielectric permittivities . @xmath67=1 ( dot - dashed line ) , 15 ( dashed line ) and 80 ( full line ) . results for run 1 , not shown here , are very close to the results for @xmath67=1 . simulation results for @xmath67=1 are scaled down by a factor of four to fit the figure dimensions . ( b ) proton - proton pair distribution function @xmath41 for different amounts of backbone polymer ( runs 2 and 6 ) . results for run 3 - 5 ( not shown here ) that correspond to intermediate amount of backbone polymer , lay between these two curves . [ fig6 ] , title="fig:",scaledwidth=60.0% ] for different box sizes @xmath79 , and hence sulfonate concentrations @xmath74 , for runs 8 and 10 - 12 . the first peak of @xmath41 for run 10 ( full line , blue in colored figure ) has a value @xmath76 , out of the scale of the plot . [ fig9],scaledwidth=90.0% ] for different sulfonate molar concentrations @xmath74 , run 13 . ( a ) : 0.16 mol / l @xmath111 1.6 mol / l . ( b ) : 1.6 mol / l @xmath111 4.8 mol / l . the case @xmath112 ( line with circles in ( b ) ) is for run 9 . the first peak of @xmath41 for @xmath113 ( @xmath114 ) mol / l ( full blue line ( dashed green line ) in ( a ) ) has a value @xmath115 ( @xmath116 ) , which is out of the scale of the plot . [ fig9a],title="fig:",scaledwidth=65.0% ] + for different sulfonate molar concentrations @xmath74 , run 13 . ( a ) : 0.16 mol / l @xmath111 1.6 ( b ) : 1.6 mol / l @xmath111 4.8 mol / l . the case @xmath112 ( line with circles in ( b ) ) is for run 9 . the first peak of @xmath41 for @xmath113 ( @xmath114 ) mol / l ( full blue line ( dashed green line ) in ( a ) ) has a value @xmath115 ( @xmath116 ) , which is out of the scale of the plot . [ fig9a],title="fig:",scaledwidth=65.0% ] mol / l ( case ( a ) ) and @xmath117 mol / l ( case ( b ) ) head groups . [ fig15],title="fig:",scaledwidth=65.0% ] + mol / l ( case ( a ) ) and @xmath117 mol / l ( case ( b ) ) head groups . [ fig15],title="fig:",scaledwidth=65.0% ] for run 1 ( solid line ) , run 2 ( dashed line ) and run 3 ( dot - dashed line ) . thick lines : @xmath97 for hydrophobic - hydrophobic ( hh ) interaction , @xmath95 for hydrophobic - hydrophilic ( hp ) and hydrophilic - hydrophilic ( pp ) interactions . thin lines : @xmath95 for hh , @xmath96 for hp and pp interactions . [ fig16],scaledwidth=80.0% ] 2 . case ( a ) : @xmath95 for hh , @xmath96 for hp and pp interactions . case ( b ) : @xmath97 for hh and @xmath95 for hp and pp interactions . in the case ( b ) the sulfonate groups are mostly on the surface of clusters.,title="fig:",scaledwidth=60.0% ] + 2 . case ( a ) : @xmath95 for hh , @xmath96 for hp and pp interactions . case ( b ) : @xmath97 for hh and @xmath95 for hp and pp interactions . in the case ( b ) the sulfonate groups are mostly on the surface of clusters.,title="fig:",scaledwidth=60.0% ] for runs 7 - 9 and @xmath104 ( a ) and @xmath105 ( b ) . @xmath100 is a fictitious charge on the head groups which causes an artificial yukawa like attraction between sulfonates . thick ( thin ) lines are for @xmath103 ( @xmath107 ) . note that the scales on @xmath58-axis for short - range and long - range behavior of @xmath41 ( shown as an inset ) are different . for other details see the text . [ fig19],title="fig:",scaledwidth=70.0% ] + for runs 7 - 9 and @xmath104 ( a ) and @xmath105 ( b ) . @xmath100 is a fictitious charge on the head groups which causes an artificial yukawa like attraction between sulfonates . thick ( thin ) lines are for @xmath103 ( @xmath107 ) . note that the scales on @xmath58-axis for short - range and long - range behavior of @xmath41 ( shown as an inset ) are different . for other details see the text . [ fig19],title="fig:",scaledwidth=70.0% ] 2 , @xmath8=0 , @xmath100=0.1 ( case ( a ) ) and @xmath100=0.5 ( case ( b ) ) . [ fig23],title="fig:",scaledwidth=70.0% ] + 2 , @xmath8=0 , @xmath100=0.1 ( case ( a ) ) and @xmath100=0.5 ( case ( b ) ) . [ fig23],title="fig:",scaledwidth=70.0% ] 2 , @xmath8=0.5 , @xmath100=0.1 ( case ( a ) ) and @xmath100=0.5 ( case ( b ) ) . [ fig24],title="fig:",scaledwidth=70.0% ] + 2 , @xmath8=0.5 , @xmath100=0.1 ( case ( a ) ) and @xmath100=0.5 ( case ( b ) ) . [ fig24],title="fig:",scaledwidth=70.0% ]

the proliferation of mobile devices and the surge of a myriad of multimedia applications has resulted in an exponential growth of the mobile data traffic . in this context , wireless caching has emerged as a powerful technique to overcome the backhaul bottleneck , by reducing the backhaul rate and the delay in retrieving content from the network . the key idea is to store popular content closer to the end users . in @xcite , a novel system architecture named _ femtocaching _ was proposed . it consists of deploying a number of small base stations ( bss ) with large storage capacity , in which content is stored during periods of offpeak traffic . the mobile users can download content from the small bss , which results in a higher throughput per user . in @xcite , it was proposed to store content directly in the mobile devices . users can then retrieve content from neighboring devices using device - to - device ( d2d ) communication or , alternatively , from the serving bs . in both scenarios , content may be stored using an erasure correcting code , which brings gains with respect to uncoded caching @xcite . the use of erasure correcting codes establishes an interesting link between distributed caching for content delivery and distributed storage ( ds ) for reliable data storage . the key difference is that in the wireless network scenario , data can be downloaded from the storage nodes ( the small bss or the mobile devices ) but also from a serving macro bs , which has always the content available . therefore , the reliability requirements in ds for reliable data storage can be relaxed . in @xcite , the placement of content encoded using a maximum distance separable ( mds ) code to small bss was investigated and it was shown that a careful placement allows to significantly reduce the backhaul rate . in @xcite , for the scenario where content is stored directly in the mobile devices , the repairing of the lost data when a device storing data leaves the network was considered . assuming instantaneous repair , the communication cost of data download and repair was investigated . in @xcite , a repair scheduling where repair is performed periodically was introduced and analytical expressions for the overall communication cost of content download and data repair as a function of the repair interval were derived . using these expressions , the overall communication cost entailed by storing content using mds codes , regenerating codes @xcite , and locally repairable codes @xcite was evaluated in @xcite and it was shown that storing content using erasure correcting code can reduce the overall communication cost with respect to the scenario where content is downloaded solely from the bs . in this paper , we consider a similar cellular network scenario as the one in @xcite , where content is stored in a number of mobile devices using an erasure correcting code . mobile devices roam in an out of a cell according to a poisson random process . however , as opposed to @xcite , where the download of a single file is considered , here we consider that users may request files , of different popularity , from a library of files . our focus is on the delay of retrieving content from the network , which was not considered in @xcite . we derive analytical expressions for the download delay if content is stored in the mobile devices using mds codes and show that mds - coded distributed caching can significantly reduce the download delay with respect to the case where content is solely downloaded from the bs and the case where uncoded caching is used . the download delay of a single file was analyzed in @xcite . the remainder of the paper is organized as follows . the system model is introduced in section [ s : system_model ] . the average download delay incurred when mds - coded distributed caching is used is analyzed in sections [ s : download ] and section [ s : d2d ] . section [ sec : numericalresults ] presents and discusses numerical results and finally some conclusions are drawn in section [ sec : conclusions ] . _ notation . _ the probability density function ( pdf ) of a random variable @xmath0 is denoted by @xmath1 and the expectation with respect to @xmath0 is denoted by @xmath2 . probability is denoted by @xmath3 and @xmath4 represents the all - ones vector of length @xmath5 . we denote by @xmath6 the stationary distribution of an @xmath7 queueing system described by a poisson birth - death process with arrival rate @xmath8 and departure rate per node @xmath9 , which is given by @xmath10 where @xmath11 . we consider a single cell in a cellular network where @xmath12 mobile devices , referred to as nodes , request files , each of size @xmath13 bits , from a library of @xmath14 files . the files have different popularities and accordingly have a given probability to be requested . depending on the placement strategy , some files are encoded and stored into @xmath15 mobile devices , referred to as storage nodes . for ease of language , the set of storage nodes is referred to as the ds network and nodes not storing any content are referred to as _ regular _ nodes . a copy of each encoded file is also available at the bs serving the cell . a node requesting a file attempts to retrieve it from the storage nodes using d2d communication , and , if the file can not be completely retrieved from the ds network , the bs assists in providing the missing data . in order to increase the system efficiency , we allow multiple d2d communications to coexist if they are sufficiently far apart in space . therefore , we divide the cell in @xmath16 virtual clusters and assume that the size of the cluster and the transmit power are properly chosen such that only one d2d communication can be established between any two nodes in the cluster and the interference across different clusters can be neglected . a similar model is considered in @xcite . _ data allocation and coding strategy . _ we adopt a deterministic allocation strategy , where the @xmath17 most popular files are cached in a distributed fashion in @xmath18 storage nodes in the cell , according to the storage capacity of the devices . these files are partitioned into @xmath19 packets , called symbols , of @xmath20 bits each and are encoded into @xmath18 coded symbols using an @xmath21 mds erasure correcting code of rate @xmath22 . we use the same code for every file in order to simplify the analysis . we assume that each storage node stores a single symbol for each of the @xmath23 most popular files . overall , @xmath24 symbols are stored in @xmath18 storage nodes and no two storage nodes store the same symbol . we model the popularity of the files in the library using the time - invariant zipf distribution @xcite . accordingly , the probability that the @xmath5th file is requested is @xmath25 where parameter @xmath26 regulates the relative popularity of the files . in the following , the set of @xmath23 files stored in the cache of the mobile devices will be referred to as the ds library . we assume that the mobile devices are free to move inside the cell . we consider a uniform spatial distribution of the nodes in the cell , and hence there are @xmath27 devices per cluster on average and among them @xmath28 storage nodes . we focus on a single cluster in isolation , and assume that the devices roam in and out of it . the arrival , departure and request model of the nodes are borrowed from @xcite . the considered scenario is shown in fig . [ f : cluster ] . _ arrival - departure model . _ we assume that nodes arrive to the cluster according to a poisson random process with exponential independent , identically distributed ( i.i.d . ) random inter - arrival times @xmath29 with pdf @xmath30 where @xmath31 is the expected arrival rate and @xmath32 is time , measured in time units ( t.u . ) . the nodes stay in the cluster for an i.i.d . exponential random lifetime @xmath33 with pdf @xmath34 where @xmath35 is the expected departure rate per node . we assume that @xmath36 , which implies that the expected number of nodes in the cluster is @xmath37 . this model corresponds to an @xmath7 queuing model and the probability that there are @xmath5 nodes in the cluster is @xmath38 . the arrival of storage nodes to the cluster can also be described as a poisson random process . in particular , the inter - arrival times @xmath39 of the set of storage nodes has pdf @xmath40 the related lifetime is described by ( [ e : departure ] ) and the probability that there are @xmath5 storage nodes in the cluster is @xmath41 .. more precisely , the only guarantee is that there are on average @xmath42 storage nodes per cluster , but there are no constraints on their instantaneous number , which can even exceed @xmath18 . on the other hand , the probability of having a high number of storage devices in one cluster is generally very low . for @xmath43 , we have @xmath44 , @xmath45 and @xmath46 . the same consideration holds for the total number of mobile devices . ] mobile devices , and @xmath42 storage nodes among them ( red circles ) , caching one different coded symbol for each of the most popular files . a device requesting a file ( pink circle ) , must collect @xmath19 symbols . it attempts to recover them by using the ds network if the requesting file is stored in the devices . it uses the bs to collect the symbols that it is not able to download from the devices . the download of a symbol from a storage node takes @xmath47 t.u . , and from the bs @xmath48 t.u .. ,width=332 ] _ ds network update . _ we assume that the nodes storing content that arrive to the cluster from neighboring clusters are not immediately available for download , but the bs serving the cell keeps track of them and periodically updates and broadcasts to all mobile devices the list of storage nodes in the cell every @xmath49 t.u .. in the sequel , parameter @xmath49 is referred to as the update interval and the set of storage nodes in the list broadcasted by the bs as the ds list . _ data delivery . _ nodes request the file at random times with i.i.d . random inter - request time @xmath50 with pdf @xmath51 where @xmath52 is the expected request rate per node . we focus on the download process . the node that requests a file attempts to retrieve it from the ds network using d2d communication . thanks to the mds property , an encoded file can be reconstructed by accessing any @xmath19 encoded symbols . if the file can not be completely retrieved from the ds network , the bs assists in providing the missing coded symbols . the download of a coded symbol from a storage node incurs @xmath47 t.u . and from the bs @xmath48 t.u .. we assume that @xmath53 due to the congestion of the bs - to - node link and the fact that d2d communication occurs over a better channel due to the reduced distance between the involved nodes . we further assume that only one d2d link at a time can be established , and that the d2d communication does not interfere with the communication between the bs and the nodes . we say that the d2d network is _ idle _ if there is no active d2d communication in the cluster . if the d2d network is not idle when one node requests the file , the whole file is downloaded from the bs . moreover , to simplify the analysis , we assume that multiple bs - to - node links can coexist . we investigate the average time that is required to retrieve one file from the wireless network , referred to as the download delay . if a requested file is stored in the ds library , the requesting node attempts to retrieve it from the ds network using d2d communication , otherwise the file is entirely downloaded from the bs . therefore , we introduce the binary random variable ( rv ) @xmath54 which describes the hitting of the ds cache , i.e. , @xmath55 if a file of the ds library is requested and @xmath56 otherwise . moreover , the d2d network can be used only if it is idle , i.e. , if there are no active d2d communications . accordingly , we introduce the binary rv @xmath57 that describes the status of the d2d network . @xmath58 if the network is idle and @xmath59 otherwise . if the d2d network is idle , the requesting node tries to collect the necessary coded symbols from the nodes of the ds list provided by the bs using d2d communication . if the requesting node is a storage node of the ds list , it needs to download @xmath60 symbols , otherwise @xmath19 symbols must be downloaded . we thus introduce the binary rv @xmath61 , which represents the type of request , i.e. , @xmath62 for requests originating from a node that belongs to the ds list and @xmath63 for the other requests . the download from the storage nodes can be either fully successful or partially accomplished , in which case the requesting node turns to the bs to recover the missing symbols . on the other hand , if the d2d network is not idle and the requested file is stored in the ds library , the node downloads @xmath19 or @xmath60 symbols from the bs , depending on the type of node . from the discussion above , the average file download delay , @xmath64 , may be formalized as the average file download for the cellular network described in section [ s : system_model ] where the @xmath23 most popular files are stored in the mobile devices using an @xmath21 mds code is @xmath65 where @xmath66 is the average number of coded symbols downloaded per request using d2d communication and @xmath67 , referred to as the average d2d download delay , is the corresponding delay . the computation of @xmath66 , @xmath67 and @xmath68 is addressed in section [ s : d2d ] . the probability of hitting the cache can be expressed as @xmath69 where the probabilities @xmath70 are given in ( [ e : zipf ] ) . it follows that @xmath71 if @xmath72 . the next step is the computation of the probability that the d2d network is idle . let @xmath73 be the status of the network at the time of the @xmath74th request . it follows @xmath75 in order to compute @xmath76 , we introduce the rv @xmath77 that denotes the time instant of the @xmath78th request . also , let @xmath79 be the time during which the d2d network is occupied by the @xmath78th request . the d2d network is idle at the time of the @xmath74th request if none of the previous requests is still using d2d communication . therefore , @xmath80 and @xmath81 assuming that if the d2d network is not idle at time @xmath82 is because of the @xmath83th request , the product in ( [ e : p_i exact ] ) reduces to the term involving the @xmath83th request only , i.e. , @xmath84 since the requests are i.i.d . with inter - request time distributed as in ( [ e : request time ] ) and on average there are @xmath37 nodes in the cluster , we can compute @xmath85 and ( [ e : p_i approx1 ] ) can be written as @xmath86 if @xmath87 , @xmath88 and @xmath89 in ( [ e : idle_ell ] ) , we used the fact that the probability of hitting the cache and the average d2d download delay are independent of the request index ( if @xmath74 is sufficiently large ) , as it is proven in lemma [ l : tdd_ell ] in section [ s : d2d ] . substituting ( [ e : idle_ell ] ) in ( [ e : idle ] ) and after some simple calculations , we obtain @xmath90 in this section , we consider the computation of the average d2d download delay @xmath67 and the average number of coded symbols @xmath66 downloaded per request using d2d communication . we assume that a node can not download in parallel from multiple nodes , but it serially tries to download the coded file symbols from the nodes in the ds list . when a node requests the file , if the d2d network is idle and the requested file belongs to the ds library , it randomly chooses one of the storage nodes from the list supplied by the bs . after each downloaded symbol , the requesting node randomly chooses another storage node from the ds list and still alive . we assume that a requesting node that has collected fewer than the @xmath19 symbols necessary to reconstruct the file turns to the bs when all the reference storage nodes left or when the download of a symbol fails , even if other storage nodes are available . to simplify the analysis , we assume that both cases ( the failed symbol download and the absence of storage nodes ) incur @xmath47 t.u . , even if the node could contact the bs earlier . we also assume that the download from the d2d network fails if the requesting node itself leaves the cluster before collecting @xmath19 symbols . in this case , the download is also completed from the bs . to derive the average d2d download delay , we introduce three rvs describing the number of nodes of different type that are present in the cluster at the instant of a request : the number of storage nodes of the ds list , the total number of storage nodes ( belonging or not to the list , the latter corresponding to the storage nodes that arrive to the cluster after the ds list update and that have not left the cluster at the time of the request ) , and the number of regular nodes . in particular , we denote by @xmath91 the rv that describes the number of storage nodes of the ds list when a request arrives . we describe by the rvs @xmath92 and @xmath93 the total number of storage nodes and the number of regular nodes at the instant of a request , respectively . moreover , we denote by @xmath94 the rv that represents the total number of storage nodes ( belonging or not to the ds list ) at the beginning of the update interval of length @xmath49 . in the following three lemmas , we give a probabilistic description of the above rvs . [ l : px1 ] the probability that there are @xmath95 storage nodes of the ds list at the time of a request is @xmath96 where @xmath97 is the probability that @xmath98 is equal to @xmath99 , given that @xmath100 storage nodes are in the cluster at the beginning of the update interval of length @xmath49 , and is @xmath101 where @xmath102 and @xmath103 . the proof is given in appendix [ app : prooflem1 ] . [ l : pq ] the probability that there are @xmath104 storage nodes in the cluster at the time of a request is given by @xmath105 the proof follows the same lines as the proof of lemma [ l : px1 ] . [ l : pv ] the probability that there are @xmath106 regular nodes in the cluster at the time of a request is given by @xmath107 the proof follows the same lines as the proof of lemma [ l : px1 ] . based on the above lemmas , we can compute the probability that the request originates from a storage node of the ds list and the probability of having a given number of storage nodes in the ds list at the time of the request conditioned to the type of request . using bayes rule , the probability that there are @xmath108 storage node of the ds list alive at the time of a request , conditioned to the type of request , is given by @xmath109 the probability @xmath110 is given in lemma [ l : px1 ] . we now compute @xmath111 and @xmath112 . we start with the probability of having one request from the ds list conditioned to the number of storage nodes in the ds list at the time of the request . for @xmath113 , it can be written as @xmath114 clearly , the condition @xmath115 implies that the request can not originate from a storage node of the ds list , therefore @xmath116 . we approximate the probability that there are @xmath117 storage nodes at the instant of the request , given the number of nodes of the ds list alive , by using the steady state probability of a poisson birth - death process with arrival rate @xmath118 and departure rate @xmath119 . in particular , we compute @xmath120 where the expectations @xmath121 and @xmath122 are obtained starting from the probabilities ( [ e : pq ] ) and ( [ e : px1 ] ) , respectively . the number of regular nodes @xmath123 is independent of the number of storage nodes at the instant of the request , therefore we finally have @xmath124 where @xmath125 is given in lemma [ l : pv ] . note that in the expression above , with some abuse of notation , we used equal sign to avoid carrying all the way the approximation sign due to the approximation introduced in ( [ e : pq_approx ] ) . starting from this result , we compute the probability that the request originates from the ds list as @xmath126 where @xmath127 is given in lemma [ l : px1 ] . the probability @xmath128 is easily computed as @xmath129 . similarly , we have @xmath130 . following the same approach for the proof of lemma [ l : px1 ] , it is easy to show that @xmath131 and @xmath132 are independent of the specific request ( when @xmath74 grows large ) , where @xmath133 is the binary rv describing the type of the @xmath74th request . the case @xmath134 represents the case of instantaneous update , where the nodes contact directly the bs when they request a file and receive the list of the storage nodes through a dedicated link . for instantaneous update , the number of storage nodes at the instant of the request and the type of request is described by the following probabilities @xmath135 the probability that there are @xmath95 storage nodes at the time of the request given the type of request can be computed by replacing the above probabilities in ( [ e : px1cond ] ) . in order to describe the d2d download , let @xmath136 be the binary rv that describes the success of the download at the first attempt . more precisely , @xmath137 represents the successful download of the coded symbol from the first contacted storage node . if the download is not successful from the first contacted storage node , @xmath138 . similarly , we denote by @xmath139 the binary rv describing the download at the @xmath78th attempt and we denote by @xmath140}$ ] , @xmath141 the random vector ( @xmath142 ) . in the following , in lemmas [ l : s1 ] , [ l : sk ] , and [ l : sj ] , we derive the probability that no symbols can be downloaded from the d2d network , @xmath143 , the probability that the content is fully recovered from the ds network , @xmath144}={\mathbf{1}}_{k - i } | r\ } $ ] , and the probability that it is only partially recovered , @xmath145}={\mathbf{1}}_{j } , s_{j+1}=0 | r\}$ ] , respectively . [ l : s1 ] the probability that no symbols are downloaded through d2d communication , conditioned to the type of request , is given by @xmath146 where @xmath147 is given in ( [ e : px1cond ] ) and @xmath148 with @xmath102 . the proof is given in appendix [ app : prooflem4 ] . [ l : sk ] the probability that the file can be completely retrieved from the ds network , i.e. , the probability that k symbols are downloaded through d2d communication when @xmath63 , or @xmath60 when @xmath62 , conditioned to the type of request , is given by @xmath149 } = { \mathbf{1}}_{k - i } |r = i\ } = e^{-(k - i)\mu { t_\mathrm{d } } } \sum_{g = 1}^\infty \sum_{d=0 } ^{g } \frac{g - d}{g } \gamma_{k - i}(g , d , i ) , i=0,1,\ ] ] where @xmath150 is defined by the recursion @xmath151 for @xmath152 and @xmath153 , with initial condition @xmath154 and where @xmath155 the function @xmath156 is given in ( [ e : theta ] ) , and @xmath157 is given in ( [ e : px1cond ] ) . the proof is given in appendix [ app : prooflem5 ] . [ l : sj ] the probability of consecutively download @xmath158 symbols and to fail the download of the @xmath159th one is @xmath160}={\mathbf{1}}_{j } , s_{j+1}=0 |r = i \ } = \gamma_{j+1}(0,0,i)a_{j+1 } + \sum_{g=1}^\infty \sum_{d=0}^d \big ( \frac{d}{g } \gamma_{j+1}(g , d , i)a_{j+1 } + \frac{g - d}{g } \gamma_{j}(g , d , i)b_{j+1 } \big ) \nonumber\ , , \ ] ] where @xmath161 , @xmath162 , and @xmath163 is given in ( [ e : gamma ] ) . the proof follows the same lines as the proof of lemma [ l : sk ] . finally , the average d2d download delay and the average number of downloaded symbols from ds network are given in the following theorem . [ t : t_ni ] consider the network described in section [ s : system_model ] , where an @xmath21 mds erasure correcting code is employed and where there are @xmath42 storage nodes in the cluster on average . let @xmath47 be the time to download a symbol through d2d communication . the average d2d download delay and the corresponding average number of downloaded symbols are given by @xmath164 where @xmath165 is the probability that the request comes from a storage node of the ds list , and @xmath166}={\mathbf{1}}_{k - i } | r = i\ } + \sum_{j=1}^{k-1-i } j \pr \{{\mathbf{s}}_{[j]}={\mathbf{1}}_{j } , s_{j+1}=0| r = i \ } \ , , \nonumber\\ { \overline{t}_i } = & { t_\mathrm{d}}\big ( { \eta_i } + c_{k , i}\pr \{\ s_{1}=0 | r = i\ } + \sum_{j=0}^{k-1-i } \pr \{{\mathbf{s}}_{[j]}={\mathbf{1}}_{j } , s_{j+1}=0 | r = i\ } \big)\ , , \qquad \qquad i=0,1 \nonumber \end{aligned}\ ] ] where @xmath167 for @xmath168 and @xmath169 otherwise . the average d2d download delay is obtained as the sum of the average d2d delays in the case of requests originated from the ds list and of requests originating from the other nodes , weighted by the probabilities @xmath170 and @xmath171 , respectively . the same approach is used for the corresponding average number of downloaded symbols . according to our model , the requesting node completes the download of @xmath172 symbols from the ds network in @xmath173 t.u . with probability @xmath144}={\mathbf{1}}_k | r = i \ } $ ] , while the partial download of @xmath174 symbols happens with probability and incurs @xmath175 t.u .. for @xmath168 , in the computation of the average d2d download delay , we also consider the case where download from the ds network completely fails . the corresponding probability is @xmath176 and the delay is @xmath47 . when the request originates from the ds list and @xmath177 , i.e. @xmath178 , no symbols need to be downloaded , therefore @xmath179 and @xmath180 are equal to zero . [ l : tdd_ell ] the average d2d download time for the @xmath74th request , @xmath181 , is independent of the specific request if the index @xmath74 is sufficiently large . similarly to the average d2d download delay , @xmath181 is @xmath182 where @xmath183}={\mathbf{1}}_{k - i } | r^{(\ell)}=i\ } + c_{k , i}\pr \{\ s^{(\ell)}_{1}=0 | r^{(\ell)}=i\ } \nonumber\\ & + \sum_{j=0}^{k-1-i } ( j+1 ) \pr \{{\mathbf{s}}^{(\ell)}_{[j]}={\mathbf{1}}_{j } , s^{(\ell)}_{j+1}=0 | r^{(\ell)}=i\ } \big)\ , , \qquad \qquad i=0,1 \nonumber \end{aligned}\ ] ] the lemma follows from the fact that the probabilities in the expressions above are independent of @xmath74 , when @xmath74 grows large . in this section , we evaluate the average download delay when content is cached using mds codes for a cluster with @xmath184 nodes on average , departure rate @xmath185 , and request rate @xmath186 . we compare the average file download delay @xmath64 of the considered network with mds - coded ds with the delay of the traditional scenario where the content is solely downloaded from the bs , denoted by @xmath187 and with uncoded caching . in the following , with no loss of generality , we set @xmath188 t.u .. we recall that each file that is cached in the ds library is divided into @xmath19 symbols , and encoded using an ( @xmath189 ) mds code , where @xmath18 is the number of storage nodes in the cell . the code parameters are chosen such that @xmath190 . in this way , the _ average storage overhead _ in a cluster , @xmath191 , is positive , which increases the probability that the content is downloaded through d2d communication only . alternatively , we can use the same @xmath192 mds code for each cluster , but in this case the bs must continuously restore the initial state of reliability of the ds network when storage nodes leave the clusters @xcite . we first consider the special case where @xmath72 , therefore the probability of hitting the cache @xmath193 is @xmath194 . in figs . [ f:0.1][f:0.001 ] , we show the gain that can be achieved using mds - coded caching , by reporting the ratio between @xmath195 and @xmath64 as a function of the update interval @xmath49 . the infinite series involved in the computation of @xmath64 are truncated to a given value @xmath32 , chosen according to @xmath196 when involving the number of storage nodes and to @xmath197 when involving the number of nodes in general . we fix the ratio @xmath198 to be @xmath199 , and consider several mds codes . moreover , we also consider an uncoded scenario where one storage node on average in the cluster stores the uncoded files . in figs . [ f:0.1 ] , [ f:0.01 ] , and [ f:0.001 ] , @xmath47 is @xmath200 , @xmath201 , and @xmath202 times , respectively , smaller than @xmath48 . in the figures , the solid lines correspond to the analytical closed - form expressions derived in the previous sections and markers correspond to simulation results . it is observed that the analytical expressions predict very well the actual performance , which shows the goodness of the approximations introduced in ( [ e : approxi ] ) and ( [ e : pq_approx ] ) . the results clearly show that mds - coded ds can greatly improve the performance in terms of content download delay with respect to the case where content is downloaded from the bs , provided that the update interval , @xmath49 , is sufficiently small . for example , for @xmath203 and @xmath204 , a speed - up factor of around @xmath205 in the download is achieved with respect to the case of downloading from the bs using a @xmath206 mds code . interestingly , the results also show that the performance improves when @xmath19 increases . in particular , simple replication ( repetition coding ) is very inefficient and much better performance are achieved using larger mds codes ( of the same rate ) . we now consider the more general case where only part of the library of files is cached in the devices . we assume that the library has a size of @xmath207 files , and each storage node stores one symbol for each of the @xmath23 most popular files . we assume that the each file is of size @xmath201 mb , which corresponds to a @xmath200-minutes video . we further assume that each storage node makes @xmath208 gb available for caching . in this case , the number of cached files @xmath23 , that corresponds to the number of symbols cached by each storage node , increases by increasing @xmath19 , since the size of one encoded symbols is @xmath201mb@xmath209 and @xmath23 is clearly obtained by dividing the storage capacity of the node by the symbol size . in fig . [ f : sigma ] , we show the download speedup factor @xmath210 as a function of the parameter @xmath26 , which regulates the relative popularity of the files . for example , a large value of @xmath26 represents the case where few popular files are responsible for the majority of the download traffic . the figure refers to the case where @xmath211 and @xmath212 . the results confirm the gain that can be achieved by mds - coded distributed caching , but highlight another important aspect . when the whole library is not cached in the network , i.e. , @xmath213 , the gain reduction is not negligible , especially for low values of @xmath26 . this fact suggests that the adopted deterministic allocation , where the @xmath23 most popular files are equally stored in the ds network , is suboptimal . . solid lines show analytical results and markers simulation results . ] . solid lines show analytical results and markers simulation results . ] . solid lines show analytical results and markers simulation results . ] . @xmath214 and @xmath212 . ] in this paper , we considered the cache of popular content in the mobile devices of a cellular network using maximum distance separable erasure correcting codes to speed - up content delivery . we derived analytical expressions for the average download delay and showed that mds - coded distributed caching may dramatically reduce the download delay with respect to the traditional case where content is always downloaded from the base station . we denote by @xmath215 the number of storage nodes available for download at the time of the @xmath74th download request , i.e. , the number of storage nodes of the ds list that have not left the cluster at the time of the request . we compute @xmath127 by averaging over an infinite number of requests , @xmath216 similarly , let @xmath217 be the number of storage nodes at the beginning of the update interval wherein the @xmath74th request arrives , denoted by @xmath218 . we have @xmath219 in @xcite , it was shown that the probability @xmath220 does not depend on @xmath74 ( when @xmath74 grows large ) , and is given by ( [ e : px_j ] ) . its derivation is based on the observation that the number of storage nodes available for download in the update interval is described by a poisson death process . the probability @xmath221 can be written as @xmath222 where in the second equality we used bayes rule . in ( [ e : y ] ) , @xmath223 is the number of storage nodes at the beginning of the update interval , which is described by a birth - death poisson process , thus @xmath224 . the probability @xmath225 is the probability that there is at least one request in @xmath218 . it depends on the inter - request time , which in turn depends on the number of nodes in the cluster . therefore , we compute @xmath226 where @xmath227 is the probability that the inter - request time is shorter than @xmath49 when @xmath228 nodes are present in the cluster . similarly , we compute @xmath229 as @xmath230 where @xmath231 is the probability that there are @xmath228 nodes in the cluster , given that there are @xmath232 storage nodes . since these probabilities are independent of the specific request , we conclude that @xmath221 is also independent of @xmath74 . substituting ( [ e : y ] ) into ( [ e : x1l ] ) , we observe that @xmath233 is also independent of @xmath74 and using ( [ e : x1 ] ) we prove the lemma . we compute the conditional probability that no symbols are downloaded by averaging over an infinite number of requests , @xmath234 where @xmath235 is the rv describing the download of the first symbol for the @xmath74th request . we now consider the computation of @xmath236 . the recovery of the first symbol fails with probability @xmath194 if the requesting node leaves the cell before completing the download . it also fails if the requesting node stays in the cluster but no storage nodes are available or if it chooses to download from a storage node which departs before @xmath47 t.u . from the start of the download let @xmath237 be the departure time of the node which places the @xmath74th request and let @xmath238 be the event that the node which places the @xmath74th request stays in the network for more than @xmath239 t.u . from the start of the download . the corresponding probability does not depend on @xmath74 and is easily computed as @xmath240 . similarly , the probability that the requesting node departs before @xmath47 t.u . from the start of the download is @xmath241 . therefore , the conditional probability that the @xmath74th request fails the first symbol download is @xmath242 let @xmath243 be the number of storage nodes useful for download for the @xmath74th request , i.e. , the number of storage nodes in the ds list that have not left the cluster at the time of the @xmath74th request , excluding the requesting node itself if it belongs to the ds list . let @xmath244 the number of departures in @xmath47 t.u . among the @xmath245 storage nodes . the probability @xmath246 can be written as @xmath247 equation ( [ e : s1cond ] ) is obtained by observing that @xmath248 the probability that the download of the first symbol fails conditioned to @xmath249 and @xmath250 is independent of the type of request , and that @xmath251 the number of useful storage nodes and the number of departures in @xmath47 t.u . among them is independent of the departure of the requesting node . the probability @xmath252 is equal to 1 if there are no useful storage nodes , i.e. , @xmath253 . otherwise , it equals the probability to choose one of the @xmath254 storage nodes that leave the cell in @xmath47 t.u . , i.e. , @xmath255 , with @xmath256 . we observe that the number of departures of useful storage nodes conditioned to their number is independent of the type of request and that the number of storage nodes useful for download @xmath249 is related to @xmath257 by @xmath258 since when the request originates from a storage node of the ds list , the requesting node itself is not counted among the useful storage nodes . therefore , the probability @xmath259 can be written as @xmath260 we denote by @xmath156 the probability @xmath261 , given in ( [ e : theta ] ) . its derivation is similar to that of @xmath262 @xcite . the probability @xmath263 is independent of the specific request and is given by ( [ e : px1cond ] ) . after simple manipulations , we finally obtain @xmath264 since the probabilities involved in ( [ e : ps1 ] ) are all independent of @xmath74 , the lemma is proved . to evaluate the probability of complete download from the ds network , we start with the following limit , @xmath265 } = { \mathbf{1}}_{k - i}| r = i \}= \lim_{l \to \infty } \frac{1}{l } \sum_{\ell=1}^l \pr\ { { \mathbf{s}}_{[k - i]}^{(\ell)}={\mathbf{1}}_{k - i}|r^{(\ell)}=i \}\ , , \ ] ] where @xmath266}^{(\ell)}= ( s_1^{(\ell ) } , \dots , s_j^{(\ell)})$ ] and @xmath267 describes the successful symbol download at the @xmath5th attempt of the @xmath74th request . we consider the @xmath74th request and , similarly to the proof of lemma [ l : s1 ] , we will find that this probability is independent of @xmath74 . we denote by the rv @xmath268 the number of storage nodes useful for download at the time of the @xmath78th attempt of the @xmath74th request , i.e. , the storage nodes of the ds list not yet contacted , excluding the requesting node if it belongs to the ds list . we denote by the rv @xmath269 the number of departures in @xmath47 t.u . among the @xmath270 nodes . we also denote by @xmath271 the event that the node which places the @xmath74th request stays in the network for more than @xmath272 t.u . from the start of the download . the corresponding probability does not depend on @xmath74 and is given by @xmath273 . the probability of complete download is zero if the requesting node departs before @xmath274 t.u . from the start of the download , therefore we can write @xmath275}^{(\ell)}={\mathbf{1}}_{k - i}|r^{(\ell)}=i \}=e^{-(k - i)\mu { t_\mathrm{d } } } \pr\ { { \mathbf{s}}_{[k - i]}^{(\ell)}={\mathbf{1}}_{k - i}|\mathcal{a}_{k - i}^{(\ell ) } , r^{(\ell)}=i \}\,.\ ] ] the probability @xmath276}^{(\ell)}={\mathbf{1}}_{k - i}|\mathcal{a}_{k - i}^{(\ell ) } , r^{(\ell)}=i \}$ ] can be written as @xmath277}^{(\ell)}={\mathbf{1}}_{k - i}|\mathcal{a}_{k - i}^{(\ell ) } , r^{(\ell)}=i \}=\nonumber\\ & = \sum_{g } \sum_{d } \pr\{s^{(\ell)}_{k - i}=1| \mathcal{a}_{k - i}^{(\ell ) } , g_{k - i}^{(\ell)}=g , d_{k - i}^{(\ell)}=d \ } \pr\ { g_{k - i}^{(\ell)}=g , d_{k - i}^{(\ell)}=d,{\mathbf{s}}_{[k - i-1]}^{(\ell)}={\mathbf{1}}_{k - i-1 } | \mathcal{a}_{k - i}^{(\ell ) } , r^{(\ell)}=i \}\nonumber \\ & = \sum_{g=1}^\infty \sum_{d=0}^{g } \frac{g - d}{g } \gamma^{(\ell)}_{k - i}(g , d , i)\ , . \label{e : gamma1 } \end{aligned}\ ] ] the last equality is obtained by observing that the probability @xmath278 for @xmath279 and @xmath280 equals the probability to choose one of the storage nodes of the ds list that remains in the cluster , i.e. , @xmath281 . moreover , we have defined @xmath282}^{(\ell)}={\mathbf{1}}_{j-1 } | \mathcal{a}_{j}^{(\ell ) } , r^{(\ell)}=i \}$ ] , that can be computed as @xmath283}^{(\ell)}={\mathbf{1}}_{j-2 } | \mathcal{a}_{j-1}^{(\ell ) } , r^{(\ell)}=i \}\ , , \nonumber\end{aligned}\ ] ] for @xmath284 . we also define @xmath285 , which is equal to one if @xmath286 and @xmath287 , and zero otherwise . the condition @xmath286 follows from the fact that the number of useful storage nodes after a successful symbol download is equal to the number of useful storage nodes still alive , , minus the storage node just used . the condition @xmath287 comes from the fact that the @xmath288th symbol download is assumed to be successful , i.e. , @xmath289 . it is easy to prove by induction that the probability in ( [ e : gamma1 ] ) does not depend on @xmath74 . by defining @xmath290 , we obtain the recursion ( [ e : gamma ] ) , with initial condition ( [ e : gammaic ] ) . the probabilities @xmath291 , @xmath158 , are equal to zero for @xmath292 . k. shanmugam , n. golrezaei , a. g. dimakis , a. f. molisch , and g. caire , `` femtocaching : wireless content delivery through distributed caching helpers , '' _ ieee trans . theory _ , vol . 59 , no . 12 , pp . 84028413 , dec . n. golrezaei , p. mansourifard , a. f. molisch , and a. g. dimakis , `` base - station assisted device - to - device communications for high - throughput wireless video networks , '' _ ieee trans . wireless commun . _ , vol . 13 , no . 7 , pp 36653676 , july 2014 . j. pedersen , a. graell i amat , i. andriyanova , and f. brnnstrm , `` repair scheduling in wireless distributed storage with d2d communication , '' in _ proc . . theory work . ( itw ) _ , jeju island , korea , 2015 . j. pedersen , a. graell i amat , i. andriyanova , and f. brnnstrm , `` distributed storage in mobile wireless networks with device - to - device communication , '' _ ieee trans . commun . _ , vol . 64 , no . 11 , pp . 48624878 , nov . 2016 . a. piemontese and a. graell i amat , `` mds - coded distributed storage for low delay wireless content delivery , '' in _ proc . turbo codes & iterative inform . ( istc ) _ , brest , france , sep .

the interest in the storage and manipulation of quantum systems have led the researchers to design strategies for preserving the coherence of such systems against the detrimental effects of the environment @xcite . in fact , the uncontrollable degrees of freedom of the environment can get entangled with the quantum system and rapidly destroy the relative phase between the components of a linear superposition state . considerable efforts have been devoted to the issue of counteracting decoherence . notable examples are quantum error - correction codes ( qecc ) @xcite and decoherence - free subspaces ( dfs ) @xcite , both based on the careful encoding of the quantum state to be protected into a wider , partially redundant , hilbert space ( see fig . [ fig : bb_control_loops ] ) . the drawback of these encoding strategies is the large amount of extra resources required @xcite . alternative approaches which avoid this hindrance have been developed and may be divided into two main categories : closed - loop control , also known as quantum feedback @xcite , and open - loop control @xcite ( see fig . [ fig : bb_control_loops ] ) . in closed - loop techniques , the system to be protected is subject to appropriate measurements , whose results are then used for a real - time correction of the system dynamics . on the contrary , in open - loop controls , the system is subject to external , suitably tailored , time - dependent pulses which do not require any measurement . the control operations are chosen so that any undesired effect of the environment , such as dissipation , decoherence , heating , can be eliminated in principle if the controls are applied faster than the environment correlation time . the physical idea behind these open - loop schemes comes from refocusing techniques in nmr spectroscopy @xcite where they are typically implemented via strong and rapid pulses known as `` bang - bang '' ( bb ) controls . open - loop decoupling strategies have been suggested in various physical contexts : to inhibit the decay of an unstable atomic state @xcite , to suppress the magnetic state decoherence @xcite and the heating in ion traps @xcite , to contrast the quantum noise due to scattering processes @xcite and the polarization mode dispersion @xcite in optical fibers . bb control has been first demonstrated in nmr systems ( see e.g. @xcite ) . then , it has been proven with nuclear spin qubits in fullerene @xcite , nuclear - quadrupole qubits @xcite and very recently also with electron spins in penning traps @xcite and polarization qubits in a plane - parallel cavity @xcite and in a ring cavity @xcite . in this paper we discuss the last of the above - mentioned applications @xcite and provide an analytical explanation of the results . photons interact very weakly with their environment and encoding quantum information in their polarization is a practical and convenient solution for many quantum communication protocols @xcite . however , the optical properties of elements like mirrors , crystals and waveplates always depend upon both polarization and frequency , realizing in this way an effective coupling between the two degrees of freedom of the photon . the detection of the final polarization state is then performed by a detector that inevitably has a nonzero integration time . hence the measurement traces out the frequency degree of freedom and if the field is not monochromatic all the information transferred to the frequency modes goes lost , inducing on the polarization qubit a dephasing process analogous to the `` transversal '' decay occurring in nmr . such a polarization decoherence can be inhibited by means of bb controls implemented in space , i.e. , along the photon path , rather than in time , by means of suitably oriented waveplates . in ref . @xcite this idea is applied to the evolution of an optical pulse circling in a triangular ring cavity . polarization decoherence occurs because of the reflection of the pulse on the cavity mirrors , which depend on both frequency and polarization . after a few cavity round trips , the initially pure polarization state becomes highly mixed . on the contrary , after inserting appropriate waveplates in the cavity realizing an effective impulsive bb , polarization decoherence is significantly inhibited . previously , a proof - of - principle of the bb impulsive dynamics was given in @xcite , by using optical pulses in a plane - parallel cavity . that experiment clearly demonstrates that bb can be applied to light pulses traveling back - and - forth between two points in space , e.g. , on a two - way optical channel @xcite . on the other hand , the present experiment including a ring cavity extends the applicability of bb to one - way channels , because it eliminates the need of a recycling mirror for suppressing decoherence , and relies entirely on bb impulsive controls . the paper is organized as follows . in sec . ii we present a brief introduction on the dynamical decoupling . in sec . iii we describe the experimental apparatus of ref . @xcite , i.e. , the preparation of the input polarization state , the triangular ring cavity within which it propagates , and its final detection . we show how polarization decoherence emerges due to the polarization dispersive properties of the cavity mirrors . we also introduce the `` compensated '' cavity obtained by placing an additional @xmath0 waveplate in the long arm of the cavity ; this is needed if we require that the cavity induces a purely decohering dynamics on the pulse polarization . in sec . iv we discuss a first experiment in which polarization decoherence is inhibited through the simple `` carr - purcell '' decoupling scheme @xcite , which is a particular case of bb control realized by adding a single wave - plate within the cavity , suitably oriented . we then consider the more general pauli - group decoupling @xcite , which is realized by employing _ two _ waveplates with orthogonal rotation axes and we show that polarization decoherence is inhibited also in this case . finally we modify our ring cavity setup in order to implement the most general form of decoherence acting on the polarization qubit . generic qubit decoherence is realized by placing a soleil babinet ( sb ) compensator with axis at 45@xmath1 with respect to the cavity plane in front of each plane mirror . we show that , in accordance with theoretical predictions @xcite , pauli - group decoupling is effective in inhibiting decoherence in this general case too . v is for concluding remarks . the open - loop control technique takes advantage of the finite time of the relaxation process induced by the environment on the system . accordingly , one is able to modify the dynamics by inducing motions on the system faster than the shortest time scale accessible by the environment . for instance , to improve the resolution in nmr spectroscopy , multiple fast pulse sequences are used in order to reverse the effect of interactions of spins with the environment @xcite . this is obtained through a careful tailoring of the amplitude , phase , and frequency of the time - dependent terms in the total hamiltonian . the main example is the `` spin echo '' refocussing technique used in nmr , extensively investigated by e. l. hahn @xcite , and later generalized by carr and purcell @xcite . in that case , the main sources of noise are the inhomogeneities in the magnetic field and the molecular spin - spin interactions , which cause the total magnetization to dephase in a typical relaxation time called @xmath2 . at the microscopic level , the dephasing is due to the slightly different larmor frequency experienced by each spin in the ensemble . in a reference frame which rotates about the @xmath3-axis at the larmor frequency , the @xmath2 process can be visualized as the spreading of the magnetization vector in the transverse plane . one can then apply suitable rf pulses at the larmor frequency to flip each spin by an angle of @xmath4 . this realizes an effective `` time reversal '' for the molecules of the sample and leads , in the ideal case , to the rephasing of the magnetization vector . this is analogous to what happens in a standard optical fiber because of the chromatic dispersion . when a light pulse travels in the fiber , all its components experience a different local birefringence , so they propagate at different velocities and the initial packet broadens . when a second fiber of opposite - sign dispersion is matched with the first one , a time - reversal of the frequency components occurs and the chromatic dispersion cancels out . a general description of an open quantum system dynamical decoupling through open - loop control was given by viola _ _ in ref . @xcite . assuming a quantum system , @xmath5 , described by the hamiltonian @xmath6 , plunged in an environment described by the hamiltonian @xmath7 , the total hamiltonian is given by : @xmath8 where @xmath9 represents the coupling between system and environment . the open - system properties like , e.g. , the decoherence time , can be modified by adding to the system hamiltonian a time - varying cyclic control , @xmath10 , with a period @xmath11 such that @xmath12 in this way the resulting dynamics of the quantum system , usually called `` stroboscopic dynamics '' , is described by an effective hamiltonian @xmath13 . if the period @xmath11 is sufficiently shorter than the memory time of the environment , @xmath14 , the effective hamiltonian can be represented by the lowest - order average hamiltonian @xcite : @xmath15\,.\ ] ] corrections at higher orders in @xmath11 can be systematically evaluated . the simplest way of realizing a quantum system dynamical decoupling is through the quantum bb control , which operates by repeatedly applying instantaneous , arbitrarily strong , control pulses ( `` kicks '' ) to a qubit , thus disrupting the environmental interaction . this allows to convert the average over the cycle into a group - theoreric average , where a discrete set of unitary control operations , @xmath16 , acting on the quantum system only , are introduced @xcite : @xmath17 intuitively the bb symmetrizes the evolution with respect to the group @xmath18 , and the system is perfectly decoupled from the environment when the interaction with the environment is symmetrized to zero . . ] fig.[fig : bb_decouplinggroup ] provides a graphical representation of the bb idea . the quantum system enters from the left side and repeatedly undergoes the periodic transformations @xmath19@xmath20 . after each complete cycle in the group @xmath18 , the properties of the quantum system are restored to their initial value ( up to corrections at the second order and higher in @xmath21 ) . this allows , in principle , to indefinitely decouple the system from the environment . at this point it is straightforward to figure out how the bb control can be adapted to the optical domain . a light pulse travels in a free - space portion of space or in an optical fiber and passes through a series of passive optical elements like waveplates or phase - shifters , disposed at regular intervals along the channel , in analogy to the schematics of fig.[fig : bb_decouplinggroup ] . given the small thickness of the optical elements , the time employed by the light to pass through them is extremely small and well approximates the bb condition of instantaneous interaction . in ref . @xcite it is showed how to inhibit decoherence through the application of shaped time - varying controls designed to be equivalent to the application of the parity operator on the system . the decoupling is realized by a sequence of very frequent `` parity kicks '' ( pk ) of duration @xmath22 after a free evolution of duration @xmath23 , @xmath24 . assuming that the external pulse is so strong that it is possible to neglect the free evolution during the pulse @xmath22 , and that the pulse hamiltonian @xmath25 and the pulse width @xmath22 can be chosen to satisfy the pk condition @xmath26 the stroboscopic dynamics of the system inhibits decoherence whenever the following general conditions hold : @xmath27 this means that the system hamiltonian is parity invariant and the interaction with the external environment anticommutes with the system parity operator . decoupling is then obtained through symmetrization with respect to the group @xmath28 so that @xmath29 . a harmonic oscillator with linear dissipation @xmath30 satisfies such conditions , and the bb control is obtained by shifting the oscillation frequency @xmath31 of @xmath32 during the time @xmath22 , under the constraint @xmath33 , for which @xmath34 : @xmath35 pk control has a strong relationship with the spin - echo technique , because in both cases the applied pulses realize a time reversal of the unwanted hamiltonian interaction and cancel its effect . we consider now a particular system - environment interaction in which the decoherence occurs along a preferred axis , e.g. , along @xmath3 . in the standard spin - boson model this is described by the following interaction hamiltonian : @xmath37 such an interaction can describe the above - mentioned pure dephasing process related to @xmath2 in nmr @xcite , when the magnetic field is aligned in the @xmath3 direction , or the birefringence experienced by a light pulse traveling in a polarization - maintaining ( pm ) optical fiber with the fast axis parallel to the @xmath3 direction . the `` fast '' and `` slow '' axes are the pm fiber eigenmodes . the pulses linearly polarized along the fast axis have a group velocity higher than those polarized along the slow axis . this implies that a pulse which is not polarized along one of these directions is split in two separate time bins with orthogonal polarizations . such an effect must not be confused with the polarization - mode dispersion ( pmd ) of a single - mode ( sm ) optical fiber , which is mainly related to a non - perfect symmetry of the fiber core . it is possible to use bb controls to reduce the decoherence induced by the interaction hamiltonian of eq.[eq : carr - purcell_interaction ] . following ref . @xcite a dynamical bb decoupling is realized by the repetition of only two steps : i ) the free evolution of the system for an interval @xmath23 represented by the unitary @xmath38 ; ii ) the flip of the qubit around the @xmath39-axis through the operator @xmath40 . after two repetitions of these basic steps the system dynamics is driven by the following sequence of operators : @xmath41 where the first equality comes from the relation @xmath42 and where @xmath43 , @xmath44 . the above sequence @xmath45 defines a particular case of dynamical decoupling called `` carr purcell decoupling '' ( cp ) . when the interaction term is like in eq . , i.e. , when there is a preferred basis for decoherence to occur ( the `` pointer basis '' @xcite ) then the cp alone is sufficient to decouple the system from the environment . more generally , the cp works perfectly when the noise acts along any axis orthogonal to the direction of the flip . intuitively the spin - flip put forward by the @xmath46 operation changes the sign of the interaction hamiltonian , eq . , which is then symmetrized with respect to the group @xmath47 this averages out the detrimental influence of the environment . it is worthwhile to mention that the cp decoupling can be easily applied to pm fibers . it suffices to connect two pm fibers of equal length in a cross configuration , so that the fast axis of the first fiber is aligned to the slow axis of the second . the cross - connector executes an operation analogous to the @xmath40-flip just described , thus removing the difference in the group velocities due to the fiber birefringence . in the spin - boson model the most general interaction of a qubit with its environment is given by @xmath48 decoherence acts along the axis parallel to the effective magnetic field @xmath49 , which is however generally unknown . therefore the simple cp - sequence is no more effective . nonetheless decoupling can still be obtained through symmetrization with respect to the complete _ pauli group _ pg : @xmath50 with @xmath51 and @xmath52 @xcite . this can be implemented using only two spin - flips around two orthogonal axes , @xmath46 and @xmath53 , because , apart an overall phase , we can write : @xmath54 with pg is possible to decouple the system from the environment , whatever the direction of the interaction hamiltonian , up to the first order in @xmath55 @xcite . however it is worth to notice that more general decoupling schemes are possible which achieve decoupling at higher orders in @xmath55 @xcite . fig.[fig : bb_decouplingexamples ] summarizes the open - loop decoupling techniques described so far , in the bb approximation of instantaneous control pulses . the inset ( a ) and ( b ) of the figure contain the schematics of pk and cp decoupling techniques respectively ; their use in the optical domain has already been discussed . the inset ( c ) represents the most general decoupling scheme , the pg . it becomes particularly important when the preferred axis of the system - environment interaction is unknown , e.g. , in an sm optical fiber . in fact , it is possible to segment an sm fiber into smaller and smaller portions until a unity of length @xmath56 is reached which features a constant birefringence ( along unknown axes ) and so constant dispersive properties . regardless of the direction of the birefringence , pg decoupling can be applied on a length scale comparable to @xmath56 to suppress decoherence on each segment of the sm fiber and , by consequence , on the fiber as a whole . all of the described decoupling scheme , in particular the pg , are suitable for a straightforward application in the optical domain . photons are ideal carriers of quantum information as they allow a fast transfer of information while interacting weakly with the surroundings . however in the practice photons are replaced by light pulses which are intrinsically non - monochromatic and are measured by detectors which do not possess an infinite integration time . hence , a vast part of the information committed to the frequency degree of freedom is traced out by the final measuring process , which is not able to discriminate each frequency component of the light pulse . a solution to this problem is to encode information in the polarization degree of freedom . however , when the light propagates in a dispersive medium , such as an optical fiber , the initial information can spread out to other unwanted degrees of freedom , like frequency , and again that information will go lost . so in this case the frequency degree of freedom acts as the environment of the polarization qubit and induces a dephasing process on it . this happens because the dispersive medium couples the two degrees of freedom with an interaction hamiltonian like the following : @xmath57 where @xmath58 is the frequency - dependent refractive index of the medium and @xmath59 are the generators of the polarization - qubit evolution . the above hamiltonian descends from the fourier transform of the one describing a light pulse propagating in an inhomogeneous medium , i.e. with an index of refraction that depends on the position of the pulse . so it is closely related to the above - mentioned pmd occurring in sm optical fibers . the similarity with eq . is apparent ; this means that pg can be employed to adverse the polarization decoherence due to pmd . in a classical communication scenario pmd adversely affect the rate of long distance communications @xcite . pmd is also substantial in a fiber - based quantum communication if one uses a qubit realized in polarization since the unavoidable coupling with the birefringent environment affect the purity and fidelity of transmitted qubits @xcite there are not many techniques to combat the pmd detrimental effects . one is of course to adopt a pm fiber , but it works only for the two linear polarizations aligned to the fiber eigenmodes . another technological solution is to spin the sm fiber so to distribute inhomogeneities symmetrically in the core @xcite ; however this increases the fiber attenuation more than 20 times for wavelengths in the third telecom window . finally , one solution is based on faraday mirrors @xcite . the principle is that of `` retracing beam '' : a light pulse travels back and forth along the communication channel ; this leads it to compensate during the backward travel all the pmd encountered in the forward travel . the only assumption in this case is that the pmd remains constant during the two trips of the pulse @xcite . this technique is clearly not applicable to all those channels where the information carriers travel in the forward direction only . given the lack of conclusive solutions to adverse the pmd on one - way communication channels , it is quite surprising that bb dynamical decoupling has not been already applied to the optical domain . in fact specific experiments on this subject are missing . the remarkable experiments reported in @xcite can not be deemed an exception in this respect , since they partly rely on the principle of retracing beam ; so their application appears to be more suitable for two - way communication channels . in the following we focus on the recent demonstration given in @xcite and show the effectiveness of the bb dynamical decoupling for suppressing the decoherence of a polarization qubit confined in a _ ring cavity_. the choice of a ring cavity has a twofold motivation : on one side it realizes the hamiltonian of eq . , which well mimic the effect of pmd in a dispersive medium . the second reason is that a ring cavity represents the perfect simulation of a one - way channel : the principle of retracing beam is removed as its root because errors tend to accumulate rather than to cancel out . the experimental apparatus of ref.@xcite is depicted in fig . [ fig : bb_setup_onlycavity ] . a laser diode with central wavelength at @xmath60 nm and bandwidth @xmath61 nm is pulsed at repetition rate of 100 khz and pulse duration @xmath62100 ps . the laser is attenuated and injected in a triangular ring cavity through a spherical mirror . ) and a quarter - waveplate ( @xmath63 ) . the laser pulses are attenuated and injected in a triangular - ring cavity constituted by a spherical mirror and two plane mirrors at 45@xmath64 . the pulses trapped in the cavity are extracted by means of a glass plate and sent into the state tomography apparatus constituted by a pbs , @xmath65 and @xmath63 . the light is finally coupled to an optical fiber and detected by a single - photon avalanche photodiode . ] the polarization state of the laser is prepared by using a polarizing beam - splitter ( pbs ) , a half - wave plate ( @xmath65 ) and a quarter - wave plate ( @xmath63 ) with a very high degree of purity . at every round trip the light is extracted from the cavity with @xmath66 probability using a @xmath62100 @xmath67 m thin glass plate . its polarization is analyzed by means of the tomographic technique @xcite , and then sent into a multimode fiber connected to a single - photon detector with quantum efficiency @xmath68 . let us now separately discuss the generation , processing , and measurement of the qubit . the attenuated pulse entering the cavity through the input spherical mirror and after the state preparation stage can be well described as a multi - mode weak coherent state , which can be written as @xmath69\right\}|vac\rangle\,.\,\,\end{aligned}\ ] ] @xmath70 is the operator creating a photon with polarization @xmath71 and frequency @xmath72 , satisfying the commutation relation @xmath73=\delta_{s , s'}\delta(\omega-\omega')$ ] , and @xmath74 denotes the vacuum state . we denote with @xmath75 ( @xmath76 ) the polarization orthogonal ( parallel ) to the plane of the cavity . the polarization of the pulse is fully characterized by its stokes parameters , which are determined by the functions @xmath77 and @xmath78 according to @xmath79 , \\ s_1 \!\!&=&\!\!\ ! \int d\omega \left[|\alpha_h(\omega)|^2- |\alpha_v(\omega)|^2\right ] , \\ s_2 \!\!&=&\!\!\ ! \int d\omega \left[\alpha_h(\omega)^*\alpha_v(\omega)+\alpha_h(\omega ) \alpha_v(\omega)^*\right],\\ s_3 \!\!&=&\!\!\ ! -i\int d\omega \left[\alpha_h(\omega)^*\alpha_v(\omega)-\alpha_h(\omega ) \alpha_v(\omega)^*\right ] . \label{eq : stokes3}\end{aligned}\ ] ] one can associate to the full state of the field of eq . ( [ eq : input - weak - cohe - state ] ) a polarization state described by the @xmath80 qubit density matrix @xmath81 where @xmath82 is the vector of the three pauli matrices , and @xmath83 is the bloch vector , completely characterizing the state of the polarization qubit . the stokes parameters of eqs . ( [ eq : stokes0])-([eq : stokes3 ] ) satisfy the condition @xmath84 , which is equivalent to the condition on the bloch vector norm @xmath85 , which is required by the fact that @xmath86 must describe a physical density matrix . this condition can be verified using the fact that eqs . ( [ eq : stokes0])-([eq : stokes3 ] ) imply @xmath87 notice that this latter equation shows that the inequality becomes an equality , i.e. , @xmath88 and the polarization state is a pure qubit state , if and only if @xmath89 , @xmath90 , which is realized if and only if the ratio @xmath91 does not depend upon @xmath72 . this is verified when the polarization state of the pulse is independent of its frequency distribution , i.e. , polarization and frequency are factorized and one can write @xmath92 ( @xmath93 ) . therefore for the multi - mode coherent state of eq . ( [ eq : input - weak - cohe - state ] ) , the polarization state @xmath86 is pure if and only if is `` disentangled '' from the frequency @xmath94 ( @xmath93 ) , with @xmath95 the amplitude spectrum of the pulse normalized such that @xmath96 on the contrary , if @xmath91 depends upon @xmath72 , the jones vector is frequency - dependent , and @xmath86 is mixed : polarization and frequency are effectively `` entangled '' . the weak coherent pulse is then injected into the cavity , where it is transformed by the action of mirrors , which are linear and passive elements and therefore can only mix @xmath97 and @xmath98 , without involving processes like creation or annihilations of photons . also the additional elements which will be later added within the cavity for the realization of the bb control , i.e. , wave - plates and soleil babinet(s - b ) compensators , are linear and passive devices , and therefore the action of a generic element within the cavity can always be described in terms of a @xmath99 unitary matrix form . the overall transformation for each round - trip in the cavity is given by : @xmath100 \rightarrow \mathbf{u}(\omega ) \left [ \begin{array}{c } \hat{a}_h(\omega)^{\dagger}\\ \hat{a}_v(\omega)^{\dagger } \end{array } \right]\,,\ ] ] and the coherent state pulse is changed into @xmath101\right\}|vac\rangle\,,\,\end{aligned}\ ] ] where @xmath102 the stokes parameters and consequently the polarization state is changed accordingly . the final attenuator ( in the experimental setup the glass plate ) , @xmath103 , decreases the intensity of the pulse , @xmath104 with @xmath105 . assuming @xmath106 , the state of the pulse at the cavity output and just at the entrance of the detection stage can be well approximated by the first order linear expansion of the exponential in eq . ( [ expa ] ) @xmath107|vac\rangle\ , . \nonumber\end{aligned}\ ] ] by post - selecting the events with at least one photon we discard the vacuum component and the state entering the detector is well approximated by a single - photon state given by the integral term on the right hand side of eq . ( [ lin ] ) . one issue that should be addressed is the use of the term `` polarization qubit '' for the description of the laser pulse traveling in the ring cavity just described . if the frequency distribution is independent of the polarization , one can write @xmath108 ( @xmath93 ) , and factorize the frequency and polarization degree of freedom @xmath109.\ ] ] if instead the polarization state is not independent of the frequency , the two degrees of freedom are entangled and one has decoherence when one looks at the polarization state only . what is relevant is that the bloch vector @xmath110 and therefore the polarization state is not changed by the extraction by the glass plate , and , obviously , by the post - selection as well . therefore the output single photon state , `` operational '' polarization qubit , perfectly describes the polarization state within the cavity , even if we have a coherent state pulse within the optical system with a non - zero probability of having more than one photon . in our experiment the state emerging from the laser is a coherent state with average photon number per pulse about equal to 1 soon after the mirror . the resulting coherent state entering the detection apparatus has a mean photon number per pulse @xmath111 , and the probability of having two or more photons in a detection event is less than 2% , as we have experimentally verified . the experimental setup detailed above then justifies a description in terms of single - photon polarization qubits , and allows us to write the state injected into the cavity , soon after the spherical mirror , as an effective single - photon state @xmath112 where @xmath113|0\rangle$ ] . the triangular ring cavity ( see fig . [ fig : bb_cavity_z ] ) is realized by a spherical mirror with radius of curvature 1 m and reflectivity @xmath62 98% . the cavity is also formed by two flat mirrors at @xmath114 with reflectivity @xmath115 99% . the aperture angle of the cavity at the spherical mirror is @xmath116 . the long arms of the cavity are @xmath117 m and the short arm is @xmath118 m determining a total cavity length of @xmath119 m. , which realize on the polarization a unitary operation , @xmath0 , and a frequency dependent unitary operations , @xmath120 , respectively . a wave - plate inserted in a long arm of the cavity implements the @xmath0 operation in order to compensate for the spherical mirror @xmath0 transformation . ] a cavity mirror generally produces a phase shift which depends upon both frequency and polarization , and therefore its action is described by the matrix @xcite ( keeping always the right - hand coordinate system before and after reflection @xcite ) @xmath121 = \mathbb{z}\exp\{-i\phi(\omega)\mathbb{z}\},\ ] ] where we have omitted an unessential global phase factor and we have defined with @xmath122 the relative phase due to the polarization - reflectivity difference . this is due to the different reflection coefficients for orthogonal and parallel polarizations with respect to the plane of the cavity ( s- and p- polarization , respectively ) . the two plane mirrors at 45@xmath64 can be assumed identical and are therefore characterized by the same matrix @xmath120 . the third concave mirror of the cavity is almost at normal incidence : this mirror does not distinguish @xmath76 and @xmath75 polarization , so that @xmath123 and therefore it acts as @xmath124 . the polarization qubit in the cavity is subjected to a phase - noise along @xmath3-axis and the unitary operator describing the polarization transformation after one cavity round - trip for a given frequency component is given by @xmath125 & = & \mathbb{z}\cdot\mathbb{m_z}(\omega)\cdot\mathbb{m_z}(\omega)\nonumber\\ & = & \mathbb{z}\exp\{-i2\phi(\omega)\mathbb{z}\}\,,\end{aligned}\ ] ] the dispersive properties of the two plane mirrors at @xmath114 , i.e. , their frequency dependence , generates polarization decoherence : in fact the frequency dependence of the matrix @xmath126 appears only if the optical element is dispersive , otherwise it is constant and it does not entangle polarization and frequency degrees of freedom . at every round trip the light is extracted from the cavity using the thin glass plate as shown in the setup of figs . [ fig : bb_setup_onlycavity ] and [ fig : bb_cavity_z ] . the output polarization state after @xmath127 round - trips is given by the reduced density matrix obtained by tracing over the frequency degree of freedom @xmath128^n |\pi\rangle_{in}\langle\pi| \mathbb{u}^\dag[\phi(\omega)]^n\,.\ ] ] experimentally the output density matrix after @xmath127 round - trips has been evaluated by the acquisition of the output signal from the detector , which stops the time conversion in a time - to - amplitude converter ( tac ) synchronized with the laser . the time delay between the input and output signals is then recorded by a multi - channel analyzer ( mca ) with a resolution of 8192 channels . the acquisition electronics has a time - resolution of @xmath129 ps . typical acquisition runs are shown in fig . [ fig : plot_hdr_log ] , showing a sequence of peaks , each corresponding to a cavity round - trip . the interval between adjacent peaks amounts to @xmath626.80 ns , in agreement with the measured length of the cavity . in all the experiments both the decoherence and the effect of bb control are studied by sending specific input polarization in the cavity and capturing the photon round trip from the cavity with respect to different tomographic measurement bases . [ fig : plot_hdr_log ] depicts the photon round trips for _ horizontal _ ( h ) , _ diagonal _ ( d ) , and _ right _ ( r ) polarization with respect to @xmath0 , @xmath40 and @xmath130 basis tomographic measurement . the decay of peaks after every round trip for h polarization input state is highly uniform with respect to all basis measurements but for d and r polarizations it is highly non - uniform , indicate that the eigenstates of the cavity unitary evolution are h and v _ vertical _ polarizations . , @xmath40 and @xmath130 , respectively . it is worth to note that for the horizontal polarization input state the vertical contribution is negligible and the outputs in the @xmath40 and @xmath130 are balanced , as expected for an eigenstate of the cavity unitary evolution . ] this is made evident by the time evolution of the stokes parameters for d and r input polarization rates , as shown by the left panels in fig . [ fig : bb_svector_compare_onlycav_longz ] . the stokes parameters are calculated as expectation values of the pauli matrices ( @xmath131 ) over the density matrices evaluated by maximum - likehood optimization procedure @xcite on the six experimental counts obtained by integration of ten time bins around each peak for the three polarization measurement bases . wave - plate in a long arm in order to compensate for the spherical mirror @xmath0 transformation , as shown in fig . [ fig : bb_cavity_z ] . note how the wave - plate inserted in a long arm of the cavity cancels the polarization - flip for diagonal and right polarization input states due to the spherical mirror transformation . the wave - plate has no effect instead in the case of a horizontal polarization input state , which is an eigenstate of the cavity transformation in both cases . the lines following the experimental data are evaluated by means of a numerical simulation with two fitting parameters : the standard deviation of the gaussian measure @xmath132 , and the phase difference for horizontal and vertical polarization , @xmath133 . ] a numerical simulation of eq . has been run assuming a gaussian frequency spectrum for the input pulse given by @xmath134{\pi\sigma^2_\omega}}\exp[-(\omega-\omega_0)^2/2\sigma_\omega^2],\ ] ] where @xmath135 represents the bandwidth of the radiation spectrum centered in @xmath31 . the frequency dependence of the integrand is determined by the fact that , with a very good approximation , the phase shift is a linear function of the frequency , @xmath136 . the integral can be therefore transformed into an average over a gaussian measure , @xmath137 , with standard deviation @xmath138 . the linear passive optical elements constituting the cavity ( mirrors and wave - plates ) have been modeled by @xmath139 unitary matrices , and the @xmath127-th round - trip polarization density matrix has been evaluated by the average over a monte - carlo simulation of the measure @xmath137 of the @xmath127-th power of the matrix describing the effect of a cavity round - trip . the lines fitting the experimental data in fig . [ fig : bb_svector_compare_onlycav_longz ] and following , are the numerical fit with two parameters : the standard deviation of the gaussian measure @xmath140 , and the phase difference between horizontal and vertical polarization , @xmath141 . in order to understand the polarization decoherence caused by cavity round - trips , we have first studied the _ cavity only _ configuration as shown in figs . [ fig : bb_setup_onlycavity ] and [ fig : bb_cavity_z ] . to quantify the decoherence , we use both purity @xmath142 and fidelity @xmath143 where @xmath144 is the output density matrix and @xmath145 is the pure input polarization state . purity @xmath146 is maximum and equal to one for pure states and to @xmath147 for maximally mixed states of dimension @xmath148 . fidelity @xmath149 measures the distance between quantum states and the definition of eq . ( [ eq : fid_def ] ) is valid in the case of the distance between a mixed and a pure state . for two density matrices @xmath150 , @xmath151 it is generalized to @xmath152 @xcite . however , one can adopt the alternative definition @xmath153 @xcite , sometimes denoted as @xmath154 and called square root fidelity . in a long arm in order to compensate for the spherical mirror @xmath0 transformation , as shown in fig . [ fig : bb_cavity_z ] . _ horizontal _ polarization is well preserved . on the contrary the _ diagonal _ and _ right _ states quickly decay to the fully mixed state ( purity and fidelity equal to @xmath155 ) . dashed lines curves are obtained by numerical simulation with the same parameters of the previous figure . ] the purity and fidelity after each cavity round trip , evaluated from the experimentally reconstructed density matrices , are shown on the left panels in fig . [ fig : bb_purityfidelity_compare_onlycav_longz ] . as expected , the cavity preserves the purity and fidelity only for h polarization input , while for d and r input states , they both quickly decay to the fully mixed state ( @xmath156 ) . assuming the amplitude spectrum of eq . , it is possible to derive analytically a gaussian decay of the purity as a function of the number of round - trips @xmath127 , @xmath157.\ ] ] the standard deviation of the gaussian distribution of the phase - shift , @xmath158 , is estimated by a best fit on the experimental data of the purity reported in the left panel of fig . [ fig : bb_purityfidelity_compare_onlycav_longz ] using the expression of eq . . the obtained value agrees with the estimated parameter from the numerical simulation and it is given by @xmath159 . differently from the purity , the fidelity depends upon both @xmath158 and the parameter @xmath141 , which is the phase difference between h and v polarization reflection at the central frequency @xmath31 , as highlighted by the slowly - varying envelope of the fidelity shown in fig . [ fig : bb_purityfidelity_compare_onlycav_longz ] . the flips of the fidelity every round - trip for d and r input polarization state , shown in the left panel of fig . [ fig : bb_purityfidelity_compare_onlycav_longz ] , are instead due to the spherical mirror transformation @xmath0 . the dashed lines following the purity and fidelity data reported in fig . [ fig : bb_purityfidelity_compare_onlycav_longz ] represent the numerical fit corresponding to the values @xmath160 and @xmath161 . in a single round trip the polarization qubit encounters two decohering elements ( the two plane mirrors represented by @xmath162 ) , and a @xmath0 transformation due to the spherical mirror . as shown by eq . , the transformation of the polarization qubit realized by a cavity round - trip is given by the product of a phase - shift @xmath163 , with gaussian - distributed random phase , and a @xmath0 operation , which flips every round trip the d and r polarization , as elucidated by the data reported in the left panels of figs . [ fig : bb_svector_compare_onlycav_longz ] , and [ fig : bb_purityfidelity_compare_onlycav_longz ] . in order to compensate these @xmath0-flips , we have inserted in one of the long arms of the cavity , before the spherical mirror , an additional wave - plate set to realize a further @xmath0 operation , as shown in fig . [ fig : bb_cavity_z ] . with this new optical element in the setup , the evolution of the polarization state after a round - trip is given by @xmath164 & = & \mathbb{z}\cdot\mathbb{z}\cdot\mathbb{m_z}(\omega)\cdot\mathbb{m_z}(\omega)\nonumber\\ & = & \exp\{-i2\phi(\omega)\mathbb{z}\}\,.\end{aligned}\ ] ] the behavior of purity and fidelity in this new configuration is shown in the right panels of fig . [ fig : bb_purityfidelity_compare_onlycav_longz ] . the flips in the fidelity decay are now completely eliminated and only the slowly - varying envelop due to @xmath141 is left . as shown in the right panels of fig . [ fig : bb_svector_compare_onlycav_longz ] , also the flips in the stokes parameters for d and r input polarizations are eliminated in the new cavity configuration . here we experimentally demonstrate the application of bb dynamical decoupling for suppressing decoherence on single - photon polarization qubits flying in the cavity . polarization decoherence takes place in localized spatial regions ( mirrors , s b , ... ) , and the bb controls will be implemented in space rather than in time by placing wave - plates along the photon path . the bb pauli group dynamical decoupling is realized by adding two control operations within the cavity : i ) a second @xmath165 wave - plate in the long arm of the cavity in addition to the one used for compensating the spherical mirror ; ii ) a s b with axis at 45@xmath64 with respect to the cavity plane and delay equal to @xmath65 in the short arm of the cavity , acting therefore as @xmath40 ( see fig . [ fig : bb_setup_cavity_shortx_longzz_2 ] ) . wave - plate is inserted in the long arm of the cavity , while a s b with axis at 45@xmath64 with respect to the cavity plane and delay equal to @xmath65 is placed in the short arm of the cavity , acting therefore as @xmath40 . ] the two controls implement every two cavity round - trips the full pauli - group decoupling for a polarization qubit @xcite , and the overall transformation for a double round - trip is given by @xmath166 & = & \left [ \mathbb{z}\mathbb{m_z}\mathbb{x}\mathbb{m_z } \right]\cdot \left [ \mathbb{z}\mathbb{m_z}\mathbb{x}\mathbb{m_z } \right ] \nonumber\\ & = & [ \mathbb{z}\mathbb{x}]\ , [ \mathbb{z}\mathbb{x } ] = ( i\mathbb{y})^2 = -\mathbb{i}\ , , \label{eq : cavitywithx_2}\end{aligned}\ ] ] where we have omitted the frequency dependence of @xmath120 . for compensating the mirror transformation ; right ) cavity with a @xmath40 in the short arm , and a second @xmath0 , as shown in fig . [ fig : bb_cavity_x ] , for implementing the pauli - group decoupling on the polarization - qubits . the dashed lines represent numerical simulation with the same parameters of the previous figures . ( adapted from ref . @xcite ) . ] therefore the polarization transformation is proportional to the identity operator , thus implying a _ perfect _ preservation of every input polarization state , i.e. , a complete suppression of decoherence . this is experimentally verified for h , d and r input polarizations in fig . [ fig : bb_purfid_generalpauli ] , where the purity @xmath146 and the fidelity @xmath149 of the output polarization state are plotted versus the number of double round - trips . as we have seen in the previous section , when bb is not performed ( left panels of fig . [ fig : bb_purfid_generalpauli ] ) , polarizations d and r decay to the fully unpolarized state ( @xmath167 ) , while h , being an eigenstate of @xmath0 , is unaffected by decoherence . on the contrary , polarization decoherence is completely suppressed when bb is applied ( right panels of fig . [ fig : bb_purfid_generalpauli ] ) . polarization decoherence caused by a cavity round - trip acts along a known axis of the bloch sphere , the @xmath3-axis , and therefore even the simplest carr - purcell decoupling scheme , based on the decoupling group @xmath168 , suffices for suppressing decoherence . we have verified this fact by inserting _ only _ the @xmath40 operation in the short arm of the cavity , by properly adjusting the s b ( see fig . [ fig : bb_cavity_x ] ) . operation in order to realize the carr - purcell decoupling scheme for polarization - qubit . ] the carr - purcell decoupling scheme requires again a double round - trip in the cavity , which gives the evolution @xmath169 = \mathbb{z}\mathbb{m_z } \cdot\mathbb{x}\cdot \mathbb{m_z}\mathbb{z}\mathbb{m_z } \cdot\mathbb{x}\cdot\mathbb{m_z } = -\mathbb{i}\,.\ ] ] the polarization transformation is again proportional to the identity operator , implying a perfect preservation of every input polarization state . in fact , when the effective noise acts along a _ known _ axis , a complete suppression of decoherence is already obtained by employing only _ one _ pauli operator , the one realizing a spin - flip around an axis _ orthogonal _ to the axis of decoherence . this is very well shown in fig . [ fig : bb_purityfidelitydouble_compare_longz_shortx ] where , starting from input polarization states h , d and r , the purity @xmath146 and the fidelity @xmath149 of the output polarization state , with ( right panels ) and without ( left panels ) carr - purcell decoupling are plotted versus the number of double round - trips . in the short arm , as shown in fig . [ fig : bb_cavity_x ] , in order to realize the carr - purcell decoupling on polarization - qubits affected only by phase - noise along the @xmath3-axis . the dashed lines represent numerical simulation with the same parameters of the previous figures . ] in the ring - cavity setup , the v and h polarization states are eigenstates of the interaction hamiltonian of the polarization qubit with its effective environment and are therefore the `` pointer states '' unaffected by decoherence @xcite . in the pointer states basis , decoherence affects only the off - diagonal elements of the polarization density matrix . in the most general case instead , the direction of the pointer state basis in the bloch poincar sphere is _ unknown_. as a consequence in a generic basis , decoherence acts _ both _ on diagonal and off - diagonal elements of the polarization density matrix . to implement such a generic noise model we have placed in front of each plane mirror a s b with axis at 45@xmath64 with respect to the cavity plane ( see fig . [ fig : bb_cavity_genericnoise ] ) . the action of the s b on the polarization state is described by @xmath170 = \exp[-{\rm i}\theta \mathbb{x}/2]\,,\ ] ] where @xmath171 , and @xmath172 is the noise delay - phase . the s - b together with a plane mirror are described by the operator @xmath173 the transformation of the polarization state after two round - trips in the presence of the s bs is therefore given by @xmath174 = [ \mathbb{n}(\omega,\theta)\mathbb{n}(\omega,\theta)]\cdot [ \mathbb{n}(\omega,\theta)\mathbb{n}(\omega,\theta)]\,.\ ] ] the free evolution ( @xmath175 ) operator @xmath176 $ ] can be rewritten as @xmath174 = \exp\left[-i \alpha_{fe}(\omega,\theta)\,\vec{s}_{fe}(\omega,\theta)\cdot \vec{\sigma}\right]\,,\ ] ] which describes a rotation in the bloch poincar sphere of an angle @xmath177 around the direction individuated by @xmath178 . therefore by varying @xmath172 and the bandwidth of radiation spectrum , i.e. the distribution of @xmath179 , one implements the generic polarization decoherence . the rotation angle is given by the implicit expression @xmath180=\sin [ \phi(\omega)/2]\cos(\theta/2),\ ] ] and @xmath181\ , , \nonumber\\ \!\!\!&&\!\!\!\sin\theta\sin\phi(\omega)\,,\nonumber\\ \!\!\!&&\!\!\!(1+\cos\theta)\sin\phi(\omega)\,\}.\end{aligned}\ ] ] pauli - group decoupling is again realized every two - round trips by adding the bb operations @xmath40 ( implemented by adding to the noise phase - delay @xmath172 of the s b present in the short arm of the cavity , a further delay of @xmath65 ) and @xmath0 in the cavity @xcite as shown in fig . [ fig : bb_cavity_genericnoise ] . the overall transformation after the two round - trips for a given frequency component is now given by @xmath182 = [ \mathbb{z}\mathbb{n}(\omega,\theta)\mathbb{x}\mathbb{n}(\omega,\theta ) ] \cdot [ \mathbb{z}\mathbb{n}(\omega,\theta)\mathbb{x}\mathbb{n}(\omega,\theta)]\,,\ ] ] which can be rewritten , as in the free evolution case , as a rotation in the bloch poincar sphere , @xmath182 = \exp[-i\alpha_{bb}(\omega,\theta)\,\vec{s}_{bb}(\omega,\theta)\cdot \vec{\sigma}]\,,\ ] ] with @xmath183 and @xmath184\ , , \nonumber\\ \!\!\!&&\!\!\ ! -\sin\theta\sin^2[\phi(\omega)/2]\}/[\sin\alpha_{bb}(\omega,\theta)].\end{aligned}\ ] ] we have performed a first test of the ability of pauli group decoupling for controlling decoherence under generic noise , by injecting in the cavity an elliptical polarization state , corresponding to an equal - weighted superposition of v , a ( _ anti - diagonal _ ) and r polarizations ( see the experimentally reconstructed density matrix and bloch - poincar sphere representation in fig . [ fig : bb_genericnoise_3.0_half ] ) . in case of no bb control , the resulting fast polarization decay of the purity , as a function of the number of double round - trips , is shown in the left panel of fig . [ fig : bb_genericnoise_3.0_half ] , for different noise delay - phases @xmath172 . the recovery of purity after the application of bb is evident from the right panel of the fig . [ fig : bb_genericnoise_3.0_half ] . for an elliptical polarization input state . the lines represent numerical simulation with the parameters of the previous figures . inset : real and imaginary parts of the density matrix of the elliptical polarization input state , also shown on the bloch poincar sphere as the point labeled @xmath185 . ( adapted from ref . @xcite ) . ] the experiment with the input elliptical polarization state already shows the effectiveness of the bb pauli group decoupling for suppressing decoherence . however , in order to be useful for any quantum information application , the scheme must work for every input polarization . therefore , in order to better highlight the usefulness of bb decoupling to combat generic birefringence - induced decoherence , we have studied the evolution of the purity and the fidelity , averaging the input state over the whole bloch poincar sphere , both with ( right panels of fig . [ fig : bb_purfid_average_compare ] ) and without bb decoupling ( left panels of fig . [ fig : bb_purfid_average_compare ] ) . each curve corresponds to a different orientation of the noise delay - phase @xmath172 of the s - b . bb again inhibits decoherence because , for each orientation of the decoherence - axis , both the average purity and the average fidelity in the presence of bb are significantly higher than the corresponding value without bb . the curve for @xmath186 reproduces the almost perfect preservation of the previous experiment of sec . instead polarization protection progressively worsens for increasing values of @xmath172 . when @xmath172 is varied , the properties of the effective environment of the polarization qubit associated with the effective birefringence caused by the mirrors changes in a nontrivial way , but a simple analytical explanation of the result can be given when the pulses are not too broad in frequency . : left ) without bang - bang control : right ) when applied the bang - bang control . the lines represent numerical simulation with the parameters of the previous figures . ( adapted from ref . @xcite ) . ] an approximate analytical description of the evolution of the polarization state as it circles within the cavity can be given under the assumption that the pulse is not too broad in frequency , @xmath187 , which is well verified in the present experiment . in this case , @xmath188 and @xmath189 ( @xmath190 ) do not vary appreciably over the range of relevant phase shifts @xmath179 and one can approximate @xmath191 with its value at the pulse center @xmath192 , while @xmath193 can be approximated by its second - order expansion around @xmath141 , @xmath194 after lengthy but straightforward calculations , one gets for the output bloch vector @xmath195 , which determines completely the density matrix as expressed in eq . , @xmath196 the matrix @xmath197 is given by @xmath198 ( @xmath199 ) , where the `` decoherence factor '' @xmath200 is given by @xmath201^{-1/4 } \exp\left[-\frac{n^2 \dot{\alpha}_0 ^ 2 \sigma_{\phi}^2 } { 1+n^2 \ddot{\alpha}_0 ^ 2 \sigma_{\phi}^4}\right],\ ] ] and the matrix @xmath202 is the orthogonal matrix describing the rotation around the @xmath203-axis of an angle @xmath204 , with @xmath205 the output purity and fidelity are respectively given by @xmath206 \left[\vec{p}(\gamma_n)\cdot \vec{s}(\phi_0)\right]\right . \nonumber \\ \!\!\!&&\!\!\ ! \left.+{\cal d}_n^2 + ( 1-{\cal d}_n)^2\left[\vec{p}_{in}\cdot \vec{s}(\phi_0)\right]^2\right\}\,,\end{aligned}\ ] ] and @xmath207 ^ 2\right\ } \,,\nonumber\end{aligned}\ ] ] with @xmath208 . these expressions are general and apply both to the bb case as well to the one without bb . the two cases differ only in the explicit expressions of @xmath203 , @xmath209 , @xmath210 , and @xmath211 . since @xmath212 , fidelity and purity tend asymptotically to the same limit @xmath213 ^ 2\}\,.\ ] ] this asymptotic situation corresponds to the existence of a pointer basis @xcite ( analogous to the principal states of polarization in fibers @xcite ) , formed by the two states with bloch vector equal to @xmath214 , which are unaffected by decoherence . all the other polarization states instead decohere , with maximum decoherence for the states on the plane orthogonal to @xmath203 . as a consequence , when averaged over the initial state , purity and fidelity tends to @xmath215 . in the @xmath216 regime , bb and no - bb case differ only in the direction @xmath203 of the pointer basis and therefore bb is not advantageous with respect to the no - bb case in this regime . however , the interesting regime of our experiment is the one corresponding to a small round - trip number @xmath127 . in that regime our ring - cavity well mimics a portion of a one - way quantum communication channel with constant dispersive properties , like a small portion of an optical fiber . for small @xmath127 one has @xmath217 ^ 2 \{1-[\vec{p}_{in}\cdot \vec{s}_j(\phi_0)]^2\}/2\,,\end{aligned}\ ] ] with @xmath218 , showing that the smaller @xmath219 the better the decoherence suppression . bb pauli - group decoupling acts just by decreasing @xmath220 : by using eqs . ( [ eq : alphano])-([eq : implicitalph ] ) and the fact that @xmath141 is quite small in our experiment , one gets @xmath221 , @xmath222 , showing that it is always @xmath223 and therefore that bb better preserves the polarization qubit for any orientation of the decoherence axis , confirming its applicability to otpical fibers polarization control . these expressions also explain the perfect preservation of the polarization qubit of fig . [ fig : bb_purfid_generalpauli ] : the latter refers to @xmath186 , implying @xmath224 and @xmath225 . we have presented the details of the experiment of ref . @xcite , which showed how bb dynamical decoupling can be efficiently implemented in the optical domain in order to suppress the polarization decoherence caused by the propagation of light pulses within birefringent media . in such media , the optical properties depend upon both frequency and polarization and the frequency degree of freedom acts as an effective environment on the photon polarization . with photons bb dynamical decoupling can be realized by placing appropriate optical elements ( i.e. , wave - plates ) implementing suitable unitary operations along the photon path . in the present proof - of - principle demonstration , a one - way communication channel is mimicked by a ring - cavity , in which an effective birefringence is introduced by two mirrors at @xmath226 degrees whose reflectivity is responsible for the decay of the density matrix off - diagonal elements , in the basis formed by h and v polarizations . in a first experiment , bb dynamical decoupling was shown to be able to perfectly suppress this kind of decoherence . in a second experiment , we modified the setup by placing two soleil babinet compensators in front of the mirrors . this allowed us to simulate the most general model of decoherence , for which the pointer basis is unknown . also in this case the pauli group bb decoupling was able to significantly reduce decoherence , as predicted by @xcite . several elements suggest that our experiment is significant for the potential extension of bb decoupling to the dynamics of polarization pulses propagating in optical fibers . first , the interaction hamiltonian governing the dynamics in the ring cavity is strictly related to the one causing pmd in optical fibers @xcite . second , the one - way nature of the ring cavity well mimics a fiber - based communication channel . third , the optical elements introducing decoherence in the ring - cavity doubtless realize a worst - case version of the fiber dynamics , because the errors they introduce accumulate systematically while in a fiber they are distributed at random . we plane to experimentally investigate the optimal length for placing the bb operations along an sm fiber , obtained from a tradeoff between the two opposite requests of minimizing the attenuation of the fiber and maximizing the decoherence suppression . we acknowledge the funding of the ec project fp6-ip - qap . m.l . is supported by the @xmath227 grant c.f .

electron transfer ( et ) is a very important process in biology , chemistry and physics @xcite . the most well known et theory is the one of marcus @xcite . of special interest is the et in configurations where a bridge ( b ) between donor ( d ) and acceptor ( a ) mediates the transfer . on this kind of et we will focus in this paper . the primary step of et in bacterial photosynthetic reaction centers is of this type @xcite and a lot of work in this direction was done after the structure of the protein - pigment complex of the photosynthetic reaction centers of purple bacteria was clarified in 1984 @xcite . many artificial systems especially self - organized porphyrin complexes have been developed to model this bacterial photosynthetic reaction center @xcite . bridge - mediated et reactions can occur via different mechanisms @xcite : incoherent sequential transfer in which the bridge level is populated or coherent superexchange @xcite in which the mediating bridge level is not populated but nevertheless necessary for the transfer . changing a building block of the complex @xcite or changing the environment @xcite can modify which mechanism is mainly at work . actually , there is still a discussion in literature whether sequential transfer and superexchange are limiting cases of one process @xcite or whether they are two processes which can coexist @xcite . to clarify which mechanism is present in an artificial system one can systematically vary the energetics of the complex . in experiments this is done by substituting parts of the complexes @xcite or by changing the polarity of the solvent @xcite . also the geometry and size of the bridge block can be varied , and in this way the length of the subsystem through which the electron has to be transfered @xcite can be changed . superexchange occurs due to coherent mixing of the three or more states of the system @xcite . the et rate in this channel depends algebraically on the differences between the energy levels @xcite and decreases exponentially with increasing bridge length @xcite . when incoherent effects such as dephasing dominate the transfer is mainly sequential @xcite , i. e. , the levels are occupied mainly in sequential order @xcite . the dependence on the differences between the energy levels is exponential @xcite . an increase of the bridge length induces only a small reduction in the et rate @xcite . this is why sequential transfer is the desired process in molecular wires @xcite . in the superexchange case the dynamics is mainly hamiltonian and can be described on the basis of the schrdinger equation . the physically important results can be obtained by perturbation theory @xcite and , most successfully , by the semiclassical marcus theory @xcite . the complete system dynamics can be directly extracted by numerical diagonalization of the hamiltonian @xcite . in case of sequential transfer the influence of an environment has to be taken into account . there are quite a few different ways how to include an environment modeled by a heat bath . the simplest phenomenological descriptions are based on the einstein coefficients or on the imaginary terms in the hamiltonian @xcite , as well as on the fokker - planck or langevin equations @xcite . the most accurate but also numerically most expensive way is the path integral method @xcite . this has been applied to bridge - mediated et especially in the case of bacterial photosynthesis @xcite . bridge - mediated et has also been investigated using redfield theory @xcite , by propagating a density matrix ( dm ) in liouville space @xcite and other methods ( e. g. @xcite ) . the purpose of the present investigation is to present a simple , analytically solvable model based on the dm formalism @xcite and apply it to a porphyrin - quinone complex which is taken as a model system for the bacterial photosynthetic reaction center . the master equation which governs the dm evolution as well as the appropriate relaxation coefficients can be derived from such basic informations as system - environment coupling strength and spectral density of the environment @xcite . in the present model relaxation is introduced in a way similar to redfield theory but in site representation not in eigenstate representation . a discussion of advantages and disadvantages these representations has been given elsewhere @xcite . the equations for the dm are the same as in the generalized stochastic liouville equation ( gsle ) model @xcite for exciton transfer which is an extension of the haken - strobl - reineker ( hsr ) model @xcite to a model with a quantum bath . here we give an analytic solution to these equations . the present equations for the dm obtained are also similar to those of ref . @xcite where relaxation is introduced in a phenomenological fashion but only a steady - state solution is found in contrast to the model introduced here . in addition the present model is applied to a concrete system . a comparison of the et time with the bath correlation time allows us to regard three time intervals of system dynamics : the interval of memory effects , the dynamical interval , and the kinetic , long - time interval @xcite . in the framework of dm theory one can describe the et dynamics in all three time intervals . however , often it is enough to find the solution in the kinetic interval for the explanation of experiments within the time resolution of most experimental setups , as has been done in ref . the master equation is analytically solvable only for simple models , for example @xcite . most investigations are based on the numerical solution of this equation @xcite . here we perform numerical as well as approximate analytical calculations for a simple model . since the solution can be easily obtained , the influence of all parameters on the et can be examined . the paper is organized as follows . in the next section we introduce the model of a supermolecule which we use to describe et processes . the properties of an isolated supermolecule are modeled in subsection [ chem - isolate ] , as well as the static influence of the environment . the dynamical influence of bath fluctuations is discussed and modeled by a heat bath of harmonic oscillators in section [ chem - sb ] . the reduced dm equation of motion ( rdmem ) describing the excited state dynamics is presented in subsection [ chem - rdm ] . in subsection [ chem - scaling ] the system parameter dependence on the solvent dielectric constant is discussed for different models of solute - solvent interaction . in subsection [ chem - parameters ] system parameters are determined . the methods and results of the numerical and analytical solutions of the rdmem are presented in section [ chem - results ] . the dependencies of the et rate and final acceptor population on the system parameters are given for the numerical and analytical solutions in subsection [ chem - super ] . the analysis of the physical processes in the system is also performed there . in subsection [ chem - solvents ] we discuss the dependence of the et rate on the solvent dielectric constant for different models of solute - solvent interaction and compare the calculated et rates with the experimentally measured ones . the advantages and disadvantages of the presented method in comparison with the gsle model @xcite and the method of davis et al . @xcite are analyzed in subsection [ chem - davis ] . in the conclusions the achievements and possible extensions of this work are discussed . the photoinduced et in supermolecules consisting of three sequentially connected molecular blocks , i. e. , donor , bridge , and acceptor , ( @xmath4 ) is analyzed . the donor is not able to transfer its charge directly to acceptor because of their spatial separation . donor and acceptor can exchange their charges only through b. in the present investigation the supermolecule consists of free - base tetraphenylporphyrin ( @xmath1 ) as donor , zinc substituted tetraphenylporphyrin ( @xmath2 ) as bridge , and p - benzoquinone as acceptor @xcite . in each of those molecular blocks we consider only two molecular orbitals ( @xmath5 ) , the highest occupied molecular orbital ( homo ) and the lowest unoccupied molecular orbital ( lumo ) @xcite . each of these orbitals can be occupied by an electron or not , denoted by @xmath6 or @xmath7 , respectively . this model allows us to describe four states of each molecular block , the neutral ground state @xmath8 , the neutral excited state @xmath9 , the positively charged ionic state @xmath10 , and the negatively charged ionic state @xmath11 . @xmath12 , @xmath13 , and @xmath14 describe the creation , annihilation , and number of electrons in orbital @xmath15 , respectively , while @xmath16 gives the number of electrons in a molecular block . the number of particles in the whole supermolecule is conserved , @xmath17 . each of the electronic states has its own vibrational substructure . as a rule for the porphyrin - containing systems the time of vibrational relaxation is two orders of magnitude faster than the characteristic et time @xcite . because of this we assume that only the ground vibrational states play a role and we do not include the vibrational substructure . a comparison of the models with and without vibrational substructure has been given elsewhere @xcite . below we consider the evolution of single charge - transfer exciton states in the system . for the full description of the system one also should include photon modes to describe for example the fluorescence from the lumo to the homo in each molecular block transferring an excitation to the electro - magnetic field . but the rates of fluorescence and recombination are small in comparison to other processes for porphyrin - type systems @xcite . when fluorescence does not have to be taken into account , all states except @xmath18 ( @xmath19 ) , @xmath20 ( @xmath21 ) , and @xmath22 ( @xmath23 ) remain essentially unoccupied , while those three take part in the intermolecular transport process . in this case the number of states coincides with the number of sites in the system and we label the states @xmath24 instead of @xmath25 . for the description of the et and other dynamical processes in the system placed in a dissipative environment we introduce the hamiltonian @xmath26 where @xmath27 describes the supermolecule , @xmath28 the dissipative bath , and @xmath29 their interaction . we are interested in the kinetic limit of the excited state dynamics here . for this limit we assume that the relaxation of the solvent takes only a very short time compared to the system times of interest . the effect of the solvent is twofold . on one hand the system dynamics is perturbed by the solvent state fluctuations , independent of the system states . @xmath29 shall only reflect the dynamical influence of the fluctuations leading to dissipative processes as discussed in the next subsection . on the other hand the system states are shifted in energy @xcite , @xmath30 due to the static influence of the solvent which is determined by the relaxed value of the solvent polarization and in general also includes the non - electrostatic contributions such as van - der - waals attraction , short - range repulsion , and hydrogen bonding @xcite . in eq . ( [ 2-hami ] ) the energy of free and noninteracting blocks @xmath31 , is given by the energies @xmath32 of orbitals @xmath33 in the independent electron approximation @xcite . the @xmath32 are chosen to reproduce the ground - state excited - state transitions e. g. @xmath34 , which change only a little for different solvents @xcite and are assumed to be constants here . to determine @xmath32 one starts from fully ionized double bonds in each molecular block @xcite , calculates the one - particle states and fills these orbitals with two electrons each starting from the lowest energy . by exciting , removing , adding the last electron to the model system one obtains the energy of the excited , oxidized , reduced molecular block in the independent particle approximation . the inter - block hopping term @xmath35\ ] ] in eq . ( [ 2-hami ] ) includes the hopping operators @xmath36 and the coherent couplings @xmath37 . we assume @xmath38 because there is no direct connection between donor and acceptor . the scaling of @xmath37 for different solvents is discussed in subsection [ chem - scaling ] . the electrostatic interaction @xmath39 scales like energies of a system of charges in a single or multiple cavity surrounded by a medium with static dielectric constant @xmath40 according to the classical reaction field theory @xcite . here we consider two models of scaling . in the first model each molecular block is in an individual cavity in the dielectric . for this case the electrostatic energy reads @xmath41 @xmath42 takes the electron interaction into account while bringing an additional charge onto the block @xmath43 and thus describes the energy to create an isolated ion . this term depends on the characteristic radius @xmath44 of the molecular block . the interaction between the ions @xmath45 depends on the distance between the molecular blocks @xmath46 . both distances @xmath44 and @xmath46 are also used in marcus theory @xcite . the term @xmath47 reflects the interaction of charges inside the supermolecule which is weakend by the reaction field according to the born formula @xcite @xmath48 in the second model , considering the supermolecule as one object placed in a single cavity of constant radius one has to use the onsager term @xcite . this term is state selective , it gives a contribution only for the states with nonzero dipole moment , i.e. , charge separation . defining the static dipole moment operator as @xmath49 we obtain @xmath50 , with onsager scaling @xmath51 one can express the dynamic part of the system - bath interaction as @xmath52 here @xmath53 denotes the field of the electrostatic displacement at point @xmath54 induced by the system transition dipole moment @xmath55 @xcite . the field of the environmental polarization is denoted as @xmath56 , where @xmath57 is the @xmath58th dipole of the environment and @xmath59 its position . only fluctuations of the environment polarization @xmath60 influence the system dynamics . averaged over the angular dependence the interaction reads @xcite @xmath61 the dynamical influence of the solvent is described with a thermal bath model . the deviation @xmath62 of @xmath63 from its mean value is determined by temperature induced fluctuations . for unpolar solvents described by a set of harmonic oscillators the diagonalization of their interaction yields a bath of harmonic oscillators with different frequencies @xmath64 and effective masses @xmath65 . in the case of a polar solvent the dipoles are interacting rotators as , e.g. used to describe magnetic phenomena @xcite . the elementary excitation of each frequency can again be characterized by an appropriate harmonic oscillator . so we use generalized coordinates of solvent harmonic oscillator modes @xmath66 for polar as well as unpolar solvents . the occupation of the @xmath67th state of the @xmath68th oscillator is defined by the equilibrium dm @xmath69 } \delta_{ij}$ ] . all mutual orientations and distances of solvent molecules have equal probability . an average over all spatial configurations is performed . the interaction hamiltonian ( [ averaged_value_of_interaction ] ) is written in a form which is bilinear in system and bath operators : @xmath70 \left [ \sum_\lambda k_\lambda ( \hat a^+_\lambda + \hat a_\lambda ) \right ] s_{\rm sb } \label{phase}\end{aligned}\ ] ] @xmath71 denotes the interaction intensity between the bath mode @xmath72 of frequency @xmath73 and the quantum transition between the lumos of molecules @xmath43 and @xmath74 with frequency @xmath75 . the scaling function @xmath76 reflects the properties of the solvent . explicit expressions for the solvent influence are still under discussion in the literature @xcite . the interaction of the system with the bath of harmonic oscillators describes the irradiative energy transfer from the system to the solvent as modeled by eq . ( [ phase ] ) . for the description of the dynamics we use the reduced dm which can be obtained from the full dm @xmath77 by tracing over the environmental degrees of freedom @xmath78 @xcite with the evolution operator technique @xcite , restricting ourselves to the second order cumulant expansion @xcite . here we apply the markov approximation , i.e. , we restrict ourselves to the limit of long times . furthermore , we replace the discrete set of bath modes with a continuous one . to do so one has to introduce the spectral density of bath modes @xmath79 . finally one obtains the following master equation @xmath80\right)_{\kappa \lambda } + 2\delta_{\kappa \lambda } \sum\limits_\mu { } \left\ { \gamma_{\mu{}\kappa } \left [ n(\omega_{\mu{}\kappa } ) + 1 \right ] + \gamma_{\kappa \mu { } } n(\omega_{\kappa \mu { } } ) \right\ } \sigma_{\mu{}\mu { } } \nonumber \\ & & - \sum\limits_\mu { } \left\ { \gamma_{\mu{}\kappa } \left [ n(\omega_{\mu{}\kappa } ) + 1 \right ] + \gamma_{\kappa \mu { } } n(\omega _ { \kappa \mu { } } ) + \gamma_{\mu{}\lambda } \left [ n(\omega_{\mu{}\lambda } ) + 1 \right ] + \gamma _ { \lambda \mu { } } n(\omega _ { \lambda \mu { } } ) \right\ } \sigma_{\kappa \lambda } \nonumber \\ & & + \left\ { \gamma _ { \lambda \kappa } \left [ 2n(\omega _ { \lambda \kappa } ) + 1 \right ] + \gamma_{\kappa \lambda } \left [ 2n(\omega_{\kappa \lambda } ) + 1 \right ] \right\ } \sigma _ { \lambda \kappa } , \label{rwa - short}\end{aligned}\ ] ] where @xmath81^{-1}$ ] denotes the bose - einstein distribution . the damping constant @xmath82 reflects the coupling of the transition @xmath83 to a bath mode of the same frequency . it depends on the density of bath modes @xmath84 at the transition frequency @xmath85 and on the transition dipole moments @xmath86 . a rdmem of similar structure was used for the description of exciton transfer in the haken , strobl , and reineker ( hsr ) model @xcite and the generalized stochastic liouville equation ( gsle ) model @xcite . the hsr method originating from the stochastic bath model , is valid only in the high temperature limit @xcite . the gsle method @xcite appeals to the quantum bath model with system - bath coupling of the form @xmath87 , which modulates the system transition frequency . in ref . @xcite the equations for exciton motion are derived using the projection operator technique . taking the different system - bath coupling we have derived the rdmem which coincides with gsle @xcite . both gsle and our rdmem are able to describe correctly finite temperatures . below we neglect the last term of eq . ( [ rwa - short ] ) corresponding to the @xmath88 term in the hsr and gsle models because the rotating wave approximation ( rwa ) is applied . for the sake of convenience of analytical and numerical calculations we replace @xmath89 and the population of the corresponding bath mode @xmath90 with the dissipative transitions @xmath91 and the corresponding dephasings @xmath92 with this one can express the rdmem ( [ rwa - short ] ) in the form @xmath93 the parameters controlling the transitions between the selected states are discussed in subsection [ chem - parameters ] . the relaxation coefficients eq . ( [ g = kv ] ) include the second power of the scaling function @xmath76 because one constructs the relaxation term of eq . ( [ rwa - short ] ) with the second power of the interaction hamiltonian . the physical meaning of @xmath94 is similar to the interaction of the system dipole with a surrounding media . that is why it is reasonable to use the onsager expression ( [ onsager - scaling ] ) for @xmath76 . in the work of mataga , kaifu , and koizumi @xcite the interaction energy between the system dipole and the media scales in leading order as @xmath95,\ ] ] where @xmath96 denotes the optical dielectric constant . from a recent paper of georgievskii , hsu , and marcus @xcite we extract @xmath97 for the multiple cavities model assuming @xmath98 . in terms of a scaling function it can be expressed as @xmath99 as we have already argued in @xcite the coherent coupling @xmath37 between two electronic states scales with @xmath40 and @xmath100 too , because a coherent transition in the system is accompanied by a transition of the environment state which is larger for solvents with larger polarity . as discussed above we neglect the vibrational substructure of each electronic state because the vibrational relaxation is about two orders of magnitude faster than the characteristic et time . but in contrast to the model with vibrational substructure the present model does not involve any reaction coordinate . to reproduce the results of the more elaborate model with vibrational substructure one has to scale the electronic couplings @xmath37 with the franck - condon overlap elements @xmath101 between the vibrational ground states of each pair of electronic surfaces @xmath102 where @xmath103 is the coupling of electronic states of the isolated molecule . for the calculation of the franck - condon factors one has to introduce the leading ( mean ) environment oscillator frequency @xmath104 . here @xmath105 is used which is similar to the frequency of the c - c stretching mode . with this scaling one implicitly introduces a reaction coordinate into the model . the dynamics of the system is controlled by the following parameters : energies of system states @xmath106 , coherent couplings @xmath37 , and damping constants @xmath89 . on the basis of the spectral data @xcite and taking reference energy @xmath107 we determine @xmath108 ev ( in @xmath109 ) . we take the energy of the state with et to @xmath3 from reference : @xmath110 @xcite . further rempel et al . @xcite estimate the coupling of initially excited and charged bridge states @xmath111 and the coupling of the two states with charge separation @xmath112 . the values of the couplings are essentially lower than the energy differences between the relevant system states @xmath113 this is the reason to remain in site representation instead of eigenstate representation @xcite . the damping constants are found with help of the analytical solution derived at the end of the next section to be @xmath114 . the typical radius of the porphyrin ring is about @xmath115 @xcite , while the distance @xmath46 between the blocks of @xmath0 reaches @xmath116 @xcite , @xmath117 , @xmath118 . the main parameter which controls et in a triad is the energy of the state @xmath119 . this state has a big dipole moment because of its charge separation and is therefore strongly influenced by the solvent . because of the special importance of this value we calculate it for the different solvents as a matrix element of the system hamiltonian ( 2 ) . the calculated values of the energies of the @xmath120 state for some solutions are shown in table [ tab:2 ] . the time evolution of the et in the supermolecule is described by solving numerically and analytically eqs . ( [ tosolve1])-([tosolve2 ] ) with the initial condition of donor excitation with a @xmath121 pulse of appropriate frequency , i.e. , the donor population is set to one . for the numerical simulation we express the system of eqs . ( [ tosolve1])-([tosolve2 ] ) in the form @xmath122 where @xmath123 is a vector of dimension @xmath124 for the model with @xmath125 system states and the super - operator @xmath126 is a matrix of dimension @xmath127 . we find an exponential growth of the acceptor population @xmath128,\ ] ] where for the solvent mthf @xmath129 and @xmath130 . the population @xmath131 which corresponds to charge localization on the bridge does not exceed @xmath132 . this means that in this case the superexchange mechanism dominates over the sequential transfer mechanism . besides it ensures the validity of characterizing the system dynamics with @xmath133 and @xmath134 dt \right\}^{-1}.\end{aligned}\ ] ] the alternative analytical approach is performed in the kinetic limit @xmath135 in laplace space the inequality ( [ kinetic - limit ] ) reads @xmath136 , where @xmath137 denotes the laplace variable . it is equivalent to replacing the factor @xmath138 in the laplace transform of eqs . ( [ tosolve1])-([tosolve2 ] ) with @xmath139 . this trick allows to substitute the expressions ( [ tosolve2 ] ) for non - diagonal elements of the dm into eq . ( [ tosolve1 ] ) . after this elimination we describe the coherent transitions to which the non - diagonal elements contribute by redefinition of the diagonal rdmem ( [ tosolve1 ] ) @xmath140 the transition coefficients @xmath141 contain dissipative and coherent contributions @xmath142^{-1}.\ ] ] now it is assumed that the bridge is not populated . this allows us to find the acceptor population in the form of eq . ( [ single - exponent ] ) , where @xmath143^{-1}.\end{aligned}\ ] ] the value of @xmath144 can be found comparing the experimentally determined et rate and eq . ( [ rate ] ) . to calculate @xmath145 would require a microscopic model . to avoid a microscopic consideration we simply take the same @xmath89 for all transitions between excited states . the value of et for @xmath0 in mthf is found by rempel et al . @xcite to be @xmath146 . if the bridge state has a rather high energy one can neglect thermally activated processes . @xmath147 is negligibly small with respect to @xmath148 . in this case our result ( [ rate ] ) reads @xmath149 with the relation @xmath150 and the experimental @xmath151 one obtains @xmath152 . the fit of the numerical solution of eqs . ( [ tosolve1])-([tosolve2 ] ) to the experimental @xmath153 in mthf gives the same value . so the damping constants are fixed for a specific solvent and for other solvents they are calculated with the scaling functions . with this method the et was found to occur with dominance of the superexchange mechanism with rates @xmath154 for cyclo and @xmath155 for @xmath156 . to discuss how the transfer mechanism depends on the change of parameters we calculate the system dynamics varying one parameter at a time . the dependencies of @xmath153 and @xmath133 on @xmath148 , @xmath147 and @xmath157 , @xmath158 are shown in figs . 2 and 3 . the change of each parameter influences the transfer in a different way . in particular , @xmath153 depends quadratically on @xmath148 from @xmath159 to @xmath160 in fig . 2 . below it saturates at the lower bound @xmath161 . this corresponds to a crossover of the et mechanism from superexchange to sequential transfer . but , due to the big energy difference between donor and bridge states the sequential transfer efficiency is extremely low . this is displayed by @xmath162 . in the region @xmath163 both mechanisms contribute to @xmath153 . the decrease of @xmath133 in this region corresponds to coherent back transfer . the et rate depends on @xmath147 in a similar way . at rather high values of @xmath148 , @xmath164 the relation ( [ inequality_w > v ] ) is no more valid . for this regime one has to use eigenstate instead of site representation because the wavefunctions are no more localized @xcite . the variation @xmath157 , @xmath158 near the experimental values shows similar behavior of @xmath165 and @xmath166 ( see fig . we independently vary @xmath157 and @xmath158 . both , @xmath165 and @xmath166 increase linear until the saturation value @xmath167 at @xmath168 is reached . there is qualitative agreement between the numerical and analytical values . in eq . ( [ fit ] ) infinite time is approximated by @xmath169 and so one can not obtain et rates lower than this limit . the physical meaning of the et rate dependence on @xmath170 seems to be transparent . at small values of @xmath170 a part of the population coherently oscillates back and forth between the states . the increase of the dephasing @xmath171 quenches the coherence and makes the transfer irreversible . so transfer becomes faster up to a maximal value . for the whole range of @xmath170 , depopulations @xmath172 , @xmath173 and thermally activated transitions @xmath174 , @xmath175 always remain smaller than the coherent couplings , therefore they do not play an essential role . next , the similarity of the dependencies on @xmath157 and @xmath158 will be discussed on the basis of eq . ( [ rate ] ) . in the limit @xmath176 thermally activated processes with @xmath177 vanish and so @xmath178 , while depopulations with @xmath179 remain constant @xmath180 . the condition @xmath181 allows us to neglect @xmath182 in comparison with @xmath183 . with these simplifications eq . ( [ rate ] ) becomes @xmath184 i.e. symmetric with respect to @xmath157 and @xmath158 . to the largest extent the mechanism of transfer depends on the bridge energy @xmath185 as presented in fig . 4 . in different regions one observes different types of dynamics . for large bridge energy @xmath186 the numerical and analytical solutions do not differ from each other . the transfer occurs with the superexchange mechanism . the et rate reaches a maximal value of @xmath187 for low bridge energies . while the bridge energy approaches the donor energy the sequential transfer starts to contribute to the et process . the traditional scheme of sequential transfer is obtained when donor , bridge , and acceptor levels are arranged in a cascade . in this region the analytical solution need not coincide with the numerical solution because the used approximations are no more valid . for equal bridge and acceptor energies @xmath153 displays a small resonance peak in fig . when the bridge energy is lower than the acceptor energy the population gets trapped at the bridge . the finite @xmath153 for @xmath188 does not mean et because @xmath189 . for the dynamic time interval @xmath190 a part of the population tunnels force and back to the acceptor with @xmath153 . the analytical solution ( [ rate ] ) gives a constant rate for the regime @xmath191 , while the numerical solution of eqs . ( [ tosolve1])-([tosolve2 ] ) is instable , because such coherent oscillations of population can not be described by eq . ( [ single - exponent ] ) and @xmath153 can not be fitted with eq . ( [ fit ] ) . in fig . 4 the regime @xmath191 occurs for small @xmath192 while @xmath193 is kept constant and for large @xmath193 while @xmath192 remains constant . the energy dependence of the final population has a transparent physical meaning for the whole range of energy . a large bridge energy ensures the transition of the whole population to the acceptor . in the intermediate case , when the bridge has the same energy as the acceptor , final population spreads itself over these two states @xmath194 . lowering the bridge even more the whole population remains on the bridge as the lowest state of the system . the dependence of the et rate on the acceptor energy @xmath195 in fig . 4 remains constant while the acceptor energy lies below the bridge energy . increase of @xmath193 up to @xmath196 gives the maximal @xmath197 . when @xmath193 increases further the acceptor becomes the highest level in the system and therefore the population can not remain on it . for the application of the results to various solvents and comparison with experiment one should use the scaling for energy , coherent couplings , and damping constants as discussed above . the combinations of the energy scaling in subsection [ chem - isolate ] and damping constants scalings in subsection [ chem - scaling ] are represented in fig . an increase in @xmath40 from @xmath198 to @xmath199 leads to an increase of the et rate , no matter which scaling is used . further increase of @xmath40 induces saturation for the onsager - mataga scaling and even a small decrease within the applied approximations an increase in the solvent polarizability and , consequently , of its dielectric constant lowers the bridge and acceptor energies and increases the system - bath interaction and , consequently , the relaxation coefficients . it induces a smooth rise of the et rate for the onsager - mataga scaling . on the other hand large @xmath40 leads to essentially different polarisational states of the environment for the supermolecule states with different dipole moment this reduces the coherent couplings , see eq . ( [ v_scale ] ) , leading for the born - marcus scaling to a small decrease of @xmath153 for large @xmath40 . the et rate with this scaling comes closer to the experimental value @xmath200 . this gives a hint that the model of individual cavities for each molecular block is closer to reality than the model with a single cavity for the whole supermolecule . below we consider born scaling eq . ( [ born - scaling ] ) for the system energies and marcus scaling eq . ( [ marcus - scaling ] ) for the damping constant to compare the calculated et rates with the measured ones . for the solvents cyclo , mthf , and @xmath156 one obtains the relative bridge energies @xmath201 , @xmath202 , and @xmath203 , respectively . the calculated et rate coincides with the experimental value @xcite for @xmath0 in cyclo , see table [ tab:2 ] . for @xmath156 the numerical et rate is approximately 30% faster than the experimental value . it has to be noted that a value for the damping rates can be chosen such that the calculated curve almost passes through all three experimental error bars . on the other hand an error in the present calculation could be due to ( i ) absence of vibrational substructure of the electronic states in the present model ; ( ii ) incorrect dependence of system states energies on the solvent properties ; ( iii ) opening of additional transfer channels not mentioned in the scheme shown in fig . each of these possibilities needs some comments . _ ( i ) : the incorporation of the vibrational substructure will result in a complication of the model @xcite giving a more complicated et rate dependence on the energy of the electronic states and dielectric constant . it should yield the maximal et rate for nonequal energies of electronic states , namely for the activationless case when the energy difference equals the reorganization energy . for a comparison of the models with and without vibrational substructure see ref . @xcite . _ ad _ ( ii ) : effects as the solvation shell @xcite do need a molecular dynamics simulation . the total influence of the solvent is , probably , reflected in an energy shift between the spectroscopically observable states @xmath204 and @xmath205 @xcite . _ ad _ ( iii ) : a solvent with large @xmath40 can bring high - lying system states closer to the ones included in fig . 1 . e.g. , because of its larger dipole moment , @xmath206 is strongly influenced by the solvent . as discussed above the rdmem are very similar to those of the gsle model . this is an extension of the hsr theory in which a classical bath is used and for which analytical solutions are available @xcite . we are not aware of any analytical solution of the gsle model as presented here . also this model has not been applied to similar et processes . the numerical steady - state method used by davis et al . @xcite is an attractive one due to its simplicity , but unlike our method it is not able to give information about the time evolution of the system . we use a similar approach derived within a redfield - like theory . but we consider dephasing and depopulation between each pair of levels . in contrast davis et al . incorporate relaxation phenomenologically only to selected levels , dephasing @xmath207 occurs between excited levels , while depopulation @xmath208 takes place only for the sink from acceptor to the ground state . the advantage of the approach of davis et al . is the possibility to investigate the et rate dependence for the bridge consisting of more than one molecular block . this was not the goal of the present work but it can be extended into this direction . we are interested in the et in a concrete molecular complex with realistic parameter values and realistic possibilities to modify those parameter . our results as well as the results of davis show that et can occur as coherent ( with the superexchange mechanism ) or dissipative process ( with the sequential transfer mechanism ) . we have performed a study of the et in the supermolecular complex @xmath209 within the dm formalism . the determined analytical and numerical et rates are in reasonable correspondence with the experimental data . the superexchange mechanism of et dominates over the sequential transfer . we have investigated the stability of the model varying one parameter at a time . the qualitative character of the transfer is stable with respect to a local change of system parameter . the crossover between the two transfer mechanisms can be induced by lowering the bridge energy . the relation of the theory presented here to other theoretical approaches to et has been discussed . the calculations performed in the framework of the present formalism can be extended in the following directions : ( i ) considerations beyond the kinetic limit . the vibrational substructure has to be included into the model as well as solvent dynamics and , probably , non - markovian rdmem . ( ii ) enlargement of the number of molecular blocks in the complex . ( iii ) initial excitation of states with rather high energy should open additional transfer channels . j. jortner and m. bixon ( eds . ) , adv . * 106&107 * ( 1999 ) . d. devault , _ quantum mechanical tunneling in biological systems _ ( cambridge university press , cambridge , 1993 ) . v. balzani and f. scandola , _ supramolecular photochemistry _ ( ellis horwood , chichester , 1991 ) . m. newton , chem . rev . * 91 * , 767 ( 1991 ) . p. f. barbara , t. j. meyer , and m. a. ratner , j. phys . chem . * 100 * , 13148 ( 1996 ) . r. a. marcus , j. chem . phys . * 24 * , 966 ( 1956 ) ; rev . mod . phys . * 65 * , 599 ( 1993 ) ; r. a. marcus and n. sutin , biochim . biophys . acta * 811 * , 265 ( 1985 ) . m. bixon , j. jortner , and m. e. michel - beyerle , biochim . biophys . acta * 1056 * , 301 ( 1991 ) ; chem . phys . , 389 ( 1995 ) . j. deisenhofer , o. epp , k. miki , r. huber and h. michel , j. mol . biol . * 180 * , 385 ( 1984 ) . m. r. wasielewski , chem . rev . * 92 * , 345 ( 1992 ) ; m. r. wasielewski , d. g. johnson , w. a. svec , k. m. kersey , d. e. cragg , and d. w. minsek , in : _ photochemical energy conversion _ , eds . j. norris and d. meisel ( elsevier , amsterdam , 1989 ) p. 135 ; m. r. wasielewski , m. p. niemczyk , d. g. johnson , w. a. svec , and d. w. minsek , _ tetrahedron _ * 45 * , 4785 ( 1989 ) . d. g. johnson , m. p. niemczyk , d. w. minsek , g. p. wiederrecht , w. a. svec , g. l. gaines iii , and m. r. wasielevski , j. am . soc . * 115 * , 5692 ( 1993 ) . w. davis , m. wasielewski , m. ratner , v. mujica , and a. nitzan , j. phys . chem . * 101 * , 6158 ( 1997 ) . s. s. scourtis and s. mukamel , chem . phys . * 197 * , 367 ( 1995 ) . m. schreiber , c. fuchs , and r. scholz , j. lumin . * 76&77 * , 482 ( 1998 ) . h. m. mcconnel , j. chem . phys . * 35 * , 508 ( 1961 ) . r. kosloff and m. a. ratner , isr . j. chem . * 30 * , 45 ( 1990 ) . u. rempel , b. von maltzan , and c. von borczyskowski , chem . * 245 * , 253 ( 1995 ) . e. i. zenkevich , v. n. knyukshto , a. m. shulga , v. a. kuzmitsky , v. i. gael , e. g. levinson , and a. f. mironov , j. lumin . * 75 * , 229 ( 1997 ) ; e. i. zenkevich , a. m. shulga , s. m. bachilo , u. rempel , j. von richthofen , and c. von borczyskowski , j. lumin . * 76&77 * , 354 ( 1998 ) ; a. chernook , u. rempel , c. von borczyskowski , a. m. shulga , and e. i. zenkevich , chem . . lett . * 254 * , 229 ( 1996 ) ; j. phys . chem . * 100 * , 1918 ( 1996 ) ; ber . bunsenges . phys . chem . * 100 * , 2065 ( 1996 ) . a. i. burstein and y. georgievskii , j. chem . phys . * 100 * , 7319 ( 1994 ) . h. sumi and t. kakitani , chem . lett . * 252 * , 85(1996 ) ; h. sumi , j. electroan . * 438 * , 11 ( 1997 ) . b. paulson , k. pramod , p. eaton , g. l. closs , and j. r. miller , j. chem . phys . * 97 * , 13042 ( 1993 ) . n. s. hush , coord . chem . rev . * 64 * , 135 ( 1985 ) . r. langen i. chang , j. p. germanas , j. h. richards , j. r. winkler , and h. b. gray , science * 268 * , 1733 ( 1995 ) ; c. j. murphy , m. r. arkin , y. jenkins , n. d. ghatlia , s. h. bossman , n. j. turro , and j. k. barton , science * 262 * , 1025 ( 1993 ) . v. mujica , m. kemp , and m. a. ratner , j. chem . * 101 * , 6849 ( 1994 ) ; * 101 * , 6856 ( 1994 ) ; m. kemp , v. mujica , and m. a. ratner , j. chem . phys . * 101 * , 5172 ( 1994 ) ; v. mujica , m. kemp , a. roitberg , and m. a. ratner , j. chem . phys . * 104 * , 7296 ( 1996 ) ; m. p. samanta , w. tian , s. datta , j. i. henderson , and c. p. kubiak , phys . rev . b * 53 * , r7626 ( 1996 ) . m. a. ratner , j. phys . * 94 * , 4877 ( 1990 ) ; r. j. miller and j. v. beitz , j. chem . phys . * 74 * , 6749 ( 1981 ) . j. w. evenson and m. karplus , j. chem 96 * , 5272 ( 1992 ) ; science , * 262 * , 1247 ( 1993 ) . w. t. pollard , a. k. felts , and r. a. friesner , adv . 93 * , 77 ( 1996 ) ; a. k. felts , w. t. pollard , and r. a. friesner , j. phys . chem . * 99 * , 2929 ( 1995 ) . w. b. davis , w. a. svec , m. a. ratner , and m. r. wasielewski , nature * 396 * , 60 ( 1998 ) . r. cave and m. d. newton , chem . * 249 * , 15 ( 1996 ) . m. bixon and j. jortner , j. chem . phys . * 107 * , 5154 ( 1997 ) . u. weiss , _ quantum dissipative systems _ ( world scientific , singapore , 1999 ) , 2nd ed . r. loudon , _ the quantum theory of light _ ( clarendon , oxford , 1973 ) ; j. r. klauder and e. c. g. sudarshan , _ fundamentals of quantum optics _ ( benjamin , new york , 1968 ) r. egger , c. h. mak , and u. weiss , phys . e * 50 * , r655 ( 1994 ) ; c. h. mak and r. egger , chem . lett . * 238 * , 149 ( 1995 ) ; n. makri , e. sim , e. makarov , and m. topaler , proc . usa * 93 * , 3926 ( 1996 ) ; e. sim and n. makri , j. phys . chem . b * 101 * , 5446 ( 1997 ) . a. k. felts , w. t. pollard , and r. a. friesner , j. phys . chem . * 99 * , 2929 ( 1995 ) . s. h. lin , m. hayashi , s. suzuki , x. gu , w. xiao , and m. sugawara , chem . phys . * 197 * , 435 ( 1995 ) . o. khn , v. rupasov , and s. mukamel , j. chem . , 5821 ( 1996 ) . k. blum , _ density matrix theory and applications _ ( plenum , new york , 1996 ) , 2nd ed . a. g. redfield , _ ibm j. res . _ * 1 * , 19 ( 1957 ) ; adv . . reson . * 1 * , 1 ( 1965 ) . v. may and m. schreiber , phys . a * 45 * , 2868 ( 1992 ) . o. khn , v. may , and m. schreiber , j. chem . phys . * 101 * , 10404 ( 1994 ) . j. m. jean , r. a. friesner , and g. r. fleming , j. chem . phys . * 96 * , 5827 ( 1992 ) ; w. t. pollard and r. a. friesner , j. chem . phys . * 100 * , 5054 ( 1994 ) ; j. m. jean and g. r. fleming , j. chem . phys . * 103 * , 2092 ( 1995 ) ; j. m. jean , j. chem . phys . * 104 * , 5638 ( 1996 ) . a. nitzan , chem . * 41 * , 163 ( 1979 ) ; r. kosloff and s. a. rice , j. chem . phys . * 72 * , 4591 ( 1980 ) . r. kosloff , m. a. ratner , and w. b. davis , j. chem . phys . * 106 * , 7036 ( 1997 ) . m. schreiber and d. kilin , in : _ proc . excitonic processes in condensed matter _ , ed . m. schreiber ( dresden university press , dresden , 1996 ) p. 331 , http://xxx.lanl.gov/abs/cond-mat/0001061 . w. b. davis , m. r. wasielewski , r. kosloff , and m. a. ratner , j. phys . a * 102 * 9360 ( 1998 ) . v. kenkre and p. reineker , _ exciton dynamics in molecular crystals and aggregates _ ( springer , berlin , 1982 ) . h. haken , p. reineker , z. phys . * 249 * , 253 ( 1972 ) . d. a. weitz , s. garoff , j. i. gersten , and a. nitzan , j. chem . phys . * 78 * , 5324 ( 1983 ) . d. kilin and m. schreiber , j. lumin . * 76&77 * 433 ( 1998 ) , http://xxx.lanl.gov/abs/cond-mat/0001052 . m. gouterman , j. mol . spectrosc . * 6 * , 138 ( 1961 ) . m. schreiber , d. kilin , and u. kleinekathfer , j. lumin . * 83&84 * , 235 ( 1999 ) , htp://xxx.lanl.gov / abs / quant - ph/9912068 . d. d. fraser , j. r. bolton , j. phys . chem . * 98 * , 1626 ( 1994 ) . v. m. agranovich and m. d. galanin , _ electronic excitation energy transfer in condensed matter _ ( north - holland , amsterdam , 1982 ) ; l. d. landau , e. m. lifshitz , _ lehrbuch der theoretischen physik , band viii _ , ( akademie - verlag , berlin 1985 ) . y. georgievskii , c .- p . hsu , r. a. marcus , j. chem . phys . * 110 * , 5307 ( 1999 ) . o. christiansen and k. mikkelsen , j. chem . phys . * 110 * , 8348 ( 1999 ) . r. p. feynman , _ the feynman lectures on physics vol . iii _ , ( addison - wesley , reading , 1963 ) . c. j. f. bttcher , _ theory of electric polarization _ ( elsevier , amsterdam , 1973 ) vol . 1 . m. karelson , g. h. f. diercksen , in : _ problem solving in computational molecular science : molecules in different environments _ , eds . s. wilson and g. h. f. diercksen ( kluwer , dordrecht , 1997 ) p. 195 . k. yosida , _ theory of magnetism _ ( springer , berlin , 1996 ) . s. v. tyablikov , _ methods of the quantum theory of magnetism _ ( plenum , new york , 1967 ) . a. apanasevich , s. ya . kilin , a. p. nizovtsev , and n. s. onishchenko , j. opt . b * 3 * , 587 ( 1986 ) . b. yoon , j.m . deutch , and j.h . freed , j. chem . phys . * 62 * , 4687 ( 1975 ) ; a. isihara , _ statistical physics _ , ( academic press , new york , 1971 ) . .energy of the charged bridge state in different solvents and corresponding et rates ( for calculations born scaling ( [ born - scaling ] ) of energy and marcus scaling ( [ marcus - scaling ] ) of dissipation are used ) . mthf denotes 2-methyl - tetrahydrofuran , and cyclo denotes cyclohexane . [ cols="^,^,^,^",options="header " , ]

one of the most important developments in the field of interstellar gas dynamics during the last half - century was the renewed perception that most processes and structures are strongly affected by turbulence . this is a paradigm shift unparalleled in many other fields of astronomy , comparable perhaps to the discovery of extrasolar planets and cosmological structure at high redshift . interstellar turbulence and its implications were commonly discussed in the 1950s , but without the range of observations that are available today , the theory of the interstellar medium ( ism ) drifted toward increasingly detailed models based largely on a preference for thermal and dynamical equilibria . standing apart were two subfields that continued to include turbulence as the primary driver of all observations : radio scintillation and cosmic ray transport . the connections between these subfields and the prevailing picture of the ism were mostly ignored because of the large discrepancy in scales . consequently , two other small - scale processes got lost in the equilibrium paradigm of the large - scale models : mixing of the elements and molecular chemistry . in this review , we discuss these four subfields in some detail . all have a well - established observational base going back several decades , but because of the complexity of turbulence and the infancy of the relevant theory , all of them are now in a state of rapid evolution . our previous review _ interstellar turbulence i _ ( this volume ) , emphasized observations in the dense neutral ism and discussed in detail the various theoretical approaches to this field . the degree to which turbulence and other transport processes mix or homogenize the gas in the face of repeated local pollutions by supernovae and other sources of fresh metals could provide an important constraint on the hydrodynamics of the ism . this general problem is called passive scalar turbulence because the particles being transported have no effect on the velocity field . a review of experiments involving incompressible turbulent scalar transport can be found in warhaft ( 2000 ) , who emphasized that passive scalar probability density functions ( pdfs ) usually have strong exponential tails . a review focussing on a particular class of theoretical approaches has been given by falkovich , gawedzki & vergassola ( 2001 ) , whereas reviews of methods to calculate the concentration pdf and its moments are in dopazo ( 1994 ) and obrien ( 1980 ) . briefer and more specialized treatments can be found in shraiman & siggia ( 2000 ) and pope ( 2000 , section 12.4 ) . here we summarize observations of abundance variations in stars and gas , and we review recent theoretical work on interstellar turbulent mixing . we do not discuss star - to - star variations in globular clusters that may be affected by stellar evolution ( yong et al . 2003 ) . recent observations of metallicity dispersions in stars of the same age and in interstellar gas suggest that the ism is very well mixed on average , with typical fluctuations about the mean of only 520% . this low value is a challenge to explain considering the spotty pollution of new elements from supernovae and other sources . in addition , there are much larger observed fluctuations in a few places , suggesting the overall distribution function for abundance has a long , non - gaussian tail . both the low dispersion and the long tail might be explained by interstellar turbulence . stars of a similar age have often shown large variations in metallicity ( a term usually referring to iron abundance relative to the sun ) . for example , edvardsson et al . ( 1993 ) and chen et al . ( 2000 ) found factor - of - four abundance variations in several elements for f and g main sequence stars . after correcting for selection effects , edvardsson et al . estimated that the dispersion at a given age is about 0.15 to 0.20 dex , i.e. , 4060% , whereas garnett & kobulnicky ( 2000 ) suggested it is smaller than 0.15 dex for the same sample . a large unbiased sample of over 550 hipparcos stars with homogeneous chromospheric age estimates and photometric metallicities showed a dispersion at a given age of only 0.13 dex ( rocha - pinto et al . reddy et al . ( 2003 ) studied 27 elements in 181 f and g stars and found a dispersion at a given age that is somewhat smaller than in edvardsson et al . ( 1993 ) and chen et al . feltzing & gonzales ( 2001 ) used an inhomogeneous sample of 5828 stars and found a larger scatter in the metallicity at a given age of 0.2 to 0.3 dex , which exceeds the observational errors . a major problem is that the ages of the stars used in these studies have large uncertainties . considering that the stellar orbits probably migrated and led to additional abundance variations from galactic radial gradients ( e.g. , wielen et al . 1996 ) , the stellar results are consistent with a dispersion in the gas of 2030% or less , the upper limit set by uncertainties inherent in field star samples . other evidence for homogeneity among field stars comes from elemental abundance ratios . reddy et al . ( 2003 ) estimated the intrinsic scatter in the ratio [ x / fe ] for element x by fitting a gaussian curve to the histogram of deviations between [ x / fe ] and the linear fit of [ x / fe ] versus [ fe / h ] . the standard deviations in the gaussian for 26 elements in 181 stars varied between 0.03 and 0.10 dex . these deviations are comparable to the errors , so the result is an upper limit to the intrinsic scatter . it does not depend on stellar age , as do the absolute abundances . reddy et al . noted that because the elements come in differing proportions from different sites of stellar nucleosynthesis , the lack of scatter implies all the ejecta were well - mixed in the gas before the stars formed . abundance fluctuations of 0.050.15 dex had also been found in studies of open clusters ( e.g. , friel & boesgaard 1992 , twarog et al . 1997 , carraro et al . 1998 ) , b stars in star - forming regions ( e.g. , cunha & lambert 1994 ) and clusters ( rolleston et al . 1994 ) , the magellanic clouds ( e.g. , olszewski et al . 1991 ) , the interstellar medium ( meyer et al . 1998 ) , galactic _ regions ( deharveng et al . 2000 ) , and _ h ii _ regions in disk ( see review in henry & worthey 1999 ) and dwarf irregular ( e.g. , thuan et al . 1995 ) galaxies . however , f stars in nearby open clusters ( friel & boesgaard 1992 and references therein ) show no evidence for star - to - star abundance variations at the level of detectability . cluster - to - cluster variations are also small , including clusters of all ages , and essentially zero ( less than 1020% ) for the four clusters younger than @xmath0 years . the latter result implies that the ism is mixed , at least to this level , for scales between @xmath11 pc ( the size of the clumps from which clusters form ) and @xmath1100 pc ( the distances between clusters ) on a timescale less than @xmath0 years . even more stringent constraints were found from the photometric metallicity study of 76 clusters by twarog et al . comparison of high - precision photometry with stellar evolution models suggested a metallicity spread less than @xmath10.03 dex in the hyades cluster ( quillen 2002 ) . paulson , sneden & cochran ( 2003 ) derived the same small [ fe / h ] spread , excluding three outlier stars , using spectroscopic abundance determinations for individual stars . these results imply fluctuations of less than @xmath17% in the gas from which the hyades cluster formed . interstellar absorption studies suggest equally low dispersions . meyer et al . ( 1998 ) , moos et al.(2002 ) , and andre et al . ( 2003 ) found oxygen abundance dispersions over distances of around one kpc that are only a few percent . a recent study of interstellar kr / h by cartledge , meyer & lauroesch ( 2003 ) found a spread about the mean of only 0.06 dex for lines of sight within the local orion spur ; this was less than the measurement error . we conclude that the most recent evidence , based on field stars of a given age , cluster stars , and the ism , suggests a metallicity dispersion that is very small , less than 20 to 30% and perhaps as small as a few percent . this small galactic dispersion is consistent with the study of 41 _ h ii _ regions in m101 by kennicutt & garnett ( 1996 ) . there are notable exceptions , though , including : ( _ a _ ) the existence of old super - metal - rich stars ( e.g. , feltzing & gonzalez 2001 ) , and a large range in abundance ratios among metal - poor stars ( fields et al . 2002 ) ; ( _ b _ ) the lack of an age - metallicity relation for five nearby clusters , which led boesgaard ( 1989 ) to conclude that local abundance fluctuations of @xmath10.20.3 dex must persist for at least @xmath2@xmath3 years ( but see friel & boesgaard 1992 for a revised result ) ; ( _ c _ ) the factor of 2 to 3 fluctuations in [ o / h ] among subgroups in orion ( cunha & lambert 1994 ) , which , however , may be a result of self - pollution ; ( _ d _ ) the study of b stars in several clusters at galactocentric distance @xmath113 kpc that indicates one cluster has [ fe / h ] larger than the rest by a factor of @xmath15 ( rolleston , dufton & fitzsimmons 1994 ) , ( _ e _ ) two stars in the carina nebula ( andre et al . 2003 ) that are separated by only a few hundred parsecs but have line - of - sight oxygen abundances different by about 40% , and ( _ f _ ) outlier stars in the cluster studies discussed earlier . these examples suggest a metallicity pdf with a small variance and possibly large higher moments , i.e. , a fat tail at large metallicities . such a pdf is similar to that found for the pdf of passive scalars in incompressible turbulence ( warhaft 2000 ) . the analogous pdf for a compressible , magnetic , and self - gravitating ism is unknown . theoretical investigations of abundance fluctuations have until recently nearly all been based on order of magnitude arguments involving characteristic spatial scales and timescales for various processes . two of the more thorough discussions are in roy & kunth ( 1995 ) and tenorio - tagle ( 1996 ) . regarding spatial scales , elmegreen ( 1998 ) pointed out that for a hierarchically structured ism with star formation operating on a local dynamical time , the size of an o and b star association is usually larger than the supernovae remnants it produces . unless there is some immediate mixing , as in hot ionized or evaporating flows , the new elements should be spotty and variable for subsequent generations of stars ( self - enrichment ) . the point about timescales was highlighted by reeves ( 1972 ) : the mean time for a gas element to get reincorporated into a star is shorter than the time for its dispersal by supernova mixing , so abundance fluctuations should be large . both results contradict the apparent uniformity of abundances inside and between clusters and in the general ism . estimates for mixing times have changed since reeves ( 1972 ) paper . edmunds ( 1975 ) found that the timescale for mixing of gas by turbulence and galactic shear is relatively short compared with the star formation time , provided the sources of metals do not cluster together making the dispersion more difficult ( kaufman 1975 ) . bateman & larson ( 1993 ) also derived a short mixing time , considering diffusion by cloud motions , as did roy & kunth ( 1995 ) , who applied the diffusion approximation to galactic shear , radial flows , turbulent diffusion , supernovae , and gas instabilities . similarly short timescales can be derived using structure functions by considering the separation of initially nearby lagrangian fluid particles ( frisch 1995 ) . this latter procedure is often used to derive the richardson 4/3 law of superdiffusive incompressible turbulence . the implication of these short mixing times is that the ism should be homogeneous . the expectation of large spatial fluctuations may be viewed in a different way too , by considering the number of overlapping regions of contamination . edmunds ( 1975 ) showed that @xmath4 supernovae in a local volume should lead to inhomogeneities of order @xmath5 if there is no subsequent mixing . adopting an average of @xmath6 , he suggested that fluctuations would be relatively small . more detailed models for overlapping contaminations were studied by roy & kunth ( 1995 ) , argast et al . ( 2000 ) and oey ( 2000 ) . oey ( 2003 ) used turbulent dispersal from bateman & larson ( 1993 ) to enlarge the effective sizes and reduce the average metallicities in the overlapping regions . low metallicity stars and galaxies should have large metallicity dispersions because the numbers of contaminating events are small ( audouze & silk 1995 ) . a problem with most of the timescale arguments is that they apply to turbulent transport and not homogenization . moving clouds around or shredding them with fluid instabilities does not homogenize the gas at the atomic level or change the frequency distribution of abundance concentrations ; viscosity and molecular diffusion are required for that . the same is true for incompressible turbulence imagined as a cascade of vortices . unless the advection process creates steep gradients on small scales , molecular diffusion can not occur in a reasonable time . the situation changes when there are large - scale gradients in the mean concentration ; then advection from distant regions can change the concentration distribution locally ( see equation [ eq : zave ] below ) . even so , homogenization requires molecular diffusion . thus , a small transport time in the ism does not imply the gas will homogenize enough to remove abundance differences before it forms stars and clusters . the size and placement of each protostellar core within the complex spatial pattern of abundance fluctuations will determine its metallicity . releasing metals into the interstellar turbulent gas is similar in basic respects to dropping a blob of ink into a turbulent fluid . figure [ fig : mix ] shows such a release into a two - dimensional ( 2d ) incompressible forced chaotic velocity field ( jullien et al . advection spreads the scalar out in space , while concentrating it locally through stretching and folding , shearing it into thin ribbons until the spatial gradients are large enough that molecular diffusion becomes significant in the third frame . by the final frame , diffusion has smeared out most of the striated structures . how does turbulence deliver gradients to small enough scales for molecular diffusion to cause homogenization in a reasonable time ? de avillez & mac low ( 2002 ) found a surprisingly long timescale for the decay of the variance in a passive scalar field that had an initial checkerboard pattern and was mixed by supernovae . their time scaled to a galaxy ( at least a few times @xmath2 years ) is larger than previous estimates and still is a lower limit because they did not include sources of new metals . de avillez & mac low also found that the mixing time is independent of the mixing length , unlike mixing modeled as diffusion , and that the mixing time decreases with increasing supernova rate as a result of the increased velocities . there may be ways other than turbulence to get gradients on small enough scales for molecular diffusion to homogenize . in tenorio - tagle s ( 1996 ) `` fountain with a spray '' model , metal - rich droplets falling onto the disk are subject to rayleigh - taylor instabilities that reduce their size even further ; diffusion will be very effective in such a situation . there are many methods for studying passive scalar turbulent transport . here we describe two approaches : prescribed artificial velocity fields and closure methods for studying restricted properties of turbulent transport . most of what is known about passive scalar turbulence comes from models using an idealized velocity field that is not a rigorous solution to the hydrodynamics equations ( see review by falkovich et al . an example would be a time - uncorrelated velocity field with gaussian statistics , a prescribed spatial correlation function , and a constraint of incompressibility . this is called a kraichnan velocity field ( e.g. , shraiman & siggia 2000 ) . the advantages of such a field are that the equations for the concentration field and its statistical properties can in some cases be solved analytically without a closure assumption , and the generic properties of the result might be valid for any velocity field that has the same variance and conservation constraints . however , there is a large danger in assuming a delta function for the time correlation function . large intermittency in the velocity gradient field means that a large fraction of the fluid contains large gradients compared with a gaussian . in incompressible turbulence ( and the ism see _ interstellar turbulence i _ ) , this fraction gets larger on smaller scales . this implies that intermittency should homogenize the gas faster than purely gaussian motions , because molecular diffusion depends on large velocity gradients at small scales . decamp & le bourlot ( 2002 ) used another model velocity field that was obtained from a wavelet reconstruction with statistical properties similar to the ism velocity field , including motions correlated in space and time . they solved the continuity equations and showed that the correlated velocity field has a faster dispersion and a larger standard deviation for passive scalars ( by factors of 2 to 4 ) than an uncorrelated gaussian velocity field . the spatial distributions for species undergoing chemical reactions were also different , suggesting a mechanism for molecular segregation in dense clouds ( section [ sect : chem ] ) . further work on turbulent mixing using synthetic velocity fields is in elliott & majda ( 1996 ) , fung & vassilicos ( 1998 ) , and boffetta et al . ( 1999 ) . for some purposes , it may be sufficient to understand how the mean , variance , and skewness of a metallicity distribution depend on the source terms and large - scale gradients . this is the moment approach , and it has the advantage that some of its results can be obtained analytically . the moment approach might explain , for example , why there is an inverse square root scaling on the supernova rate for the decay timescale of the abundance variance ( de avillez & mac low 2002 ) . the evolution equation for the one - point probability distribution of metallicities was derived rigorously for incompressible turbulence by dopazo , valino & fueyo ( 1997 ) , who also derived the moment equations for incompressible flow ( see also obrien 1980 ; chen & kollman 1994 ; dopazo 1994 ; pope 1994 , 2000 ) . perhaps the simplest question that can be asked about scalar turbulence concerns the way in which the average relative concentration @xmath7 of a local patch of contaminant ( relative to hydrogen , for example ) will spread in space as a function of time . the usual place to start is to examine the mean square distance , @xmath8 , traveled by a marked lagrangian fluid particle . this distance can be related to an integral over the lagrangian correlation function , assuming only statistically stationary and isotropic turbulence . without knowing the correlation function , however , this relation is of little use . an approximation is to use the fact that the correlation is unity at time zero and approaches zero for large separations . then for times much smaller than the correlation time @xmath9 , @xmath10 whereas at @xmath11 , @xmath12 . here , @xmath13 is the two - time correlation function at a given point in the flow for lag @xmath14 ; i.e. , @xmath15 for flow speed @xmath16 and the average @xmath17 is over all pairs of points separated by @xmath14 . normalization to the variance , @xmath18 , makes @xmath13 dimensionless with values between @xmath19 and @xmath20 . in this formulation , the particle behaves ballistically at short times and has an uncorrelated random walk at large times . these results are usually referred to as taylor s theorem ( taylor 1921 ; see mccomb 1990 ; pope 2000 ) . the result is important for illustrating the nature of turbulence as a correlated random walk and the analogy between turbulent and molecular diffusion , where in the latter case @xmath21 would be the mean collision time . klessen & lin ( 2003 ) recently showed that taylor s theorem is correct in 3d - compressible turbulence . mccomb ( 1990 ) , piterbarg & ostrovskii ( 1997 ) and others have pointed out that taylor s theorem does not solve the problem of turbulent dispersal and transport , even for the mean field because what is required is information in the lab frame , not following the particle . klessen & lin ( 2003 ) suggest that the result can be used to develop a phenomenological mixing - length or diffusion model for turbulent transport in the ism in which the integral correlation time is related to the mean time between shock passages . another fundamental relationship is richardson s law ( richardson 1926 ) for the rate at which pairs of points separate in space in the inertial range of turbulence . one form of this law states that the time dependence of the rms separation is @xmath22 ( falkovich et al . 2001 , boffetta & sokolov 2002 , nicolleau & vassilicos 2003 ) . work on pair dispersion using synthetic velocity fields has resulted in at least two important insights : ( _ a _ ) the dynamics of particle pairs is sensitive to initial conditions but their rms separation increases algebraically , not exponentially ( unlike low - dimensional chaotic systems ) . ( _ b _ ) particle pairs travel together for long times and then separate explosively when they encounter straining regions around hyperbolic points in the flow ( e.g. , nicolleau & vassilicos 2003 ) . a traditional approach that could be applied to the spreading of newly produced elements from sources in the ism is to ignore the detailed dynamics and use an equation for the evolution of the mean concentration field alone . in general , the concentration @xmath23 of an element evolves in a given velocity field @xmath24 as @xmath25 where @xmath26 is the mass diffusivity and @xmath27 is the rate at which sources inject the element . the second term on the left represents the effect of the turbulent velocity field in advecting the concentration field . for the interstellar medium this velocity field will consist of complex compressions and vortical motions . the first term on the right represents molecular diffusion , which operates only on small scales in regions where the gradient of the concentration field is large . this equation illustrates how turbulent advection of a passive scalar spreads it out in space while increasing the local gradients through stretching , compressing , and folding . molecular diffusion then operates where the gradients are high . timescales for molecular diffusion have been discussed recently by oey ( 2003 ) for a uniform gas . the result is sensitive to temperature , so most of the homogenization of elements in the ism may occur in the hot regions , like superbubbles ( silich et al . 2001 ) , as well as those with the steepest velocity gradients . to obtain the time evolution of the average concentration , one might choose to ignore all the details of the turbulence and spatially average equation [ eq : z ] . ignoring also the molecular diffusion and source terms , we get @xmath28 for mean concentration and velocity @xmath7 and @xmath29 and fluctuating parts @xmath30 and @xmath24 . the last term is zero for incompressible turbulence , leaving only @xmath31 , which contains the unknown ensemble average of the scalar flux , @xmath32 , representing interactions between the fluctuating parts of the velocity and concentration fields . this is a classic closure problem . we could derive a differential equation for the scalar flux , but it would contain more unknown correlations , including triple correlations , and further manipulation only leads to higher and higher order unknowns . in the incompressible case , the simplest treatment is to imagine that the turbulent transport behaves diffusively , like molecules in a gas . then the scalar flux is proportional to the mean concentration gradient and given by @xmath33 where @xmath34 is a scalar diffusivity given by some timescale times the square of some velocity . often these are taken to be the correlation time and the rms velocity , although a variety of other forms have been used to fit specific experimental results ( piterbarg & ostrovskii 1997 ) . using equation [ eq : flux ] , equation [ eq : zave ] becomes the usual diffusion equation , and the patch of concentration , if initially a point source , would spread according to a gaussian whose width grows @xmath35 . in fact , turbulent transport can not be described by classical diffusion because triple correlations are as important as quadratic correlations , or , more physically , turbulent motions are not random independent steps but correlated bulk transports over a wide range of scales . there have been many attempts , all unsuccessful to various degrees , to develop closure techniques for passive scalar turbulence . many of these are described in reviews by obrien ( 1980 ) and dopazo ( 1994 ) . perhaps the most popular in the physics community is the mapping closure suggested by chen , chen & kraichnan ( 1989 ) . a more complicated closure method is based on the eddy - damped quasinormal markovian approximation ( lesieur 1990 ) . an application of this closure to the passive scalar problem with reference to earlier work is in herr , wang & collins ( 1996 ) . a promising approach that is not too complex yet captures the importance of the triple correlations was proposed by blackman & field ( 2002 ) in connection with turbulent dynamo theory and then applied to the scalar turbulence problem ( blackman & field 2003 ) . the idea is to model the time derivative of the scalar flux as a normal diffusion term with an eddy viscosity plus a damping term @xmath36 , where @xmath14 is a relaxation time . this approach is similar to a closure method used in terrestrial atmospheric turbulent transport ( lewellen 1977 ) . when this is substituted in the equation for @xmath7 , a little manipulation results in a damped wave equation for the mean concentration . brandenburg , kapyla & mohammed ( 2004 ) did extensive comparisons with numerical simulations of passive scalars in mildly compressible turbulence ( rms mach number @xmath10.2 ) and found good agreement with this model for the time evolution of the spreading size and the non - gaussianity ( kurtosis ) of @xmath7 . the moment approach can derive the variance of the abundance distribution directly from equation [ eq : z ] by writing the velocity , abundance , and source fields as the sum of mean and fluctuating parts ( e.g. , @xmath37 ) , multiplying the resulting equation for @xmath38 by @xmath30 , and then ensemble averaging to obtain @xmath39 where @xmath40 is the fluctuating part of the source term ( @xmath41 ) . the first term on the right shows how the scalar flux ( @xmath42 ) interacts with any large - scale gradient in the abundance , increasing variance if @xmath42 is proportional to a mean gradient ( as in equation [ eq : flux ] ) . this is analogous to how stellar orbit diffusion in the presence of a radial gradient produces fluctuations in the stellar abundances . the second term indicates that any gradient in the mean square fluctuations will be advected with the mean flow ( e.g. , radial flows or infall ) . the third term in equation [ eq : z2 ] shows how turbulence can affect a concentration variance even with no gradient in @xmath7 , zero mean velocity , and without diffusion . writing this term as @xmath43 shows that it consists of the mean divergence of the flux of scalar variance , as might be expected , minus a term that is only nonzero in the presence of compressibility . differentiation and ensemble averaging commute , so the first term of these two is zero for a homogeneous fluid because ensemble averages do not depend on spatial position . the fourth term represents the net production of scalar variance by the fluctuating nucleosynthetic sources . `` diffusion '' in this context is shorthand for the fairly complex set of second - order derivatives involving mass diffusion at the molecular level . as a simple illustration , assume there are no large scale flows and the abundance is linear in galactocentric distance with constant of proportionality @xmath44 . for a general gradient term , we would have the triple correlation whose closure is discussed by blackman & field ( 2003 ) and brandenburg , kapyla & mohammed ( 2004 ) . then we can also derive an equation for the correlation @xmath42 by multiplying equation [ eq : z ] by @xmath24 and ensemble averaging . the result is , for @xmath45 , @xmath46 the first term on the right , which is @xmath47 for an isotropic velocity field , represents the production of correlations by interaction of the velocity fluctuations with the mean metallicity gradient . we can assume that we are given some steady - state isotropic turbulent velocity field so that @xmath48 is constant . the second term on the right , often called pressure scrambling , is a serious closure problem for compressible turbulence . the source term is important if the metals come from young stars whose formation is associated with the turbulence itself , but for type i supernovae or any other source that has time to decorrelate from the flow , @xmath49 . this indicates the importance of using a consistent model for the formation of metal - producers from the turbulent field . equations [ eq : z2 ] and [ eq : zu ] can now be solved to give the evolution of the standard deviation of the abundance distribution for a particular velocity and source field . if we neglect pressure scrambling in equation [ eq : zu ] and the compressibility term in equation [ eq : z2 ] , and also replace the triple correlation term with a simple gradient - diffusion closure , then the result would show that the concentration fluctuation variance increases @xmath50 , even in the absence of sources . the turbulence is transporting high - z gas into low - z gas ( and vice versa ) due to the existence of the gradient . obviously , after some time the large fluctuations imply steep gradients at very small scales , and so the microscopic diffusion term will cause the variance to level off . in addition , the gradient itself will be washed out by the turbulence unless there is some mechanism continually producing it . without a sustained mean gradient , such equations could be used to understand how the variance in elemental fluctuations should decrease with time under the action of only advection and diffusion , giving a simplified model to understand the numerical decay results of de avillez & mac low ( 2002 ) . with some assumed closure for the source terms , or with prescribed velocity field statistics , the solution of equation [ eq : z2 ] would show how the abundance variance depends on the source rate . the source terms in the transport equations depend on how the metals in expanding supernovae and superbubbles are released into the turbulent background . this is not a trivial problem , as discussed by tenorio - tagle ( 1996 ) . simulations of superbubbles ( korpi et al . 1999 ) and supernova remnants ( balsara et al . 2001 ) expanding into a turbulent ism may provide guidance . some metal - producing outflows should be jet - like , leading to the release of metals by shear flow along the walls of a chimney . begelman & fabian ( 1990 ) and slavin , shull & begelman ( 1993 ) proposed that such turbulent layers could mix hot and warm gas , but they could just as well mix freshly - produced metals . tenorio - tagle ( 1996 ) suggested that superbubbles blow out of the disk and rain down metal rich droplets that break apart by rayleigh - taylor instabilities . this process could disperse metals over several kpc in the disk with no assistance from disk turbulence . these processes are also relevant for the redistribution of elements ejected by galactic blowout into the intergalactic medium ( igm ) . several papers ( e.g. , ferrara et al . 2000 , aguirre et al . 2001 and references therein ) have noticed that it is difficult to mix metals on scales comparable to the mean distance between galaxies , requiring an additional ( unknown ) mechanism . turbulence in the igm could provide such a mechanism . the advection of the grains by turbulent gas can lead to spatial clustering relative to the gas , and to a spatially inhomogeneous dust - to - gas ratio . the degree of segregation depends on the grain size . local variations in grain - grain velocities affect grain growth and shattering . early estimates for the efficiency of this process were given by cameron ( 1973 ) , burke & silk ( 1976 ) , scalo ( 1977 ) , and draine ( 1985 ) , but the conception of how grain advection occurs differs greatly in these papers . more rigorous derivations , aimed at grain collisions in protostellar disks but also applicable to the ism , were given by vlk et al . ( 1980 ) and markiewicz , mizuno & vlk ( 1991 ) . recent work has been directed toward trapping grains in protoplanetary disks by anticyclonic vortices ( e.g. , bracco et al . 1999 ) , which can persist for many turnover times . large grain concentrations can result , accelerating the formation of planetesimals . grains might also become trapped in the general ism because of drag forces , as proposed by falgarone & puget ( 1995 ) . a review of experiments and simulations on turbulent gas - grain segregation was presented by fessler , kulick & eaton ( 1994 ) . more recent theoretical investigations ( e.g. , sigurgeirsson & stuart 2002 , lopez & puglisi 2003 ) have demonstrated several interesting properties of this clustering . all of this work relies on synthetic velocity fields because inertial particle advection in real turbulent flows would require resolution much finer than the dissipation scale . nevertheless , the essential physics of particle drag is becoming more clear . the particle motion is dissipative even if the advecting flow is incompressible , and hence , dissipationless ( balkovsky , falkovich & fouxon 2001 ) . this leads to simplified treatments in which the stokes drag effect is modeled as a small compressible component to the velocity field of the inertial particles . a detailed study of inertial particle dynamics in incompressible flows ( bec et al . 2003 ) shows that particles smaller than a critical size form fractal clusters whereas larger sizes fill space with a nonuniform density . an extension to highly compressible flows is desirable . the motion of charged dust grains in anisotropic magnetohydrodynamic ( mhd ) turbulence was considered by yan & lazarian ( 2002 ) , who made a distinction between the directions parallel and perpendicular to the field . grains with a gyration time around the field that is shorter than the gas drag time will follow the turbulent field fluctuations in the perpendicular direction . yan & lazarian ( 2003b ) calculated the acceleration of charged grains that resonate with mhd fluctuations . another reason to study stirring by interstellar turbulence is to understand the origin of galactic radial abundance gradients ( carraro et al . 1998 , hou et al . 2000 , rolleston et al . 2000 ) . models typically account for these gradients using a radially dependent timescale for infall and star formation ( e.g. , chiappini et al . 2001 ) . a problem is that turbulence should reduce large - scale gradients , so studies of this turbulent washout could give an estimate of the timescale on which the gradient must be replenished . order of magnitude estimates along the lines of those mentioned in section [ sect : howlarge ] indicate replenishment timescales larger than characteristic infall times , suggesting that contributions from a radially dependent stellar initial mass function ( imf ) or other effects are not required . a peculiar abundance gradient has been observed in the outer galaxy , where clusters ( twarog et al . 1997 ) and cepheids ( caputo et al . 2001 , andrievsky et al . 2002 , luck et al . 2003 ) suggest a sudden decrease in metal abundance by about 0.2 dex setting in at 10 kpc . this discontinuity appears over a radial distance less than 1 kpc . it is hard to see how such a feature could survive turbulent washout for more than @xmath2 years using timescale arguments of the type summarized earlier . either we are missing something fundamental about turbulent transport ( e.g. , maybe differential rotation or magnetism suppress turbulent motions in the radial direction ) , or this metallicity near - discontinuity does not really exist . the coupling of interstellar chemistry to the dynamics of interstellar gas has focused for a long time on collapsing or slowly contracting clouds ( e.g. , gerola & glassgold 1978 ; prasad , heere & tarafdar 1991 ; shematovich et al . 2003 ) and single shocks or mhd waves ( e.g. , hollenbach & mckee 1979 , 1989 ; draine & katz 1986 ; charnley 1998 ; flower & pineau des forts 1998 ) . the collapse models explained why early - time solutions for static chemistry often agreed better with observations than late - time solutions ( e.g. , prasad et al . 1991 ; nejad & wagenblast 1999 ; but see turner et al . the result that nondynamical time - dependent abundances agreed with observations for different ages is a strong motivation for the influence of turbulence because turbulent timescales are often smaller than the chemical equilibrium times . turbulent transport can prevent the chemistry from attaining a steady state ( e.g. , phillips & huggins 1981 ; xie , allen & langer 1995 ; willacy , langer & allen 2002 ) . turbulence can affect ism chemistry in three ways : ( _ a _ ) by continually transporting material between regions with different physical conditions , like ambient ultraviolet ( uv ) radiation flux , temperature , and density , ( _ b _ ) by creating localized heating where temperature - sensitive reactions , especially those involving endothermic reactions , will be enhanced , and ( _ c _ ) by magnetically forcing ions to move faster than the thermal speed where they can enhance the temperature - sensitive ion - neutral reactions . progress in this field is difficult because of limitations in the dynamical models and because of the large number of nondynamical chemical effects that may occur , like neutral - neutral reactions at low temperature , variable c and o depletion , three - body channels for dissociative recombination , the presence of cosmic - ray - induced uv photons , chemical phase transitions , and many more , as reviewed by herbst ( 1999 ) , hollenbach & tielens ( 1999 ) , van dishoeck & hogerheijde ( 1999 ) , and langer et al . ( 2000 ) . for more recent work on chemistry covering a variety of applications not necessarily related to turbulence , see roberts & herbst ( 2002 ) , doty et al . ( 2002 ) , stantcheva & herbst ( 2003 ) , shematovich et al . ( 2003 ) , and charnley & markwick ( 2003 ) . the classic observation that implied turbulent transport was the large amount of carbon in the gas phase of dense molecular cores , considering the short freeze - out time on dust . boland & dejong ( 1982 ) proposed that turbulent circulation could explain this discrepancy by bringing grains to the outer layers for uv desorption . modern observations ( kramer et al . 1999 ) show more depletion , however , and photodissociation region ( pdr ) models suggest carbon comes from clump surfaces at a range of depths ( howe et al . 2000 ) . turbulent transport has other applications in modern chemical models . pijpers ( 1997 ) considered a diffusive model of turbulent grain transport and obtained a long migration time . ragot ( 1998 ) pointed out that turbulent transport is probably neither ballistic nor diffusive , and used a model based on levy flights to return to the smaller timescales . this transport effect is important because the degree to which co is released from grains into the gas can affect the rest of the chemistry ( co deactivates much chemistry by depleting the h@xmath51 ion ; nejad & wagenblast 1999 ) . a detailed study of turbulent transport including 87 species and 1100 reactions was given by xie et al . ( 1995 ) , who assumed that radial flux is proportional to the spatial gradient of a species and used a mixing length model for turbulent diffusion . they found that most carbon - bearing species and several other important molecules were strongly affected by turbulence . the model was extended recently by willacy , langer & allen ( 2002 ) , who showed how the _ h i_/h@xmath52 ratio could provide a sensitive signature of turbulent transport . this occurs because h@xmath52 forms when h reacts on dust grains . the formation rate depends on the density of atomic h and the total dust cross section , whereas the destruction rate depends on the uv field . if turbulence cycles material between regions with different uv optical depths on a timescale comparable to the formation time , then h will penetrate deeper into the cloud and h@xmath52 will be closer to the surface . because many other simple molecules are formed from h@xmath52-based intermediaries ( especially h@xmath51 ) , turbulent transport could control the chemistry for many species . timescales for reaching chemical equilibrium depend on density , temperature , and ionization fraction , and are typically @xmath53@xmath54 years ( van dishoeck & hogerheijde 1999 ) . for the atomic part of a photodissociation region , the longest part of the cycle c@xmath55 ch@xmath56 co@xmath57 c @xmath58c@xmath59 is the formation time of ch@xmath60 through radiative association ( tielens & hollenbach 1985 ) with a timescale @xmath110@xmath61 years for @xmath62 particle density in units of @xmath63 @xmath64 . for turbulent transport , the fastest possible timescale is ballistic , @xmath65 , for turbulent speed @xmath66 and length @xmath67 . using the ism scaling relations for illustration ( see _ interstellar turbulence i _ ) , this minimum time is @xmath68 years . the chemical and transport timescales are comparable at @xmath63 @xmath64 , but the transport time becomes larger than the chemical time at higher density and then turbulence would seem to have little effect . because cosmic ray ionization drives the chemistry , and each ionized h@xmath52 molecule eventually forms co or some other important molecule , the chemical timescale in the cores of dense clouds is the ratio of the relative abundance of co , @xmath110@xmath69 , to the cosmic ray ionization rate , @xmath14@xmath70 s@xmath71 ; this gives @xmath110@xmath72 years ( d.j . hollenbach , private communication ) . another way in which turbulence may affect chemistry is through localized heating in shocks ( flower & pineau des for@xmath73ts 1998 ) , vortices ( joulain et al . 1998 ) , ambipolar diffusion ( padoan , zweibel & nordlund 2000 ) , and magnetic reconnection ( lazarian & vishniac 1999 ) . there are many reactions that seem to require higher than normal temperatures . for diffuse clouds , the large abundance of ch@xmath59 and molecules like hco@xmath59 that can form from ch@xmath59 , and the large abundance of oh ( gredel 1997 , lucas & liszt 1997 ) , all suggest that high temperature ( @xmath110@xmath74 k ) reactions take place , as in shock fronts ( elitzur & watson 1980 , draine & katz 1986 ) . however , the correlation between ch@xmath59 and the lower - temperature molecule c@xmath52 ( gredel 1999 ) , and the similar radial velocities of ch@xmath59 and ch rule out ch@xmath59 production in single large shocks or cloud surfaces . instead , ch@xmath59 could be made in numerous unresolved shocks that blend on a line of sight with cooler gas at the same average velocity ( gredel , pineau des for@xmath75ts & federman 2002 ) . part of the appeal of turbulence is its great range of scales that can yield such an effect . temperature fluctuations of several hundred degrees could play a major role in the chemistry of many species because of the inverse temperature dependence of radiative association and the range of activation energies for collisional dissociation , neutral - neutral reactions , and some ion - molecule reactions . neutral - neutral reactions may be important even in dark clouds ( bettens et al . prasad & huntress ( 1980 ) long ago pointed out the importance of temperature variations in radiative association reactions and some ion - molecule reactions . also , in dense shielded clouds , warm regions drive most of the o , oh , and o@xmath52 to h@xmath52o through reactions like o + h@xmath76 oh@xmath77h and oh+h@xmath76h@xmath52o+h ( e.g. , charnley 1998 ) , not only enhancing the h@xmath52o abundance , but also preventing the formation of molecules like so@xmath52 , which rely on the oxygen species for their formation . the cn abundance should be enhanced because its destruction by reaction with o and o@xmath52 is suppressed ; this can lead to enhanced production of hc@xmath78n through reaction of cn with c@xmath52h@xmath52 . there have been several attempts to model the effects of local turbulent heating on chemistry . black & van dishoeck ( 1991 ) pointed out that if turbulence extends down to the collisional mean free path , then the tail of the particle velocity distribution could be enhanced . reaction rates are averages over the particle relative velocity distribution , so energy - dependent cross - sections could have greatly enhanced rates . spaans ( 1996 ) noted that the distribution of velocity differences in turbulent flows at small scales has a large tail and suggested this could result in non - maxwellian relative speeds between molecules , especially those involved with ch@xmath59 production . this assumption is questionable because the reaction scales are the mean free paths , of order 1au@xmath79 , and these are smaller than the viscous dissipation scale of incompressible turbulence . the validity of the fluid approximation for neutrals is also questionable at these scales ( see _ interstellar turbulence i _ ) , although mhd turbulence in the ionic component could enhance the ion - neutral collision rate . another approach to the problem was presented in a series of papers by falgarone and coworkers . falgarone & puget ( 1995 ) , falgarone et al . ( 1995 ) , and joulain et al . ( 1998 ) considered the diffuse cloud abundances of ch@xmath59 and several other molecules such as oh and hco@xmath59 . they focused on the solenoidal part of the velocity field in which molecular viscosity on 10 au scales produces numerous local hot spots ( up to several times @xmath80k ) . falgarone et al . ( 1995 ) showed that the resulting chemical column densities were consistent with observations . joulain et al . ( 1998 ) used a burgers - vortex model for the dissipation to explain other features , such as the similarity of ch and ch@xmath59 line centroid velocities , when the number of dissipation regions on the line of sight is large . decamp & le bourlot ( 2002 ) approximated turbulence by a stochastic 1d velocity field correlated in space and time , and solved the continuity equation for the density of each species . for a small reaction network , the abundances developed different patterns on different scales , showing how species segregation can occur in a turbulent flow . species that do not normally peak at the same time for a nonturbulent model can coexist in a turbulent model , which means that the age of a region can not be inferred from the abundances . the issue of chemical differentiation is important because it is well - established and unexplained in a number of starless cores such as tmc-1 , l134n , and l1544 , and observed as 110 au variations in co , h@xmath52co , and oh relative to h@xmath52 toward extragalactic background sources ( marscher , moore & bania 1993 ; moore & marscher 1995 ; for more recent work , see pan , federman & welty 2001 ; rollinde et al . 2003 ; stanimirovic et al . although some of these variations may be attributed to freeze - out of molecules on dust grains ( bergin et al . 2002 ) , most are probably also affected by time - dependent chemistry ( langer et al . one important timescale is cloud contraction ( shematovich et al . 2003 ) , but another is turbulent cycling between different physical conditions over a range of timescales ( decamp & le bourlot 2002 ) . the observed variations at scales so small that gravity is unimportant seem to favor turbulence . the calculations by decamp & lebourlet also develop a bistable chemical pattern in which dual equilibria are possible . however , it is not known how many of the effects result from the imposed artificial velocity field , which is not able to react to the changing density field through pressure . the only truly hydrodynamical turbulence simulation that included a small but relevant chemical reaction subset was by pavlovski et al . ( 2002 ) , who generalized previous work by mac low and coworkers to include a large number of coolants and a wide range of possible temperatures and densities . the reaction network in pavlovski et al . included h , c , o , and five molecules formed from them , and it included h@xmath52 formation on grains . they followed the decay of hypersonic turbulence without self - gravity in a dense ( @xmath81 @xmath64 ) , small ( @xmath82 cm ) region . their surprising result was that after h@xmath52 dissociated in a shock , the molecules reformed in a pattern of filaments , clumps , and diffuse gas yet the abundances were fairly uniform after only 100 years . there are probably other processes that produce localized regions of hot chemistry . for example , ambipolar diffusion heating depends on a high - order derivative of the magnetic field and so can produce heating rates that exceed the average rate by orders of magnitude on very small scales ( padoan , zweibel , & nordlund 2000 ) . magnetic reconnection ( lazarian & vishniac 1999 ) is another possibility for localized hot gas . the chemistry induced by these processes has not yet been investigated . cosmic rays scatter off magnetic waves and mhd turbulence as the particles propagate along magnetic field lines , and they generate waves and turbulence if they stream much faster than the local alfvn speed . observations of cosmic rays at the earth and of radio and x - ray emissions produced by cosmic rays at their sources and throughout the galaxy suggest the ism is filled with irregularities on the scale of a particle gyroradius . this ranges from 1 au to 1 pc for protons with energies between a few gev and the knee of the cosmic ray energy distribution at @xmath83 ev . this section reviews the connection between cosmic rays and turbulence . it is not known if the turbulence that scatters cosmic rays has the same origin as the turbulence on larger scales , i.e. , whether it is part of an energy cascade from these larger scales . it could be generated locally , for example , by the cosmic rays themselves or in other small scale instabilities ( see also section 3 , _ interstellar turbulence i _ ) , and it could be more similar to scintillation turbulence , which has the same small scale ( section [ sect : scint ] ) than molecular cloud turbulence , which is presumably generated on much larger scales . other reviews of cosmic rays are in cesarsky ( 1980 ) , berezinskii et al . ( 1990 ) , drury et al . ( 2001 ) , jokipii ( 2001 ) , ptuskin ( 2001 ) , and schlickeiser ( 2002 ) . the important point about cosmic rays from the perspective of this review is that they scatter frequently , diffusing rather than streaming through the galaxy . their total path length determined from the ratio of secondary to primary nuclei is equivalent to @xmath84 galactic disk crossings at gev energies ( engelmann et al . 1990 , higdon & lingenfelter 2003 ) . their flight time determined from isotope ratios ranges from @xmath115 my for a leaky - box model with a small halo ( simpson & garcia - munoz 1988 , yanasak et al . 2001 ) to @xmath2 years for a diffusion model in a @xmath15 kpc halo ( ptuskin & soutoul 1998 ) . they scatter so frequently that their flux at the earth is isotropic to within @xmath85 ( cutler & groom 1991 , munakata et al . 1997 ) , even though there are only a handful of nearby sources ( i.e. , geminga , vela , lupus loop , loop iii see ptuskin & soutoul 1998 , jones et al . 2001 , taillet & maurin 2003 ) . the scattering distance , @xmath86 , is the step size in a random walk over the galactic thickness @xmath87 ( achterberg , blandford & reynolds 1994 ) . for @xmath88 random isotropic scatterings , @xmath89 , so for total distance travelled @xmath90 with @xmath91 particle speed , we get @xmath92 at gev energies . this is a very small distance on galactic scales , but still much larger than a gyroradius , @xmath93 for rest mass @xmath94 , relativistic @xmath95 , and @xmath96 , magnetic field strength @xmath97(gauss ) and charge @xmath98 with @xmath99 esu . in the expression for @xmath100 , the angle of particle motion with respect to the field line is @xmath101 . taking @xmath102 and @xmath103 , we get @xmath104 cm for hadron particle energy @xmath105 in gev and field strength @xmath106 in microgauss . thus , the mean scattering distance for gev energies is @xmath110@xmath72 gyroradii and the gyroradius has the same scale as the solar system . the relative amplitude of the magnetic irregularities @xmath107 that have the size @xmath100 may be determined from the random walk of angular deflections , @xmath108 , which build up in proportion to the square root of the number @xmath109 of gyrations as @xmath110 for small @xmath111 . the scattering frequency is @xmath112 , and the number of gyrations is @xmath113 for gyrofrequency @xmath114 . thus , @xmath115 and @xmath116 when @xmath1170.2 pc . evidently , magnetic fluctuations of less than one percent on 0.1 au scales can isotropize cosmic rays with gev energies ; in general , a wide spectrum of fluctuations is needed to scatter the full energy range . this level of irregularity is much smaller than observed on @xmath118 pc scales from pulsar dispersion measures , which suggest @xmath119 g and @xmath120 g ( rand & kulkarni 1989 ) . it is also smaller than the integrated fluctuations found by scintillation on scales less than @xmath121 pc , which give @xmath122 g and @xmath123 g ( minter & spangler 1997 ) , and smaller than @xmath124 from optical polarization ( heiles 1996 , fosalba et al . 2002 ) . a spectrum of magnetic energy fluctuations like the kolmogorov or kraichnan spectra will make @xmath111 much smaller on the scale of the gyroradius than these observed field irregularities . such a spectrum will also give the required rigidity dependence for the diffusion coefficient ( strong & moskalenko 1998 ) . the rigidity of a cosmic ray is @xmath125 . cosmic rays can also scatter off strong magnetic irregularities having a scale comparable to or larger than r@xmath126 ( jokipii 1987 ) . inverting the above expression and integrating over a spectrum of waves @xmath127 , the mean free path is @xmath128 ( schlickeiser 1989 ) . for @xmath124 and a particle gyroradius that fits inside the compressed region , @xmath129 , this gives @xmath130 . if the scattering arises in a network of strong shocks in a supersonically turbulent medium , then the mean free path for diffusion will be about the shock separation . because the gyroradius of gev protons is less than the collision mean free path for atoms and ions , collisional mhd models of compressible turbulence do not apply to these small scales ( see _ interstellar turbulence i _ and section [ sect : mirror ] below ) . the field line fluctuations that scatter cosmic rays could be coherent waves , stochastic turbulence , or shock fronts . one common view of waves is that they are weak fluctuations compared with the mean field , with wave travel times much longer than an oscillation period and a well - defined dispersion relation between frequency and wavenumber . particles in such a wave field interact successively with many crests , travelling rapidly along the field until their cumulative pitch angle deviation becomes large and they begin to interact with a different wave train . a turbulent medium is not so regular but particle interactions with a broad wave spectrum can be similar as long as the particle sees a magnetic perturbation with the right wavelength after each gyration . in both cases the field can be treated as static for the fast - moving particles . much of the history of cosmic ray theory is based on observations of the solar wind , which contains a mixture of weak and strong fluctuations , depending on frequency ( drge 1994 ; goldstein , roberts & matthaeus 1995 ) . perhaps @xmath115% of it is in the form of slab - like alfvn waves ( @xmath131 ) and @xmath185% is two - dimensional ( 2d ) magnetic turbulence ( @xmath132 , bieber , wanner & matthaeus 1996 ) . here , @xmath133 and @xmath134 are wavenumbers perpendicular and parallel to the mean field . the solar wind observations led to the quasilinear theory of cosmic ray propagation ( jokipii 1966 ) , which assumes the field is uniform in space and unchanging in time over a particle gyration . interstellar applications of this theory may be appropriate at high spatial frequencies where weak electron perturbations scatter radio waves ( section [ sect : scint ] ) , or near shock fronts where weak waves on either side scatter relativistic particles back and forth ( section [ sect : cracc ] ) , or for waves that are excited by the cosmic rays themselves ( chandran 2000a ) . the waves cause cosmic rays to diffuse in both momentum and space ( e.g. , schlickeiser 1994 ) . momentum diffusion means that velocities change and isotropize ; space diffusion means that cosmic rays spread out slowly to more uniform densities . the two diffusion coefficients are inversely related : higher momentum diffusion means lower spatial diffusion as the particles isotropize their velocities more easily . there are two types of waves that could be important for cosmic ray scattering . alfvn waves are transverse oscillations with a restoring force from field line tension and a frequency less than the ion gyro frequency . their dispersion relation between frequency @xmath135 and wavenumber @xmath26 is @xmath136 for angle @xmath137 between the wave propagation vector and the average field . their group and phase velocity parallel to the mean field is the alfvn speed @xmath138 for ion density @xmath139 . when thermal pressure with sound speed @xmath140 contributes to the restoring force there are also magnetosonic waves with the dispersion relation @xmath141^{1/2}\right).\ ] ] the positive and negative signs are for fast and slow waves . for oblique propagation of magnetosonic waves , when @xmath142 , the phase speed of the fast mode can be large along the field lines : @xmath143 in the low @xmath144 limit ( @xmath144 is the ratio of thermal to magnetic pressures , @xmath145 ) and @xmath146 in the high @xmath144 limit . this speed is important for cosmic rays because they also move quickly along the field lines and will resonate with the magnetic fluctuations in these waves . oblique magnetosonic waves also have a magnetic pressure that varies slightly along the field as a result of the successive convergences and divergences of the velocity and field , which are in phase for fast waves and out of phase for slow waves . these field line changes create magnetic mirrors that enhance the scattering ( schlickeiser & miller 1998 , ragot 2000 ) . the condition for resonance between a cosmic ray and a wave is ( jokipii 1966 , hall & sturrock 1967 , hasselmann & wibberenz 1968 ) @xmath147 this expression says that the doppler shifted frequency of the magnetic wave , as viewed by the particle moving parallel to the field at speed @xmath148 for pitch angle @xmath101 , equals an integer number , @xmath4 , of the gyration frequency . for an alfvn wave , @xmath149 , and for a fast mode at low @xmath144 , @xmath150 . the case @xmath151 is gyroresonance . for wave spectra that decrease with @xmath26 as a power law , only the lowest - order resonances ( @xmath152 ) are important ( cesarsky & kulsrud 1973 ) . magnetosonic waves have an additional resonance for @xmath153 ( lee & vlk 1975 , fisk 1976 , achterberg 1981 ) , when particles stay between two crests by moving along the field at the parallel wave speed . such a particle slows down when it hits the compressive part of the wave ahead of it ( by the mirror effect ) , and it speeds up if the compressive part hits the particle from behind . as a particle oscillates between crests , it can have a parallel velocity that resonates with another wave moving faster along the field , and then jump over to become trapped between two new crests . for a broad wave spectrum , the particle jumps from one wave to the next , gaining energy on average . this process is called transit time acceleration of particles or transit time damping ( ttd ) of waves ( miller 1997 , schlickeiser & miller 1998 ) . ttd increases the parallel component of cosmic ray momentum by mirror scattering in the wave frame . the mirror force depends on the square of the perpendicular component of the velocity ( section [ sect : mirror ] ) , so ttd must operate with the pitch - angle scattering of gyroresonance to maintain momentum isotropy ( miller 1997 ) . a particle that interacts with an alfvn wave can resonate only with the wavelength that is equal to the parallel gyrolength in the doppler shifted frame , @xmath154 . a particle that interacts with magnetosonic waves can respond to the whole wave spectrum because it can choose the wave direction @xmath137 where the parallel wave speed equals its own ( schlickeiser & miller 1998 , schlickeiser & vainio 1999 ) . in practice , such a wave field may not have the appropriate @xmath137 because wave damping is larger at higher @xmath137 . ttd was originally thought to be such an important energy loss for the waves that the fast mode would not be important for cosmic rays ( schlickeiser 1994 , minter & spangler 1997 , tsap 2000 ) . damping is large for high @xmath144 because the thermal particles resonate with the waves and only highly parallel modes exist ( holman et al . 1979 , foote & kulsrud 1979 ) . however , @xmath155 in the solar wind and in much of the cool dense phase of the ism , and observations of the solar wind show that oblique and fast modes dominate over plane - parallel alfvn waves ( tu , marsch & thieme 1989 ; bieber , wanner & matthaeus 1996 ; matthaeus , goldstein & roberts 1990 ) . thus , ttd particle acceleration could be important in parts of the ism where @xmath144 is small . the waves still have to be nearly isotropic to get effective scattering , though , and they should not be damped by other processes , such as ion - neutral or viscous friction . such damping is important in most of the neutral medium where @xmath156 is otherwise favorable for ttd ( kulsrud & pearce 1969 , felice & kulsrud 2001 ) . this leaves few places where wave damping is low enough that ttd might be important . alfvn wave scattering is questionable too because of the extreme anisotropy of the fluctuations that may cascade down from pc scales to the gyroradius . anisotropic turbulence produces tiny perpendicular irregularities that average out over a particle gyration ( chandran 2000a , 2001 ; lerche & schlickeiser 2001 ; yan & lazarian 2002 ; teufel , lerche & schlickeiser 2003 ) . the result is a scattering coefficient that is many orders of magnitude less than in the isotropic case . for this reason , yan & lazarian ( 2003a ) reconsidered ttd scattering by fast modes at moderate @xmath144 , emphasizing the isotropy of the fast waves and considering ism regions where wave damping at particular cosmic ray energies is relatively small . the resolution of this damping problem for wave scattering is far from clear . the energy of fast cosmic ray gyrations is adiabatically invariant during slow changes in the magnetic field amplitude , so an increasing field puts more energy into the gyromotion while removing it from the parallel motion . this change , combined with momentum conservation , causes fast particles with high pitch angles to bounce off converging field lines . the mirror force is @xmath157 for magnetic moment @xmath158 with particle mass @xmath94 . this type of scattering involves very long wavelengths parallel to the field ( ragot 1999 , 2000 ; felice & kulsrud 2001 ; lu et al . the mirror sources could also be molecular clouds and clumps ( chandran 2000b ) . electric fields can also scatter cosmic rays . the flow of plasma at speed @xmath159 transverse to a magnetic field generates an electric field , @xmath160 , that can be important in the parallel direction when the magnetic field has at least 2d structure , and it is convected around by turbulent motions ( fedorov et al . 1992 , le roux et al . 2002 ) . transport transverse to the mean field occurs as individual field lines wander ( jokipii & parker 1969 , bieber & matthaeus 1997 , michalek & ostrowski 1998 , giacalone & jokipii 1999 , kta & jokipii 2000 ) . this is important for radial diffusion in the solar wind , which has a spiral field ( chen & bieber 1993 ) , for diffusion in oblique shock fronts ( duffy et al . 1995 ) , and for diffusion in the vertical direction out of the galaxy . magnetic fields with irregularities smaller than the gyroradius cause particles to skip over field lines , giving anomalous diffusion ( parker 1964 ; chuvilgin & ptuskin 1993 ; casse , lemoine & pelletier 2002 ; otsuka & hada 2003 ; erlykin et al . braided field lines can lead to sub - diffusion ( getmantsev 1963 ) , when the mean squared particle position in the cross field direction increases as the square root of time , instead of directly with time as in normal diffusion . compound diffusion has both cross - field diffusion and diffusion along the field lines ( kirk et al . cross diffusion requires a lot of perpendicular structure to the field lines over a diffusion length for the particle motion parallel to the field ( qin , matthaeus & bieber 2002 ) . for ism turbulence , the cross - field diffusion coefficient is 0.10.2 of the parallel coefficient ( chandran & maron 2004a ) . shocks provide another site for cosmic ray scattering . blandford & ostriker ( 1980 ) considered an ism made of cleared hot cavities from supernovae and showed how the resulting shocks could scatter and accelerate cosmic rays . bykov & toptygin ( 1985 ) extended this model to include secondary shocks that arise in the turbulence caused by the supernovae . klepach , ptuskin & zirakashvili ( 2000 ) included stellar wind shocks . the composite spectrum of cosmic rays from an ensemble of shocks in a supersonically turbulent medium was determined by schneider ( 1993 ) . in such a medium , the summed energy distribution for cosmic rays from all of the shocks is close to a power law , but there can be flat parts or bumps , unlike the energy distribution from a single shock ( see also achterberg 1990 , bykov & toptygin 1993 ) . cosmic rays gain energy as they scatter off randomly moving parts of the ism ( fermi 1949 ) . this is momentum diffusion and called the second - order fermi mechanism because the energy gained per collision depends on the second power of the rms turbulent speed . it is reminiscent of the thermalization of a star cluster with the light particles gaining speed as they approach equilibrium with the heavy particles . scott & chevalier ( 1975 ) first applied this mechanism to supernova remnants , using random motions inside the remnant for scattering sites . supernovae are a likely source for cosmic rays because the cosmic ray energy density , @xmath11 ev @xmath64 , divided by a @xmath115 my lifetime in the galaxy , is about 5% of the total supernova power ( baade & zwicky 1934 , ginzburg & syrovatskii 1964 ) . cosmic ray acceleration by momentum diffusion occurs throughout a compressibly turbulent ism , not just in supernova remnants ( kulsrud & ferrari 1978 , ptuskin 1988 , dolginov & silantev 1990 , bykov & toptygin 1993 , webb et al . 2003 , chandran & maron 2004b ) . in a system of many shocks , particles that are trapped in the weak magnetic turbulence of the ambient medium also get accelerated every time these regions are compressed ( jokipii 1987 ) . multiple shock crossings at the edge of a supernova remnant also accelerate cosmic rays ( see review in blandford & eichler 1987 ) . if the field is parallel to the shock direction , then magnetic turbulence created ahead of the shock by the outward streaming particles ( wentzel 1974 , skilling 1975 ) scatters these particles back into the shock where they encounter more turbulence . the post - shock turbulence was formerly pre - shock turbulence that got compressed and amplified ( schlickeiser , campeanu & lerche 1993 ; vainio & schlickeiser 1998 ) . if the field is oblique to the shock direction , then gyromotions ( jokipii 1987 ) and field line wandering cycle particles through the front ( ragot 2001 ) . each time a particle cycles through the front it gains energy from the converging flow . this is called the first - order fermi mechanism ( fermi 1954 ) because the energy gained per crossing depends on the first power of the shock velocity . about @xmath161 of the incoming thermal particles are injected into cosmic rays ( pryadko & petrosian 1997 ; kang , jones & gieseler 2002 ; bamba et al . 2003 ) , with an efficiency depending on obliqueness ( ellison , baring & jones 1995 ; kobayakawa , honda , samura 2002 ) . generally , the first - order mechanism dominates in strong shocks ( axford et al . 1977 , krymsky 1977 , bell 1978 , ostrowski 1994 ) , although the second mechanism is more important downstream than upstream ( vainio & schlickeiser 1998 ) . shock acceleration explains the cosmic ray energy spectrum up to the `` knee '' at @xmath110@xmath162 ev . the particles that stay in the shock longest end up with the most energy . the spectrum is a distribution function in the number of shock crossings , considering the continuous loss of particles that are trapped in the downstream flow ( e.g. , kato & takahara 2003 ) . observations of the edge sharpness of supernova remnants suggest the amplitude of mhd waves near the shock is @xmath160 times the average ism value ( achterberg , blandford & reynolds 1994 ) . simonetti ( 1992 ) observed a factor of at least 10 in the magnetic wave amplitude from faraday rotation irregularities in a supernova remnant compared with the adjacent line of sight . the x - ray synchrotron emission from supernova remnants is direct evidence for acceleration of relativistic electrons ( koyama et al . 1995 , aschenbach & leahy 1999 ) . a collection of particles with a high enough density can not stream along a magnetic field much faster than the alfvn speed because they generate magnetic irregularities that scatter them ( lerche 1967 ; wentzel 1968a , 1969 ; kulsrud & pearce 1969 ; tademaru 1969 ) . the growth rate of the instability at wavenumber @xmath26 for isotropic wave generation is ( cesarsky 1980 ) @xmath163 the first term in the parenthesis represents the stabilizing effect of wave damping from ion gyromotions . the second term drives the instability with a rate proportional to the ratio of the streaming speed to the alfvn speed . the influence of collective effects is in the ratio of the density of those cosmic rays that resonate with wavenumber @xmath26 to the background density of thermal ions . the size of the magnetic irregularity produced is about @xmath164 times the particle gyroradius ( @xmath110@xmath165 cm for gev energy and @xmath166gauss field ) , so more energetic particles make larger - scale field distortions . these distortions are not at the bottom of a turbulent cascade , so they should not be anisotropic ( chandran 2000a ) . they could cascade to give the ism scintillation . the upper limit to the cosmic ray energy that can generate waves sufficient for their own scattering is around 100 to 1000 gev ( cesarsky 1980 , yan & lazarian 2003a ) . much higher energies are possible in a galactic wind model where the boundary between diffusion and advection depends on energy and the streaming instability landau damps in a highly ionized galactic halo ( ptuskin et al . 1997 ) . hall ( 1980 ) suggested that waves generated by cosmic rays would damp quickly and proposed instead that scintillation - scale structures result from mirror and firehose instabilities in the hot ism phase . both require pressure anisotropies , which hall notes should arise at the level of 0.01 - 0.1 if the hot medium is turbulent . the firehose instability requires @xmath167 for parallel and perpendicular pressures @xmath168 and @xmath169 ( lerche 1966 , wentzel 1968b ) , and the mirror instability requires @xmath170 . these conditions are satisfied for anisotropy @xmath1710.010.1 if the plasma @xmath172 is very large , as might be the case in the hot intercloud medium if supernovae continuously sweep it free of gas and field . one of the earliest indications that the ism is turbulent came from scintillation observations of electron density fluctuations on very small scales . these fluctuations cause diffraction and refraction of radio signals from pulsars and a few extragalactic sources . diffraction broadens pulsar images , spreads out the pulse arrival times , and narrows the frequency interval over which the pulses have a coherent behavior . the relative motion of the diffracting medium also modulates the pulsar intensity on a timescale of minutes . refraction from larger structures causes the images to split or wander in position and to vary in intensity on timescales of days to months . scintillation effects like these offer many diagnostics for electron density structures in the ionized interstellar medium , including _ regions , hot bubble edges and the ionized and hot intercloud media . still , the origin of this turbulence is not clear . because the scales are very small , typically @xmath83 cm or below , it could be part of a cascade from larger structures past the viscous length and down into the collisionless regime of mhd turbulence ( section 4.11 in _ interstellar turbulence i _ ) , or it could be generated locally by cosmic ray streaming or other small - scale instabilities or by low - mass stellar winds and wakes ( sections 3 and [ sect : cr ] in _ interstellar turbulence i _ ) . here we review ism scintillations . more extensive reviews may be found in rickett ( 1977 , 1990 ) , hewish ( 1992 ) , and cordes , rickett & backer ( 1988 ) . the index of refraction for radio waves is ( nicholson 1983 ) @xmath173 with electron density @xmath174 , classical electron radius @xmath175 cm , and wavenumber of the radio signal @xmath176 for wavelength @xmath86 . the phase change of a signal that passes through a clump of size @xmath177 with an excess index of refraction @xmath178 compared with a neighboring average region is @xmath179 . the transmitted signal adds constructively to the neighboring signal at a relative propagation angle @xmath180 if the difference in their path lengths , @xmath181 , multiplied by @xmath26 , equals the phase change . this gives @xmath182 for small @xmath180 . this phase change is random for each clump the signal meets , so after @xmath183 such clumps on a path length @xmath184 , the root mean square angular change of the signal from diffraction is @xmath185 . the most important clump size has a cumulative scattering angle @xmath186 equal to the clump diffraction angle , @xmath187 . evaluation of @xmath188 requires some knowledge of how @xmath189 depends on @xmath177 . for a turbulent medium , this relation comes from the power spectrum of electron density fluctuations , @xmath190 where @xmath191 and @xmath192 for 3d turbulence with a kolmogorov spectrum ( we use @xmath193 here for the wavenumber of electron density fluctuations to distinguish it from the wavenumber @xmath26 of the radio radiation used for the observation ) . the mean squared electron density fluctuation is obtained from the integral over phase space volume , @xmath194 . for logarithmic intervals , this is approximately @xmath195 . thus , we have @xmath196 . using the wave equation , cordes , pidwerbetsky & lovelace ( 1986 ) got essentially the same result with @xmath197 replacing @xmath198 . the full theory more properly accounts for differences in the cumulative effects of the line - of - sight and transverse density variations . a recent modification considers power - law rather than gaussian statistics for the clump size distribution ( boldyrev & gwinn 2003 ) . for a reference set of parameters , e.g. , ghz frequencies , kpc distances , and typical @xmath199 meters@xmath200 , the characteristic scale is @xmath201 cm . this small size implies that only pulsars and a few extragalactic radio sources can produce detectible diffraction effects at radio frequencies . the smallness compared with the collisional mean free path of electrons also implies that the fluctuations have to be collisionless ( see _ interstellar turbulence i _ ) . many observable properties of small radio sources give information , either directly or indirectly , about @xmath177 , which can be used to infer the strength and power index of the electron density fluctuation spectrum and its distribution on the line of sight . for example , the logarithm of the visibility of an interferometer is @xmath202 where @xmath40 is the baseline length and @xmath203 is the scattering measure . spangler & cordes ( 1998 ) found @xmath204 using this baseline dependence for 4 pulsars . very long baseline interferometry ( vlbi ) observations of pulsar image sizes measure angular broadening as a function of frequency ( lee & jokipii 1975a ; cordes , pidwerbetsky & lovelace 1986 ) and determine @xmath205 ( spangler et al . for the reference parameters , @xmath206 mas . angular broadening gives a spread in path lengths for the radio waves , and this corresponds to a spread in arrival times of pulses , @xmath207 ( lee & jokipii 1975b ) . such pulse broadening scales as @xmath208 , so the frequency dependence gives @xmath101 and the absolute value gives @xmath205 for an assumed screen depth @xmath184 . the flux density is correlated only over a small range of frequencies , @xmath209 ( salpeter 1969 , lee & jokipii 1975b ) . pulsar amplitudes vary on a timescale @xmath210 several minutes as a result of pulsar transverse motions @xmath211 km s@xmath71 ( cordes 1986 ) . using the relations above , this can be rewritten @xmath212 . if pulsar distances and proper motions are also known , along with the dispersion measure , then the distribution of scattering material on the line of sight can be modeled ( harrison & lyne 1993 , cordes & rickett 1998 ) . longer time variations ( sieber 1982 ) result from the changing refraction of radio signals in moving structures that have the diffraction angular size @xmath213 or larger ( rickett , coles & bourgois 1984 ; blandford & narayan 1985 ) . this angular size corresponds to a physical size for the refracting elements @xmath214 . for typical @xmath215 mas , @xmath216 cm at @xmath217 kpc . the corresponding refraction scintillation time is @xmath218 days . time variations over months are also observed from larger interstellar structures . the relative rms amplitude of the source is the modulation index , @xmath219 . for diffraction this can be 100% , but for refraction it is typically less than @xmath130% ( stinebring & condon 1990 , labrecque et al . 1994 , stinebring et al . the modulation index depends on the strength of scattering , which involves a length scale equal to the geometric mean of the scale for angular broadening and the dominant scale for electron density fluctuations , @xmath220 . this is the fresnel length , @xmath221 , which is the transverse size of some object at distance @xmath184 that is just small enough to show diffraction effects . when @xmath222 is high , there are many diffracting elements of size @xmath223 inside each refracting element of size @xmath224 , so diffraction is strong . for weak diffraction , the total phase change on the line of sight is small and the dominant clump size is the fresnel length itself ( lovelace et al . 1970 , lee & jokipii 1975c , rickett 1990 ) . in this case , the modulation index scales approximately with @xmath225 ( lovelace et al . the modulation index measures interstellar properties in this weak limit because then @xmath226 , giving a diagnostic for @xmath227 and @xmath101 ( rickett 1977 ) . small @xmath94 usually corresponds to high frequencies , where diffraction is relatively unimportant . time variations can be visualized with a dynamic spectrum , which is a gray scale plot of intensity on a coordinate system of time versus frequency ( figure [ fig : dynamspect ] ) . diffraction alone gives a random dynamic spectrum from the motion of unresolved objects of size @xmath177 , but larger refractive structures , which disperse their radio frequencies over larger angles , @xmath228 ( from the above expression @xmath229 ) , sweep a spectrum of frequencies past the observer ( hewish 1980 ) . the intensity peaks on a dynamic spectrum ( averaged over many pulses ) then appear as streaks with slope @xmath230 ( cordes , pidwerbetsky & lovelace 1986 ) ; this can be used as a diagnostic for @xmath231 . the right - hand part of figure [ fig : dynamspect ] is the 2d power spectrum of the dynamic spectrum . stinebring et al . ( 2001 ) and hill et al . ( 2003 ) suggest that the arcs arise from interference between a central image and a faint scattering halo 2030 times larger . strong refraction can result in multiple images that produce interference fringes on a dynamic spectrum ( cordes & wolszczan 1986 ; rickett , lyne & gupta 1997 ) . multiple images imply that refraction , which determines the image separation , bends light more than diffraction , which determines the size . because large - scale fluctuations dominate refraction , this observation implies either @xmath232 , so that small wavenumbers dominate ( cordes , pidwerbetsky & lovelace 1986 ; romani , narayan & blandford 1986 ) , or there is additional structure on @xmath233 cm scales that is not part of a power law power spectrum ( rickett , lyne & gupta 1997 ) . radio scintillation has been used to determine many properties of electron density fluctuations in the ism . diffraction and refraction come from the same electrons , so the ratio of their strengths is proportional to the ratio of the amplitudes of the electron density fluctuations on two different scales . from this ratio armstrong , rickett & spangler ( 1995 ) derived the power spectrum of the fluctuations spanning @xmath234 orders of magnitude in scale . they determined @xmath235 m@xmath200 and @xmath192 between @xmath81 cm and @xmath236 cm . they also suggested from rotation measures ( rm@xmath237 ) that the same power law extends up to @xmath238 or @xmath239 cm . cordes , weisberg & boriakoff ( 1985 ) found @xmath240 using the relation between @xmath241 and @xmath86 . they mapped the spatial distribution of @xmath227 from @xmath241 observations of 31 pulsars , suggesting the galaxy contains both thin and thick disk components . for the thin component , @xmath242 m@xmath200 assuming @xmath192 and for the thick component , with @xmath243 kpc , @xmath244 m@xmath200 . the rms level of electron density fluctuations at the dominant scale @xmath201 cm , which comes from the integral over the power spectrum ( see above ) , was found to be @xmath245 @xmath64 for the high latitude medium , for which @xmath246 @xmath64 from dispersion measures , and @xmath247 @xmath64 for the low latitudes . bhat , gupta & rao ( 1998 ) observed 20 pulsars for three years to average over refraction variations and determined @xmath227 in the local ism from @xmath241 . they observed an excess of scattering material at the edge of the local bubble , which also produced multiple images of one of these pulsars ( gupta , bhat & rao 1999 ) . the edge of the local bubble may also have been seen by rickett , kedziora - chudczer & jauncey ( 2002 ) as a source of scattering in a quasar . bhat & gupta ( 2002 ) found a similar enhancement at the edge of loop i , where @xmath248 pc m@xmath200 and the density enhancement is a factor of @xmath1100 over the surrounding gas . they also found excess scattering for more distant pulsars from the sagittarius spiral arm . lazio & cordes ( 1998a , b ) used the angular sizes of radio sources and other information on lines of sight to the galactic center and outer galaxy to suggest that the scattering material is associated with the surfaces of molecular clouds . ionized cloud edges were also suggested by rickett , lyne & gupta ( 1997 ) to explain multiple images . spangler & cordes ( 1998 ) observed @xmath249 from small sources behind six regions in the cygnus ob1 association and found an excess in sm that was correlated with the emission measure , indicating again that scattering is associated with _ h ii _ regions . this result is consistent with the high cooling rate required by the dissipation of this turbulence , which implies high temperatures ( zweibel , ferriere & shull 1988 ; spangler 1991 ; minter & spangler 1997 ) . the relation @xmath250 combined with pulsar distances and proper motions led gupta ( 1995 ) to determine a 1 kpc scale height for the scattering layer from long - term observations of 59 pulsars . cordes & rickett ( 1998 ) also used this method to find that scattering is rather uniformly distributed toward two pulsars , but for three lines of sight it was concentrated toward the pulsar , including vela where the supernova remnant is known to contribute to the scattering ( desai et al . 1992 ) ; six other pulsars in that study had significant scattering from either a foreground spiral arm or _ h ii _ regions . bhat , gupta & rao ( 1999 ) compared @xmath251 obtained from @xmath241 with the angular size of the refraction pattern , @xmath252 from dynamic spectra . they showed that @xmath253 for all 25 pulsars that had this data , implying that diffraction dominates refraction and therefore @xmath254 . they also compared @xmath227 from @xmath255 on the refractive scale with @xmath227 from @xmath241 on the diffractive scale to determine the slope of the power spectrum directly ( as did armstrong et al . 1995 ) . they found the kolmogorov value @xmath192 to within the accuracy in most cases , but six pulsars , mostly nearby , gave slightly higher @xmath256 . the longest - term variations had significantly more power than an extrapolation of the kolmogorov spectrum , however , and the modulation indices were large , leading them to suggest an additional component of scattering electrons at scales of @xmath257 to @xmath83 cm . the same data led bhat , rao & gupta ( 1999 ) to derive @xmath258 m@xmath200 and @xmath259 , indicating strong scattering . lambert & rickett ( 2000 ) looked at the correlation between the modulation index for long - term variations from refraction and the relative decorrelation bandwidth , @xmath260 , which comes from diffraction . the correlation at 610 mhz for 28 sources fit kolmogorov scaling ( @xmath192 ) better than a shock - dominated model ( @xmath261 ) model but at 100 mhz the kolmogorov fit was not as good . they suggested that the minimum turbulent length was large , @xmath262 to @xmath233 cm instead of @xmath263 cm ( armstrong , rickett & spangler 1995 ) . an excess of electron density structure at large scales could also explain the discrepancy . stinebring et al . ( 2000 ) monitored the modulation index of 21 pulsars for five years and found low values ( @xmath264% ) that bracket @xmath265 with no perceptible inner ( smallest ) scale for most of the pulsars . for the crab , vela , and four others , enhanced modulation indices were consistent with an inner scale of @xmath262 cm . rickett , lyne & gupta ( 1997 ) suggested this excess scattering could be from au - sized ionized clumps at cloud edges . an inner scale of @xmath266 cm was found by spangler & gwinn ( 1990 ) after noting that interferometer baselines shorter than this had scintillation with @xmath2674 whereas longer baselines had @xmath101 near the kolmogorov value , @xmath268 . they suggested the inner scale is caused by a lack of turbulence smaller than the ion gyroradius , @xmath269 for thermal speed @xmath270 , or by the ion inertial length , @xmath271 , whichever is larger . this value of the inner scale is consistent with an origin of the scattering in the warm ionized medium or in _ h ii _ regions , but not in the hot coronal medium , which has much larger minimum lengths . shishov et al . ( 2003 ) studied pulsar psr b0329 + 54 over a wide range of frequencies and found a power spectrum for electron density fluctuations with a slope of @xmath272 for lengths between @xmath2 cm and @xmath273 cm . this spectrum is expected for turbulence in the iroshnikov - kraichnan model ( see _ interstellar turbulence i _ ) . shishov et al . noted how other lines of sight though the galaxy gave different spectra and suggested that the nature of the turbulence varies from place to place . they also found refraction effects on a scale of @xmath274 cm corresponding to electron density fluctuations of strength @xmath275 @xmath64 and suggested these were neutral clouds with an overall filling factor of @xmath10.1 . scattering image anisotropy suggests anisotropic turbulence ( see _ interstellar turbulence i _ ) . lo et al . ( 1993 ) observed 2:1 anisotropy in images of sgr a@xmath276 , wilkinson , narayan & spencer ( 1994 ) found scale - dependent anisotropic images of cygnus x-3 , frail et al . ( 1994 ) observed 3:1 anisotropy in scattering of light from galactic center oh / ir stars , whereas trotter , moran & rodriguez ( 1998 ) observed axial ratios of 1.21.5 for quasar light that scatters through a local _ h ii _ region . a ratio of 4:1 was found for another quasar by rickett , kedziora - chudczer & jauncey ( 2002 ) . spangler & cordes ( 1998 ) observed anisotropic scattering with axial ratio of 1.8 surrounding the cygnus ob association and desai & fey ( 2001 ) observed axial ratios of @xmath11.3 toward the same region . the actual anisotropy of local fluctuations can not be determined from these observations because many different orientations blend on the line of sight ( chandran & backer 2002 ) extreme scattering events are observed in some extragalactic radio sources and a few pulsars . their modulation is strong , @xmath277% , they can last for several months , and they have light profiles that are flat - bottom with spikes at the end , or smooth bottom with no spikes ( fiedler et al . 1994 ) . they may result from supernova shocks viewed edge - on ( romani , blandford & cordes 1987 ) or ionized cloud edges in the galaxy halo ( walker & wardle 1998 ) . some appear correlated with the edges of local radio loops ( fiedler et al . 1994 , lazio et al . 2000 ) , in support of the shock model . the actual scattering process could be a combination of refractive defocusing , during which an intervening electron cloud produces a lens that diverges the light and makes the source dimmer ( romani et al . 1987 , clegg et al . 1998 ) , or stochastic broadening by an excess of turbulence in a small cloud ( fiedler et al . lazio et al . ( 2000 ) found an excess @xmath278 kpc m@xmath200 associated with an event , corresponding to @xmath279 m@xmath200 for ratio @xmath280 of the line - of - sight extent ( @xmath184 ) to the transverse extent . fiedler et al . ( 1994 ) suggested @xmath281 @xmath64 . this high level of scattering along with an observed increase in angular size during the brightness minimum ( lazio et al . 2000 ) suggests the scattering cloud is not part of a power spectrum of turbulence but is an additional au - size feature ( lazio et al . 2000 ) . the amplitude , slope , and anisotropy of the power spectrum of interstellar electron density fluctuations have been observed by scintillation experiments . estimates for the inner scale of these fluctuations range from @xmath54 cm to @xmath262 cm or more . the shorter of these lengths is about the ion gyroradius in the warm ionized medium ( spangler & gwinn 1990 ) . this is much smaller than the mean free path for electron collisions so the fluctuations are collisionless ( see _ interstellar turbulence i _ ) . a lower limit to the largest scale of @xmath110@xmath282 cm was inferred from rotation measures assuming a continuous power law for fluctuations , but this assumption is uncertain . the slope of the power spectrum is usually close to the kolmogorov value , @xmath268 , and distinct from the slope for a field of discontinuities , which is @xmath283 . deviations from the kolmogorov slope appear in some studies , and may be from a large inner scale , @xmath284 cm , an excess of scattering sites having size @xmath1110 au , or a transition in the scaling properties of the turbulence . scintillation arises from both the low - density , diffuse , ionized ism and the higher - density _ h ii _ regions , ionized cloud edges , and hot shells , where the amplitude of the power spectrum increases . on average , the relative fluctuations are extremely small , @xmath285 on scales that dominate the diffraction at ghz radio frequencies , @xmath286 cm ( cordes et al . 1985 ) . for a kolmogorov spectrum , they would be larger on larger scales in proportion to @xmath287 ; if the spectrum is continuous up to pc scales , the absolute fluctuation amplitude there would be @xmath11 ( lee & jokipii 1976 ) . anisotropy of scattered images is at the level of 50% , which is consistent with extremely large intrinsic anisotropies ( @xmath288 ) from mhd turbulence ( sections 4.11 and 4.13 in _ interstellar turbulence i _ ) if many orientations on the line of sight are blended together . the turbulence that is observed directly on resolvable scales in the ism also has important effects on very small scales , down to the level of atomic diffusion , mixing , and viscosity , and continuing far below to the thermal ion gyroradius . the resolvable turbulence in the neutral medium was reviewed in the first part of this series ( _ interstellar turbulence i _ ) , along with general theory and simulations . the smaller scale implications were reviewed here . turbulence helps to disperse the elements made in supernovae and other stellar sources by stretching and folding the contaminated gas until the gradient length approaches the collision mean free path . then atomic diffusion does the final step of interatomic mixing . this mixing and homogenization may occur partly at the source , in the shock fronts and contact discontinuities around the expanding flow following dynamical instabilities , and it may occur partly in the ambient ism after the expansion subsides . observations suggest that the dispersion in elemental abundances among stars inside clusters and field stars of the same age , and the abundance differences between hii regions and diffuse clouds , are only a few percent up to perhaps 30% . observations of clusters also suggest that most of the mixing occurs within several times @xmath2 years . this homogenization has to occur between the injection scale of supernovae and superbubble explosions and the star formation scale containing on the order of a solar mass . the corresponding range of spatial scales is @xmath289 for gas at the ambient density . homogenization can not be significant on scales much larger than a superbubble , because then it would diminish the galactic radial abundance gradient over a hubble time . these gradients enter the problem in another way too because turbulence mixes the gas over galactic radius to increase the elemental dispersion locally at the same time as it mixes this gas locally and leads to homogenization . while the distribution of elemental abundances appears to be fairly narrow , with a dispersion on the order of 10% , it may also have a fat tail caused by occasional odd stars and diffuse clouds with very different abundances . this tail is reminiscent of other fat tails in the distribution functions for turbulent media , and these other tails seem to originate with intermittency . in the case of elemental dispersion , this means that the homogenization process is spotty so some regions survive for long periods of time with very little mixing . several methods for studying turbulent mixing have been employed . because of the wide range of scales involved , these studies often employ simplifying assumptions that appear to capture the essential physics . they include artificial stochastic velocity fields and closure methods using moment equations . direct numerical simulations of ism elemental dispersion have only just begun . the theory suggests that a turbulent medium mixes passive scalars like elemental abundances faster than a gaussian velocity field because of the long - range correlations that are associated with turbulence . turbulence affects interstellar chemistry by mixing regions with different properties , heating the gas intermittently on the viscous scale , and enhancing ion - neutral collisions in regions with strong magnetic field gradients . turbulent mixing is more far - reaching than diffusive mixing because of long - range correlated motions in a turbulent flow . such mixing spreads out each chemical species over a large radial range inside a cloud . turbulent heating promotes temperature - sensitive reactions inside otherwise cold clouds . applications to the formation of ch@xmath59 and oh in diffuse clouds look promising . chemical reaction networks that include these turbulent processes are only beginning to understand some of the implications , and direct simulations of chemistry in a turbulent medium are limited to only a few studies so far . the scattering and acceleration of cosmic rays depends strongly on the presence of turbulence in the ism . the scale for this turbulence is the gyroradius , which is less than the collisional mean free path for most cosmic rays , which have energies less than 1 gev . thus , scattering relies entirely on magnetic irregularities in collisionless plasma turbulence . cosmic ray acceleration is by two processes , both of which involve turbulence : the first - order fermi mechanism accelerates cosmic rays by cycling them through shock fronts where they pick up a relative velocity kick comparable to the shock speed at each passage . this cycling occurs because magnetic irregularities on each side of the front scatter the out - streaming cosmic rays back into the front . the second - order fermi mechanism accelerates cosmic rays through turbulent diffusion . each collision with a randomly moving magnetic irregularity turns the cosmic ray around with a reflection speed in the moving frame that is comparable to the incident speed . over time , the result is a transfer of energy from the turbulence to the cosmic rays . cosmic ray scattering occurs in several ways . fast moving particles can interact weakly but resonantly with numerous magnetic irregularities , and successive interactions of this type can randomly change the pitch angles of their helical motions along the field . this eventually leads to a complete reversal in the direction of motion . cosmic rays can also interact strongly with large magnetic irregularities , as would occur in shock fronts and at the edges of dense cloud complexes . these interactions change the particle directions significantly each time . the nature of cosmic ray diffusion in both momentum and space is not understood well because the structure and strength of the important magnetic irregularities are not observed directly . if mhd turbulence is highly anisotropic on the scale of cosmic ray scattering , with transverse irregularities much stronger than parallel as suggested by theory , then resonant scattering processes during motions along the mean field can be very weak . the structure of magnetic waves below the mean free path is also unclear , as the usual fast and slow magnetosonic modes do not exist . cosmic rays can also generate turbulence by streaming instabilities following particle - wave resonances or by fire - hose and mirror instabilities that operate even without resonances . the expected anisotropy of alfvn wave turbulence diminishes the first of these mechanisms significantly , however . the others are problematic because they require strong cosmic ray pressure anisotropies . as a result , the impact of cosmic rays on turbulence is currently not understood . radio wave scintillation is indirect evidence for interstellar plasma turbulence . these radio observations span a very wide range of spatial scales through a combination of diffraction and refraction effects . the scales are mostly below the collision mean free path , and they are far below the limits of angular resolution . the first of these limits makes it difficult to understand the origin of the density irregularities . unlike the turbulence that scatters cosmic rays , which requires only magnetic irregularities and no density structure , the turbulence that causes scintillation requires small - scale density irregularities in the ionized medium . the associated magnetic irregularities are not observed , and the connection to cosmic ray scattering is unclear , even though the length scales are about the same . the origin of density structures below the collision mean free path is unknown . atomic diffusion should smooth them out on a sound crossing time unless magnetic field irregularities hold them in place . in that case they could be the result of slight temperature variations , with the cooler regions having higher electron densities , all divided up and mixed together by transverse magnetic motions . the nature of these motions below the mean free path is unclear , because , as mentioned above , the usual fast and slow compressional modes do not exist , nor do the usual thermal and pressure - regulated processes . the second of the two limits on spatial scale imply that the geometrical properties of scintillation turbulence are difficult to observe . the scintillation is clearly anisotropic , but whether it is in sheets or filaments , for example , is unknown . the importance of scintillation observations for studies of ism turbulence is that they give the power spectrum of electron density fluctuations fairly accurately . this is usually close to the kolmogorov spectrum of incompressible turbulence . rarely are the spectra so steep that the medium can be interpreted as a superposition of sharp edges , like shock fronts . one recent observation with fairly high precision obtained the relatively shallow iroshnikov - kraichnan power spectrum , leading to the conjecture that the slope varies from region to region . as discussed in _ interstellar turbulence i _ , this variation may arise from a variation in the relative strength of the magnetic field compared to the turbulent motions , with the iroshnikov - kraichnan spectrum present in regions of relatively strong fields . observations suggest that scintillation arises in a distributed fashion from the ambient ionized medium and also from discrete high - density places like the edges of local bubbles , ionized molecular clouds , and hii regions . these discrete regions are likely to be highly turbulent and they also have a juxtaposition of hot and cool gas , which is necessary for isentropic mixing and electron density structure . there are evidently many uncertainties in the nature of ism turbulence on small scales , even though the evidence for this turbulence is pervasive . part of the problem is that none of the features of this turbulence have been observed directly : not the densities , magnetic fields , temperatures , or motions . still , the density irregularities are revealed indirectly through scintillation , the magnetic field irregularities through cosmic ray scattering , the temperature fluctuations through chemistry , and the motions through elemental and chemical mixing . an additional problem is that many small scale effects of ism turbulence rely on details of the theory that are independent of the usual scaling relations , such as viscous heating and elemental diffusion , which arise at the bottom of the cascade in the neutral medium . turbulence in the ionized medium is also below the collisional mean free path , where pressure and thermal effects are relatively unimportant . moreover , the small scale ism processes are often strongly dependent on the large scale processes , such as turbulent shock formation , energy and metal injection , and galactic - scale gradients . this means that direct simulations of small scale turbulence are impossible without simplifying assumptions about the large - scale medium assumptions that require more knowledge about ism turbulence on the large scale than is presently available . while the observations and theory of ism turbulence have come a long way from the first efforts in the 1950s , the details of this new information have led to a growing awareness that the complete problem is far too large to solve any time soon . we rely on future generations of astronomers and physicists to continue this work , and hope that they find this field as intriguing and challenging as we do today . we are grateful to a. brandenburg , d. lambert , f. matteucci , s. oey , and g. tenorio - tagle for helpful comments on section 2 ; e. falgarone , d. hollenbach and j. le bourlot for helpful comments on section 3 ; b. chandran , r. jokipii , a. lazarian , v.s . ptuskin , and r. schlickeiser for helpful comments on section 4 ; b. rickett , s. spangler , and d. stinebring for helpful comments on section 5 and to a. hill for figure [ fig : dynamspect ] .

wireless networks are ubiquitous and are on their way to become even more prevalent , e.g. , with the advent of internet - of - things . wireless communication is , however , particularly challenging to model algorithmically . in two crucial interrelated aspects , wireless networks on the ground differ from models typically assumed in algorithmic studies . one is the _ communication modeling _ : when is a transmission successfully decoded , as a function of the environment and the _ interference _ from other transmissions . the other is _ variability with time _ : wireless networks are particularly susceptible to changes . both of these are hard to capture accurately with well - defined , clear - cut rules . we aim in this paper to address core information dissemination problems _ local broadcast _ and ( global ) _ broadcast _ in dynamic distributed networks , under very weak assumptions on the communication . to this end , we propose a communication model with significant flexibility that allows for adversarial control , generalizing essentially all known analytic wireless models . it allows for more general interference relationships than treated before . the network can experience adversarial dynamic behavior , both edge changes ( change of signal strengths ) and node insertions / deletions . wireless communication has traditionally been modeled theoretically by graphs , either geometric or general . interference is then also transmitted on graph edges ( precluding a node from receiving a message from a neighbor if another neighbor is transmitting ) , but is sometimes represented by a supergraph . _ fading channel _ or _ physical _ models common in communication engineering , on the other hand , consider interference as cumulative , decreasing with distance but adding up . they have been popular in recent algorithmic studies , adding more realism to the formulation . the standard assumption of _ geometric signal decay _ ( that signal decreases inverse polynomially with distance ) in the sinr model , the prototypical fading channel , is though equally at odds with experimental evidence . ultimately , it may prove futile to hope for a clean deterministic model , or even a purely stochastic one , without a significant dose of unpredictability and non - determinism . wireless communication is commonly closely linked to mobility , as the transceivers are more often than not on the move . dynamic changes to reception conditions have many causes other than node mobility , since almost any changes in the environment affect transmissions due to reflections of signals over multiple paths , antenna characteristics , scattering , and diffraction . these changes are by nature hard to predict , even when assuming a `` mobility model '' . the most robust approach would then be to assume a non - trivial adversarial component . while the study of algorithms in dynamic networks has a long history , little has been done in cases where interference plays a role . * setting and model*. nodes are distributed and autonomous . there is no built - in structure and the nodes have no information besides bounds on model parameters and an upper bound on the number of nodes . communication is locally synchronous , but there is no global clock . we will assume a very general model for when nodes successfully communicate . nodes are located in space , with separation between points given by the relative decrease in strength of signal ( or interference ) sent between the points . this induces a metric space , when these decays are raised to the appropriate power actually , it is a _ quasi - metric _ , since symmetry need not hold . for distributed computation to be possible , the quasi - metric must have _ bounded independence _ ( to be defined precisely ) . _ edge changes _ can occur , with some restrictions , which are changes in the signal strength between the pair of points . the rule for when communication is successful is only partially pre - specified : transmission succeeds on a `` clear channel '' . that is , if a sender is within a _ communication radius _ from the receiver , if no other node transmits within a ( larger ) radius , and if the combined interference from all other transmitting nodes is ( quite ) small , then the transmission succeeds . otherwise , success is up to the adversary , or it can be further specified by the particular model assumptions desired . this captures essentially all known algorithmic wireless models ( including quasi - unit disc graphs , unit - ball graphs , bounded - independence graphs , @xmath0-hop extensions , and sinr ) . the only exception is the radio network model with general graphs , which can not be extended to involve comprehensive interference without a major hit in time complexity . the generality of our model is a key feature . given the vagaries of actual wireless environments , it is preferable for robustness reasons to make minimal assumptions about the communication model . a conservative approach is then to seek algorithms that work in most established models rather than depending on model - specific factors that simplify the life of the algorithm designer . [ [ dynamic - network - model ] ] dynamic network model + + + + + + + + + + + + + + + + + + + + + the network is dynamic , both in terms of nodes and links . nodes can come and go , potentially adversarially . this includes the situation where nodes are wakened by other nodes . we only expect nodes that have been awake for long enough ( logarithmic number of rounds ) to participate in the computation , i.e. , the _ stable _ nodes . in the _ local broadcast _ problem , all stable nodes should inform all of their ( stable ) neighbors ( i.e. , within the communication radius ) . in the _ broadcast _ problem , the nodes in the network are to be informed of a message that initially resides at a source node . the network need not be connected at any instant in time , but we need to make assumptions about what transient paths are usable by an algorithm operating under interference . this leads to a definition of a _ dynamic diameter _ of the network , that becomes the competitive objective for the time complexity . * our approach and results*. the key algorithmic technique is a natural randomized _ contention balancing _ procedure , where a node continuously adjusts its transmission probability based on the interference that it senses . it allows nodes to stabilize quickly from any initial conditions , or after waking up . this routine is a variation on an old story , a simple backoff procedure to manage local contention : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if a node @xmath1 senses contention in a given round beyond a fixed threshold , then @xmath1 halves its transmission probability in the next round and otherwise doubles it . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ our main technical contribution is to show logarithmic - round convergence of this method to a steady state of nearly balanced contention , from an arbitrary starting configuration and in the presence of network changes . it proves also to be surprisingly tolerant of different communication models . the higher level algorithms are then built on top of this primitive . a crucial component is the use of _ carrier sense _ to detect the cumulative amount of signals in the air . since it is supplied by the cheapest available hardware today as rss ( received signal strength ) readings , we posit that carrier - sense capability should be the default assumption in wireless algorithmics ( while exploring the necessity of different assumptions is interesting theoretically ) . as carrier - sense indicators can provide fine - grained information , we are interested in restraining its use and identifying which aspects are necessary to achieve the results obtained . to this end , we identify several _ primitives _ that carrier - sense can supply , restrict the algorithm to use only a subset of the primitives , and examine which of these are truly necessary . the local broadcast algorithm simply runs the contention balancing procedure , with nodes bowing out when they are sure to have completed their transmission . the broadcast algorithms are based on sparsifying the instance , so that only nodes of constant density actually participate in the global broadcast action . the former is achieved in @xmath2 time , where @xmath3 is the maximum number of neighbors that a node can have , while the latter takes @xmath4 rounds , where @xmath5 is a dynamic diameter . these dissemination algorithms are efficient enough to improve on some of the results known for static versions of the problems . the local broadcast algorithm is strongly optimal , or within constant factors on every instance . in the standard setting ( static , spontaneous case ) , the algorithm is _ uniform _ , in that it need not know the network size . the broadcast algorithm is also optimal and uniform in the same setting , while in the non - spontaneous setting it is faster by a logarithmic factor than the previous algorithm of @xcite that however does not require carrier sense . * closely related work*. there are two largely disjoint bodies of work of wireless algorithmic results , with work on fading models like sinr slowly catching up with the better studied graph - based models . one approach for capturing more realism in sinr model is to move beyond euclidean metrics @xcite , even to general ones @xcite . one can view relative signal decrease as implicitly defining a quasi - distance metric @xcite . link scheduling problems can be formulated on edge - weighted interference graphs @xcite that properly generalize both graph - based and sinr models , linked by a graph - theoretic parameter . distributed dissemination problems , however , necessarily require metric restrictions , such as doubling or _ fading _ metrics @xcite , and limits on communication abilities in order to capture both types of models . the local broadcast and global broadcast problems have been extensively studied in both graph - based radio network models @xcite and the sinr model @xcite . for local broadcast , the best results known in the radio network model are both @xmath6 with and without knowing ( an upper bound on ) @xmath3 @xcite . in the sinr model , with knowledge of @xmath3 , the local broadcast can be accomplished in the same time bound as in the radio network model @xcite . if @xmath3 is not known , the best result is @xmath7 @xcite , which is improved to @xmath2 with free acknowledgments @xcite . this can be further improved to @xmath8 in the spontaneous case , when @xmath3 is known @xcite . the time complexity of non - spontaneous broadcasting in the radio network model is @xmath9 @xcite without collision detection . with collision detection , this lower bound was recently broken in @xcite , where a solution of @xmath10 was given . broadcasting has also been treated in the sinr model under a variety of assumptions . some are stronger than ours ( location information @xcite , power control @xcite ) , while others relax the assumption about the connectivity property , incurring necessarily much higher complexity @xcite . results in our setting , but without carrier sensing , include time complexity of @xmath11 @xcite ( see also @xcite ) , where @xmath12 denotes the maximum ratio between distances of stations connected in the communication graph ; and @xmath13 @xcite . in the spontaneous setting , where the nodes can build an overlay structure along which the message is then propagated , scheideler et al . @xcite used carrier sense to give a dominator algorithms , which can be applied to solve broadcast in @xmath14 rounds . yu et al . @xcite solved the problem in the same time bound using power control , while the algorithm in @xcite that requires neither power control nor carrier sense runs in time @xmath15 . these problems have also been treated in dynamic networks . in the _ unstructured _ model @xcite , where nodes may wake up asynchronously ( modeling the node insertion ) , the local broadcast problem is well studied , even in the sinr setting @xcite , but this model does not consider node deletion . in the _ dual graph _ model @xcite ( originally due to @xcite ) , both the local broadcast @xcite and global broadcast @xcite problems are studied . but this model involves only edge behavior and not node changes ( churn ) . hence , the impact of dynamicity on wireless information dissemination is still largely unexplored . considering wireless interference , there are two classes of wireless network models : graph - based and physical models . basically , the graph - based models define a local and binary type of interference , while the physical models consider fading effect of signal on wireless channels . there are many widely used graph - based models , such as the unit disc graph ( udg ) model @xcite ( unit ball graph ( ubg ) model @xcite ) , the quasi unit disc graph ( qudg ) model @xcite , the protocol model @xcite , the bounded - independence graph ( big ) model @xcite and the @xmath0-hop variants of the above models @xcite . physical models , also known as sinr models @xcite , capture the missing practical issues in wireless interference , such as fading and cumulativeness of wireless interference . most of current works in the sinr model focus on networks embedded in euclidean space , while many of the results hold also for doubling or `` bounded growth '' metrics @xcite . those doubling metrics constrain growth at every ( or arbitrarily small ) granularity , while ours only bounds regions proportional to the transmission range , in particular capturing bounded independence graphs ( big ) . all above models are included by our proposed communication model as shown in sec . [ sec : primitiveproofs ] . for more on wireless models , please refer to @xcite . more detailed related work is introduced in sec . [ drw ] . * our contributions*. we have obtained generalized and improved algorithms for two of the most fundamental dissemination problems , in some cases improving the best results known in static settings . beyond these specific results , we identify the following technical contributions : _ unified model of wireless networks_. the model proposed appears to be the first that allows for the development of pan - model distributed dissemination algorithms . this hopefully prompts further studies crossing the artificial boundary between graph- and fading - based models . _ dynamic networks under interference_. this appears to be the first work to address dynamic networks in the presence of comprehensive interference . _ uniform algorithms_. our algorithms in the static spontaneous setting appear to be the first in fading models that work independent of instance parameters ( number of nodes , max . degree ) . _ primitives for carrier - sense . _ we introduce several primitives or _ capabilities _ that can be implemented using environmental sensing , and propose to study the power of such primitives . _ stabilization mechanisms . _ we identify contention adaptation as a fundamental ability in wireless networks , that appears to be of crucial value to implement other distributed tasks . * roadmap*. the formal model and basic definitions , including the definitions of communication model and carrier sensing primitives , are given in sec . [ sec : model ] . [ sec : balancingcontention ] contains the core technical part of the paper that includes the basic contention balance routine and its analysis . the main results concerning local and global broadcast problems are presented in sections [ sec : localbroadcast ] and [ sec : broadcast ] , respectively . due to space constraints , most proofs are relegated to appendices . we consider a dynamic network of point - size wireless devices ( nodes ) . nodes can transmit messages in time slots / rounds that are sufficiently long to allow a transmission of a single message . no global clock or synchronization of rounds is required , but the clocks of different nodes run at a similar rate , i.e. , the length of a round differs between nodes at most by a factor of 2 . nodes may arrive and leave the network at any time . unless specified otherwise , the nodes are assumed to work _ non - spontaneously _ : they can initially be in sleep state and join the execution of an algorithm only after receiving a message . we say a node is _ alive _ at some point in time if it is present in the network . we assume the total number of nodes in the network is polynomially bounded by a number @xmath16 in each round . we use @xmath17 to denote the set of alive nodes at any fixed point in time and also use @xmath16 to denote the current number of nodes , i.e. @xmath18 . we assume all nodes use the same transmission power @xmath19 for communication in all rounds . _ metrics_. the _ signal strength _ or _ interference _ , depending on context of transmitting node @xmath20 on a node @xmath1 is @xmath21 , where @xmath22 is the _ path loss _ from @xmath20 to @xmath1 . the _ metricity _ of a space @xmath23 is the smallest number @xmath24 such that for every triplet @xmath25 , @xmath26 @xcite . we define @xmath27 if @xmath20 and @xmath1 are different nodes and @xmath28 when @xmath29 . note that @xmath30 is a _ quasi - metric _ , as all metric axioms except symmetry hold . in the rest of the paper , we assume that in each round the metricity of the network is bounded by a fixed constant @xmath24 and will work with values @xmath31 instead of @xmath32 . we assume the quasi - metric @xmath30 has _ bounded independence _ , defined below , roughly stating that there can not be many nodes each causing high interference to a _ fixed _ node , while having low mutual interferences . first , some notations . the _ ball _ with radius @xmath33 centered at @xmath20 is defined as @xmath34 . the _ in - ball _ with radius @xmath33 centered at @xmath20 is defined as @xmath35 ; clearly , @xmath36 . a set @xmath37 is a @xmath33-_packing _ for set @xmath38 if balls of radius @xmath33 centered at nodes in @xmath39 are contained in @xmath38 and are disjoint . @xmath39 is a @xmath40-_cover _ for @xmath38 if the union of balls of radius @xmath40 centered at nodes in @xmath39 contains @xmath38 . note that any maximal @xmath33-packing is a @xmath41-cover , and thus one can bound sizes of covers by packings . we say that @xmath30 has @xmath42-_bounded independence _ , for given @xmath43 and @xmath44 , if for every @xmath45 and every in - ball @xmath46 of radius @xmath47 , the size of a maximum cardinality @xmath48-packing of @xmath46 is at most @xmath49 , where @xmath50 is a constant , possibly depending on @xmath51 . for instance , the euclidean plane is @xmath52-bounded independent , for every @xmath33 . _ neighborhoods , communication graph and dissemination problems_. let @xmath53 denote the maximum transmission distance possible when no other node transmits . as the latter event is arguably very rare , we define the _ communication radius _ @xmath54 as a slightly smaller distance , where @xmath55 is a _ precision _ parameter . we will drop the parameter @xmath55 whenever it is fixed and clear from the context . fix a round @xmath56 . the _ neighborhood _ of a node is @xmath57 , describing who @xmath20 can communicate with directly . the basic operation of interest is when @xmath20 broadcasts a message to its neighbors @xmath58 . the _ communication graph _ is a directed graph @xmath59 , where @xmath60 if and only if @xmath61 . thus , the sequence @xmath62 defines a _ dynamic graph_. the _ vicinity _ of @xmath20 refers to a larger region , @xmath63 , for a parameter @xmath64 . the data dissemination problems that we consider are defined below . we say that a node @xmath20 _ mass - delivers _ in round @xmath56 if it transmits and all its neighbors ( @xmath58 ) receive the message . in the _ local broadcast _ problem , given a node @xmath20 , it is required to minimize the time from the beginning of the algorithm until node @xmath20 mass - delivers at least once , assuming it stays alive during that time . in the _ ( global ) broadcast _ problem , given a distinguished _ source _ node that initially holds a message , the goal is to minimize the time needed to deliver the message to every node in the network through multihop transmissions . _ one hop communication_. when is a transmission successfully received ? suppose a node @xmath20 transmits in round @xmath56 , and let @xmath39 be the set of concurrently transmitting nodes . let @xmath65 and @xmath66 be parameters that depend on the precision @xmath55 . ( , success on a clear channel ) if no other node in @xmath67 transmits _ and _ the total interference at node @xmath20 is at most @xmath68 ( i.e. @xmath69 and @xmath70 ) , then the transmission of node @xmath20 is successfully received by _ all _ its neighbors ( @xmath58 ) . otherwise , the reception is under adversarial control . _ randomized algorithms_. we mainly consider randomized algorithms of the following form : in each round @xmath56 , node @xmath1 makes a transmission with probability @xmath71 , independent of other nodes transmissions in that round . an important notion for the analysis of such algorithms is _ local contention _ , the sum of the transmission probabilities in a close region . the contention in the _ close neighborhood _ of a node @xmath1 in round @xmath56 is @xmath72 , where the radius @xmath73 allows all pair of nodes in @xmath74 to potentially communicate . also , let @xmath75 denote the contention in the larger _ vicinity _ of @xmath1 in round @xmath56 ( @xmath76 will be fixed later ) . we will also use the notation @xmath77 to denote the interference at @xmath1 from nodes outside its vicinity ( in @xmath78 ) in round @xmath56 . the expected value of @xmath77 is @xmath79 . _ sensing primitives_. we assume the nodes have abilities to sense activity on the channel . namely , we assume the nodes are able to detect high and low contention in their vicinity , detect ( under some conditions ) whether their transmission in a given round succeeded and detect a single very near transmission . in the following , we formalize these notions in three primitives : , and . we show in sec . [ sec : primitiveproofs ] how _ all _ these primitives can be implemented with _ basic _ physical carrier sensing and possibly also with other means . ( ) . contention can be probabilistically deduced from measured level of radio activity . we want to relax this ability and will use the following variant , where the outcome of is one of the two values : busy or idle channel . formally , for each node @xmath1 and round @xmath56 : if contention among close neighbors is high ( @xmath80 ) then they _ all _ detect busy channel in round @xmath56 with probability at least @xmath81 , for given @xmath82 , where @xmath83 is a constant , if the contention in the vicinity of @xmath1 is low ( @xmath84 ) _ and _ the interference on @xmath1 from outside its vicinity is above a threshold ( @xmath85 ) then @xmath1 detects idle channel in round @xmath56 with probability at least @xmath86 , for given @xmath87 , where @xmath88 and @xmath89 are constants . if a node @xmath20 has the = ( @xmath55 ) primitive ( depending on the precision parameter @xmath55 ) then : if @xmath20 transmits in round @xmath56 , the interference at @xmath20 is bounded by @xmath90 and the transmission is received by all nodes in @xmath91 , then the outcome of is 1 , where @xmath90 is a parameter . if the transmission is not received by a node @xmath92 then the outcome is @xmath93 . otherwise , the outcome is @xmath93 or @xmath94 , adversarially . ( ) . with = ( @xmath55 ) primitive , a node is able to detect if a transmitter is very close , assuming that it receives the transmitted message . the outcome of is @xmath94 for node @xmath1 in round @xmath56 if @xmath1 _ receives _ a transmission from a node @xmath20 , @xmath95 . otherwise , the outcome of is @xmath93 . this can also be made approximate . _ dynamicity_. we consider a dynamic network where the topology may change adversarially in each round due to node churn ( node arrivals / departures ) and edge changes . we assume that arriving nodes start running the algorithms from an initial configuration , so we do not limit the rate of churn . with edge changes , existing nodes that were not neighbors before , may become neighbors ( e.g. due to mobility ) . the new neighbors may cause too much interference in a too short time , so the edge changes should be limited . we assume the amount of edge changes is bounded for each node @xmath1 , as follows . consider a time interval @xmath96 of length @xmath97 . we require that the number of new neighbors of @xmath1 during @xmath96 ( not counting churn ) is bounded by @xmath98 , where @xmath99 denotes the number of rounds in @xmath96 and @xmath100 is a constant , to be fixed later . we further assume the fraction of rounds in @xmath96 when there are more than @xmath101 new neighbors of @xmath1 is bounded by @xmath102 for every @xmath103 , where @xmath104 . note that there is no restriction on distance changes inside the neighborhood of @xmath1 ( e.g. it is fine for node @xmath1 if its neighbors move , as far as they remain neighbors ) . note that the edge changes may affect the underlying metric , but we require that the upper bounds on metricity and independence are maintained . _ requirements and assumptions_. for the convenience of the reader , we gather together all of our assumptions and requirements in a single place . communication is assumed to succeed in a clear channel ( ) . we assume constant metricity @xmath105 and that @xmath30 has @xmath42-bounded independence with @xmath106 . for the local broadcast problem , we assume that @xmath107 , and for the broadcast problem that @xmath108 . as is standard in fading models , the communication radius @xmath109 is necessarily an @xmath110-fraction of the maximum transmission distance in a clear channel . besides the knowledge required by the primitives that are needed for a particular algorithm , the nodes are assumed to know the precision parameter @xmath55 . a polynomial estimate on the number of nodes , @xmath16 , is needed in dynamic and non - spontaneous algorithms , but not in the static spontaneous problems . knowledge of approximations of model parameters are needed to implement primitives , including @xmath111 , @xmath105 , @xmath53 , @xmath112 , and @xmath68 . knowledge of the maximum degree @xmath3 is not needed . synchronous operation is only assumed in the broadcast algorithm . aspects not defined or constrained are assumed to be under ( adaptive ) adversarial control , including when transmissions that fail are successful , or when nodes appear or disappear from the network . the extent of increases in edge strengths over a period is restricted , as detailed above , while decreases are not and neither are node changes . the dissemination problems are only expected to function with the set of nodes that are sufficiently stable , as detailed in the respective section . in order to keep the contention in the network balanced , we propose a basic procedure called try&adjust , which will be the main building block in our algorithms . the idea is to let each node adapt its transmission probability to the contention detected using the assumed primitive . the parameter @xmath113 describes the passiveness of the newly arriving nodes . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ try&adjust(@xmath114 ) : each node @xmath1 maintains transmission probability @xmath115 in each round @xmath56 , initialized as @xmath116 when @xmath1 enters the network . in round @xmath117 , @xmath1 does : + 1 . transmit with probability @xmath71 , and + 2 . set @xmath118 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the aim for controlling contention is , of course , to ensure that transmissions made have a fair chance of being successful , which means they sufficiently overpower the interference experienced at intended receiver from all other transmissions made in that round . we account for this interference in two ways : the _ contention _ captures the expected interference from the nodes neighbors , while the _ interference _ integrates also the interference from nodes further away . we will measure the contention in the _ vicinity _ of each node @xmath1 , i.e. in @xmath119 , where @xmath76 is a large enough constant . we specify a threshold @xmath120 ( recall @xmath121 from definition ) for measuring contention : if @xmath122 then round @xmath56 is a _ high contention round _ for node @xmath1 and is a _ low contention round _ , otherwise . we further specify a threshold @xmath123 for interference : if @xmath124 then round @xmath56 is a _ high interference round _ for @xmath1 and is _ low interference round _ , otherwise . these thresholds are chosen so as to ensure that in a low - contention / interference round , node @xmath1 will be likely to succeed if it transmits . however , requiring all or most rounds to be low contention for all nodes will lead to high delays instead , it turns out that most rounds will have _ bounded contention _ and low interference , which allows for good progress ; we say node @xmath1 experiences _ bounded contention _ in round @xmath56 if @xmath125 , where @xmath126 is a large enough constant , to be specified later . we say that a round @xmath56 is _ good _ for node @xmath1 if @xmath56 is both bounded contention and low interference round ( in which case , some node in @xmath1 s vicinity has a good chance of successfully transmitting ) . we analyze the properties of try&adjust using the notion of _ a phase _ , the shortest time in which at least @xmath127 rounds occur for all nodes , where @xmath128 is sufficiently large ( given in prop . [ pr : pgoodroundl ] ) . we use @xmath129 to denote a general phase and also the set of rounds in that phase . @xmath130 denotes the number of rounds in a phase @xmath129 ( i.e. @xmath131 ) . the fundamental property of try&adjust ( prop . [ pr : pgoodroundl ] ) is that , for each node and each phase , most of the rounds in the phase are good for that node . this property is then used to show that : if most of the good rounds in a phase have low contention , then node @xmath1 detects idle channel in most of those rounds . otherwise , during a constant fraction of the rounds , a node in the vicinity of @xmath1 mass - delivers . by choosing the parameters carefully , i.e. requiring _ low enough _ contention , we can make sure that during a phase with mostly low contention , the node detects idle channel in _ most of the rounds _ ( more than half ) in a phase , thus leading to an increase of transmission probability by the end of the phase , which , after sufficiently many phases ensures message delivery , w.h.p . on the other hand , during a phase with mostly high contention , there will be many nodes in the vicinity of @xmath1 that successfully transmit , leading to lower contention . these ideas are applied in thms . [ thm : localbroadcast ] and [ thm : broadcast ] . the core idea behind this analysis is based on @xcite . [ pr : pgoodroundl ] let @xmath132 . if constants @xmath133 and @xmath134 are large enough then for each node @xmath1 and phase @xmath129 , with probability @xmath135 , a @xmath136-fraction of the rounds in @xmath129 are good . the proof is rather technical and is deferred to sec . [ sec : contentiondetails ] but the intuition is as follows . the contention in each neighborhood must be bounded most of the time , because when it becomes large , it has a high chance of being decreased due to busy channel . moreover , we show that in expectation , the contention in _ all _ local neighborhoods is bounded , which is then combined with a geometric argument to show that the expected interference at each node is low most of the time . we derive from the fundamental property two useful propositions . the first says that if contention is high , then nodes in the vicinity deliver the message . [ pr : pdecrease ] assume that constants @xmath137 are large enough . for each node @xmath1 and phase @xmath129 , if at least @xmath138-fraction of the rounds of @xmath129 are of high contention , then @xmath139 nodes in @xmath140 mass - deliver , with probability @xmath135 . to this end , we first show that if a round is good for node @xmath20 and a node in its vicinity transmits , then it mass - delivers with constant probability , utilizing both metric assumptions and the properties of good rounds . we then argue that since most rounds are good ( by prop . [ pr : pgoodroundl ] ) and most rounds have by assumption sufficient contention , many rounds will be both good and with sufficient contention , and in each of those , a node in the vicinity of @xmath1 is likely to transmit and succeed . when contention is low in a lot of rounds of a phase , the node will detect idle channel by the primitive in many rounds . this will actually happen during many good rounds , which have the low local contention and low external interference to allow for this detection . [ pr : idle ] assume that @xmath137 are large enough . for each node @xmath1 and phase @xmath129 , if at least @xmath141-fraction of the rounds of @xmath129 are low contention rounds , then with probability @xmath135 , in at least @xmath142-fraction of the rounds of @xmath129 , @xmath1 will detect idle channel and have low contention and low interference . we propose an algorithm for asynchronous local broadcast in dynamic networks . the algorithm is an extension of the try&adjust procedure , where the nodes try to balance the contention in the network and stop transmitting as soon as they deliver their message . we assume the nodes are powered with and primitives . note that the passiveness parameter is set to @xmath143 , which means that the transmission probability of nodes does not get below @xmath144 . localbcast : each node @xmath1 executes try&adjust(1 ) with the following additional step : if @xmath1 transmits and detects , it stops ( i.e. @xmath145 for @xmath146 ) . we will estimate the performance of the algorithm using the notion of _ dynamic degree _ , defined as follows . given a parameter @xmath147 , we denote @xmath148 for node @xmath1 and rounds @xmath149 with @xmath150 , where @xmath151 denotes the in - ball @xmath119 in round @xmath33 . below we prove that if there are not too many node insertions in the neighborhood of a node @xmath1 , then @xmath1 mass - delivers ( delivers to all its neighbors ) in time comparable to its dynamic degree with @xmath76 a constant . the main tools for proving the bound are props . [ pr : pdecrease ] and [ pr : idle ] . first we argue that if there is a phase of mostly low contention , then node @xmath1 will deliver its message , w.h.p . then we show that if the insertions are not too intensive then the contention around @xmath1 will decrease and a phase with mostly low contention will happen . [ thm : localbroadcast ] there is a constant @xmath147 , such that a node @xmath1 performing localbcast asynchronously in a time interval @xmath152 $ ] with @xmath153 mass - delivers , w.h.p . , provided that @xmath154 . note that the assumption @xmath154 is needed only for making the claim w.h.p . : it can be relaxed to higher degree polynomials by only increasing constant factors . let us fix constants @xmath155 so that props . [ pr : pdecrease ] and [ pr : idle ] with @xmath156 . we partition @xmath96 into phases ( for node @xmath1 ) and classify them into two types : ( type a ) phases @xmath129 where at least @xmath138-fraction of rounds are high contention rounds ( i.e. @xmath157 ) , and ( type b ) phases @xmath129 where at least @xmath141-fraction of rounds are low contention rounds ( i.e. @xmath158 ) . node @xmath1 mass - delivers in a type b phase , w.h.p . let @xmath159 be the low contention and interference rounds during phase @xmath129 where @xmath1 detects idle channel . by prop . [ pr : idle ] @xmath160 > 1-o(n^{-3})$ ] . assume for now that the latter happens . for each @xmath161 we have @xmath162 . let us call this operation doubling . the value of @xmath71 at the beginning of the phase is at least @xmath144 , so @xmath163 doubling operations are sufficient to raise it to @xmath164 . the probability can be further halved during the phase at most @xmath165 times . thus , we may assume we have at most @xmath166 halving and at least @xmath167 doubling operations applied to an initial value @xmath164 . if @xmath168 , then the total number of halving operations is less than @xmath169 . it follows that @xmath1 has @xmath170 in at least @xmath171 low contention / interference rounds . by lemma [ le : broadgood ] , in each such round , @xmath1 mass - delivers with probability at least @xmath172 ; hence , if @xmath128 is large enough , @xmath1 mass - delivers in @xmath129 , w.h.p . it remains to argue that there will be a type b phase during time interval @xmath96 . consider a type a phase @xmath129 . [ pr : pdecrease ] implies that with probability @xmath135 , @xmath173 nodes in @xmath119 deliver their message and stop during phase @xmath129 . thus , with probability @xmath174 , there are at most @xmath175 type a phases with @xmath176 a constant , as there are at most @xmath177 nodes in @xmath140 during the time interval @xmath96 ( also recall that @xmath178 ) . we conclude that if @xmath96 consists of at least @xmath179 phases , it will contain a type b phase and @xmath1 will deliver its message w.h.p . * implications for static networks*. in static networks , the parameter @xmath177 is at most @xmath180 if @xmath76 is constant , where @xmath181 is the maximum size of a neighborhood in the network . thus , we obtain the following optimal result ( up to constant factors ) for static networks , as @xmath3 and @xmath163 are lower bounds even when running in the spontaneous mode @xcite . when running localbcast in a static asynchronous network , each node @xmath1 completes local broadcast in @xmath182 rounds , w.h.p . * remark . * in the special case when the nodes can start executing the algorithm simultaneously , i.e. in the _ spontaneous mode _ , the nodes need not know an upper bound on the size of the network . indeed , each node may start running try&adjust with initial probability set to an arbitrary value and with no lower limit . the first phase will be spent for stabilization and can be ignored , while the argument for the rest of the phases is nearly identical to the one in thm . [ thm : localbroadcast ] . for the broadcast problem , we assume nodes communicate in synchronized rounds of equal length . each round consists of two slots . the idea is to use the first slot of each round for disseminating the message with try&adjust and the second slot for notifying nodes which have no uninformed neighbors . the latter is accomplished by using higher precision primitives , namely ( @xmath183 ) and ( @xmath183 ) , when executing try&adjust . this helps to detect a transmission that is successfully received by all nodes in @xmath184 of a node @xmath1 . upon detecting such a transmission , node @xmath1 resends the message in the second slot , in order to inform nodes @xmath20 with @xmath185 that their neighborhood ( @xmath186 ) has been covered . a node @xmath20 can detect that @xmath185 using . the algorithm is presented below . we assume the passiveness parameter @xmath114 of try&adjust is large enough , to be defined later . note that the algorithm works for the non - spontaneous mode , as nodes act only after receiving the message . bcast(@xmath114 ) : initially , only a source node @xmath187 has the message . a node @xmath1 , upon receiving a message , starts executing try&adjust(@xmath114 ) in the first slot of rounds . in addition , in each round @xmath56 , + 1 . if @xmath1 detects in the first slot , it retransmits in the second slot and restarts try&adjust(@xmath114 ) , + 2 . if @xmath1 receives a message in the first slot and detects in the second slot , it restarts try&adjust(@xmath114 ) . in order to evaluate the progress of the algorithm , we use a notion of a _ dynamic distance _ , as defined below . let @xmath188 be a parameter . a sequence @xmath189 is called a _ stable @xmath187-@xmath1 path _ if there is a sequence @xmath190 of time intervals with @xmath191 $ ] , such that @xmath192 , @xmath193 and nodes @xmath194 and @xmath195 are both alive and @xmath196 during @xmath197 . the _ time - length _ of a stable @xmath187-@xmath1 path is @xmath198 . the _ stable @xmath187-@xmath1 distance _ @xmath199 is defined as the minimum time - length of a stable @xmath187-@xmath1 path . note that a stable path need not be connected at any fixed point in time . moreover , most of the nodes might be missing at any given point in time . the core idea behind the analysis of the following theorem is similar to the case of local broadcast : we show that as soon as a neighbor @xmath20 of a node @xmath1 has the message and @xmath20 and @xmath1 keep being neighbors for @xmath200 rounds , @xmath1 will receive a transmission of @xmath20 during those rounds . [ thm : broadcast ] assume the edge change rate @xmath100 is sufficiently small . there are constants @xmath201 , such that when running bcast(@xmath114 ) in the synchronous mode , each node @xmath1 receives the message in @xmath202 rounds w.h.p . let us fix constants @xmath155 so that props . [ pr : pdecrease ] and [ pr : idle ] hold with interference threshold @xmath203 , where @xmath204 and @xmath205 are the parameters corresponding to precision @xmath183 . we set the passiveness parameter of try&adjust to @xmath206 and the stable distance parameter @xmath207 . consider a node @xmath20 that receives the message in round @xmath56 . let @xmath1 be such that @xmath208 in the time interval @xmath209 $ ] ( and both stay alive during @xmath96 ) . the theorem follows by an induction and union bound from the claim below . node @xmath1 gets the message of @xmath20 during @xmath96 , with probability @xmath210 . we prove the claim by contradiction , and assume that @xmath1 can not get the message during @xmath96 . let us split @xmath96 into phases ( for @xmath20 ) and , similar to the proof of thm . [ thm : localbroadcast ] , classify the phases @xmath129 into types : ( type a ) at least @xmath138-th of rounds in @xmath129 are high contention rounds for @xmath20 @xmath211 , and ( type b ) at least @xmath141-th of rounds in @xmath129 are low contention rounds for @xmath20 @xmath212 . we show that with probability @xmath135 , all phases in @xmath96 are of type b. consider a type a phase @xmath129 . we know from prop . [ pr : pdecrease ] that , with probability @xmath135 , there is a set @xmath39 of @xmath139 nodes in @xmath140 that deliver their messages and restart try&adjust during @xmath129 . we split @xmath39 into three subsets , @xmath213 those inserted by churn during @xmath129 , @xmath214 those inserted by edge change during @xmath129 and @xmath215 the rest . consider @xmath214 first . we know that in each local neighborhood , there are at most @xmath216 nodes arriving due to edge changes . note that @xmath217 can be covered with @xmath218 local neighborhoods ; hence , the total number of nodes arriving in @xmath217 due to edge changes is @xmath219 . since @xmath220 , setting @xmath221 a small enough constant gives @xmath222 . now consider @xmath215 . by the primitive , for each node @xmath223 , all nodes in @xmath224 also restart try&adjust , setting their transmission probability to @xmath225 . by the setting of @xmath226 , the probability that a node restarting try&adjust in phase @xmath129 transmits again in @xmath129 is at most @xmath227 . thus , with probability at least @xmath135 , the nodes in @xmath214 constitute an @xmath228-packing of @xmath140 and are at most @xmath229 . as for @xmath213 , the probability that a newly arriving node transmits during phase @xmath129 is at most @xmath227 , so @xmath230 w.h.p . these bounds lead to a contradiction @xmath231 , if constant @xmath128 is large enough . thus , each phase @xmath129 of @xmath96 is of type b with probability @xmath135 . it remains to recall that @xmath96 contains constant number of phases . now assume that @xmath129 is of type b. in a similar way as in theorem [ thm : localbroadcast ] , we can apply prop . [ pr : idle ] and show that if @xmath232 at the beginning of @xmath129 , then by the end of @xmath129 , the effect on @xmath71 is equivalent to applying at least @xmath233 of @xmath234 operations on @xmath235 . since @xmath236 at the beginning of @xmath96 is at least @xmath225 , it will take at most 11 phases to increase the probability to the value @xmath164 . then , a single phase will suffice for @xmath20 to deliver its message ( including to node @xmath1 ) , with probability at least @xmath210 . * implications for static networks*. when the network is static , theorem [ thm : broadcast ] can be reformulated in terms of hop - distance @xmath237 in the communication graph , which is defined as the length of the shortest directed @xmath187-@xmath1 path in @xmath238 : we have that @xmath239 for any node @xmath1 . note also that in this setting nodes that succeeded transmitting or detected need not continue the algorithm , so they stop transmitting . in this case , setting the passiveness parameter to @xmath143 suffices . we call this variant of the algorithm bcast@xmath240 . when running bcast@xmath240 in synchronous non - spontaneous mode in a static network with source @xmath187 , each node @xmath1 receives the message in @xmath241 rounds w.h.p . when the communication graph is strongly connected , the broadcast from any source node is completed in @xmath242 rounds , where @xmath243 is the diameter of the communication graph . in the spontaneous mode , the bound above can be further improved to @xmath244 ; see appendix [ sec : spontaneous ] . this is based on finding a constant - density dominating set in @xmath200 time @xcite and simultaneously propagating along the dominators in @xmath244 time . we can therefore extend the approach based on @xcite to uniform algorithms in bounded - independence metrics . these results are close to best possible . we show below that in order to obtain bounds of that magnitude , it is necessary to have the primitive . to this end , we extend the lower bound construction of ( * ? ? ? * thm . 7 ) for `` compact sinr '' to our setting . this construction leverages the property of our model that there can be arbitrarily many nodes that are mutually close to each other . namely , the bounded - independence metric is strictly more relaxed than the standard euclidean metrics . indeed , there is a @xmath245-round broadcast algorithm for the sinr model that does not need or other carrier sensing primitives @xcite . what the lower bound then illustrates is that to obtain such results , one must depend on opportune traits of the sinr model that we have tried to avoid and are not necessary for problems like local broadcast . thus we can observe concrete tradeoffs depending on model assumptions . [ thm : brlowerbound ] for every ( possibly randomized ) broadcast algorithm @xmath246 that uses neither node coordinates nor primitive , there is a @xmath247-bounded - independence metric space where @xmath246 needs @xmath248 rounds to do broadcast in a @xmath249-broadcastable network , even if the nodes have and primitives and operate spontaneously . 10 n. alon , a. bar - noy , n. linial , and d. peleg . . in _ stoc _ , pages 274285 , 1989 . n. alon , a. bar - noy , n. linial , and d. peleg . a lower bound for radio broadcast . , 43(2):290298 , 1991 . r. bar - yehuda , o. goldreich , and a. itai . on the time - complexity of broadcast in radio networks : an exponential gap between determinism and randomization . in _ podc87 _ , 1987 . l. barenboim and d. peleg . nearly optimal local broadcasting in the sinr model with feedback . in _ sirocco _ , pages 164178 , 2015 . m. h. bodlaender and m. m. halldrsson . beyond geometry : towards fully realistic wireless models . in _ podc _ , pages 347356 , 2014 . m. h. bodlaender , m. m. halldrsson , and p. mitra . . in _ podc _ , pages 355364 . acm , 2013 . i. chlamtac and s. kutten . on broadcasting in radio networks problem analysis and protocol design . , 33(12):12401246 , 1985 . b. s. chlebus , d. r. kowalski , and s. vaya . distributed communication in bare - bones wireless networks . , 2015 . b. n. clark , c. j. colbourn , and d. s. johnson . unit disk graphs . , 86(1):165177 , 1990 . a. e. clementi , a. monti , and r. silvestri . round robin is optimal for fault - tolerant broadcasting on wireless networks . , 64(1):8996 , 2004 . a. czumaj and w. rytter . broadcasting algorithms in radio networks with unknown topology . in _ focs03 _ , 2003 . s. daum , s. gilbert , f. kuhn , and c. newport . . in _ disc _ , pages 358372 , 2013 . b. derbel and e. talbi . . in _ icdcn10 _ , pages 155166 , 2010 . s. dolev , s. gilbert , m. khabbazian , and c. newport . . in _ disc _ , pages 252267 , 2011 . a. fanghnel , t. kesselheim , and b. vcking . . , 412(24):26572667 , 2011 . m. ghaffari , s. gilbert , c. newport , and h. tan . optimal broadcast in shared spectrum radio networks . in _ opodis12 _ , 2012 . m. ghaffari , b. haeupler , and m. khabbazian . randomized broadcast in radio networks with collision detection . in _ podc13 _ , 2013 . m. ghaffari , b. haeupler , n. lynch , and c. newport . bounds on contention management in radio networks . in _ disc _ , 2012 . m. ghaffari , e. kantor , n. lynch , and c. newport . multi - message broadcast with abstract mac layers and unreliable links . in _ podc14 _ , 2014 . m. ghaffari , n. lynch , and c. newport . the cost of radio network broadcast for different models of unreliable links . in podc13 _ , 2013 . o. goussevskaia , t. moscibroda , and r. wattenhofer . . in _ dialm - pomc 08 _ , pages 3544 , 2008 . p. gupta and p. kumar . , 46(2):388404 , 2000 . m. m. halldrsson . . , 9(1):7:17:20 , dec . m. m. halldrsson , s. holzer , and n. a. lynch . a local broadcast layer for the sinr network model . in _ podc _ , pages 129138 , 2015 . m. m. halldrsson , s. holzer , p. mitra , and r. wattenhofer . . in _ soda13 _ , pages 15951606 , 2013 . m. m. halldrsson and p. mitra . . in _ icalp _ , pages 625636 , 2011 . m. m. halldrsson and p. mitra . . in _ podc _ , 2012 . m. m. halldrsson and p. mitra . . fomc _ , 2012 . m. m. halldrsson , y. wang , and d. yu . leveraging multiple channels in ad hoc networks . in _ podc15 _ , 2015 . m. hoefer , t. kesselheim , and b. vcking . approximation algorithms for secondary spectrum auctions . in _ proceedings of the twenty - third annual acm symposium on parallelism in algorithms and architectures _ , spaa 11 , pages 177186 , new york , ny , usa , 2011 . r. impagliazzo and v. kabanets . constructive proofs of concentration bounds . in _ approx - random _ , pages 617631 , 2010 . t. jurdzinski , d. kowalski , m. rozanski , and g. stachowiak . distributed randomized broadcasting in wireless networks under the sinr model . in _ disc _ , pages 373387 , 2013 . t. jurdzinski , d. kowalski , m. rozanski , and g. stachowiak . on the impact of geometry on ad hoc communication in wireless networks . in _ podc 14 _ , 2014 . t. jurdzinski , d. r. kowalski , and g. stachowiak . . in _ fct13 _ , pages 195209 , 2013 . t. jurdzinski , d. r. kowalski , and g. stachowiak . distributed deterministic broadcasting in wireless networks of weak devices . in _ icalp13 _ , 2013 . t. kesselheim and b. vcking . . in _ disc _ , pages 163178 , 2010 . d. kowalski and a. pelc . broadcasting in undirected ad hoc radio networks . in _ podc03 _ , 2003 . f. kuhn , n. lynch , and c. newport . brief announcement : hardness of broadcasting in wireless networks with unreliable communication . in _ disc _ , 2009 . f. kuhn , n. lynch , c. newport , r. oshman , and a. richa . . in _ podc _ , pages 336345 . acm , 2010 . f. kuhn , t. moscibroda , and r. wattenhofer . . in _ mobicom _ , pages 260274 . acm , 2004 . f. kuhn and r. oshman . dynamic networks : models and algorithms . , 42(1):8296 , 2011 . f. kuhn , r. wattenhofer , and a. zollinger . ad hoc networks beyond unit disk graphs . , 14:715729 , 2008 . e. kushilevitz and y. mansour . . , 27(3):702712 , 1998 . c. lenzen and r. wattenhofer . distributed algorithms for sensor networks . , 370(1958 ) , 2012 . t. moscibroda and r. wattenhofer . . , 21(4):271284 , 2008 . c. newport radio network lower bounds made easy . in _ disc14 _ , 2014 . c. scheideler , a. richa , and p. santi . . in _ mobihoc 08 _ , pages 91100 , 2008 . s. schmid and r. wattenhofer . algorithmic models for sensor networks . in _ wpdrts06 _ , 2006 . j. schneider and r. wattenhofer . . in _ podc _ , pages 210219 , 2009 . j. schneider and r. wattenhofer . in _ disc _ , pages 133147 . springer - verlag , 2010 . d. yu , q .- s . hua , y. wang , and f. lau . . dcoss 12 _ , pages 132139 , 2012 . d. yu , q .- s . hua , y. wang , h. tan , and f. lau . . in _ sirocco12 _ , pages 111122 , 2012 . d. yu , q .- s . hua , y. wang , j. yu , and f. lau . . in _ infocom13 _ , pages 24272435 , 2013 . d. yu , y. wang , q .- s . hua , and f. c. m. lau . . in _ distributed computing in sensor systems and workshops ( dcoss ) , 2011 international conference on _ , pages 18 , 2011 . d. yu , y. wang , y. yan , j. yu , and f. lau . speedup of information exchange using multiple channels in wireless ad hoc networks . in _ infocom15 _ , 2015 . _ wireless models : _ considering wireless interference , there are two classes of wireless network models : graph - based and physical models . basically , the graph - based models define a local and binary type of interference , while the physical models consider fading effect of signal on wireless channels . the most classical graph - based model is the radio network model @xcite . in this model , the network is modeled using a communication graph , where each pair of nodes that can communicate with each other is connected by an edge . it defines the interference just from direct neighbors , and a transmission can succeed if and only if there is only one neighbor of the receiver transmitting . there are many widely used variants of the classical radio network models , including : 1 ) the @xmath0-hop model where the interference comes from @xmath0-hop neighbors @xcite ; 2 ) unit disc graph ( udg ) model @xcite which defines the neighborhood using a unit disc ; 3 ) quasi unit disc graph ( qudg ) model @xcite which just defines all pairs of nodes with distance at most @xmath76 for some given @xmath250 $ ] are adjacent , and leave the ` grey ' area in @xmath251 $ ] being determined by an adversary ; 4 ) protocol model @xcite , where each node has a transmission range and an interference range , and a successful transmission occurs if a node falls into the transmission range of a transmitter and outside the interference ranges of all other transmitter ; 5 ) bounded - independence graph ( big ) model @xcite , which defines abstractly and requires that the size of the maximal independent set in the @xmath33-hop neighborhood of each node is bounded by a polynomial function with @xmath33 . though the graph - based interference models miss certain crucial aspects of actual wireless networks , the simple definition of these models can help derive novel insights into distributed solutions to wireless problems . physical models , also known as sinr models @xcite , capture the fading and cumulative features of receptions in actual wireless environments . the default assumption is that interference fades with a polynomial of the distance , and transmission succeeds only if the received signal strength is sufficiently larger than the total interference plus noise . recently , the sinr model has attracted great attentions in the distributed community @xcite . most of these works focus on networks embedded in euclidean space , while many of the results hold also for doubling or `` bounded growth '' metrics @xcite . those doubling metrics constrain growth at every ( or arbitrarily small ) granularity , while ours only bounds regions proportional to the transmission range , in particular capturing bounded independence graphs ( big ) . for more on wireless models , please refer to @xcite . _ local broadcast _ : in the radio network model , probably the first local broadcast result was a randomized algorithm of alon et al . @xcite in a synchronous model , running in @xmath6 rounds . derbel and talbi @xcite later generalized their algorithm to work without knowledge of @xmath3 and their proposed algorithm can accomplish local broadcast in @xmath7 rounds . the decay strategy also yields an @xmath6 time algorithm for local broadcast without knowledge of @xmath3 @xcite . goussevskaia et al . @xcite gave the first results for local broadcast in the sinr model , running in time @xmath6 and @xmath252 with and without knowledge of @xmath3 , respectively . the latter was improved in @xcite and further improved , independently , to @xmath7 time @xcite . with free acknowledgements , this was improved to @xmath253 @xcite . when additionally @xmath3 is known , this was further improved recently to @xmath8 in the spontaneous setting @xcite . the speedup of multiple channels on local broadcast was considered in @xcite _ broadcast _ : the complexity of broadcasting is well understood in graph - based models . in the radio network model , bar - yehuda et al . @xcite presented the decay protocol which can accomplish non - spontaneous broadcast in @xmath254 rounds , where @xmath46 is the diameter . this result was improved to @xmath9 independently by czumaj and rytter @xcite , and kowalski and pelc @xcite . these algorithms can be viewed as clever optimizations of the decay protocol and match the lower bound @xcite . with collision detection , this lower bound was recently broken in @xcite , where a solution of @xmath10 was given . broadcast in multi - channel radio networks was considered in @xcite . for the udg model , an @xmath255 time algorithm was given in @xcite in the spontaneous setting . broadcasting has been treated in the sinr model under a variety of assumptions . some are stronger than ours ( location information @xcite , power control @xcite ) . some relax the assumption about the connectivity property , but incur necessarily super - linear complexity @xcite . results in our setting , but without carrier sensing , include time complexity of @xmath256 @xcite ( see also @xcite ) , where @xmath12 denotes the maximum ratio between distances of stations connected in the communication graph ; and @xmath13 @xcite . in the spontaneous setting , where the nodes can build an overlay structure along which the message is then propagated , scheideler et al . @xcite solved the problem in @xmath257 rounds using physical carrier sensing . yu et al . @xcite solved the problem in @xmath258 rounds using power control , allowing stations to decide the strength of a transmitted signal in each step . and the algorithm in @xcite then runs in time @xmath254 . _ distributed models of temporal variability _ : dynamic networks have been studied extensively in recent years ( see @xcite for a survey ) , but generally not in the presence of interference . the _ dual graph _ model @xcite ( originally due to @xcite ) was designed to capture inherent _ unreliability _ in wireless networks , much of which can be due to dynamicity . the main focus of that work is on extending the radio network model in general graphs . importantly , the dual graph model does not distinguish between interference and communication edges ; only that the unreliable edges can transmit both interference and the usual communication , but their availability is under adversarial control . thus , there is no way to capture interference from further away nodes . most problems become extremely difficult against a powerful adversary , and to get good result , one must assume a much weaker one @xcite . this model only involves edge behavior and not node changes ( churn ) . both the local broadcast @xcite and global broadcast @xcite problems are studied in the dual graph model . a dynamic model that considers node insertion is the _ unstructured _ model @xcite , which admits arbitrary wake - up mode and asynchronous communication . this model was first proposed in the unit - disc setting , and then extended to bounded independence graphs ( big ) @xcite and sinr @xcite . it has been widely used in the solution of a variety of distributed wireless problems @xcite , including local broadcast @xcite , but there are no known global broadcast results . the model neither consider node deletion nor edge changes . hence , the impact of dynamicity on global communication is largely unexplored . [ [ modeling - communication ] ] modeling communication + + + + + + + + + + + + + + + + + + + + + + our communication model captures most known algorithmic wireless models , as it is demonstrated below on the example of sinr , disk - graph based and protocol models . note that we assume below the distance @xmath259 to be symmetric , but the results hold also for `` almost symmetric '' functions , i.e. when there is a constant @xmath260 such that @xmath261 for all @xmath262 . _ sinr model . _ consider a network in a metric space . in the sinr model of communication , if @xmath39 is the set of simultaneously transmitting nodes in the network , a node @xmath1 receives the transmission of node @xmath20 if and only if @xmath263 where constants @xmath113 and @xmath264 denote the minimum sinr threshold and the ambient noise , respectively . note that @xmath265 in this setting . we can implement with parameters @xmath266 and @xmath267 , as shown in the proposition below . [ pr : scc ] if the interference at a node @xmath1 is less than @xmath68 in round @xmath56 , then @xmath1 will deliver its message if it transmits . note that if the interference at node @xmath1 is not more than @xmath68 then there is no node @xmath268 transmitting in round @xmath56 , as otherwise the interference at @xmath1 would be at least @xmath269 ; hence , only nodes in @xmath270 can transmit . consider an arbitrary node @xmath271 . for each node @xmath272 , @xmath273 . then , the interference at @xmath274 is at most : @xmath275 which implies that node @xmath274 receives @xmath1 s transmission : @xmath276 . _ the udg and ubg models . _ these models are described by geometric graphs : a node @xmath20 receives a message from another node @xmath1 if and only if @xmath1 is the only transmitting neighbor of @xmath20 . in the unit disk graph ( udg ) and unit ball graph ( ubg ) models the nodes are located in a metric space and two nodes are connected by an edge if and only if their distance is at most @xmath277 . the functionality of can be modeled as follows : the transmission of a node @xmath1 is received by all its neighbors if there is no other node at distance less than @xmath278 from @xmath1 transmitting simultaneously , i.e. we can set the parameters to @xmath279 and @xmath280 . _ the quasi - udg model . _ the quasi - udg model is an extension of the udg model : a ) if @xmath281 then @xmath20 and @xmath1 are connected by an edge , b ) if @xmath282 then they are disconnected , c ) otherwise , @xmath20 and @xmath1 may be connected or not . in this case may be implemented by setting @xmath279 and @xmath283 , with the adversary constrained to follow the specific static situation captured by the qudg . _ the protocol model . _ in the protocol model , the nodes are in a metric space and there are two radii : @xmath277 the communication radius , and @xmath284 the interference radius . a node @xmath1 receives the transmission of a node @xmath20 if and only if : 1 ) @xmath1 is in the communication range of node @xmath20 , i.e. @xmath281 , and 2 ) there is no transmitting node @xmath274 such that @xmath1 is in the interference range of @xmath274 : for each transmitting node @xmath285 , @xmath286 . may be implemented here by setting @xmath279 and @xmath283 . _ the big model . _ in the bounded independence graph ( big ) model , for a parameter @xmath287 , we are given a graph on the nodes with the property that for every node @xmath1 and every @xmath288 , the maximum independent set in the @xmath0-neighborhood of @xmath1 is @xmath289 . the shortest - path distance metric on the graph is now naturally a @xmath290-bounded independence metric . to fit in our model , the growth parameter @xmath287 must be less than @xmath105 . _ @xmath0-hop variants . _ these graph models can be naturally generalized to a model on interference , where nodes of distance at most @xmath0 cause interference , for some @xmath291 . we capture this by extending @xmath112 as needed . [ [ implementing - primitives - with - physical - carrier - sensing ] ] implementing primitives with physical carrier sensing + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + one way of implementing the primitives mentioned in this paper is to use physical carrier sensing , i.e. we assume the nodes have technology to detect if the interference ( plus noise ) is higher than a given threshold . we show below how to implement , and primitives using carrier sensing . * primitive . * the primitive can be implemented using a carrier sensing threshold @xmath292 and setting the parameter @xmath293 ; busy channel is detected if and only if the interference is at least @xmath96 . we will need the following technical fact . [ ieq ] for every @xmath294 $ ] , @xmath295 , it holds that @xmath296 if @xmath297 in round @xmath56 then all nodes in @xmath298 detect busy channel with probability at least @xmath299 . in particular , we can take @xmath300 if @xmath301 . by the setting of @xmath96 and the definition of @xmath74 , if two nodes in @xmath74 transmit in round @xmath56 , then all nodes in @xmath74 will detect busy channel . hence , the probability of all nodes in @xmath74 detecting busy channel is at least the probability of more than one node transmitting in round @xmath56 . the probability of no node transmitting is @xmath302 by lemma [ ieq ] . the probability of exactly one node transmitting is : @xmath303 where we used the assumption that @xmath304 and lemma [ ieq ] . thus , the probability of detecting busy channel is at least @xmath305 . for every @xmath147 , if @xmath306 and @xmath307 then @xmath1 detects idle channel with probability at least @xmath308 . by the setting of the threshold @xmath96 , if there is no node transmitting in @xmath119 then node @xmath1 will detect idle channel . thus , the probability that @xmath1 detects idle channel is at least @xmath309 using lemma [ ieq ] . * primitive . * in order to detect successful transmission , we can use interference threshold @xmath310 and set @xmath311 , where @xmath312 and @xmath313 are the parameters of . if a node @xmath1 senses that the interference is no higher than @xmath96 , it knows that : 1 . there is no node in @xmath314 transmitting , as otherwise the interference would be at least @xmath315 , 2 . the interference is at most @xmath68 ; thus , it knows that its transmission has been received by all neighbors in @xmath316 by . * primitive . * if a node @xmath1 receives a message from a node @xmath20 then @xmath1 can separate the signal from the interference and measure the received signal strength . as the nodes use uniform power assignment , node @xmath1 knows that @xmath95 if the received signal is stronger than @xmath317 . [ [ implementing - primitives - by - other - means ] ] implementing primitives by other means + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the primitives can frequently be implemented in other ways , often with the logarithmic blowup that explains the differences with the best carrier - sense - free results . * * in an asynchronous system , it may be impossible to implement by other means than carrier sense . in a synchronized system , however , we can be achieved with logarithmic or polylogarithmic factor overhead . consider a given round . for each probability @xmath318 , repeat @xmath319 times : the senders in the original round transmit with probability @xmath320 . using concentration bound with @xmath50 sufficiently large , one can infer the contention within an small approximation , with high probability . such a strategy has been applied , e.g. , in @xcite . * primitive . * a simple strategy is to work with only probabilistic guarantees of a transmission being received by all neighbors . then , simply repeat the protocol until this has been achieved @xmath319 times , which gives an guarantee , w.h.p . this approach underlies , e.g. , the local broadcast algorithms without carrier sense @xcite . * primitive . * this primitive , which is essential for dominator - based strategies for broadcast , can be implemented using power control : by lowering the power on all units appropriately , one can ensure that nodes further away ( by a small constant factor ) will not be able to hear the message due to the ambient noise term , see e.g. @xcite . alternatively , one can assume that distances can be determined in other ways , such as by gps @xcite . recall that we need to prove the following . we assume that at the beginning of the _ first _ phase under consideration , the contention in the whole network is bounded by a constant . this holds for all algorithms in this paper , as the initial probability of nodes is always at most @xmath321 . * * proposition [ pr : pgoodroundl ] . * * let @xmath132 . if constants @xmath133 and @xmath134 are large enough then for each node @xmath1 and phase @xmath129 , with probability @xmath135 , a @xmath136-fraction of the rounds in @xmath129 are good . recall that in a good round , there should be both bounded contention and low interference . the proof is split into two parts , each handling one of these properties . [ pr : pgoodroundl ] follows by simply combining those two parts . we will need the following concentration bounds . @xcite[le : chernoff ] consider a collection of binary random variables @xmath322 , and let @xmath323 . if there are probabilities @xmath324 with @xmath325\leq \prod_{i\in s}p_i$ ] for every set @xmath326 , then it holds for @xmath327 and @xmath328 that @xmath329\leq\left(\frac{e^{\delta}}{(1+\delta)^{1+\delta}}\right)^\mu\leq e^{-\frac{\delta^2\mu}{2(1+\delta/3)}}.\ ] ] if , on the other side , there are probabilities @xmath324 with @xmath325\geq \prod_{i\in s}p_i$ ] for every set @xmath326 , then it holds for @xmath327 and @xmath330 that @xmath331\leq\left(\frac{e^{-\delta}}{(1-\delta)^{1-\delta}}\right)^\mu\leq e^{-\delta^2\mu/2}.\ ] ] first , we show that for each fixed node @xmath20 , the contention in the _ local _ neighborhood @xmath332 is bounded in most of the rounds of a phase . to this end , we show that in each round , the contention is either already low or will be halved with significant probability , then apply a concentration bound to show the claim . recall the constant @xmath333 from the definition of . [ le : phi ] let @xmath129 be a time interval and assume that @xmath334 at the beginning of @xmath129 . then for every @xmath335 , @xmath336 happens at most @xmath337 times during @xmath129 , with probability @xmath338 , where @xmath339 and @xmath340 is the edge change parameter . by the assumptions on edge changes , we have that during phase @xmath129 , the fraction of rounds where more than @xmath341 nodes become a neighbor of @xmath20 because of edge changes is @xmath102 . let @xmath159 denote the remaining set of rounds . we have @xmath342 . in each of those rounds , the contribution of edge changes in @xmath343 is clearly at most @xmath344 . next we bound the number of rounds in @xmath159 , where @xmath345 . consider such a round @xmath161 . let @xmath346 . by the definition of , all nodes in @xmath332 detect busy channel in round @xmath56 and halve their transmission probabilities with probability at least @xmath347 . thus , @xmath348\ge 1-\mu'$ ] , where the additive @xmath164 accounts for the sum of probabilities of the nodes that just join the ball @xmath332 due to node churn ( recall that each of them has an initial probability at most @xmath349 ) and @xmath344 is an upper bound on the contention due to edge changes ( by the definition of @xmath159 ) . for each round @xmath161 , let us define a binary random variable @xmath350 as follows : @xmath351 if @xmath352 or @xmath353 and @xmath354 ( recall that @xmath355 ) and @xmath356 otherwise . by the discussion above , we have : @xmath357&=pr[p_{t}(u)\leq\phi]+pr[p_{t}(u)>\phi\wedge p_{t+1}(u)\leq\frac{5}{6}p_t(u)]\\ & = pr[p_{t}(u)\leq\phi]+pr[p_{t}(u)>\phi]\cdot pr[p_{t+1}(u)\leq\frac{5}{6}p_t(u)|p_{t}(u)>\phi]\\ & \geq pr[p_{t}(u)\leq\phi]+pr[p_{t}(u)>\phi]\cdot ( 1-\mu')\\ & \geq 1-\mu ' . \end{aligned}\ ] ] this implies that for each round @xmath358 and each subset @xmath359 of earlier rounds ( @xmath360 for @xmath361 ) , @xmath362\geq 1-\mu'$ ] and @xmath363\leq \mu'$ ] . thus , for each subset @xmath364 we get @xmath365\le \mu'^{|s| } , $ ] and if we denote @xmath366 , then @xmath367\le \mu ' |h'|=\mu |h|$ ] , where we denote @xmath368 . thus , we can apply chernoff bound ( lemma [ le : chernoff ] ) with @xmath369 to bound @xmath370 with high probability : @xmath371\le 2^{-\mu|h|}.\ ] ] using this bound , we obtain a bound on the number of rounds @xmath56 with @xmath372 . let @xmath373 denote this number . consider a maximal interval @xmath374\subseteq h'$ ] such that @xmath345 for all @xmath375 . for each round @xmath375 , @xmath376 if @xmath351 and @xmath377 otherwise . by maximality of @xmath378 , if @xmath379 _ is not _ the first round of phase @xmath129 , then we have @xmath380 . then a simple calculation shows that @xmath356 for at least @xmath381 part of the rounds of @xmath378 , i.e. @xmath382 . on the other hand , if there is a unique maximal interval @xmath378 starting at the first round of phase @xmath129 and such that @xmath383 , then it must hold that @xmath384 , where @xmath385 . combining these observations , we get that @xmath386 . by combining this bound with ( [ eq : xupperl ] ) , we have : @xmath387 \le pr[x\ge \delta\mu|h| ] < 2^{-\mu|h|}.\ ] ] thus , with probability @xmath388 , in at most @xmath389 fraction of rounds in @xmath129 there can be @xmath345 . recall that for node @xmath1 we need to show that @xmath390 for most rounds in a phase . we prove this by taking a _ constant size _ @xmath73-cover of @xmath119 and applying lemma [ le : phi ] to all nodes in the cover _ [ pr : lowc ] for every @xmath132 and @xmath147 , there are constants @xmath391 and @xmath392 , such that for every node @xmath1 and phase @xmath129 , with probability @xmath135 , at least @xmath136-fraction of rounds in @xmath129 are bounded contention rounds , i.e. @xmath393 . it will be sufficient to choose @xmath128 and @xmath394 such that @xmath395 and @xmath396 . by the definition of @xmath287 and bounded independence , there is a @xmath73-cover @xmath39 of @xmath119 of size @xmath397 . set @xmath398 . let us fix a node @xmath399 . if @xmath129 is the first phase or @xmath400 then we set @xmath401 . otherwise , consider the phase @xmath159 of @xmath127 rounds preceding @xmath129 . by lemma [ le : phi ] , if we set @xmath402 then , with probability @xmath403 , there is a round in @xmath159 where @xmath404 . let @xmath405 be such a round and set @xmath406 $ ] , where @xmath407 is the last round of @xmath129 . clearly , @xmath408 . now we can apply lemma [ le : phi ] for node @xmath20 with @xmath101 as above and conclude that with probability at least @xmath403 , there are at most @xmath409 rounds in @xmath378 ( note that @xmath410 ) where @xmath345 , so there are at most @xmath411 rounds in @xmath129 where @xmath345 . now we can apply the same argument for all nodes in @xmath412 simultaneously and conclude that with probability at least @xmath135 , there are at most @xmath413 rounds in @xmath129 where @xmath345 for _ every _ @xmath399 , if we take @xmath395 to be large enough . in the remaining @xmath414 rounds we have @xmath415 . we extract the following result from the proof of lemma [ le : phi ] , to use it later . the proof is similar to the proof of prop . [ pr : lowc ] . [ co : pexpectation ] for every @xmath335 and each node @xmath1 , the expected number of rounds in each phase @xmath129 where @xmath416 , is @xmath411 . let @xmath378 be the shortest time interval containing @xmath129 such that @xmath417 at the beginning of @xmath378 . let @xmath418 denote the event that @xmath408 . as in the proof of prop . [ pr : lowc ] , we have @xmath419 > 1-n^{-3}$ ] . let @xmath370 denote the number of rounds @xmath375 with @xmath80 . we know from the proof of lemma [ le : phi ] that @xmath420 = o(\phi^{-k})\cdot|\hat h| = o(\phi^{-k})\cdot |h|$ ] and @xmath367=e[x|\mathcal{e}]\cdot pr[\mathcal{e } ] + e[x|\bar{\mathcal{e}}]\cdot pr[\bar{\mathcal{e}}]= o(\phi^{-k})\cdot |h|$ ] . this completes the proof as @xmath421 . next we show that if @xmath76 is appropriately chosen then for each node @xmath1 , @xmath422 happens most of the time during each phase w.h.p . in order to show this , first we split the set of nodes in @xmath423 into local neighborhoods . we show that if in a given round the contention in each local neighborhood is bounded by an appropriate threshold , then the expected interference is small . for each local neighborhood , the expected number of rounds when the contention is higher than its threshold is bounded using corollary [ co : pexpectation ] . combining these results into one we get a bound on the expected number of rounds when @xmath424 is small . it then remains to apply a concentration bound to conclude the proof . [ pr : pinterferenceboundl ] for every @xmath132 and @xmath425 , there are constants @xmath426 and @xmath427 , such that for each node @xmath1 and phase @xmath129 , with probability @xmath135 , at least @xmath136-fraction of rounds in @xmath129 are low interference rounds , i.e. @xmath428 . let @xmath429 denote the in - ball of radius @xmath430 centered at node @xmath1 for @xmath431 . let @xmath432 be an @xmath73-cover of @xmath429 of size @xmath433 for a constant @xmath434 , which exists by bounded independence . let @xmath435 for @xmath436 , where @xmath437 is a large enough constant to be defined later . [ le : pexpectal ] if @xmath438 holds for all @xmath439 and @xmath440 and @xmath441 then @xmath442 . we have , by the definition of expected interference ( explanations below ) , @xmath443 where the first inequality follows from the definition of sets @xmath444 and the assumption of the claim , the second one follows by @xmath435 and @xmath445 , the third one is a rearrangement of the sum , the fourth one follows by the fact that @xmath446 for every @xmath447 and @xmath448 ( here , @xmath449 ) , and the last one follows because @xmath450 . thus , if @xmath441 then the claim follows . let us put the nodes in an arbitrary order and let @xmath451 denote the number of rounds @xmath452 s.t . @xmath453 , where @xmath454 is the @xmath455-th node in @xmath456 . by corollary [ co : pexpectation ] , @xmath457= o(\phi_i^{-k})\cdot|h| $ ] holds for each @xmath458 . let @xmath459 denote the number of rounds in @xmath129 when @xmath460 . by claim [ le : pexpectal ] , @xmath461 $ ] can be bounded as follows , @xmath462\leq\sum_{i}\sum_{j}e[z_{i , j } ] \leq \sum_i|s_i\setminus s_{i-1}|\cdot o(\phi_i^{-k})\cdot |h| \leq \frac{\sigma}{3}|h|,\ ] ] where the last inequality holds if @xmath463 for a large enough constant @xmath464 ; we have @xmath465 and @xmath466 with @xmath467 ( by definition ) , so we have @xmath468 . each @xmath451 can be considered as the sum of binary random variables similar to that in the proof of lemma [ le : phi ] . thus , @xmath459 can also be seen as the sum of binary random variables . moreover , the binary random variables satisfy the conditions of the chernoff bound in lemma [ le : chernoff ] . thus , by taking @xmath469 and @xmath470 , we have : @xmath471\leq 2^{-\sigma|h|/3}\le n^{-3}. $ ] [ le : broadgood ] let @xmath472 , @xmath473 and @xmath474 , where @xmath475 are the parameters of . let @xmath56 be a round such that for a node @xmath20 , @xmath476 and @xmath477 . + if a node in @xmath332 transmits in round @xmath56 then it mass - delivers , with probability at least @xmath478 . let @xmath1 be the transmitting node , which could possibly be @xmath20 itself . the probability that no other node in @xmath479 transmits is @xmath480 where we used lemma [ ieq ] with the assumption that @xmath481 and the bounded contention assumption . on the other hand , we have that @xmath482\ge 9/10 $ ] . indeed , @xmath483\ge 9/10 $ ] follows by the bounded interference assumption and markov inequality , so it suffices to show that @xmath484 , which we have because for each node @xmath485 , @xmath486 , implying that @xmath487 thus , we have that with probability @xmath478 , @xmath1 transmits , no other node in @xmath217 transmits and the interference from @xmath488 at @xmath1 is at most @xmath68 , so @xmath1 delivers its message , by . * proposition [ pr : pdecrease ] * _ assume that constants @xmath137 are large enough . for each node @xmath1 and phase @xmath129 , if at least @xmath138-fraction of the rounds of @xmath129 are of high contention , then @xmath139 nodes in @xmath140 mass - deliver , with probability @xmath135 . _ let @xmath489 . let us choose constants @xmath490 such that prop . [ pr : pgoodroundl ] holds for @xmath491 and @xmath492 and such that @xmath493 is large enough , as stated below , where @xmath475 are the parameters of . we will use the fact that in prop . [ pr : pgoodroundl ] , @xmath394 has the following dependency on @xmath494 : @xmath495 , where @xmath0 is the edge change parameter ( see the proof of prop . [ pr : lowc ] ) . we set @xmath496 . namely , if we choose @xmath137 large enough ( satisfying the relation above ) , then we have that with probability @xmath135 , there are at least @xmath497 rounds @xmath358 such that @xmath498 and @xmath499 . let @xmath56 be such a round . for each node @xmath500 , it holds that @xmath501 and @xmath502 . let us fix a node @xmath500 . first note that @xmath503 , implying that @xmath504 . it remains to bound the expected interference at @xmath20 by nodes in @xmath488 . this interference can be split into two parts : 1 . the interference by nodes in @xmath505 and 2 . the interference by nodes in @xmath506 . the first part can be bounded by @xmath507 using a computation as in the proof of lemma [ le : broadgood ] . the second part can be bounded as follows : the contention in the area @xmath506 is at most @xmath394 and the distance of those nodes to node @xmath20 is at least @xmath76 , so the expected interference by nodes in @xmath506 is at most @xmath508 , and is less than @xmath509 if @xmath76 is large enough , as @xmath510 . here we used the assumption that @xmath511 . recall that there are at least @xmath512 high contention rounds for node @xmath1 , i.e. rounds where @xmath122 . let @xmath56 be such a round . by the bounded independence property , there is a @xmath513-cover @xmath39 of @xmath119 of constant size . by the pigeonhole principle , there is a node @xmath399 such that @xmath514 . by summarizing the above said , we get that there are at least @xmath497 rounds @xmath56 s.t . for all nodes @xmath500 , @xmath515 and @xmath516 . on the other hand , there are at least @xmath512 rounds where there is a node in @xmath119 that has @xmath517 . thus , there are at least @xmath518 rounds @xmath56 with the following property : there is a node @xmath500 such that @xmath519 and @xmath516 . fix such round @xmath56 and node @xmath20 . by lemma [ le : broadgood ] , there is a node @xmath520 that transmits ( the probability of this is at least @xmath521 ) and delivers its message with probability @xmath522 . thus , the expected number of nodes in @xmath119 that deliver their message during @xmath129 is @xmath523 . the proof now follows by applying a chernoff bound with the assumption that the constant @xmath128 is large enough : @xmath524 . * proposition [ pr : idle ] * _ assume that @xmath137 are large enough . for each node @xmath1 and phase @xmath129 , if at least @xmath141-fraction of the rounds of @xmath129 are low contention rounds , then with probability @xmath135 , in at least @xmath142-fraction of the rounds of @xmath129 , @xmath1 will detect idle channel and have low contention and low interference . _ let @xmath525 . assume that @xmath137 large enough , such that prop . [ pr : pgoodroundl ] holds for @xmath526 and @xmath527 , where @xmath528 is the parameter of . let @xmath418 denote the event that there are at least @xmath529 good rounds for @xmath1 in phase @xmath129 . by prop . [ pr : pgoodroundl ] , we have @xmath419 = 1-o(n^{-3})\ge 19/20 $ ] , if @xmath128 is large enough . given @xmath418 , there are at least @xmath530 rounds @xmath56 that are both good and low contention for @xmath1 . let @xmath159 denote the set of such rounds and let @xmath350 be a binary random variable with value 1 if and only if @xmath1 detects idle channel in round @xmath161 . let @xmath531 . for each round @xmath161 , since @xmath56 is a good round and @xmath532 , it holds with probability at least @xmath141 that the interference at node @xmath1 from nodes in @xmath119 is at most @xmath528 . given this and the fact that in low contention rounds we have @xmath533 , @xmath1 will detect idle channel with probability at least @xmath534 in round @xmath56 . thus , in each round @xmath161 , irrespective of prior rounds , we have @xmath535\ge 81/100 $ ] . this implies : @xmath420\ge ( 9 ^ 2/10 ^ 2)|h'|\ge \frac{9 ^ 2}{10 ^ 2}\cdot \frac{9 ^ 2}{10 ^ 2}|h|$ ] and @xmath536 \ge ( 19/20)\cdot\frac{9 ^ 4}{10 ^ 4}|h| > 0.62|h|.\ ] ] just as in the proof of lemma [ le : phi ] , we can now apply chernoff bound to obtain that @xmath537 < n^{-3}$ ] if @xmath128 is large enough . this completes the proof by recalling that all rounds of @xmath370 are low contention and low interference rounds for node @xmath1 . * theorem [ thm : broadcast ] * _ assume the edge change rate @xmath100 is sufficiently small . there are constants @xmath201 , such that when running bcast(@xmath114 ) in the synchronous mode , each node @xmath1 receives the message in @xmath202 rounds w.h.p . _ let us fix constants @xmath155 so that props . [ pr : pdecrease ] and [ pr : idle ] hold with interference threshold @xmath203 , where @xmath204 and @xmath205 are the parameters corresponding to precision @xmath183 . we set the passiveness parameter of try&adjust to @xmath206 and the stable distance parameter @xmath207 . consider a node @xmath20 that receives the message in round @xmath56 . let @xmath1 be such that @xmath208 in the time interval @xmath209 $ ] ( and both stay alive during @xmath96 ) . the theorem follows by an induction and union bound from the claim below . node @xmath1 gets the message of @xmath20 during @xmath96 , with probability @xmath210 . we prove the claim by contradiction , and assume that @xmath1 can not get the message during @xmath96 . let us split @xmath96 into phases ( for @xmath20 ) and , similar to the proof of thm . [ thm : localbroadcast ] , classify the phases @xmath129 into types : ( type a ) at least @xmath138-th of rounds in @xmath129 are high contention rounds for @xmath20 @xmath211 , and ( type b ) at least @xmath141-th of rounds in @xmath129 are low contention rounds for @xmath20 @xmath212 . we show that with probability @xmath135 , all phases in @xmath96 are of type b. consider a type a phase @xmath129 . we know from prop . [ pr : pdecrease ] that , with probability @xmath135 , there is a set @xmath39 of @xmath139 nodes in @xmath140 that deliver their messages and restart try&adjust during @xmath129 . we split @xmath39 into three subsets , @xmath213 those inserted by churn during @xmath129 , @xmath214 those inserted by edge change during @xmath129 and @xmath215 the rest . consider @xmath214 first . we know that in each local neighborhood , there are at most @xmath216 nodes arriving due to edge changes . note that @xmath217 can be covered with @xmath218 local neighborhoods ; hence , the total number of nodes arriving in @xmath217 due to edge changes is @xmath219 . since @xmath220 , setting @xmath221 a small enough constant gives @xmath222 . now consider @xmath215 . by the primitive , for each node @xmath223 , all nodes in @xmath224 also restart try&adjust , setting their transmission probability to @xmath225 . by the setting of @xmath226 , the probability that a node restarting try&adjust in phase @xmath129 transmits again in @xmath129 is at most @xmath227 . thus , with probability at least @xmath135 , the nodes in @xmath214 constitute an @xmath228-packing of @xmath140 and are at most @xmath229 . as for @xmath213 , the probability that a newly arriving node transmits during phase @xmath129 is at most @xmath227 , so @xmath230 w.h.p . these bounds lead to a contradiction @xmath231 , if constant @xmath128 is large enough . thus , each phase @xmath129 of @xmath96 is of type b with probability @xmath135 . it remains to recall that @xmath96 contains constant number of phases . now assume that @xmath129 is of type b. in a similar way as in theorem [ thm : localbroadcast ] , we can apply prop . [ pr : idle ] and show that if @xmath232 at the beginning of @xmath129 , then by the end of @xmath129 , the effect on @xmath71 is equivalent to applying at least @xmath233 of @xmath234 operations on @xmath235 . since @xmath236 at the beginning of @xmath96 is at least @xmath225 , it will take at most 11 phases to increase the probability to the value @xmath164 . then , a single phase will suffice for @xmath20 to deliver its message ( including to node @xmath1 ) , with probability at least @xmath210 . the basic observation is that if all the nodes start running bcast@xmath240 simultaneously , a _ constant density dominating set _ can be computed in @xmath200 rounds ( cf . @xcite ) , where an @xmath33-dominating set of density @xmath538 is a set @xmath39 of nodes such that for each node @xmath20 , @xmath539 . the dominator algorithm is as follows : all nodes run bcast@xmath240 simultaneously , and 1 . if a node @xmath1 stops by then it is a _ dominator _ , if a node @xmath20 stops by detecting of node @xmath1 then it is dominated by @xmath1 . having formed a constant density dominating set , it remains to disseminate the message using only dominators . in order for the set of dominators to derive the connectivity properties of the original graph , we need that @xmath30 forms a @xmath540-bounded independence _ metric _ space . in the static spontaneous setting , there is a uniform algorithm that performs broadcast in @xmath541 rounds , w.h.p . [ thm : static - spont - bcast ] the algorithm for spontaneous broadcast consists of two stages : 1 . compute a constant density @xmath542-dominating set @xmath543 , 2 . transmit the message using only the nodes in @xmath543 . [ [ dominating - set . ] ] dominating set . + + + + + + + + + + + + + + + the dominating set is constructed by running bcast@xmath240 in the spontaneous mode , i.e. all nodes start running the algorithm simultaneously . let @xmath543 be the set of nodes that stop the algorithm by . since the dominated nodes use @xmath544 as a stopping condition , @xmath543 is a @xmath542-dominating set . moreover , @xmath543 is a @xmath545-packing , which implies that each in - ball @xmath46 of radius @xmath546 contains at most @xmath547 dominators . in particular each node @xmath20 is dominated by at most constant number of dominators . the proof the algorithm terminates after at most @xmath200 rounds is almost identical to the proof of theorem [ thm : broadcast ] . [ [ broadcast . ] ] broadcast . + + + + + + + + + + recall that we assume that the communication graph is connected and has diameter @xmath243 . also , since @xmath548 is a metric , the graph is undirected . the broadcast part is as follows : in the first round , the source node transmits the message to its neighbors ; each dominator @xmath20 , upon receiving the message , transmits it in each round with probability @xmath232 until detecting ( @xmath183 ) , where @xmath235 is a small enough constant . note that as soon as all dominators successfully transmit at least once , all nodes will get the message . moreover , if the constant @xmath235 is small enough , then the two algorithms can be run simultaneously : the key point is that the number of dominators in each @xmath549-neighborhood is bounded by a constant ( not depending on @xmath235 ) . this ensures that the nodes _ need not know @xmath16 _ , in order to coordinate the two algorithms . it remains to show that the message will get to all dominators in @xmath244 rounds using the broadcast algorithm . consider a graph @xmath129 defined over the dominating set @xmath543 , where for every pair of nodes @xmath550 , @xmath551 form an edge if @xmath552 . the diameter of @xmath129 is at most @xmath243 . note that for every pair of nodes @xmath553 with @xmath554 , the corresponding dominators @xmath555 and @xmath556 are adjacent in @xmath129 . indeed , by the definition of the dominating set , we have @xmath557 let @xmath558 be two arbitrary nodes in @xmath543 and let @xmath19 be the path of length at most @xmath243 in @xmath238 connecting nodes @xmath555 and @xmath556 . by the observation above , if we replace each node @xmath559 with its dominator @xmath560 , we obtain a path @xmath561 of length @xmath46 in @xmath129 , connecting @xmath555 and @xmath556 . the completes the proof of the claim . for every transmitting node @xmath562 , the probability that @xmath274 delivers its message to all its neighbors in @xmath129 is @xmath563 in each round @xmath56 , if constant @xmath235 is sufficiently small . let @xmath39 denote the set of neighbors of node @xmath274 in @xmath129 , i.e. the nodes of @xmath543 that are at distance at most @xmath564 from @xmath274 . it suffices to show that all nodes in @xmath39 receive the message in a fixed round @xmath56 where @xmath274 transmits with @xmath563 probability . this event holds if the interference at @xmath274 is no more than @xmath68 event @xmath565 , and no other node transmits in @xmath566 event @xmath567 , in round @xmath56 . by the properties of the dominating set , the contention in a ball of radius @xmath568 is at most @xmath569 . similarly to the proof of prop . [ pr : pinterferenceboundl ] , it can be shown that the expected interference by nodes in @xmath570 is @xmath571 . thus , by markov inequality , @xmath572 if @xmath235 is small enough . on the other hand , since the contention in @xmath566 is @xmath571 , @xmath573 follows by lemma [ ieq ] . thus , @xmath574 . given that each informed node in @xmath543 has probability at least @xmath575 of successfully broadcasting the message , the rest of the proof essentially follows along the lines of the proof of ( * ? ? ? * lemma 6 ) . we present a sketch of the proof for completeness of the argument . let @xmath576 be the set of nodes in @xmath543 that have been informed by round @xmath56 , where @xmath577 contains only the source node . let us fix a node @xmath578 and let @xmath579 be the distance ( in graph @xmath129 ) from @xmath1 to the nearest node in @xmath576 . note that @xmath580 . the difference @xmath581 is the progress made in round @xmath56 and is a bernoulli random variable with @xmath582\ge \eta$ ] for all @xmath56 when @xmath583 and @xmath584 otherwise . let @xmath585 be a random variable that has the same distribution as @xmath586 when @xmath583 and is an i.i.d . bernoulli random variable with mean @xmath587 otherwise . let @xmath588 . note that node @xmath1 has been informed by round @xmath56 iff @xmath589 . thus , we need to bound the probability @xmath590 . let @xmath591 . it is easy to show that the sequence @xmath592 is a submartingale . moreover , for a round @xmath593 and constant @xmath594 , @xmath595 implies that @xmath596 . since the sequence @xmath592 is a submartingale , we can apply azuma - hoeffding bound to show that @xmath597 . now the theorem follows by union bound over all nodes , by choosing the constant @xmath260 large enough . * theorem [ thm : brlowerbound ] * _ for every ( possibly randomized ) broadcast algorithm @xmath246 that uses neither node coordinates nor primitive , there is a @xmath247-bounded - independence metric space where @xmath246 needs @xmath248 rounds to do broadcast in a @xmath249-broadcastable network , even if the nodes have and primitives and operate spontaneously . _ assume that @xmath598 . we present the construction using a distance function @xmath548 : it is then straightforward to construct the corresponding path - loss matrix . denote @xmath599 . first , assume the nodes operate non - spontaneously , i.e. a non - source node may need to receive a message in order to start participating in a protocol . recall that @xmath600 . consider @xmath16 points @xmath601 , such that for every @xmath602 , it holds that @xmath603 , @xmath604 , @xmath605 and @xmath606 , where @xmath607 ( see the diagram in fig . [ fig : nonspon ] ) . clearly , this set of points forms a @xmath608-bounded independence metric space . we place @xmath16 wireless nodes at distinct points @xmath601 uniformly at random . let @xmath195 be the node at @xmath609 . we assume , further , that communication only happens according to ; in particular , if the interference at a node is more than @xmath68 then _ none _ of its neighbors receives its transmission . note that broadcast in this network can be completed in 2 steps , starting at any point . note also that @xmath610 can not be directly reached from nodes @xmath195 with @xmath611 . moreover , if at least @xmath612 nodes @xmath195 with @xmath611 transmit simultaneously , _ no node _ receives a message , including the potential communication between @xmath613 and @xmath610 . this follows by observing that @xmath614 : the signal power at the neighboring nodes must be more than the interference threshold , otherwise a node could receive two signals of the same power . assume @xmath195 with some @xmath611 is the source . at the first round of the algorithm , nodes @xmath615 receive the message . let @xmath616 denote the event that @xmath613 transmits and no more than @xmath617 nodes @xmath195 with @xmath611 transmit in round @xmath56 . by the observations above , @xmath610 will not receive the message until @xmath616 happens . moreover , during the subsequent steps of the algorithm , until @xmath616 happens for the first time , all nodes at @xmath615 will have the same history and will be symmetric with respect to and primitives . it remains to bound the expected time until @xmath616 happens for the first time . we can assume w.l.o.g . that in each round there are no more than @xmath187 nodes transmitting . as discussed above , if @xmath610 is not informed in such a round , then the nodes that transmitted in the given round can learn at best that none of them is at @xmath613 and stop transmitting thereafter . thus , we can assume that after each unsuccessful round there are at most @xmath618 nodes that stop transmitting and the probability @xmath619 of success at round @xmath620 ( if no success occurred before round @xmath33 ) is at most @xmath621 where the numerator counts the number subsets of an @xmath622-element set that contain a fixed element ( the node @xmath613 ) and no more than @xmath187 elements in total , and the denominator counts all subsets containing no more than @xmath187 elements . a further straightforward calculation shows that the expected number of rounds until the first success is @xmath248 ( see e.g. ( * ? ? ?

scale invariance is a well - known symmetry @xcite that has been studied in many physical contexts . a strong physical motivation for incorporating scale symmetry in fundamental physics @xcite@xcite@xcite comes from low energy particle physics . namely , the _ classical _ action of the standard model is already consistent with scale symmetry if the higgs mass term is dropped . this invites the idea , which many have considered , that the mass term may emerge from the vacuum expectation value of an additional scalar field @xmath0 in a fully scale invariant theory . another striking hint of scale symmetry occurs on cosmic scales : the ( nearly ) scale invariant spectrum of primordial fluctuations , as measured by wmap and the planck satellite @xcite . this amazing simplicity seems to cry out for an explanation in terms of a fundamental symmetry in nature , rather than as just the outcome of a scalar field evolving along some particular potential given some particular initial condition . in this paper , we consider the incorporation into fundamental physics of _ local conformal symmetry _ ( or weyl - symmetry ) : that is , classical local scaling symmetry in an action that includes the standard model coupled to gravity . a new result of our approach @xcite-@xcite is that using conformal symmetry we are able to solve the classical frw equations across big crunch - big bang transitions , thus obtaining the _ full set of geodesically complete cosmological solutions _ of our conformal standard model given in eq.([action1 ] ) . this follows from the new properties of the standard theory whose couplings to gravity includes _ all patches of field space _ that are required for geodesic completeness of all cosmological solutions for all times and any set of initial conditions ) as described in section ( [ confcosm ] ) . ] . the models we describe contain no mass scales no gravitational constant , no mass for the higgs field , no cosmological constant and no mass parameters for the quarks , leptons or gauge bosons . all of these are prevented by the local conformal symmetry combined with the su@xmath1su@xmath2u@xmath3 gauge symmetry of the standard model . there is only one source of mass which follows from gauge fixing ( in some sense spontaneously breaking ) the weyl symmetry through a scalar field which is a singlet under su@xmath1su@xmath4u@xmath3 or a combination of both su@xmath5su@xmath6u@xmath3 singlet and non - singlet fields , depending on the choice of gauge . this source , which is associated with the emergence of the planck scale , drives the spontaneous breakdown of the electroweak symmetry , which in turn generates the other known masses of elementary constituents , with ratios of masses related to dimensionless constants . our approach to conformal symmetry has new features with significant physical consequences that were not considered before . while we shall work in @xmath7 dimensions throughout , there is a close connection with field theories and gravity or supergravity in @xmath8 dimensions in the context of 2t - physics @xcite-@xcite as will be pointed out occasionally throughout this paper . stated directly in @xmath7 dimensions , some of the novel features and their consequences are as follows : * in section ( [ general ] ) we provide the most general form of coupling gravity or supergravity with any number of scalar fields , fermions and gauge bosons , while maintaining local conformal symmetry . only a specialized form of our general formalism in section ( [ general ] ) coincides with previously known methods for local conformal symmetry . we find that in a gauge symmetric theory , if there is only one physical scalar field ( for example , the higgs field ) _ after all gauge choices are exhausted _ , then there is a unique gauge - invariant and weyl - invariant form for the lagrangian representing the geodesically complete coupling to gravity . * the minimal , realistic and locally conformal standard model coupled to gravity includes the higgs doublet @xmath9 and an additional su@xmath10su@xmath11u@xmath12 singlet @xmath0 . taking into account all of the local symmetry , there is a single physical scalar field identifiable with the observed higgs field . therefore , according to the statement above , this is a geodesically complete conformal standad model provided it utilizes our unique coupling to gravity as shown in eq.([action1 ] ) and as proven in section ( [ generalgrav ] ) . the relative minus sign between @xmath13 and @xmath14 that appears in this action is obligatory and is related to geodesic completeness as will be explained later . the unique structure of local scale symmetry has led to full analytic control of cosmological solutions and , in certain limits , enabled us to obtain the complete set of homogeneous cosmological solutions for all possible initial conditions of the relevant fields . this is discussed in section ( [ confcosm ] ) . * in a particular gauge of the spontaneously broken local scale symmetry ( called @xmath15-gauge ) we recover the usual renormalizable field theory of the standard model in the flat space limit , including a mass scale for the higgs . in this gauge , low energy physics of our model coincides with the familiar form of the standard model unless more fields are included beyond those already observed , such as right handed neutrinos and dark matter which can be accommodated consistently with conformal symmetry following our methods . thus , low energy physics at lhc scales is not sensitive to the conformal structures suggested in our approach . * on the other hand , physics at cosmological scales can be quite sensitive to the conformal structures discussed in this paper as becomes apparent in certain gauges . in section ( [ confcosm ] ) we outline some of the phenomena that follow from our unique conformal coupling . in the minimal conformal standard model of eq.([action1 ] ) , guided by our analytic solutions ( supplemented , where necessary , with numerical methods ) , we propose a _ cyclic conformal cosmology _ driven only by the higgs field with no recourse to other scalar fields such as an inflaton . the higgs - driven cosmological scenario we propose is strongly motivated by the metastability of the higgs vacuum as implied by the renormalization - group flow of the minimal standard model to large higgs vev , with its parameters fixed by the lhc measurements @xcite . the main discussion in this paper is at the classical field theory level , but it is worth commenting briefly on the important question of quantum corrections . there are two enormous hierarchies in the standard model coupled to gravity : the weak scale is around @xmath16 and the dark energy scale is around @xmath17 of the planck scale . it is hard to understand how such tiny numbers enter fundamental physics , and why quantum corrections would not spoil these fine tunings . in both cases , conformal symmetry has been suggested as a solution . we will not explore this here in any detail , except to note that the stability of such hierarchies in the perturbative standard model , which is what attracted attention in the past for the higgs field , should not be in jeopardy from quantum corrections since dimensionless constants in a conformal theory are logarithmically divergent as opposed to the quadratic divergence of a bare higgs mass term . to fully grasp how this can work at the quantum level for fundamental scalars and in particular the higgs , requires a better understanding of how to perform regularization and renormalization consistent with local conformal symmetry in @xmath7 dimensions . elaborating on earlier suggestions @xcite@xcite@xcite , there has been recent progress @xcite in developing a renormalization theory consistent with weyl symmetry . according to this new work , local conformal symmetry remains as a valid symmetry at the quantum level as anticipated in @xcite . the local scale invariance survives quantization even though there is a trace anomaly in the stress tensor of the matter sector of the theory ; this point that caused confusion in the past is clarified by noting that the trace of the _ matter _ stress tensor is _ distinct _ from the generator of local scale transformations that includes the additional fields , such as @xmath18 that implement and establish the local conformal symmetry . in passing , we note that if @xmath7 dimensional conformal symmetry is treated consistently within 2t - physics in @xmath8 dimensions , then scale ( dilation ) symmetry is a part of a _ linearly realized _ so@xmath19 in the flat limit which is presumably not anomalous just as its lorentz subgroup so@xmath20 in @xmath7 dimensions can not be anomalous . this quantum aspect of 2t - physics is under investigation directly in @xmath8 dimensions and this is expected to shed additional light on these issues . a plan of this paper is as follows . in sec . ii , we will review and expand on the motivation for global and local symmetry and show how to recast ( or lift ) the standard model plus gravity into a locally gauge invariant , weyl- symmetric theory by adding new fields and , at the same time , introducing compensating gauge symmetries , in such a way that the theory is _ both weyl invariant and geodesically complete_. the gauge fixed version of this theory reverts back to the familiar minimal form of the standard model at low energies wherever gravity is negligible . however , in regions of spacetime where strong gravity effects are important , such as in cosmology , especially close to the singularity , the geodesically complete weyl lifted version plays a crucial role , as described in sec . in section iii , following @xcite we present a general form for a lagrangian consistent with weyl symmetry for any number of scalar fields . we argue that , if the weyl invariant theory is equivalent to having only a single physical scalar field after gauge fixing ( like the physical higgs ) , then by field redefinitions it is always possible to recast the theory into our version given in eq.([action1 ] ) which includes all patches in field space so as to insure geodesic completeness . as another example of our general formalism , we lift the bezrukov - shaposhnikov higgs inflation model @xcite to a fully scale invariant model , making it logically and physically consistent . based on our formalism for constructing general weyl - invariant actions , we show that the model is not unique . in fact , there are many single - field cosmological higgs models with the same properties , all consistent with conformal symmetry and producing the same inflationary outcome . however , this inflationary scenario is not geodesically complete , furthermore , in view of our complete set of solutions , it is extremely unlikely . in sec . iv , we turn more generally to cosmology . we explain that a weyl - lifted theory can resolve the singularity and enable the geodesic completion of cosmological spacetimes , while also indicating the presence of new phenomena . our previous work in @xcite-@xcite elaborated in some detail on the properties of geodesically complete cosmological analytic solutions in the context of solvable examples of what was supposed to be toy models . however the same models now re - appear in our conformal standard model and hence our previous solutions are now the full set of cosmological solutions for all initial conditions of the familiar standard model . in particular , we point out a model independent attractor mechanism discovered in @xcite that may help to determine the likely initial conditions of our universe just after the big bang . the weyl - invariant formulation also provides a natural framework for incorporating cyclic cosmology @xcite driven only by the higgs field as introduced . in the case that the higgs field is metastable , the higgs inflation model is inoperable , but the minimal single - higgs model is naturally compatible with the cyclic picture @xcite . in general , the weyl - invariant approach also hints at the intriguing possibility that the minimal , electroweak higgs may have played a central role in cosmic evolution as explained in detail in @xcite . there is a deep connection between the ideas presented in this paper and theories in @xmath21-space and @xmath22-time dimensions . many of the ideas discussed here for a global or local conformally symmetric theory emerged progressively from developing the @xmath8 dimensional formalism since 1996 ( for a recent overview see @xcite ) , and then for the standard model as given in @xcite , for gravity in @xcite , for susy in @xcite and supergravity in @xcite . the formulation given in @xcite , with many scalars coupled to gravity and supergravity in @xmath8 and @xmath7 dimensions was the precursor of the general formalism presented in section iii , while the simplest specific models were analyzed cosmologically in some detail in @xcite-@xcite . there are some recent similar examples @xcite that overlap with our conformal symmetry vision in @xmath7 dimensions ; these seem to be oblivious to our previous publications that introduced at an earlier stage the crucial conformal structures with restrictions on scalars in @xmath7 dimensions as predictions from 2t - physics @xcite . the @xmath8 theory has more predictions of hidden symmetries and dualities in @xmath7 dimensions which are mainly understood in classical and quantum mechanical contexts @xcite and are also partially developed in field theory @xcite ; these go well beyond conformal symmetry in their implications for unification and the meaning of spacetime and are bound to play an interesting role in future progress . we emphasize that , despite the name , the physics content in the 2t - physics formalism in @xmath23 dimensions is same as the physics content in the standard 1t - physics formalism in @xmath24 dimensions except that 2 t physics provides a holographic perspective and , due to a much larger set of gauge symmetries , naturally makes predictions that are not anticipated in 1t - physics . these additional gauge symmetries are in phase space rather than position space , and can be realized only if the formalism is developed with two times . nevertheless , the gauge invariant sector of 2t - physics is equivalent to a causal 1-time spacetime without any ghosts . in this paper , we will stick to @xmath7 dimensions , advocating a coherent overall picture of a conformally invariant formulation of fundamental physics and cosmology . however , we will occasionally remind the reader that these outcomes naturally follow from @xmath8 dimensions with appropriate ( but unusual ) gauge symmetries . in this section , we will describe the motivation for and construction of simple theories with global and local scale invariance . an important application is the lift of the standard model plus gravity into a weyl- invariant theory that is also geodesically complete . we begin by examining a scale invariant extension of the standard model of particles and forces . for example consider the usual standard model with all the usual fields , including the doublet higgs field @xmath9 coupled to gauge bosons and fermions , but add also an su@xmath25u@xmath3 singlet @xmath0 ( plus right handed neutrinos and dark matter candidates ) , and take the following purely quartic renormalizable potential involving only the minimal set of scalar fields@xmath26 this model , discussed at some length in section vi of @xcite , is the minimal extension of the standard model that is fully scale invariant at the classical level , globally . see also @xcite-@xcite that use a similar field , and the different approach that also adds a weyl vector field @xcite leading to different physical consequences . the field @xmath0 , which we call the dilaton , absorbs the scale transformations and is analogous to the dilaton in string theory . in the current context , it has a number of interesting features : due to su@xmath1su@xmath25u@xmath3 gauge symmetry , the singlet @xmath13 is prevented from coupling to all other fields of the standard model - except for the additional right handed singlet neutrinos or dark matter candidates . these features of @xmath18 that prevent it from interacting substantially with standard visible matter except via the higgs in eq.([potential ] ) , suggest naturally that @xmath13 itself could be a candidate for dark matter @xcite . the only parameters associated with @xmath18 that are relevant for our discussion , are @xmath27 and @xmath28 . for positive @xmath29 the minimum of this potential occurs at @xmath30 accordingly , the vacuum expectation value of the higgs may fluctuate throughout spacetime , depending on the dynamics of @xmath31 without breaking the scale symmetry . however , if for some reason ( e.g. driven by quantum fluctuations or gravitational interactions ) @xmath13 develops a vacuum expectation value @xmath32 which is constant in some region of spacetime , then the higgs is dominated by a constant vacuum expectation value @xmath33 with @xmath34 fixed by observation to be approximately @xmath35 the higgs vacuum @xmath36 provides the source of mass for all known elementary forms of matter , quarks , leptons , and gauge bosons ( while @xmath32 may be the source of majorana mass for neutrinos ) . the observation at the lhc of the higgs particle , which is just the small fluctuation on top of the vacuum value @xmath34 , has by now solidified the view that this is how nature works in our region of the universe , at least up to the energy scales of the lhc . fig . 1 illustrates how the higgs field slowly relaxes to the spontaneously broken symmetry vacuum described by eq . ( [ v - value ] ) beginning from large oscillations shortly after the big bang . here , as is the case throughout the paper , the solutions are in the limit of negligible gauge and top quark mass coupling so that the higgs evolution is described by classical equations of motion . this time - dependent behavior of the higgs is driven by the evolution of the field @xmath13 , as anticipated in @xcite for this simple model . in fact , it is realized as the generic solution for all homogeneous cosmological solutions if the higgs vacuum is stable . there are other interesting theoretical structures worth noting about this scale invariant setup . instead of supersymmetry , conformal symmetry may explain the stability of the hierarchy between the low mass scale of @xmath37 versus the planck scale of @xmath38 gev , as suggested in @xcite . the conformal protection of the hierarchy is not as clear - cut as susy s protection and requires better understanding of regularization and renormalization techniques consistent with conformal symmetry as outlined in the introduction . what about the dilaton ( fluctuations in @xmath13 ) that emerges in the broken scale invariance scenario above ? at least from the perspective of only the standard model , the dilaton is a massless goldstone boson due to the spontaneous breaking of the global scale invariance . as discussed in some detail in section vi of @xcite , the original doublet field @xmath14 is the only field coupled to known matter , while @xmath13 is decoupled . however , after the spontaneous breaking , @xmath14 must be rewritten as a mixture of the mass eigenstates of the model , which include the observed massive higgs particle and the massless dilaton ( or maybe low - mass dilaton if the scale symmetry is broken by some source ) . the mixing strength is controlled by the dimensionless parameter @xmath39 that appears in eq.([v - value ] ) . therefore , through this mixing , the ( pseudo- ) goldstone dilaton must couple to all matter , just like the higgs does , namely with a coupling given by the mixing angle @xmath40 times the mass of the particle divided by @xmath34 . the largest coupling is to the top quark . if @xmath32 is much larger than @xmath41 gev@xmath42 it is possible for the dilaton to hide from accelerator experiments due to the weak coupling of @xmath27 in eq.([alpha ] ) . such a ( pseudo ) goldstone boson has many other observable consequences , including a long range force that competes with gravity and contributions to quantum effects as a virtual particle in feynman loop diagrams . if such an additional massless ( or low mass scalar ) is observed in experiments , it would be strong evidence in favor of the scale invariant scenario . however , if it is not observed , this could be interpreted merely as setting a lower limit on the scale @xmath43 eventually it must be understood what sets the scale @xmath32 . the model as presented above has no mechanism at the classical level to set the scale @xmath32 ; its equations of motion are self consistent at the minimum of the potential , @xmath44 only if @xmath45 @xcite ; then @xmath32 remains undetermined due to a flat direction in the remaining term in the potential ( [ potential ] ) . the value for @xmath32 would then be obtained phenomenologically from experiment , without a theoretical explanation . quantum corrections , such as those discussed in @xcite , may alleviate this problem by removing the flat direction . however , if the quantum corrections are small , there is the danger that @xmath32 would be so low that it is not possible to obtain the small value of @xmath27 in eq.([alpha ] ) required to protect the dilaton from current experimental limits . since at present there is no sign of the dilaton in low energy physics , suppose it does not exist at all as a degree of freedom . is this incompatible with the idea that conformal symmetry underlies fundamental physics ? not at all , because any possible phenomenological problems associated with a goldstone boson fluctuation of @xmath13 can be overcome if the scaling symmetry is a _ local _ gauge symmetry , known as the weyl symmetry . then the massless fluctuations of the dilaton can be eliminated by fixing a unitary gaugeu@xmath3 scalar field as part of the general theory that we will discuss in section ( [ generalgrav ] ) . but for simplicity in this section we will first concentrate on the minimal model that contains only the confirmed observed degrees of freedom up to now , thus allowing the fluctuations of a single @xmath13 to be eliminated by a gauge symmetry in this minimal model . ] . the standard model decoupled from gravity has no local scale symmetry that could remove a goldstone dilaton . but such a gauge symmetry can in fact be successfully incorporated as part of the standard model provided it is coupled to gravity in the right way . coupling the standard model to einstein gravity in the conventional way makes no sense because the dimensionful newton constant explicitly breaks scale invariance . if scale invariance is already broken in one sector of the theory , then there is no rationale for requiring that it be a good symmetry in another part of the theory . at best , it would occur as an accidental symmetry of low energy physics and only when gravity is negligible . this is not the scenario we have in mind ; we argue for a fully scale invariant approach to all physics , a natural outcome of the larger gauge symmetries in @xmath8 dimensions , as formulated in 2t - physics . happily , the idea of an underlying @xmath8 dimensions with appropriate extra gauge symmetry fits all known physics in @xmath7 dimensions , from dynamics of particles and field theory @xcite-@xcite all the way to supergravity @xcite . so , consistency with this larger underlying structure is well motivated . in fact , there is a locally scale invariant field theory in @xmath7 dimensions , compatible with 2t - physics in @xmath8 dimensions , that couples the standard model and gravity with no dimensionful constants . we do not mean conformal gravity which has ghost problems , but rather the non - minimal conformal coupling of the curvature @xmath46 to scalar fields @xcite which is invariant under weyl transformations as a gauge symmetry . in the next section we will discuss a generalized weyl invariant coupling with many scalar fields also predicted by 2t - physics , but in this section we begin with the well known method of conformally coupled scalars , as follows @xcite@xcite,@xmath47 these two terms form an invariant unit ( up to a total derivative ) under _ local _ scale transformations , @xmath48 , with a local parameter @xmath49 when there are more scalar fields , the most general way of achieving local conformal symmetry is discussed in section ( [ genweyl ] ) . however , when there is only one additional scalar field beyond @xmath18 which is in a single representation of a yang - mills gauge group , there is a unique way to also achieve geodesic completeness as explained in more detail in sections ( [ generalgrav],[confcosm ] ) . hence , _ for geodesic completeness _ we require that not only the field @xmath18 but also the doublet higgs field be a set of conformally coupled scalars consistent with su@xmath2u@xmath3 . namely using the unit @xmath50 which is also locally invariant ( up to a total derivative ) , we can lift the globally scale invariant standard model of the last section into a locally invariant one , while also being coupled to gravity in a gedesically complete theory . all other terms present in the usual standard model , namely all fermion , gauge boson and yukawa terms , when minimally coupled to gravity are already automatically invariant under the local weyl symmetry . hence a _ weyl invariant _ action @xmath51 that describes the coupling of gravity and the standard model is given by @xmath52{c}\frac{1}{12}\left ( \phi^{2}-2h^{\dagger}h\right ) r\left ( g\right ) \\ + g^{\mu\nu}\left ( \frac{1}{2}\partial_{\mu}\phi\partial_{\nu}\phi - d_{\mu } h^{\dagger}d_{\nu}h\right ) \\ -\left ( \frac{\lambda}{4}\left ( h^{\dagger}h-\alpha^{2}\phi^{2}\right ) ^{2}+\frac{\lambda^{\prime}}{4}\phi^{4}\right ) \\ + l_{\text{sm}}\left ( \begin{array } [ c]{c}\text{{\small quarks , leptons , gauge bosons,}}\\ \text{{\small yukawa~couplings to~}}{\small h\&\phi,~}\text{{\small dark~matter.}}\end{array } \right ) \end{array } \right ] \label{action1}\ ] ] the only yukawa couplings of @xmath13 allowed by su@xmath5su@xmath6u@xmath3 are to the right handed neutrinos for which it becomes a source of mass . note the relative minus sign between @xmath13 and @xmath14 terms , which is required , as explained below . here @xmath53 is the well known standard model lagrangian minimally coupled to gravity , except for the higgs kinetic and potential terms , which are now modified and explicitly written out in the first three lines . this action is invariant under weyl rescaling with an arbitrary local parameter @xmath54 as follows@xmath55{c}g_{\mu\nu}\rightarrow\omega^{-2}g_{\mu\nu},~\phi\rightarrow\omega \phi,\;h\rightarrow\omega h,\;\\ \psi_{q , l}\rightarrow\omega^{3/2}\psi_{q , l},\;a_{\mu}^{\gamma , w , z , g}\rightarrow\omega^{0}a_{\mu}^{\gamma , w , z , g } , \end{array } \label{weylrescaling}\ ] ] where @xmath56 are the fermionic fields for quarks or leptons , and @xmath57 are the gauge fields for the photon , gluons , @xmath58 and @xmath59 note that the gauge bosons do not change under the weyl rescaling . we note that the lagrangian ( [ action1 ] ) is the one obtained in the second reference in @xcite from the @xmath8 version of the standard model @xcite@xcite coupled to 2t- gravity @xcite , by a method of gauge fixing and solving some kinematical equations associated with constraints related to the underlying gauge symmetries . in that approach we learned that the weyl symmetry is not an option in @xmath7 dimensions , it is a prediction of 2t - physics : the @xmath8 dimensional theory is not weyl invariant , but yet the local weyl symmetry in @xmath7 dimensions emerges as a remnant gauge symmetry associated with the general coordinate transformations as they act in the extra @xmath60 dimensions . so the weyl symmetry is a required symmetry in @xmath7 dimensions as predicted in the @xmath8 dimensional approach ; this symmetry carries information and imposes properties related to the extra @xmath60 space and time dimensions @xcite . the weyl gauge symmetry of the action in ( [ action1 ] ) does not allow any dimensionful constants : no mass term in the higgs potential , no einstein - hilbert term with its dimensionful newton constant , or any other mass terms . this is a very appealing starting point because it leads to the emergence of all dimensionful parameters from a single source . that source is the field @xmath0 that motivated this discussion in the previous section , and the only scale is then generated by gauge fixing @xmath13 to a constant for all spacetime @xmath61 in the gauge fixed version of the lagrangian where @xmath62 we can express the physically important dimensionful parameters , namely , the newton constant @xmath63 , the cosmological constant @xmath64 associated with dark energy , and the electro - weak scale @xmath34 , in terms of @xmath32 : @xmath65 since the field @xmath0 in this gauge ceases to be a degree of freedom altogether , the massless dilaton is absent and the potential problem with global scaling symmetry is avoided . nevertheless , there still is an underlying hidden conformal symmetry for the full theory . we can explain now why it is necessary to have the opposite signs for @xmath13 and @xmath14 in the first two lines of the action in eq.([action1 ] ) . the positive sign for @xmath13 is necessary in order to obtain a positive gravitational constant in ( [ dimensionfulconstants ] ) . however , conformal symmetry then forces the kinetic term for @xmath13 to have the wrong sign , so @xmath13 is a ghost . this can be seen to be a gauge artifact , though , since in a unitary gauge the ghost @xmath13 is fixed to a constant or expressed in terms of other degrees of freedom . this also explains why @xmath14 must have the opposite sign , since otherwise @xmath14 would be a real ghost . this relative sign has important consequences in cosmology as follows . in flat space @xmath66 where experiments such as those at the lhc are conducted , the lagrangian above becomes precisely the usual standard model , including the familiar tachyonic mass term for the higgs . furthermore , in weak gravitational fields , at low energies , the gravitational effect of the higgs field coupling to the curvature , @xmath67 is ignorable since @xmath14 ( order of @xmath68 gev ) is tiny compared to the planck scale ( @xmath6910@xmath70 gev ) . actually , the gravitational constant measured at low energies is corrected by the electroweak scale , namely @xmath71 @xmath72 instead of ( [ dimensionfulconstants ] ) , but in practice this is a negligible correction since @xmath73 so , at low energies there is no discernible difference between our weyl - lifted theory and the usual standard model . the practically isolated standard model appears as a renormalizable theory decoupled from gravity . however , in the cosmological context , the conformally coupled @xmath14 can and will be large at some stages in the evolution of the universe . working out the dynamics of cosmological evolution , it turns out that , for generic initial conditions , @xmath74 typically grows to large scales over cosmological times , hitting @xmath75 in the vicinity of a big bang or big crunch singularity . this behavior plays an essential role in cosmological evolution , as well as in the resolution of the singularity via geodesic completeness . in our conformal theory , the standard model is not isolated from gravity , and we posit that the higgs field can play a bigger role in nature than originally anticipated . we may take the higgs doublet in the unitary gauge of the su@xmath25u@xmath3 gauge symmetry @xmath76{c}0\\ \frac{1}{\sqrt{2}}s\left ( x\right ) \end{array } \right ) . \label{unitary}\ ] ] the field @xmath77 , where @xmath78 is the higgs field fluctuation on top of the electroweak vacuum , is then identified with the generic field @xmath79 that appeared in our cosmological papers @xcite-@xcite . the action for @xmath80 has the exact same form in those papers as here . thus , our previous analytical cosmological solutions can now be applied to investigate the cosmological properties of the higgs field . when the parameter @xmath27 vanishes , our previous work provides the full set of geodesically complete analytic cosmological solutions for all initial conditions of the relevant fields in the standard model coupled to gravity . the non - zero @xmath27 is then easily taken into account with numerical methods . this much analytic control over cosmological properties of a realistic theory is unprecedented . this became a very valuable tool that led us into a new cosmological scenario driven only by the higgs field , as outlined in section ( [ confcosm ] ) and completed in @xcite . the gauge symmetries derived from 2t - gravity and 2t - supergravity in @xmath8 dimensions lead to the general weyl invariant coupling described below for any number of scalar fields in @xmath7 dimensions @xcite . this generalizes the possible forms of conformally coupled scalar theories beyond those encompassed by eq.([confscalar ] ) and allows for richer possibilities for model building consistent with local scale invariance . we ignore the spinors and gauge bosons whose minimal couplings to gravity are already automatically weyl symmetric . we begin with gravity without supersymmetry . in the next subsection we will indicate the additional constraints that emerge in supergravity . we assume any number of _ real _ scalar fields @xmath81 labeled by the index @xmath82 if there are complex fields , we can extract their real and imaginary parts and treat those as part of the @xmath83 we introduce a weyl factor @xmath84 a sigma - model - type metric in field space @xmath85 and a potential @xmath86 the general lagrangian takes the form @xmath87 the results given in @xcite are the following constraints on these functions . * @xmath88 must be homogeneous of degree two , @xmath89 and @xmath90 must be homogeneous of degree four@xmath91 and @xmath92 must be homogeneous of degree zero@xmath93 * the following differential constraints must also be satisfied . these may be interpreted as homothety conditions on the geometry in field space@xmath94 the second and third equations follow from the first one and the homogeneity requirements . * physics requirements also include that @xmath95 can not have more than one negative eigenvalue because the local scale symmetry is just enough to remove only one negative norm ghost . however , if more gauge symmetry that can remove more ghosts is incorporated , then the number of negative eigenvalues can increase accordingly . the gauged r - symmetry in supergravity ( which is automatic in the @xmath8 approach ) is such an example . in @xcite these rules emerged from gauge symmetries in 2t- gravity . since weyl symmetry is an automatic outcome from @xmath8 dimensions , we can check that these same requirements follow directly in @xmath7 dimensions by imposing weyl symmetry on the general form in eq.([genweyl ] ) . as an example , it is easy to check that all these conditions are automatically satisfied by the action given in eq.([action1 ] ) , with one @xmath13 and four real fields in the doublet @xmath96 using the symbol @xmath97 as @xmath98 we write it in the form @xmath99 where @xmath100 is a flat minkowski metric in the 5-dimensional field space with a single negative eigenvalue - this reduces to a 2-dimensional @xmath101 in the unitary gauge of eq.([unitary ] ) since indeed there is only a single physical field @xmath102 . in our work @xcite-@xcite generally we let the potential @xmath103 to be an arbitrary function of its argument @xmath104 as allowed by the weyl symmetry conditions above . in eq.(1 ) we have a purely quartic renormalizable potential , @xmath105 . when quantum corrections are included , @xmath103 contains logarithmic corrections , but by re - examining the underlying symmetry , the effective quantum potential can again be rewritten in the form @xmath106 consistent with the local conformal symmetry . we use the quantum corrected potential in our discussion of higgs cosmology . next we give the general solution of the requirements above in a convenient parametrization for @xmath107 fields labelled with @xmath108 namely , @xmath109 where @xmath110 is distinguished , while @xmath111 with @xmath112 are all the other scalar fields . then the general solution to the conditions above for weyl gauge symmetry takes the following form @xmath113 where the ratio @xmath114 is gauge invariant . thus , after using the chain rule , @xmath115 the general weyl invariant action becomes @xmath116{c}\frac{1}{12}u\left ( \phi , s\right ) r\left ( g\right ) -v\left ( \phi , s\right ) \\ + \frac{1}{2}\left ( u - s^{i}\frac{\partial u}{\partial s^{i}}-s^{k}s^{l}g_{kl}\right ) \left ( \partial^{\mu}\ln\phi\right ) \left ( \partial_{\mu}\ln\phi\right ) \\ -\frac{1}{2}g_{ij}\left ( \partial^{\mu}s^{i}\partial_{\mu}s^{j}\right ) + \left ( 2g_{ij}s^{j}+\frac{\partial u}{\partial s^{i}}\right ) \partial^{\mu } s^{i}\partial_{\mu}\ln\phi \end{array } \right ) \label{weylgrav}\ ] ] with the @xmath117 in eqs.([sol - u]-[solg ] ) . it should be noted that @xmath118 are not determined by weyl symmetry alone . various other symmetries in a given model could restrict them . any choice consistent with additional symmetries is permitted in the construction of physical models . for example , the model in eq.([action1 ] ) has the su@xmath1su@xmath25u@xmath3 symmetry . in particular , local superconformal symmetry gives more severe restrictions by relating @xmath95 and @xmath119 from the beginning , as discussed below . as the number of scalar fields increases , the restrictions imposed by weyl symmetry become less severe . for example , an additional su@xmath120su@xmath6u@xmath3 gauge singlet field beyond @xmath13 that would be needed to reproduce the phenomenology of the dilaton - like singlet discussed in section ( [ gscaleinv ] ) can be included with slightly more freedom on its coupling parameters . keeping such possibilities in mind for more general phenomenological considerations , we the define the minimal model to include only the standard higgs doublet and the singlet @xmath13 as in the previous section . hence , for clarity we will write out the general weyl invariant lagrangian for the case of only the higgs doublet field @xmath14 plus @xmath121 this generalizes eq.([action1 ] ) , but we suppress the other fields in our discussion . furthermore , we will work directly with the gauge fixed version of the higgs field in eq.([unitary ] ) , so the higgs doublet is reduced to a single field @xmath79 as in ( [ unitary ] ) . in that case , from eqs.([sol - u]-[sol - g00 ] ) we obtain @xmath52{c}\frac{1}{12}\phi^{2}u\left ( s/\phi\right ) r\left ( g\right ) -\phi ^{4}f\left ( s/\phi\right ) \\ + g^{\mu\nu}\left ( \begin{array } [ c]{c}\frac{1}{2}\left ( u-\frac{s}{\phi}u^{\prime}-\frac{s^{2}}{\phi^{2}}\bar { g}\right ) \partial_{\mu}\phi\partial_{\nu}\phi\\ -\frac{1}{2}\bar{g}\partial_{\mu}s\partial_{\nu}s+\left ( u^{\prime}+2\frac { s}{\phi}\bar{g}\right ) \partial_{\mu}\phi\partial_{\nu}s \end{array } \right ) \end{array } \right ] \label{action2}\ ] ] where @xmath122 reduces to @xmath123 for the single field . generally , @xmath124 in eq.([action2 ] ) are three arbitrary functions of the higgs field , @xmath125 which may be used for model building . as an example , if we take @xmath126 and @xmath127 we obtain a relatively simple kinetic term with an arbitrary potential@xmath52{c}\frac{1}{12}\left ( \phi^{2}+\xi s^{2}\right ) r\left ( g\right ) -\phi ^{4}f\left ( s/\phi\right ) \\ + \left ( \begin{array } [ c]{c}\frac{1}{2}\left ( 1-\frac{s^{2}}{\phi^{2}}\left ( 1+\xi\right ) \right ) \partial_{\mu}\phi\partial^{\mu}\phi\\ -\frac{1}{2}\partial_{\mu}s\partial_{\nu}s+2\frac{s}{\phi}\left ( 1+\xi\right ) \partial_{\mu}\phi\partial^{\mu}s \end{array } \right ) \end{array } \right ] \label{xi}\ ] ] this lagrangian is weyl invariant for any value of the constant parameter @xmath128 but we see now a simple generalization of the special simplifying role played by @xmath129 that corresponds to the conformally coupled scalars of eq.(1 ) . in the general single-@xmath102 lagrangian ( [ action2 ] ) , field redefinitions of the gauge invariant variable @xmath130 ( or @xmath131 ) may be used to map one of these three functions , @xmath132 to any desired function of @xmath133 without changing the form of the lagrangian in ( [ action2 ] ) . for example it may be convenient to take @xmath127 and still obtain all single-@xmath102 weyl - symmetric models by using all possible @xmath134 . another option is to take all possible @xmath135 with a fixed @xmath136 as in eq.([action1 ] ) , which has certain advantages for understanding geodesic completeness @xcite of all the fields @xmath137 in cosmological spacetimes , as discussed in sec . if we choose @xmath138 and demand renormalizability of the action in the limit of @xmath139 ( where gravity effectively decouples , as seen in the gauge @xmath140 with @xmath141 , we are forced to @xmath142 and @xmath143 being a quartic polynomial . the last form is actually quite unique . we now give a proof that the general model of eq.([action2 ] ) with a single @xmath102 ( which corresponds to our minimal realistic model ) can always be written in the geodesically complete form that we advocated in our previous work @xcite-@xcite and implemented in proposing the action in eq.([action1 ] ) . we begin in the notation of eq.([genweyl ] ) with only two fields @xmath144 labelled by @xmath145 using well known results of geometry in 2 dimensions , the metric in field space @xmath146 can always be diagonalized by general field reparametrizations ( as in general relativity ) and put into the conformally flat form @xmath147 where @xmath100 is the flat metric in 2 dimensions . using further field reparametrizations with an overall rescaling consistent with weyl transformations , the factor @xmath148 can also be set equal to the constant @xmath149 having arrived at @xmath150 as still the most general metric in field space , the only possible form of @xmath88 that is consistent with the local weyl symmetry is @xmath151 this completes the general proof that the action in eq.([action1 ] ) is the most general weyl invariant theory without losing generality . then the complete analysis of homogeneous cosmological solutions provided in our previous work @xcite-@xcite shows that this form is the general geodesically complete version of the theory . certain other forms for @xmath88 arrived at by field reparametrizations , in particular purely positive forms of @xmath152 always end up putting restrictions in field space inadvertently , and such specialized restrictions on fields is what leads to geodesic incompleteness . hence , the a class of well motivated models that we used in many of our studies amounts to @xmath153 @xmath154 and the low energy renormalizable quartic potential @xmath155 @xmath156 with @xmath157 this is the model that now is identical to the conformally symmetric standard model we proposed . its homogeneous cosmological equations have been solved analytically exactly in @xcite for all possible initial conditions of the fields , including radiation and curvature , and all possible values of the parameters @xmath158 but with @xmath159 . after this much analytic control , a small @xmath27 is easily handled with numerical methods and still have a full understanding of all the cosmological solutions of the standard model . this forms the basis of our further work in cosmology that is discussed in the following sections . for further discussion , a useful gauge choice is @xmath160 for all spacetime as in eq.([c - gauge ] ) . this is the gauge called the @xmath15-gauge in our previous work . because we use also other gauges , we attach the letter @xmath15 to each field in this gauge , thus @xmath161 will remind us that we are in the @xmath15-gauge , where @xmath162 in @xmath15-gauge we rename @xmath163 to recall that in this gauge we obtain the simplest connection to the higgs field @xmath164 at low energy , in nearly flat spacetime , @xmath165 as discussed in the paragraphs before eq.([unitary ] ) . the lagrangian in eq.([action2 ] ) with general @xmath166 takes the following greatly simplified form in the @xmath15-gauge @xmath167 . \label{action3}\ ] ] if we apply a weyl transformation to go to the einstein frame after an integration by parts of the last term ( or by dropping a total derivative ) we obtain the given result . ] , @xmath168 then we obtain@xmath169 \label{e - frame}\ ] ] without loss of generality , a particularly simplifying choice for @xmath170 , namely @xmath171 is the one that yields a canonically normalized higgs field . in this form , the theory can be interpreted directly as an einstein frame formulation of the standard model with a canonically normalized higgs field it is possible to do a field redefinition such that @xmath172 is written in terms of a canonically normalized field @xmath173 where the relation between @xmath174 and @xmath175 is given by the first order differential equation @xmath176=u\left ( h/\phi_{0}\right ) .$ ] then the lagrangian rewritten in terms of the canonically normalized @xmath174 is again of standard form with the same potential @xmath177 as eq.([veff ] ) , except for expressing @xmath172 in terms of @xmath178 this complicated procedure is avoided by the simple choice of @xmath63 in eq.([gchosen ] ) . [ fieldredef ] ] and an effective higgs potential @xmath179 given by @xmath180 this potential can be crafted by various choices of @xmath103 and @xmath181 to fit cosmological observations , while still being consistent with an underlying local conformal symmetry . however , since this gauge led to an overall positive coefficient in front of the ricci scalar , it could not recover all field configurations that are demanded in the geodesically complete theory ; inevitably the _ generic _ solutions in this gauge are geodesically incomplete as we have learned in our previous work . for example , the best fit inflaton potentials based on recent planck satellite data @xcite are plateau " models @xcite . using the geodesically incomplete form of the theory in ( [ e - frame ] ) it is possible to construct examples with plateaus at large @xmath182 , by choosing @xmath183 and @xmath184 such that the potential @xmath185 is very slowly varying and approaches a constant at large @xmath172 and by requiring @xmath179 becomes approximately the familiar higgs potential needed to fit low energy physics for @xmath186 . the bezrukov - shaposhnikov ( bs ) model for higgs inflation @xcite is a special case of our more general form in eq.([action3 ] ) , namely their proposal@xmath187 , \label{shoposh}\ ] ] follows from eq.([action3 ] ) by taking @xmath188and @xmath189 this leads to an effective potential @xmath190 of the type above in eq.([veff ] ) , @xmath191 in the bs model , because @xmath63 is chosen as @xmath192 , the field @xmath172 is not canonically normalized , so @xmath172 must be replaced through a field redefinition @xcite to obtain a canonically normalized higgs , @xmath173 as described more generally in footnote ( [ fieldredef ] ) . hence , the bs model has a fully weyl symmetric formulation which was not noticed before . its presentation in the literature has included various ambiguities and inconsistencies , with clashing ideas on scaling symmetries at the classical level and quantum corrections . for example , in some cases , the coupling to gravity has a dimensionful newton constant that is inconsistent with the scaling symmetry of the rest of the theory ; a massless dilaton is said to exist in cases where global scaling symmetry is explicitly broken at low energies ; unimodular gravity has been introduced but this is inconsistent with scale symmetry ; and there is ambiguity about which renormalization scheme is appropriate for computing quantum corrections . these issues are fully resolved with the underlying weyl symmetric formulation discussed here . also , now that we have recast the bs model into a fully conformally invariant form , we can see in the gauge fixed einstein frame ( [ e - frame]-[veff ] ) that it is not unique . rather , it is just a special case of a larger set of conformally invariant models including a range that have similar plateau properties . furthermore , since @xmath193 is purely positive , it means the fields in the bs model form a basis that is geodesically incomplete , and hence its generic solutions can describe only a subsector of the available field space . we make further comments in sec . ( [ confcosm ] ) including the effects of geodesic incompleteness . we do not know if supersymmetry ( susy ) is a property of nature , but theoretically it is an attractive possibility . therefore , it is of interest to investigate whether it is compatible with an underlying local conformal symmetry . as a superconformal local symmetry , the generalization of the weyl invariant formalism of the previous section produces stronger constraints on scalar fields . this was derived in @xcite from the gauge symmetry formalism in @xmath8 dimensions . it is possible to arrive at the results given below directly in @xmath7 dimensions by requiring supergravity with a local superconformal symmetry , but provided the usual einstein - hilbert term is dropped ( which is unusual in the supergravity literature ) , and instead a weyl symmetric formulation like the previous sections is implemented . in the @xmath8 dimensional approach of 2t - physics there is no option : it is a prediction that the emergent @xmath7 dimensional theory is automatically invariant under a local symmetry su@xmath194 where su@xmath195so@xmath19 is the connection to @xmath8 dimensions , the subgroup so@xmath196so@xmath197so@xmath198so@xmath199 is the weyl subgroup that acts on the extra @xmath60 dimensions , and the supersymmetrization in @xmath8 dimensions promotes su@xmath200 to su@xmath201 for @xmath202 supersymmetries @xcite , with @xmath203 in the present case @xcite . the scalar - field sector of the emergent @xmath7 dimensional locally superconformal theory is presented here briefly in a streamlined fashion , including some simplifications and extensions . first note that a scalar field in a chiral supermultiplet must be complex . thus susy requires the singlet @xmath13 to be complexified so that it is a member of a supermultiplet . then to describe all the scalar fields we use a complex basis @xmath204 and denote their complex conjugates as @xmath205 with the barred index @xmath206 the notation is reminiscent of standard supergravity as reviewed in @xcite . however now there is no einstein - hilbert term ; instead there is a weyl plus an additional @xmath207 gauge symmetry that can gauge fix a complex scalar @xmath13 into a dimensionful real constant @xmath32 that plays the role of the planck scale . recalling the discussion about removing the ( real ) ghost scalar field @xmath13 in sec . iib , it should be noticed that the complex @xmath13 amounts to two real ghosts , and therefore two gauge symmetries are needed to remove them . this role is played by the weyl symmetry so@xmath208 that acts on the extra 1 + 1 dimensions and the local r - symmetry u@xmath3 , both of which are included in the local su@xmath209 outlined above . the local u@xmath3 is a crucial ingredient as a partner of the weyl symmetry to remove the additional ghost from a complexified dilaton field @xmath13 that is demanded by susy . like before , we have @xmath210 but susy requires that the metric @xmath211 be derived from the derivatives of @xmath212 like a khler metric@xmath213 the metric must be non - singular and @xmath214 can not contain any more than one negative eigenvalue ( because no more than one complex ghost can be removed ) . we denote its inverse formally as @xmath215 a simple quadratic example similar to eq.([action1 ] ) is @xmath216 in this example , all scalars in the theory are conformally coupled , and u and g are automatically invariant under a global su(1,n ) symmetry . this global symmetry , which will continue to be a hidden symmetry in some gauges , may be broken explicitly by some terms in the potential @xmath217 , depending on the model considered . nevertheless , keeping track of this ( broken ) symmetry in physical applications can be useful . the scalar field sector of the supergravity theory with local weyl symmetry is given as follows . @xmath218 .\ ] ] the khler metric and khler potential structure is reminiscent of general supergravity @xcite , however the absence of the einstein - hilbert term , and the corresponding scale invariance is the important difference . here , @xmath219 is homogeneous of degree two and satisfies homothety constraints similar to eq.([homothety]),@xmath220 the potential energy has two parts @xmath221 the potential @xmath222 must be derived from an analytic superpotential @xmath223 that depends only on @xmath204 and not on @xmath205 , and is homogeneous of degree three so that @xmath224 is homogeneous of degree four @xmath225 the potential @xmath226 is derived from another independent analytic function @xmath227 with adjoint group indices @xmath228 that appears in yang - mills terms as , @xmath229 . this @xmath227 should be homogeneous of degree zero @xmath230 then @xmath231 takes the form @xmath232 where @xmath233 is the matrix representation of the yang - mills group as it acts on the scalars @xmath234 the fermionic terms are added consistently with the usual rules of supergravity @xcite . this is the general setup for the weyl invariant matter coupled to weyl invariant supergravity as derived from the @xmath8 dimensional theory . the simplest example that corresponds to the supersymmetric generalization of eq.([action1 ] ) would be the minimal supersymmetric standard model extended with an additional singlet supermultiplet , whose scalar component is the complex field @xmath121 then @xmath119 takes the quadratic form suggested in @xcite @xmath235 where @xmath236 are the two higgs doublets needed in the supersymmetric version of the standard model . the superpotential @xmath237 and the matrix @xmath238 are chosen to fit the usual practice of supersymmetric model building @xcite . an example is @xmath239 , and @xmath240 where @xmath241 is the only su@xmath6u@xmath3 invariant which is analytic in both @xmath242 and @xmath243 . in an effective ( rather than renormalizable ) low energy theory , these can be modified by replacing the dimensionless coupling constants @xmath244 by an arbitrary function of the ratio @xmath245 and similarly for @xmath246 when the complex @xmath0 is gauge fixed to the real constant @xmath247 this approach generates all the dimensionful parameters from the same source @xmath32 which is of order of the planck scale . then we see that the modification of @xmath244 and @xmath248 by arbitrary functions of @xmath249 is negligible at low energies . more general scale invariant models with more fields ( such as generalizations of the minimal susy model ) are easily constructed by using the rules on scalar fields given in this section . with special forms of the superpotential @xmath223 , it is possible to construct so called no - scale models @xcite-@xcite in which the cosmological constant is guaranteed to be zero at the classical level even after spontaneous breakdown of symmetries . simple examples of no - scale models , that are lifted to be fully weyl invariant , are given in @xcite . in this paper we will not explore any further the general superconformal structures discussed above although this could be of interest in some future applications . some of our examples that are very similar to the weyl invariant supersymmetric cases previously discovered in @xcite were later explored in a cosmological context in @xcite . however their discussion , which is focussed on very specialized initial conditions that are so non - generic , is difficult to be convincing especially in the face of the complete set of solutions that we now understand much better . in this paper we have emphasized the geodesically complete nature of our conformal standard model . in this section we first describe what we mean by geodesic completeness and the related notion of completeness in field space , and then discuss some applications of our conformal standard model in cosmology . there are two required properties for geodesic completeness : the first is smooth local geodesic continuation through singularities across all space - time patches ; the second is infinite action for geodesics that reach arbitrarily far into the past so that unnatural initial conditions are avoided . both properties are satisfied in our theory which is insured to also be complete in field space . geodesics @xmath250 are computed by extremizing the particle action @xmath251 where @xmath252 is the mass and @xmath253 is a gravitational field . in our weyl invariant standard model ( [ action1 ] ) all particle masses are generated by the higgs field , so @xmath252 is proportional to the higgs field @xmath172 , @xmath254 where @xmath255 is the dimensionless coupling constant for the corresponding particle as prescribed by the standard model . then this particle action is locally invariant under weyl transformations @xmath256 and @xmath257 for our cosmological discussion , geodesics @xmath258 in a spatially homogeneous space are computed when both homogeneous fields that appear in the particle action @xmath259 , @xmath260 and @xmath261 where @xmath262 is the conformal time , correspond to consistent solutions of the field equations of our conformal standard model in ( [ action1 ] ) . the full solution to the geodesic equation ( neglecting curvature and anisotropy for simplicity ) is @xmath263 where @xmath264 is the conserved spatial component of the particle momentum vector , @xmath265 while @xmath266 , and @xmath267 is the initial position . these expressions are invariant under reparametrizations of the affine parameter @xmath268 since they depend only on the physical spacetime variables @xmath269 then , the local continuation property of a geodesic @xmath270 in a homogeneous space is satisfied automatically when one insures that all the fields , including @xmath271 and @xmath272 are given in all the patches of homogeneous field space and that the fields are smoothly continued through singularities . in @xcite@xcite , we showed how local weyl invariance , in the special weyl frame displayed in eq.([action1 ] ) that covers all the patches , are essential ingredients to insure that all homogeneous cosmological solutions of our conformal standard model ( [ action1 ] ) have this continuation property . the global action for a geodesic ( [ geodesicx ] ) is given by @xmath273 the combination @xmath274 that appears in @xmath275 is invariant under weyl transformations , so it can be computed in any convenient gauge ( see e.g. eqs.([egaugeu]-[detxu ] ) ) . in particular in the @xmath276-gauge used in the solutions in @xcite@xcite , where @xmath277 we have shown that in all cyclic solutions in @xcite , @xmath278 oscillates with a fixed maximum amplitude many times within a cycle , and for an infinite number of times over an infinite number of cycles , leading to an infinite action , since @xmath279 receives the same positive finite contribution in each cycle . for these solutions the average magnitude of @xmath280@xmath281@xmath280 in the @xmath282-gauge ( usual einstein frame , see eq.([egaugeu ] ) ) is the same in each cycle . recently , we have also constructed @xcite solutions with increasing entropy in each future cycle such that the average @xmath280@xmath281@xmath280 in the @xmath282-gauge increases in each future cycle . for these solutions , @xmath275 diverges to the past @xcite even faster as compared to our solutions in @xcite . hence , unlike inflation scenarios @xcite , we may expect that the universe described by our conformal standard model has no unnatural initial conditions . next , we discuss various potential applications of weyl invariant theories to cosmology . early universe cosmology near the singularity is a natural place to look for uses because the interactions imposed by weyl invariance ( more profoundly @xmath8 dimensional gauge symmetries ) produce their most significant changes relative to conventional physics in the limit of strong gravity , as noted in some of the examples above . black hole phenomena is another interesting arena for study , but we will not discuss them here . the first important application was to use weyl invariance to construct geodesically complete cosmological solutions . conventional cosmological analysis is usually confined to theories coupled to einstein gravity with a dimensionful newton s constant . in this framework , all cosmological solutions of interest are geodesically incomplete . however , we have observed that a weyl - invariant theory can be cast in einstein frame , as illustrated by eq.([e - frame ] ) , and in other frames . the observation made in @xcite-@xcite is that geodesically incompleteness is an artifact of an unsuitable frame choice : geodesically incomplete solutions in einstein frame may be completed in other frames , even though the theories are entirely equivalent away from the singularity . the einstein frame can always be reached directly from a weyl invariant theory in ( [ action2 ] ) by making the einstein - gauge choice @xmath283 we label this the @xmath282-gauge and mark the fields in this gauge with the subscript @xmath282 ( as in @xmath284 ) to distinguish them from the @xmath15- gauge fields @xmath285 note that the @xmath282-gauge can be valid only in a patch in field space since restrictions must be imposed on the fields to require that @xmath286 is positive . the same theory can easily be transformed to other gauges . for example , the relations between the @xmath282- and @xmath15-gauge fields can be easily derived by considering the weyl gauge invariants such as @xmath287 and @xmath288 @xmath289 etc . , for example we deduce @xmath290 from these , we can express @xmath291 and @xmath292 in terms of the single field @xmath293 so that the gauge condition ( [ egaugeu ] ) is satisfied . inserting these expressions into the gauge invariant action ( [ action2 ] ) we arrive at the same @xmath282-frame action as eq.([e - frame ] ) . as argued in our work @xcite-@xcite , classically geodesically complete solutions can be obtained for all single - scalar theories that can be cast in the form of eq.([action2 ] ) . but , for all patches of field space @xmath294 to be included , as demanded by the geodesics derived from @xmath179 in the einstein frame , @xmath295 must be brought to an appropriate form by using the field redefinitions discussed below eq.([xi ] ) . the patches of field space @xmath294 that are missing in the einstein frame can then be added in order to obtain a geodesically complete space . we have argued in @xcite , and provided a proof in section ( [ generalgrav ] ) , that geodesic completeness is accomplished when we bring @xmath119 to the form @xmath296 where @xmath297 is the sum of all scalars other than @xmath121 geodesics remain incomplete when they hit the singularity in all frames in which @xmath119 is always positive , or the equivalent einstein frame , such as the one in eq.([shoposh ] ) . when we rewrite those , by field redefinitions , in terms of new fields in which @xmath296 then the patches of field space @xmath298 are equivalent to the einstein frame , or to the other frames with positive @xmath119 . geodesic completion is achieved by allowing all regions of field space in the parametrization that has @xmath299 for this form , @xmath119 is allowed to smoothly go negative . in these coordinates the gauge invariant vanishing point of @xmath300 given by the gauge invariant expression @xmath301 has a special significance ; it corresponds to the singularity of the scale factor of the universe in the einstein gauge . this is seen by equating the gauge invariant @xmath302 in the einstein gauge , in which @xmath303 and the unimodular gauge ( labelled with @xmath276 ) , in which @xmath304 as follows @xmath305 at spacetime points or regions where @xmath306 which is where the gauge invariant quantity @xmath307 hits unity in all gauges , @xmath308 @xmath309 the geometry in the einstein gauge fails completely since @xmath310 , and this is how the cosmological singularity occurs at some point in time @xcite . then , as we can see in the @xmath15-gauge , the region @xmath311 ( higgs of planck size ) is the region of the big crunch or big bang singularity , where @xmath312 in any gauge . the behavior of the universe in this region is governed by a universal attractor mechanism that is independent of the scalar potential , therefore independent of the details of the model @xcite . although in the classical theory , cosmological singularities of frw - type are typically resolved in the weyl - lifted theory with @xmath313 in suitable weyl gauges , one should still worry about quantum gravity corrections . when @xmath119 vanishes , the coefficient of the ricci scalar in the gravitational action vanishes so there is no suppression of metric fluctuations and quantum gravity corrections should become large . however , it is notable that for certain types of cosmic singularities , including realistic ones , the metric and fields possess a unique continuation around the singularity in the complex time plane . a complex time path can be chosen to remain far from the singularity so that @xmath119 is always large and gravity remains weak . thus , we are able to find an analytic continuation of our classical solutions connecting big crunches to big bangs along which quantum gravity corrections are small . this issue is under active investigation and we defer further discussion to future work . a particularly interesting application of weyl - invariance is to higgs cosmology , which the geodesically complete solutions show can be much more interesting than conventionally assumed . in fact , there is the possibility that the higgs field alone may be sufficient to explain the large - scale features of the universe , as suggested by the bezrukov - shaposhnikov higgs inflation model @xcite and our recent work on the higgs in cyclic cosmology @xcite@xcite . in considering higgs cosmology or other models of the early universe , we believe that the cosmological solutions found for geodesically complete theories provide some important new insights on the question of the likely initial conditions just after the big bang . for example , the _ generic _ behavior of the higgs after the bang when @xmath314 is dramatically different than the contrived initial conditions that are commonly assumed in higgs inflation scenarios . this is easily seen by examining the analytic work in @xcite@xcite , where for the simple model in eq.([action1 ] ) with @xmath315 all homogeneous solutions including curvature and radiation are obtained and the effects of anisotropy are determined @xcite . emphasizing that this is not only the contrived solutions , but all solutions , it serves as an example of the richness of the phenomena that occur in a model even with a simple potential . the results make clear a point that is obvious , but often forgotten : for a given scalar potential , there is an enormous range of cosmological solutions . by comparison , it is clear that the slow - roll initial conditions frequently assumed in the analysis of cosmic inflation are very special and unlikely . for example , by transforming the inflationary solution in the bezrukov - shaposhnikov higgs inflation model @xcite to a field basis in which @xmath313 , it may be possible to trace cosmic evolution right back to the singularity and to judge whether the inflation is likely in the space of geodesically complete cosmological solutions . fig.(1 ) is an illustration of our solution for the _ generic _ cosmological behavior of the higgs field just after the big bang if the higgs potential has a stable non - trivial minimum , as in eq.([action1 ] ) , as usually assumed and as required for the bezrukov - shaposhnikov higgs inflation model @xcite . this figure describes the generic cosmological evolution of the higgs field , that must start with fluctuations of planck size and energy ( due to the universal attractor near the singularity @xcite ) , and quickly reduce its amplitude by losing energy to the gravitational field ; then after a phase transition , settle down to an almost constant value at the electroweak scale @xmath34 determined by the dimensionless parameter @xmath27 in eq.([alpha ] ) . cosmohiggs-stable.eps + fig.(1 ) - after the big bang the higgs field oscillates initially around zero with a large amplitude of the order of the planck scale . it slowly loses energy to the gravitational field , causing its amplitude to diminish . as it approaches the time or energies of the electroweak scale , it undergoes the phase transition seen in the figure , and then slowly settles to a constant vacuum value @xmath34 at a stable minimum of the potential . [ cosmohiggs-stable.eps ] the solution of fig.(1 ) changes drastically if the vacuum is metastable after including quantum corrections , which is a possibility suggested by the most recent lhc data for the higgs and the top quark masses @xcite , and assuming no new physics up to the planck scale . metastability is incompatible with higgs inflation and generally causes problems for inflation because the higgs will typically escape from the metastable phase right after the big bang and roll to a state of negative energy density that can prevent inflation of any sort from occurring . the exact solutions of the weyl - invariant theory suggest an alternative cosmology in this case . the generic solution at first behaves as in fig.(1 ) after the big bang , all the way through the electroweak phase transition . but after some time ( order of the lifetime of the universe ) the higgs oscillations in the electroweak vacuum grow larger and larger , like the mirror image of fig.(1 ) , taking away energy from the gravitational field and eventually causing a collapse of the universe to a big crunch , while the higgs does a quantum tunneling to a lower state of the potential . at that stage our exact analysis near the singularity given in @xcite takes over to describe interesting new phenomena that occur just after the crunch and before another rebirth of the universe with a big bang . the result is a regularly repeating sequence in which the higgs field is trapped in its metastable state after a big bang , remains there for a long period of expansion followed by contraction , escapes as the universe approaches the big crunch , passes through to a big bang and becomes trapped again . the evolution can be considered a higgs - driven cyclic theory of the universe . the details are presented in a separate paper @xcite . this research was partially supported by the u.s . department of energy under grant number de - fg03 - 84er40168 ( ib ) and under grant number de - fg02- 91er40671 ( pjs ) . research at perimeter institute is supported by the government of canada through industry canada and by the province of ontario through the ministry of research and innovation . the work of nt is supported in part by a grant from the john templeton foundation . the opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the john templeton foundation . ib thanks cern and pjs thanks perimeter institute for hospitality while this research was completed . f. englert , e. gunzig , c. truffin , p. windey , conformal invariant general relativity with dynamical symmetry breakdown , phys.lett . b57 ( 1975 ) 73 ; f. englert , c. truffin and r. gastmans , conformal invariance in quantum gravity , nuclear physics * b117 * ( 1976 ) 407 . i. bars , s .- h . chen , p. j. steinhardt and n. turok , complete set of homogeneous isotropic analytic solutions in scalar - tensor cosmology with radiation and curvature , phys . * d86 * ( 2012 ) 083542 [ arxiv:1207.1940 ] . i. bars , traversing cosmological singularities , complete journeys through spacetime including antigravity , arxiv:1209.1068 , in _ beyond the big bang , competing scenarios for an eternal universe _ , ed . rudy vaas , springer [ 2013 ] , isbn 978 - 3 - 540 - 71422 - 4 . i. bars , gravity in 2t - physics , phys . * d77 * ( 2008 ) 125027 [ arxiv:0804.1585 ] ; i. bars , s. h. chen geometry and symmetry structures in 2 t gravity , phys . rev . * d79 * ( 2009 ) 085021 [ arxiv:0811.2510v2 ] . i. bars and y - c kuo , field theory in 2t - physics with @xmath202=1 supersymmetry , phys . * 99 * ( 2007 ) 041801 [ hep - th/0702089 ] ; supersymmetric field theory in 2t - physics , phys . rev . * d76 * ( 2007 ) 105028 [ hep - th/0703002 ] ; @xmath202=2,4 supersymmetric gauge field theory in 2t - physics , phys . rev . * d79 * ( 2009 ) 025001 [ arxiv:0808.0537 ] ; super yang - mills theory in 10 + 2 dimensions , the 2t - physics source for n=4 sym and m(atrix ) theory , arxiv:1008.4761 . k. meissner and h. nicolai , conformal symmetry and the standard model , phys . lett . * b648 * ( 2007 ) 312 , [ arxiv : hep - th/0612165 ] . effective action , conformal anomaly and the issue of quadratic divergences , phys.lett . * b660 * ( 2008 ) 260 [ arxiv:0710.2840 ] . itzhak bars , at conference on conformal nature of the universe , perimeter institute , may 2012 ; neil turok , cern colloquium , july 2012 ; t.w.b . kibble 80th birthday conference , imperial college , march 2013 ; yale physics department colloquium , april 2013 ; paul steinhardt , at davis conference , may 2013 . s. ferrara , r. kallosh , a. linde , a. marrani and a. van proeyen , superconformal symmetry , nmssm , and inflation , phys . * d83 * ( 2011 ) 025008 [ arxiv:1008.2942 ] ; r. kallosh and a. linde superconformal generalization of the chaotic inflation model arxiv:1306.3211 and superconformal generalizations of the starobinsky model , arxiv:1306.3214 . i. bars , s - h chen and g. quelin , dual field theories in @xmath316 emergent spacetimes from a unifying field theory in d+2 spacetime , phys . rev . * d76 * ( 2007 ) 065016 [ arxiv:0705.2834 ] ; i. bars and g. quelin , dualities among one - time field theories with spin , emerging from a unifying two- time field theory , phys . rev . * d77 * ( 2008 ) 125019 [ arxiv:0802.1947 ] . r. foot , a. kobakhidze and r. r. volkas , electroweak higgs as a pseudo - goldstone boson of broken scale invariance , phys . lett . * b655 * ( 2007 ) 156 [ arxiv:0704.1165 ] . r. foot , a. kobakhidze , k. mcdonald , r.volkas , r. foot , a. kobakhidze , k. mcdonald , r.volkas , neutrino mass in radiatively - broken scale - invariant models , arxiv:0706.1829 ; a solution to the hierarchy problem from an almost decoupled hidden sector within a classically scale invariant theory , arxiv:0709.2750 . h. cheng , the possible existence of weyl s vector meson , phys . rev . lett . * 61 * ( 1988 ) 2182 ; h. nishino and s. rajpoot , broken scale invariance in the standard model , hep - th/0403039 ; _ ibid . _ class . & gr . * 28 * ( 2011 ) 145014 ; aip conf . proc . * 881 * ( 2007 ) p. 82

the first data system requiring dark energy ( de ) concerned cosmic acceleration , detected through high - redshift supernovae @xcite . cmb @xcite and deep sample @xcite data supported also the de case , showing that the density parameter for non relativistic matter @xmath1 , while the total density parameter @xmath2 in the most popular scenario , de is ascribed to a cosmological constant @xmath3 . alternative options include a self interacting scalar field , @xmath4 ( quintessence or dynamical de @xcite ) and modifications of general relativity @xcite . it is known that models with @xmath3 ( @xmath5cdm ) apparently accommodate all available data systems . the problem is the physical origin of @xmath6 which can be a false vacuum ; this however causes well known _ fine tuning _ and _ coincidence _ problems . the former problem is partially eased in dynamical de ( dde ) scenarios , when self interaction is due to a _ tracking _ potential @xmath7 @xcite . if @xmath7 is sugra @xcite , the fit with data is at least as good as for @xmath3cdm @xcite . in the attempt to ease the coincidence problem , dm de interaction ( e.g. , @xcite ) was also considered , yielding an energy transfer between the dark components , so allowing a ( quasi)parallel scaling of dm and de from a fairly high redshift until the present . while laboratory data set no significant constraint on dm de interaction strength , parametrized by @xmath8 ( see below ) , recent works placed constraints on possible couplings , by using snia data @xcite or the redshift evolution of the hubble parameter , @xmath9 @xcite . accordingly , @xmath100.15 @xcite is hardly consistent with observations . unfortunately , such a low coupling level no longer eases the coincidence problem @xcite , but , once the genie has come out from the lamp , it is hard to put it back inside . the point is whether low values of @xmath8 , as allowed by current data , can interfere with future data analysis . in particular , when we allow for non zero @xmath8 , how do errors on other parameters behave ? in this work we tried to answer this question by using a fisher matrix technique . we considered two different models , set by similar values of cosmological parameters , without and with coupling . in the latter case , we took @xmath11 starting from these models , we evaluated the expected errors on cosmological parameters , as obtained when data concern just cmb anisotropy and polarization or include tomographic weak lensing ( wl ) . as a matter of fact , in coupled models , the time evolution of the dark components is non standard . if such models are considered in a newtonian approximation , it is as though dm particles had a @xmath4dependent mass . also for quite low @xmath8 s , this anomalous scaling leaves an imprint on both the expansion history of the universe , and the growth of ( matter ) fluctuations , at the linear and non linear levels ( e.g. @xcite ) . however , any detected evolution of @xmath9 can be reproduced through a suitable redshift dependence of de density @xmath12 and state parameter @xmath13 , when @xmath4 approaches @xmath14 ( the planck mass ) . a risk is that , if matter and dark energy are coupled , fitting observations leads to an estimate of a phantom equation of state ( @xmath15 ) , even if @xmath16 at all redshifts @xcite . in principle , this risk can be excluded if the redshift dependence of the growth factor @xmath17 is also tested , through the increase in number and concentration of bound systems . data providing information both on @xmath18 and @xmath17 are therefore able to discriminate between coupled and uncoupled models . experiments , or combinations of experiments , probing @xmath18 and @xmath17 are then needed . cmb data , used to constrain coupling @xcite , place only upper limits on @xmath8 . the analysis of ly@xmath19 and the matter power spectrum of the 2df and sdss surveys @xcite does not lead to great improvements . at the available sensitivity level , such data systems provide just weighted integrals of @xmath18 and @xmath17 , which remain consistent with a rather wide set of options . on the contrary , gravitational lensing , alone or in combination with cmb data , was already shown to be a powerful tool for the analysis of de . wl tomography probes the power spectrum @xmath20 at different redshifts and is thus well suited to constrain @xmath21 in this work we aim to put these conceptual points on a more quantitative basis and to deepen the case of coupling , by performing a fisher analysis of future wl surveys and cmb experiments . the outline of this work is as follow . in sec . [ sec : def ] we review the basic properties and definitions of dde models and wl , in sec . [ sec : fore ] we show the results of the fisher analysis , in sec . [ sec : discu ] we discuss them and in sec . [ sec : summ ] we summarize our findings and draw our conclusions . we consider a cosmological model where the de field @xmath4 interacts with the cold dm component . the model requires the specification of the potential @xmath7 and the function @xmath22 characterizing the coupling . the equation of motion for @xmath4 then reads @xmath23 here dots denote ordinary time differentiation , @xmath24 and @xmath25 is dm energy density . in turn , its evolution is governed by @xmath26 this equation can be integrated and gives : @xmath27 for @xmath28 eqs . ( [ eq1 ] ) , ( [ eq : rho ] ) return ordinary dde equations . the equations for the other components remain unchanged . in a generic coupled model , then , the ratio between the energy densities of cold dm and baryons is not fixed , but evolves in time according to @xmath22 . however , it is always possible to define an effective de component of density @xmath29 + \rho_{de } ( a)~.\ ] ] in general , @xmath30 is not guaranteed to be positive and detecting @xmath31 would be a clear indication that our description of the dark sector is not adequate . lacking such clear giveaway , however , experiments probing @xmath18 can hardly discriminate between de dm interaction and an _ ad hoc _ de component . if its effective state parameter @xmath32 is @xmath33 , data may appear to require phantom de ( see @xcite for discussion ) . coupling affects also fluctuation growth . in the newtonian limit , _ i.e. _ well below the horizon , and neglecting the contribution of radiation , baryons and dm fluctuations grow according to the equations ( e.g. @xcite ) : @xmath34 , \end{array } \label{eq : delta}\end{aligned}\ ] ] where we defined @xmath8 so that @xmath35 while @xmath36 . therefore , baryons and dm perturbations grow at different rates and , even soon after recombination , a growing mode @xmath37 no longer exists ( see fig . [ fig : lingrowth ] ) , leading to a bias between baryon and dm perturbations . analytical models of spherical collapse have shown that this differential growth results in a baryon dm segregation , with baryons occupying the outer regions of collapsed objects @xcite . eqs ( [ eq : delta ] ) also show that the growth equations explicitly depend on both @xmath9 and @xmath25 ; therefore , if we measure the growth from data , a possible anomalous scaling can no longer be masked though an _ ad hoc _ definition of an effective de density . it is then licit to conclude that experiments probing the rate of growth of fluctuations are in principle well suited to test coupling between the dark components . the growth suppression rate @xmath38 for cdm ( solid line ) and baryons ( dashed line ) in a coupled model . curves refer to a sugra potential with exponential coupling . it can be noticed that the cold dark matter evolution rapidly diverges from the standard cold dark matter solution , @xmath39 ( dotted line ) , even well into the matter dominated era . for comparison , we also plot @xmath40 for the same sugra model with coupling turned off ( dot dashed line).,width=377 ] in this work we are interested to combining wl and cmb data , which can not be accurately described using current parametrisations @xcite . therefore , we follow a more conventional approach and choose the functional forms @xmath41 the sugra @xcite potential @xmath7 depends on the slope @xmath19 and the energy scale @xmath3 . fixing de density today and @xmath5 ( @xmath19 ) , however , determines a unique value of @xmath19 ( @xmath3 ) . cmb , snia and deep sample data yield @xmath42gev @xcite , in the absence of coupling . here we focus on the sugra potential as it naturally arises in the context of supergravity theories and is an example of tracking potential characterized by a rapid time variation of the equation of state , when de becomes dominant . therefore , assuming a constant @xmath43 for this class of potential may lead to misleading results . different choices for the potential are clearly possible . the coupling function @xmath22 depends on @xmath8 , and @xmath44 is the field value today . in this work we assume a constant @xmath45 ( see however @xcite for a different approach ) ; data place the upper limit @xmath46 @xcite . for reasonable values of the cosmological parameters and of @xmath3 , we expect coupling effects not to be relevant for @xmath47 , so that the dynamically interesting values for the coupling lies in the range @xmath48 . among the cosmological probes allowing the analysis of the nature of de , the cosmological wl has been earning a fundamental role ( see @xcite for a thorough review ) . in fact , next generation wl surveys will cover a significant fraction of the sky and observe galaxies at deeper redshift , making wl a powerful tool to study the properties of the universe . furthermore , wl tomography will allow to significantly increase the cosmological information that can be recovered from such surveys ( see @xcite and @xcite for details about lensing tomography ) . the power spectrum for the wl convergence between the @xmath49th and @xmath50th redshift bin is given by : @xmath51 here @xmath52 is the non linear matter power spectrum , at the redshift @xmath53 and wave number @xmath54 ; the quantity @xmath55 is the radial comoving distance between @xmath53 and @xmath56 , @xmath57 while @xmath58 is the window function which weights different redshift bins according to the factor @xmath59 , @xmath60 being @xmath61 the i th redshift bin . the function @xmath62 is the normalized redshift distribution of the source galaxies falling in the @xmath49-th photometric redshift bin , per unit solid angle , @xmath63^{-1},\end{aligned}\ ] ] where @xmath64 the last expression codifies the effect of errors in photometric redshifts of source galaxies @xcite . in fact , due to the large number of observed galaxies with future surveys , one needs to refer to their photometric redshifts , even if the determination of these redshifts may be not so accurate as the spectroscopic ones . in the model given by , the mapping between the photometric @xmath65 and the spectroscopic redshift @xmath53 is obtained convolving the overall galaxy distribution per unit solid angle , @xmath66 , with a probability distribution @xmath67 in @xmath65 at a given @xmath53 . we choose a gaussian function at each redshift for the distribution of photometric redshifts , i.e. @xmath68}\end{aligned}\ ] ] while the overall distribution of source galaxies is chosen to have the parametrized form : @xmath69 } , \label{eq : distro}\ ] ] where @xmath70 , @xmath71 and @xmath72 are the parameters . one can easily check that equation turns into : @xmath73,\end{aligned}\ ] ] with @xmath74 and @xmath75 the error function . the cosmic shear power spectrum will receive a shot - noise contribution from the random intrinsic ellipticities of source galaxies and measurement error @xcite . thus , the observed power spectrum between redshift bins @xmath49 and @xmath50 can be expressed as : @xmath76 where @xmath77 is the rms shear due to intrinsic ellipticity and measurement noise ( we assume @xmath78 @xcite ) and @xmath79 \left(\frac1{60}\frac{\pi}{180}\right)^{-2}\hat{n}_i\end{aligned}\ ] ] is the average number density of galaxies per steradians in the @xmath49-th redshift bin , @xmath80 being the number of galaxies per square arcminute and @xmath81 the fraction of sources belonging to the bin . the covariance between the power spectra @xmath82 and @xmath83 is approximately given by @xmath84= \frac{\delta_{\ell\ell ' } } { ( 2\ell+1)\delta\ell f_{sky } } \left[p^{\rm obs}_{(im)}(\ell)\,p^{\rm obs}_{(jn)}(\ell ' ) + p^{\rm obs}_{(in)}(\ell)\,p^{\rm obs}_{(mj)}(\ell ' ) \right ] \label{covar}\end{aligned}\ ] ] where @xmath85 is the sky fraction covered by the survey and @xmath86 is the bin width centred at @xmath87 . the above expression assumes that the power spectrum in each multipole bin is very flat , in order to replace the value of the spectrum evaluated at the bin center with the average of spectrum over each bin ( see [ append : a ] for the complete expression ) . in addition , we have not included the non - gaussian term , due to the contribution of the shear trispectrum @xcite . the fisher matrix formalism @xcite provides lower limits to the error bars of the cosmological parameters one wishes to measure . the basic tool in fisher s method is the likelihood function , yielding the probability that a model gives the set of data @xmath88 . suppose we want to test an hypothesis , _ i.e. _ a cosmological model set by @xmath89 parameters @xmath90 . the likelihood function @xmath91 $ ] is often a complicated function of @xmath92 ; the value @xmath93 corresponding to the peak of @xmath94 defines the _ maximum likelihood estimator _ which , in the limit of large data sets , becomes the _ best unbiased estimator _ of the actual parameter set . thus , the likelihood can be taylor expanded to second order ( the first non vanishing term ) around @xmath93 , being so approximated with a multivariate gaussian distribution @xmath95 here @xmath96 is a positive semi - definite non - singular matrix , dubbed _ covariance matrix _ of the @xmath97 . we remind that holds just in a sufficiently small neighborhood around the maximum @xmath93 . in turn , the _ fisher information matrix _ reads @xmath98 the average @xmath99 is taken over all possible data realizations , given the model parameters . the feature making fisher s formalism so significant is the cramr - rao theorem . it states that the parameter variance about any unbiased estimator value owns a lower bound : @xmath100 , if the other parameters are estimated from the data as well , @xmath101 , if all the other parameters are known . therefore , the fisher information matrix components are the expectation values of @xmath102 . accordingly , the inverse of the fisher matrix is an estimate of the covariance matrix of the parameters @xmath103 . a convenient way to re write the fisher matrix is computing the derivatives of the likelihood function using the following chain rule @xcite : @xmath104 thus , the fisher matrix can be expressed as : @xmath105 where @xmath106 and @xmath107 , respectively , are the fisher and the covariance matrix for the observables @xmath88 . the region in the @xmath89dimensional space of the parameters , defined by @xmath108 , is a hyper ellipsoid of constant probability density for the function . marginalizing over the other parameters , one can project this ellipsoid in the two parameter subspace , yielding a two dimensional ellipse . the analytical expression for the projected ellipse for the two parameters @xmath109 and @xmath110 is given by @xcite : @xmath111^{-1 } \left ( \begin{array}{c } \delta\theta_{\alpha}\\\delta\theta_{\beta } \end{array } \right)=\delta\chi^2(n=2,\sigma)\end{aligned}\ ] ] this can be interpreted as an estimate of the confidence region within a given confidence level @xmath112 for the two parameters @xmath109 and @xmath110 . we present here the results of the fisher analysis of future experiments , considering both wl and cmb measurements . for definiteness we assume a fiducial wl survey with characteristics similar to those of the recently proposed dune project @xcite . we assume a redshift distribution of the form with @xmath113 , @xmath114 and @xmath115 , corresponding to a median redshift of the survey @xmath116 ( see @xcite ) , and a mean surface density of galaxies @xmath117 . the full survey , covering half of the sky ( @xmath118 ) , is divided into @xmath119 redshift bins , with @xmath67 given by equation and @xmath120 . we consider lensing multipoles up to @xmath121 , since we find that results do not depend significantly on larger @xmath87 . however , one should bear in mind that when considering scales @xmath122 there could be some non linear and baryonic effects on the matter power spectrum , and so on the wl spectrum @xcite . these effects , not yet well understood , could be important for forecasts . however , in this work we suppose these effects to be negligible . for cmb data , we consider an ideal experiment with characteristics based on the 143ghz planck channel : angular resolution @xmath123 and sensitivity @xmath124 , @xmath125 . the cosmological model we consider is characterized by 7 parameters with fiducial values : @xmath126 here @xmath127 represents the current total ( cdm + baryons ) matter density in units of the critical density ; @xmath128 and @xmath129 are the physical baryons and total matter densities , respectively ; @xmath130 is the slope of the primordial power law spectral index of density fluctuations ; @xmath131 is the rms mass fluctuation in spheres of 8@xmath132 mpc radius while @xmath133 and @xmath8 were defined in sec . [ sec : de ] . let us notice that the class of de models considered here reduces to @xmath3cdm for @xmath134 and @xmath135 . moreover , the fiducial values of de parameters @xmath136 and @xmath137 are chosen in order to reproduce at @xmath138 an effective equation of state which mimics the case of @xmath3cdm model , @xmath139 . finally , when dealing with cmb data , we also need to fix the value of the optical depth to reionization , @xmath140 . we compute the cmb anisotropies ( temperature and polarisation ) power spectra and the transfer functions , used to calculate linear matter power spectrum , using a modified version of camb @xcite . to evaluate the non linear matter power spectrum , @xmath141 , we employ the prescription by smith _ this is only tested for model with a cosmological constant ; as we are concerned here with fisher matrix estimates assume that the results of @xcite can be extended to coupled models simply by taking into account the non standard scaling of @xmath25 ( eq . [ eq : rho ] ) . numerical derivatives were evaluated considering a @xmath142 stepsize , except for @xmath143 , where we adopted a @xmath142 stepsize on @xmath144 . @lllllll & & & + & & & & & & + & & & & & & + 100@xmath145 & 0.016 & 0.019 & 0.5 & 0.9 & 0.011 & 0.012 + @xmath146 & 0.002 & 0.006 & 0.016 & 0.03 & 0.0004 & 0.0005 + @xmath127 & 0.05 & 0.12 & 0.002 & 0.0014 & 0.0011 & 0.0014 + @xmath147 & 0.004 & 0.005 & 0.012 & 0.018 & 0.0014 & 0.0021 + @xmath131 & 0.07 & 0.13 & 0.0026&0.0029 & 0.0017 & 0.0016 + @xmath148 & 7.2 & 9.5 & 0.89 & 1.1 & 0.28 & 0.28 + @xmath8 & & 0.04 & & 0.018 & & 0.0016 + @xmath149 & 0.005 & 0.006 & & & & + table [ tab : errors ] lists the estimated errors on the various parameters considered . for each data set , we compare forecasts for the target model with results for a sugra model with the same values of the relevant parameters . the table clearly shows that a planck like experiment is able to provide a measurement of a direct de dm interaction at 68% confidence level , even for moderate values of the coupling strength @xmath8 . however , we expect that at 90% confidence level data will still be compatible with @xmath135 . in any case , allowing for a direct interaction strongly degrades the experimental sensitivity on the parameters characterizing the matter density and the normalization of the primordial spectrum of density fluctuations . errors on these quantities increase by a significant amount . figure [ fig : cmb ] shows the joint @xmath150 confidence regions between @xmath8 and each of the other parameters , except for @xmath149 , considering only cmb data . in each plot we marginalized over the parameters not shown . @xmath8 is strongly correlated with most parameters considered here , with the exception of @xmath151 ( and @xmath149 ) , thus introducing additional degeneracies in actual data analysis . forecasts of joint 1@xmath112 confidence regions on the coupling parameter @xmath137 and selected parameters , for a planck like experiments , after full marginalization over the remaining parameters.,width=377 ] a detailed characterization of these degeneracies would require a different approach than that followed here ( e.g. monte carlo markov chains simulations ) . we just point out that they can be understood recalling that the heights of the acoustic peaks of cmb spectra are sensitive to the total matter density and to baryon / dark matter ratio at last scattering . in coupled models , these quantities are not univocally determined by their present day value , but strongly depend on @xmath8 . in addition , the total growth between the last scattering epoch and today is strongly sensitive to @xmath8 , resulting in a clear degeneracy between @xmath8 and @xmath131 . let us notice that , as we are concerned with a combination of cmb and wl data , the parameter set is not optimized for cmb experiments . using a different parametrisation can alter error estimates and/or the degeneracies between the various parameters . in particular , cmb data are better described in terms of the angle subtended by the acoustic horizon at recombination , @xmath152 , and the amplitude of the primordial spectrum of density fluctuations @xmath153 , rather than in terms of @xmath154 and @xmath131 . this results in large errors on the latter parameters , when cmb data alone are considered . adopting a set of parameters better suited to the analysis of cmb data results in slightly lower error estimates overall , but the effects of coupling are largely unchanged . forecasts of joint 1@xmath112 confidence regions on the coupling parameter @xmath137 and selected parameters , for a dune like experiments , after full marginalization over the remaining parameters . notice the change of scales with respect to , width=377 ] comparison between the 1@xmath112 confidence regions of a coupled sugra model with @xmath137 ( red ) and a non coupled sugra model with @xmath135 ( blue ) for a wl experiment . it is clearly possible to distinguish between the two models.,width=245 ] joint 1@xmath112 confidence regions on @xmath8 and @xmath143 after marginalization over the remaining parameters , for different number of galaxies @xmath80 . on the left sugra coupled model with @xmath137 , on the right sugra coupled model with @xmath135 and @xmath8 derivatives calculated only on one side , for positive values of the parameter.,title="fig:",width=245 ] joint 1@xmath112 confidence regions on @xmath8 and @xmath143 after marginalization over the remaining parameters , for different number of galaxies @xmath80 . on the left sugra coupled model with @xmath137 , on the right sugra coupled model with @xmath135 and @xmath8 derivatives calculated only on one side , for positive values of the parameter.,title="fig:",width=245 ] inverse error as a function of the maximum multipole for the wl survey , for different numbers of redshift bins . left panes show results for @xmath8 , right panel refers to @xmath148.,title="fig:",width=245 ] inverse error as a function of the maximum multipole for the wl survey , for different numbers of redshift bins . left panes show results for @xmath8 , right panel refers to @xmath148.,title="fig:",width=245 ] is analogous to for our target weak lensing survey . together with table [ tab : errors ] , these results show the great potential of wl surveys in constraining interacting de models . marginalized errors on @xmath8 and @xmath148 are of the order of @xmath155 and @xmath156 ; these figures represent a factor of 2 , or more , improvement over planck estimates . wl data alone can clearly distinguish the target model from a non coupled model or a cosmological constant even at the 3@xmath112-level , _ viceversa _ assuming a reference sugra model with @xmath135 , we can expect to put an upper limit @xmath157 , at the same confidence level ( see ) . as expected , wl surveys perform significantly better than cmb experiments also with respect to parameters specifying the current matter density , @xmath154 and @xmath131 . moreover , constraints on these parameters are not significantly affected by the coupling degrees of freedom . errors on the remaining parameters , instead , increase by a factor of @xmath158 . finally , we consider a combination of cmb and wl data . cmb and wl probe very different epoch of the universe and are sensitive to different combination of cosmological parameters . considering both cmb and wl data allows to constrain the de parameters with a few percent accuracy , and significantly reduces the degeneracies introduced by de coupling . in this case , the errors on the cosmological parameters are very similar in both models considered , with the exception of @xmath147 . next we considered how our results depend on the characteristic assumed for the target survey . in the impact of mean surface density of galaxies on the determination of @xmath8 and @xmath148 . with @xmath159 constraints on @xmath8 degrades by @xmath160 , while @xmath161 gives only a marginal improvement on expected errors ; constraints on @xmath148 are similarly affected . however , even in the worst case considered here , next generations wl survey will provide an improvement over the information that we are likely to obtain from planck data . lastly , we consider the dependence of our results on the number of bins and the multipoles range considered . in we plot the inverse of the expected variance on @xmath8 and @xmath148 as a function of the maximum multipole considered in the analysis and for different number of bins . with 3 redshift bins , the precision on both parameters depends mostly on multipole up to a few thousands ; smaller scales do not provide a significant contribution . dividing the survey in 5 bins strongly improves the constraints on both parameters and allows to exploit information from multipoles up to @xmath162 . for a dune like survey , a further increase in the number of bins does not lead to significant improvements on the constrains on coupled models parameters . it must be outlined that these results assume that the theoretical framework used to predict the matter power spectrum on intermediate ( @xmath163 ) and small ( @xmath164 ) scales can accurately account for the effects of baryons on non linear structures . in general , the fitting formulas used to predict the non linear power spectrum are calibrated using dissipationless n body simulations and , therefore , do not properly describe baryonic structures . while baryons make up @xmath165 of the matter in the universe and on large scales are expected to trace the dm field , their distribution inside halos is significantly different from dm . in turn , this alters the shape of the non linear power spectrum on the corresponding scales , and the possibility of extracting precision constraints from @xmath166 hinges on our capability of accurately modeling baryon physics @xcite . however , simulations do not yet have the accuracy required for precision constraints and the problem is even more serious for the coupled models considered in this work . modelling non linear stages though spherical growth , mainini @xcite shew that baryons and dm will be however differently distributed , even independently of the onset of gas dynamics . n body simulations of cde models were performed @xcite , by using a ratra peebles @xcite potential ; hydro simulations , instead , were never produced . should accurate prediction be still unavailable for the analysis of a dune like experiment , a more conservative cutoff of @xmath167 would be required . shows that in this case the expected errors on @xmath3 and @xmath8 would increase by a factor @xmath158 . all previous analysis shows that , even if we admit quite a little dm de coupling , we open a pandora s box , leading to a severe degradation in our capacity to deduce cosmological parameters from a given set of measures . as a matter of fact , coupling destroys our trust that the period between the recombination and the start of de relevance is under control . if coupling is absent , during such period scdm is a fair approximation . let us then remind what happens to the growth factor , as soon as coupling is onset : figure [ fig : lingrowth ] shows that : ( i ) deviations from scdm are significant already when @xmath168 ; ( ii ) they are then different for dm and baryons ; ( iii ) they work in the opposite direction , in respect to the effects of a de components . as far as the growth factor is concerned , a tiny coupling is able to overwhelm a huge de amount , with compensation occurring for @xmath1690.4@xmath170 , however keeping always @xmath40 at values greater by @xmath17115@xmath172 . altogether , growth is faster in coupled models . hence , if we do not include the information that coupling is zero in the fit , we can find an agreement between data and a wider range of de amounts . scale dependence of comoving distances in coupled or uncoupled sugra cosmologies , compared with @xmath3cdm . the case @xmath173 , corresponding to a rather strong dm de coupling is also shown.,width=321 ] similar points can be made for the comoving radial distance @xmath174 ( see eq . [ rzzs ] ) . in figure [ fig : distar ] we compare comoving distances for @xmath3cdm with various cosmologies . the figure shows that dde , in the absence of other parameter shifts , sets the last scattering band ( lsb ) closer to the observer . once again , a mild coupling acts in the opposite direction and tends to re - set the lsb at the distance it had in @xmath175 . in the figure we consider the behavior of distances also for a rather strong coupling , @xmath173 . then the distance behavior is different in the period when de density can be neglected , in respect to the epoch when de and dm have similar densities . the key point , however , is that the lsb becomes then farther from the observer . when fitting cmb data to such models , in order to compensate such effect , the value of @xmath176 tends to be increased . strong coupling therefore yields a large hubble parameter estimate . figures [ fig : omega ] finally show the scale dependence of the density parameters in the different models . once again , when de is mildly coupled , a behavior more similar to @xmath175 is recovered . on the contrary , when considering a greater coupling strength , we see that de and dm keep similar densities up to a fairly large redshift . this was indeed the initial motivation of cde cosmologies . scale dependence of the density parameters of the various components in @xmath3cdm . compared with other models . in the upper panel uncoupled and weakly coupled sugra models are considered . for the sake of comparison , in the lower panel we also show the effects of choosing a stronger coupling . , title="fig : " ] .6truecm scale dependence of the density parameters of the various components in @xmath3cdm . compared with other models . in the upper panel uncoupled and weakly coupled sugra models are considered . for the sake of comparison , in the lower panel we also show the effects of choosing a stronger coupling . , title="fig : " ] altogether , these figures indicate that adding a small coupling reduces the effects of the very passage from @xmath175 to dde ; owing to the excellent fit that @xmath175 cosmologies have with data , this tells us that only highly refined cmb data will be able to test the possibility that a mild dm de coupling exists . on the contrary , a stronger coupling , although easing the coincidence problem , displaces several observables in a unacceptable way . forecasts of joint 1@xmath112 confidence regions on the coupling parameter @xmath137 and selected parameters , for a combination of a planck like and a dune like experiment , after full marginalization over the remaining parameters . , width=377 ] future wl surveys will certainly put more stringent constraints on cosmological parameters and will be crucial to break quite a few degeneracies between parameters . within this context , in this paper we focus on coupled de models with a twofold aim . detecting a signal of dm de coupling would be certainly decisive to fix the nature of the dark components . henceforth determining the level of sensitivity needed to appreciate such an effect is crucial in setting the appeal of forthcoming projects . there is however a complementary aspect which deserves much attention . in order to convert raw data into physical information , a set of parameters , spanning a variety of models , is to be fixed ; a bias on parameter selection , however , can lead to an optimistic estimate of the confidence level for the best fitting model , far from reality . in this paper , we focused on this kind of danger , when we open the option of dm de coupling . even if such coupling is absent or quite weak , we showed that just considering its possibility may widen the error bars for a number of parameters , also apparently unrelated to the coupling itself . in a sense , when a new degree of freedom is opened , such an effect is natural and expected . coupling , however , has really a major impact , affecting different parameters for the different observables considered here ; moreover , its impact is drastically reduced when we work out parameter values by using simultaneously both observables . before outlining our main conclusions it is however worth reminding a technical limit we had to face . while cmb predictions depend on linear spectra , the wl spectrum is limited to its non linear shape . for the purpose of the present analysis , we assumed that prescription for @xmath3cdm models @xcite can be trivially extrapolated to coupled models , so enabling us to estimate the non linear spectrum once the linear spectrum is known . let us however remark that the shift was estimated from quite a wide set of @xmath3cdm simulations . simulations of cosmological models with state parameters @xmath177 , although performed by several authors , are still not so extensively studied as @xmath3cdm ; let alone coupled de simulations : in this case the only available simulations are due to @xcite and deal with a potential @xmath7 different from sugra . however , differences between prescriptions for @xmath3cdm and dde are small @xcite and it seems however clear that model differences can most affect the rate of evolution of halo concentration , slightly shifting the scale where non linearity effects become significant . the use of more precise prescriptions can therefore only cause minor variation on the estimated errors and , although welcome , such simulations are not expected to interfere substantially with our conclusions . our estimates were based on assuming that a photometric survey is available , with @xmath120 and a median redshift @xmath178 , covering half of the sky ( @xmath118 ) . these features are similar to the recently proposed dune experiment @xcite . we also compared and combined results from wl with the constraints expected for an ideal planck like experiment . the basic results of our calculations are quoted in table [ tab : errors ] which is one of the main results of this work . a first set of conclusions concerns an ideal cmb experiment considered by itself . in this case , introducing coupling degrees of freedom is crucial for the error estimates on some of parameters , also apparently unrelated to coupling . in particular , while the error on @xmath179 keeps @xmath1801% , the errors on @xmath181 and @xmath131 increase from 1.3% to 3.9% and from 9% to 16% , respectively . from a physical point of view , the option opened by coupling is that expansion rate and fluctuation growth , from the last scattering band to the observer s site , is non standard ( _ e.g. _ , the proportionality law @xmath182 could be mildly violated ) . although cmb data themselves set stringent limits on such deviations , this widens the volume of the parameter space consistent with a given data set ; in particular , it increases the likelihood of values of @xmath183 that would otherwise be negligible and , because of the intercorrelation amongst parameters , this reflects immediately on @xmath181 and @xmath154 estimates . similar effects occur in wl experiments , although involving different parameters . such experiments are a direct test of @xmath154 , whose estimated error is reduced by a factor @xmath184 in respect to a cmb experiment . when the coupling option is opened , the error on @xmath154 does not increase ; on the contrary , it becomes easier to attribute raw data uncertainties to other parameters and , although marginally , the error on @xmath154 becomes somehow smaller . a completely new situation occurs if both cmb and wl measurements are simultaneously used . in this case , the opening of the coupling option causes just a marginal increase of the errors on most parameters . this is a clear indication of the complementarity of the cmb and wl measurements , as described in @xcite and one of our conclusion is that the combination of these observables , besides of providing parameter values independent of @xmath8 , can set a ( nearly)final word on the coupling option . as a matter of fact , by comparing , and , we see that , when joining cmb and wl results , the degeneracies , between @xmath8 and @xmath131 , as well as between @xmath8 and @xmath154 , disappear . breaking degeneracies is the main aim when different observables are simultaneously considered . we see that , from this point of view , the efficiency of using both cmb and wl measures can be hardly overestimated . let us then focus on the case of the spectral index of scalar fluctuations . when both cmb and wl data are used to constrain @xmath147 , a sharp reduction of errors occurs . no surprise that cmb data , by themselves , reflecting the state of the universe before the onset of non linear processes , were more efficient to constrain @xmath147 , with or without coupling , than wl measures . joining together the two observables , we then see errors to decrease from 0.4% and 1.2% down to 0.16% , in the uncoupled case ; in the presence of coupling we have a similar behavior , with errors passing from 0.5% and 1.8% down to 0.25% . the error level achieved , in both cases , is exceptional , even for precision cosmology , and clearly suggests to relieve the constraint of a single @xmath147 value , so inspecting its possible scale dependence , with realistic possibilities to find a direct insight into the nature of the inflationary potential . in conclusion , future wl surveys could really allow a significant step forward in the comprehension of the dark cosmic side ; we can adfirm that , when they will be available , the endeavour to put the genie back inside the lamp will approach a full success . lplc is supported by nasa grant nnx07ah59 g and jpl planck subcontract no . 1290790 . we acknowledge the use of the camb package . lplc and glv thank s. bonometto for insightful discussion and help with the preparation of the manuscript . in order to determine the convergence power spectrum covariance , one can introduce the so called `` flat - sky '' approximation and treat the sky as flat , replacing spherical harmonic sums with fourier transforms ( ft ) . of course , this approximation is acceptable just for small angular scales we also consider the tomographic case and use greek letters as superscripts to denote quantities belonging to different redshift bins . the ft of the convergence field can be defined as : @xmath185 while the convergence power spectrum and trispectrum are , respectively : @xmath186 @xmath187 here @xmath188 is the dirac function . the value of the lensing power spectrum in correspondence of a multipole @xmath189 can be estimated as the mean over a multipole bin of width @xmath190 , centered at @xmath189 : @xmath191 where @xmath192 is the area of the shell of width @xmath190 corresponding to @xmath189 , while @xmath193 is the area of the survey . quite in the same way , for the trispectrum we have : @xmath194 let us then consider the following expression : @xmath195 the 4-point function at the r.h.s . can be decomposed in its connected parts : @xmath196 replacing their expression in eq . , one can easily recognize the contribution of the trispectrum , using eq . : @xmath197 where @xmath198 . owing to the definition of covariance , @xmath199 \equiv \langle \mathcal{p}^{\alpha\beta}_{\it{l } } \mathcal{p}^{\gamma\delta}_{\it{l ' } } \rangle -\langle \mathcal{p}^{\alpha\beta}_{\it{l}}\rangle \langle\mathcal{p}^{\gamma\delta}_{\it{l ' } } \rangle,\label{eq : a.10}\end{aligned}\ ] ] and using , one can then argue that the integrals in make sense only if they correspond to the same @xmath189bin ; the same can be claimed for . this property can be described introducing a kronecker delta function @xmath200 . thus , the expression becomes : @xmath201 & = & \frac{1}{a}\mathcal{t}_{\it{ll'}}^{\alpha\beta\gamma\delta } + \delta_{\it{ll'}}\frac{\left ( 2\pi \right)^2}{a^2 } \int_{s_{\it{l}}}\frac{\rmd^2\mathbf{l}_1}{a_{\it{l}}^2 } \left [ p_{l_{1}}^{\alpha\gamma } \langle\kappa^{\beta}\left ( -\mathbf{l}_{1 } \right ) \kappa^{\delta } \left ( \mathbf{l}_{1 } \right ) \rangle + p_{l_{1}}^{\alpha\delta } \langle\kappa^{\gamma } \left ( \mathbf{l}_{1 } \right ) \kappa^{\beta } \left ( -\mathbf{l}_{1 } \right ) \rangle \right]\nonumber\\ & = & \frac{1}{a}\left\{\mathcal{t}_{\it{ll'}}^{\alpha\beta\gamma\delta } + \delta_{\it{ll ' } } \frac{\left ( 2\pi \right)^2}{a_{s_{\it{l } } } } \int_{s_{\it{l}}}\frac{\rmd^2\mathbf{l}_1}{a_{s_{\it{l } } } } \left [ p_{l_{1}}^{\alpha\gamma}p_{l_{1}}^{\beta\delta } + p_{l_{1}}^{\alpha\delta}p_{l_{1}}^{\gamma\beta}\right]\right\}\\ & \thickapprox&\frac{1}{4\pi\,f_{sky } } \mathcal{t}_{\it{ll'}}^{\alpha\beta\gamma\delta}+ \frac{\delta_{\it{ll'}}}{2l\delta l f_{sky } } \left [ p_{l}^{\alpha\gamma}p_{l}^{\beta\delta } + p_{l}^{\alpha\delta}p_{l}^{\gamma\beta}\right].\end{aligned}\ ] ] in the last line , we have supposed the lensing power spectrum to be smooth enough to treat it as a constant within each bin width . 50 riess a g , kirshner r p , schmidt b p , jha s , et al . 1998 _ astrophys . j. _ * 116 * 1009 ; perlmutter s , aldering g , goldhaber g , et al . 1999 _ astrophys . j. _ * 517 * 565 ; riess a g , strolger l - g , tonry j , et al . 2004 _ astrophys . j. _ * 607 * 665 de bernardis p , ade p a r , bock j j , et al . 2000 _ nature _ * 404 * 955 ; padin s , cartwright j k , mason b s , et al . 2001 _ astrophys . j. _ * 549 * l1 ; kovac j , leitch e m , pryke c , et al . 2002 _ nature _ * 420 * 772 ; scott p f , carreira p , cleary k , et al . 2003 _ mon . not . soc . _ * 341 * 1076 ; spergel d n , bean r , dor et al . 2007 _ astrophys . j. _ * 170 * 377 _ preprint _ astro ph/0603449 colless m m , dalton g b , maddox s j , et al . 2001 _ mon . not . _ * 329 * 1039 ; colless m m , peterson b a , jackson c , et al . 2003 _ preprint _ astro ph/0306581 ; loveday j ( the sdss collaboration ) 2002 _ contemporary phys . _ * 43 * 437 ; tegmark m , blanton m , strauss m et al . 2004 _ astrophys . j. _ * 606 * 702740 ; adelman mccarthy j k , agueros m a , allam s s , et al . 2006 _ astrophys . j. suppl . _ * 162 * 3848 caldwell r r , dave r and steinhardt p j 1998 1582 ; wetterich c 1988 b*302 * 668 brax p and martin j 1999 b*468 * 40 ; brax p and martin j 2000 d*61 * 103502 ; brax p , martin j and riazuelo a 2000 d*62 * 103505 carroll s m , duvvuri v , trodden m , et al . 2004 d*70 * 043528 ; nojiri s and odintsov s d 2003 d*68 * 123512 zlatev i , wang l and steinhardt p j 1999 896899 ; steinhardt p j , wang l and zlatev i 1999 d*59 * 123504 schimd c et al a&a * 463 * ( 2007 ) 405 colombo l p l and gervasi m jcap * 10 * ( 2006 ) 001 ellis j. , kalara s. , olive k.a . & wetterich c. , 1989 , phys . b228 , 264 ; wetterich c. , 1995 , a&a 301 , 321 ; amendola l. , 2000 , phys.rev . d62 , 043511 ; gasperini m. , piazza f. & veneziano g. , 2002 , phys.rev . d65 , 023508 ; comelli , d. pietroni m. amd riotto , phys . b * 571 * ( 2003 ) 115 ; chimento l.p . , jakubi a.s . , pavon d. & zimdahl w.,2003 , phys.rev d67 , 083513 guo z k , ohta n and tsujikawa s 2007 phys . rev . d * 76 * 023508 ; abdalla e , abramo l r w , sodre l j and wang b _ preprint _ arxiv:0710.1198 [ astro - ph ] ; manera m and mota d f 2006 _ mon . not soc . _ * 371 * 1373 j. khoury and a. weltman , phys.rev.lett . * 93 * ( 2004 ) 17110 ; s.s . gubser and j.khoury , phys.rev.d * 70 * ( 2004 ) 104001 ; p. brax et al . , phys.rev.d * 70 * ( 2004 ) 123518 majerotto e , sapone d and amendola l ( 200 ) astro - ph/0410543 amendola l , campos g.c . and rosenfeld r ( 2006 ) astro - ph/0610806 wei h and zhang s n 2006 phys . lett . b*644 * 7 wei h and zhang s n 2007 phys . lett . b*654 * 139 _ preprint _ astro - ph/0704.3330 mainini r , colombo l p l and bonometto s a _ astrophys . j. _ * 632 * ( 2005 ) 691 di porto c and amendola l ( 2007 ) astro - ph/0707.2686 m. doran , j. jaeckel , phys.rev . d * 66 * ( 2002 ) 043519 ; amendola l , quartin m , tsujikawa s , and waga i , phys.rev . d * 74 * ( 2006 ) 023525 amendola l phys.rev . d * 62 * ( 2000 ) 043511 ; amendola l phys.rev . d * 69 * ( 2004 ) 103524 huey g and wandelt b d phys.rev . d * 74 * ( 2006 ) 023519 ; das s , corasaniti p s and khoury j , phys.rev . d * 73 * ( 2006 ) 083509 amendola l and quercellini c phys.rev . d*68 * ( 2003 ) 023514 olivares g , atrio - barandela f and pavon d ( 2007 ) astro - ph/0706.3860 das s , corasaniti p s and khoury j phys.rev . d*73 * ( 2006 ) 083509 mainini r and bonometto s a jcap ( 2007 ) astro - ph/0703303 mainini r phys.rev.d 72 ( 2005 ) 083514 bartelmann m and schneider p 2001 phys.rept . * 340 * 291 refregier a _ ann . rev . _ 2003 * 41 * 645 van waerbeke l and mellier y _ preprint _ astro - ph/0305089 schneider p _ preprint _ astro - ph/0509252 amara a and refregier a _ preprint _ astro - ph/0610127 hu w 1999 _ astrophys . j. _ * 522 * l21 huterer d 2002 _ phys . _ d * 65 * 063001 hu w and jain b 2004 _ phys . _ d * 70 * 043009 ma z m , hu w and huterer d 2005 _ astrophys . j. _ * 636 * 21 kaiser n 1998 _ astrophys . j. _ * 498 * 26 cooray a and hu w 2001_astrophys . j. _ * 554 * 56 takada m and jain b 2004 _ mon . not . * 348 * 897 sivia d s 1996 _ data analysis : a bayesian tutorial _ ( oxford : oxford university press ) fisher r a 1935 _ j. roy . * 98 * 39 tegmark m , taylor a and heavens a 1997 _ astrophys . j. _ * 480 * 22 oh s p , spergel d n and hinshaw g 1999 _ astrophys . j. _ * 510 * 551 matsubara t and szalay a s _ preprint _ astro - ph/0203358 . refregier a et al . _ preprint _ astro - ph/0610062 rudd d h , zentner a r and kravtsov a v _ preprint _ astro - ph/0703741 . zentner a r , rudd d h and hu w 2008 _ phys . _ d * 77 * 043507 http://www.camb.info/ smith r e , et al . 2003 _ mon . not . * 1311 mainini r 2005 _ phys . rev . _ d * 72 * 083514 maccio a v quercellini c mainini r amendola l and bonometto s a 2004 _ phys . _ d * 69 * 123516 ratra b and peebles p j e 1998 _ phys _ d * 37 * 3406 ma z m _ preprint _ arxiv : astro - ph/0610213 . eisenstein d j hu w and tegmark m 1999 _ astrophys . j. _ * 518 * 2

an important drawback of the conventional single - electron transistor @xcite ( set ) is its relatively large output resistance which should be much larger than the quantum resistance @xmath1 . this limits the operation frequency of the prospective integrated single - electron circuits @xcite and imposes a severe frequency limitation for individual sets used nowadays as electrometers . estimating the total capacitance of a wire delivering set signal from a cryostat to outside electronics as 1 nf , we get the time constant on the order of @xmath2 ; therefore the operating frequency is limited to few khz , which is a typical value achievable by conventional set setups.@xcite while the operating frequency can be significantly increased by placing a preamplifier in a close vicinity of the set , @xcite the preferred at present solution of the problem is the use of the radio - frequency set @xcite ( rf - set ) which in many instances has already replaced the traditional set setup . the principle of the rf - set operation is somewhat similar to the operation of the radio - frequency superconducting quantum interference device @xcite ( rf - squid ) and is based on the microwave reflection from a tank ( lc ) circuit containing the set ( fig . [ fig1 ] ) ; another possibility is to use the transmitted wave . @xcite the measured set input signal changes effective set resistance and affects the intensity and the phase of the reflected ( or transmitted ) wave that is later sensed by either homodyne detection or simple rectification ( to separate incoming and reflected waves a directional coupler can be used ) . the high operation frequency of the rf - set is due to the signal propagation by the microwave , so that the set does not need to charge the whole output wire , while the tank circuit tuned in resonance works as an impedance transformer providing a better match between the effective set differential resistance @xmath3 ( @xmath4 ) and the microwave cable wave impedance @xmath5 ( typically @xmath6 ) . for a good matching the `` unloaded '' quality factor @xmath7 of the tank circuit consisting of inductance @xmath8 and capacitance @xmath9 should be comparable to @xmath10 . while a much lower value , @xmath11 , was used in the first experiment , @xcite the values close to the matching condition ( sometimes even higher ) are typically used at present . increase of the @xmath0-factor obviously decreases the rf - set bandwidth limited by @xmath12 where @xmath13 is the `` carrier '' microwave frequency and @xmath14 is the `` loaded '' @xmath0-factor , which also takes into account the effect of the set ( see below ) . the straightforward design @xcite shown in fig . [ fig1 ] can be somewhat modified to reduce the bandwidth further for a dense multiplexing;@xcite however , in this paper we will consider only the original design . the rf - set bandwidth as wide as 100 mhz has been demonstrated @xcite using a relatively high carrier frequency @xmath15 and relatively low @xmath0-factor @xmath11 . however , to improve experimental rf - set sensitivity , it happens to be advantageous to reduce the carrier frequency by few times ( to reduce the noise contribution from the amplifier ) and also increase @xmath0 ( closer to the impedance matching regime ) , so that a practical bandwidth at present is about 10 mhz ( for example , the bandwidth of 7 mhz for the carrier frequency of 332 mhz has been reported in ref . @xcite ) . the high operation frequency of the rf - set makes it easily possible to avoid the 1/@xmath16 noise limitation of the set sensitivity which is typically dominant at frequencies @xmath17 , and so work in the region of the shot noise limited sensitivity.@xcite though experimentally the contribution from the amplifier noise is still comparable or larger than the set noise , a relatively rapid improvement of the rf - set charge sensitivity from @xmath18 at 1.1 mhz in the first experiment @xcite to the value @xmath19 ( 4.8 @xmath20 in energy units ) at 2 mhz reported in ref . @xcite assures that the pure shot noise limit will be achieved pretty soon . ( the above numbers can be compared with the sensitivity @xmath21 at 4.4 khz of a purely conventional set reported in ref . @xcite , the value @xmath22 at 10 hz for the `` stacked '' set , @xcite and the sensitivity @xmath23 at 45 hz for the set made of carbon nanotubes . @xcite ) one of important potential applications of the rf - set is for the readout of the charge qubits in a solid - state quantum computer . @xcite the possibility of a single - shot qubit measurement requires fast enough distinguishing between two charge states to avoid significant qubit evolution during measurement . this requires sufficiently wide rf - set bandwidth and most importantly good enough charge sensitivity . the estimates @xcite show that the single - shot qubit measurement is almost within the reach of present - day rf - set performance ; however , reliable measurement still requires significant improvement in sensitivity . this makes very important the question of ultimate ( theoretical ) rf - set sensitivity . the ultimate rf - set sensitivity is also a crucial parameter for monitoring quantum dynamics of nanomechanical resonators @xcite and for a variety of rf - set applications as an electrometer in classical single - electron devices . in spite of significant experimental activity on rf - sets , we are aware of only few theoretical papers on the rf - sets . the basic theory of the shot noise limited charge sensitivity of the rf - set has been developed in ref . a similar theory has been applied to the sensitivity analysis for the rf - set - based micromechanical displacement detector . @xcite some theoretical analysis of the transmission - type rf - set can be found in ref . the theory of a somewhat related device , radio - frequency bloch - transistor , has been developed in ref . @xcite . in our opinion , the rf - set definitely requires further theoretical attention , since many questions about rf - set performance has not yet been answered theoretically . in this paper we extend the theory of ref . @xcite to the case of arbitrary @xmath0-factor of the tank circuit , removing the assumption ( strongly violated in the present - day experiments ) of @xmath0 being much smaller than the impedance matching value . we calculate the response and sensitivity of the normal - metal rf - set and optimize these magnitudes numerically over the rf wave amplitude and the set background charge . then we study the dependence of the optimized rf - set response and sensitivity on the tank @xmath0-factor , operation temperature , set resistance , carrier frequency , and set asymmetry due to asymmetric biasing . some results of this paper have been presented earlier in a short form . the schematic of the rf - set used in our analysis is shown in fig . the set consists of two tunnel junctions with capacitances @xmath24 and @xmath25 and resistances @xmath26 and @xmath27 . the set is coupled via gate capacitance @xmath28 to the measured charge source ( for example , a single - electron box or a similar structure ) , which is characterized by the charge @xmath29 and capacitances @xmath30 and @xmath31 to the set leads ( the total source capacitance is @xmath32 ) . assuming constant @xmath29 ( neglecting backaction from the set ) , it is easy to show that the set coupled to the charge source is equivalent to the simple double - junction set structure with parameters @xmath33 where @xmath34 and @xmath35 are effective junction capacitances ( the total set island capacitance is @xmath36 ) , and the total induced charge of the set island @xmath37 is the sum of the initial background charge @xmath38 and the contribution from the measured charge source . for numerical results we have used the `` orthodox '' model @xcite for a normal - metal set ( see appendix ) . = 3.3 in the current @xmath39 through the set affects the quality factor of the tank circuit consisting of inductance @xmath8 and capacitance @xmath9 , while the contribution to the tank circuit capacitance is neglected , assuming @xmath40 $ ] . ( this condition is usually well satisfied experimentally ; it also allows us to neglect the effect of single electron jumps on the tank circuit oscillations . ) the quality factor of the tank circuit is also affected by the wave impedance @xmath5 of the cable . for a simple linear analysis let us substitute the set by an effective differential resistance @xmath3 . then , assuming the case of weak damping , we can simply add the contributions from @xmath5 and @xmath3 , so that the total ( `` loaded '' ) quality factor of the tank circuit is @xmath41 where the `` unloaded '' quality factor @xmath0 corresponds to absence of the set , @xmath42 , while @xmath43 corresponds to the damping by the set only , @xmath44 . notice that @xmath0 is fixed by the rf - set design , while @xmath43 and therefore @xmath14 depend on the operating conditions ; so even though @xmath14 is a more physically meaningful quantity than @xmath0 , in this paper we consider unloaded @xmath0 as an independent parameter and call it as @xmath0-factor . for the incoming voltage wave @xmath45 , the reflected wave @xmath46 is determined by the complex reflection coefficient @xmath47 : @xmath48 since the measured charge signal changes the set resistance @xmath3 , the rf - set response is proportional to @xmath49 . using the first order approximation close to the resonant frequency @xmath50 of the tank circuit : @xmath51 ( in particular , this approximation neglects the shift of the resonant frequency , which is a second - order effect ) it is easy to get @xmath52 for fixed values of @xmath5 and @xmath3 , this expression increases with @xmath0 for small @xmath0 , thus showing the need of @xmath0-factor increase to achieve a better rf - set response . the maximum is reached at @xmath53 which is the case of practically matched impedances , @xmath54 , at frequencies close to resonance , and also corresponds to the condition @xmath55 . notice that in this case the reflection practically vanishes , @xmath56 . actually , eq . ( [ dadr ] ) is not really relevant , because the rf amplitude @xmath57 should depend on @xmath0 in order to maintain approximately constant amplitude @xmath58 of the set bias voltage oscillations , which is determined by the voltage scale of the coulomb blockade . taking into account the relation @xmath59 we obtain @xmath60 which is somewhat similar to eq . ( [ dadr ] ) and also reaches maximum at @xmath61 . notice that this condition optimizes only the rf - set response , while the shot - noise - limited sensitivity can ( and as will be seen later does ) have completely different dependence on @xmath0 . the linear analysis discussed above can be used only as an estimate because of the significant nonlinearity of the set current - voltage ( @xmath62 ) dependence . for a more exact analysis @xcite we use the kirchhoff s rules taking into account the current @xmath39 through the set . let us separate the voltage @xmath63 at the end of the rf cable into the dc component @xmath64 ( which can be supplied via the bias - tee @xcite ) and rf component @xmath65 , where @xmath66 is the incoming wave ( from now on we do not use complex representation ) and @xmath67 is the outgoing reflected wave . the differential equation for the rf component @xmath68 is @xmath69 , \label{difeq}\end{aligned}\ ] ] where @xmath70 is the current through the set averaged over time much longer than @xmath71 , and the time dependence of the set current @xmath39 can be found self - consistently from the time dependence of its bias voltage @xmath72\ , q/\omega_0 . \label{vb}\ ] ] in this paper we assume that the rf frequency is small compared to the frequency of electron tunneling through the set , @xmath73 , so that the set shot noise is a small contribution compared to the deterministic part of the set current @xmath39 , which is calculated using dc @xmath62 curve and @xmath74 ( actually , it is still ok if this condition is not satisfied during some fraction of the period due to coulomb blockade , since the small current does not affect the oscillations significantly ) . in a steady state ( assuming that the induced set charge @xmath37 does not change with time ) the reflected wave can only contain the incoming frequency @xmath75 and its overtones : @xmath76 . \label{vout}\ ] ] we separate the term @xmath77 mainly to follow the notations of ref . @xcite , even though it really makes sense only for a low-@xmath0 case at @xmath78 , when all @xmath79 and @xmath80 are small . using the substitution @xmath81 $ ] in eq . ( [ difeq ] ) , we find the coefficients @xmath79 and @xmath80 as @xmath82 where @xmath83 ( averaging is over the oscillation period ) , @xmath84 is the normalized frequency , @xmath85 is the kronecker symbol , and the current @xmath39 is calculated self - consistently using the set bias voltage @xmath86 . \label{vbiter}\end{aligned}\ ] ] because of the resonant behavior of the tank circuit at @xmath87 , the contribution of overtones ( @xmath88 ) in the reflected signal is always small if @xmath78 ( since the overtones are far from resonance ) . this can be easily seen from eqs . ( [ xn])([yn ] ) , especially in the case @xmath89 , when they are significantly simplified : @xmath90 therefore , a linear ( one - frequency ) approximation in which only @xmath91 and @xmath92 are taken into account , works very well in this case . in our numerical analysis we used the result of the linear approximation as a starting point of the iterative procedure [ @xmath93 to solve selfconsistently eqs . ( [ xn])([vbiter ] ) taking into account few ( typically 3 - 5 ) overtones . we have checked numerically that the account of overtones typically gives a small correction in the case of a reasonably large @xmath0-factor and @xmath78 . in the linear approximation the set bias voltage has only one frequency component : @xmath94 and therefore for the calculation of @xmath95 and @xmath96 [ see eq . ( [ anbn ] ) ] the set can be simply replaced by the effective resistance @xmath97 where @xmath98 is the set current - voltage dependence . [ notice that @xmath99 , i.e. there is no effective reactance contribution . ] hence , this linear ( one - frequency ) approximation is completely equivalent to the case of a resistor instead of the set , considered in the previous subsection . the only new condition is a selfconsistent relation between the effective resistance @xmath3 and the amplitude @xmath100 of the set bias voltage . the amplitude @xmath100 ( which depends on @xmath3 ) can be calculated either using eq . ( [ vinvb ] ) in which @xmath101 or using eq . ( [ vbiter ] ) , which gives @xmath102 ^ 2 + ( x_1+q\tilde{\omega}y_1)^2 } , \ ] ] while the components @xmath91 and @xmath92 are given by equations @xmath103 ^ 2 } { ( q^2+\tilde{r})^2 } } \ , v_{in } , \label{x1}\\ & & y_1= \frac{2\tilde\omega q}{q^2+\tilde{r } } \frac{-\frac{q^2[1+(1-\tilde\omega^2)\tilde{r}]}{q^2+\tilde{r } } + ( 1-\tilde\omega^2)\tilde{r } } { \tilde\omega^2+\frac{q^2[1+(1-\tilde\omega^2)\tilde{r}]^2 } { ( q^2+\tilde{r})^2 } } \ , v_{in } , \label{x2}\end{aligned}\ ] ] where @xmath104 ( these equations can obviously be rewritten in a shorter way ; however , they become less transparent to analyze ) . the equations ( [ x1])([x2 ] ) significantly simplify in the case @xmath89 : @xmath105 from which it is clear that @xmath106 for @xmath107 , and therefore @xmath108 so far we have implicitly assumed that the set current @xmath39 depends on time only because of the periodic time dependence of the set bias voltage @xmath74 . however , @xmath39 has also a small noise component , the magnitude of which depends on the bias voltage and therefore also has a periodic time dependence . the shot noise of the set current leads to the fluctuations of the parameters @xmath109 and @xmath110 defined by eq . ( [ anbn ] ) and consequently to the fluctuations of the reflected wave quadratures @xmath79 and @xmath80 . since the noise of @xmath79 and @xmath80 can be meaningfully discussed only at frequencies less than @xmath111 , which is much less than the typical frequency @xmath112 of electron tunneling in the set , it is sufficient to consider the low - frequency limit of the set shot noise . the low - frequency spectral densities of @xmath109 and @xmath110 , and their mutual spectral density can be calculated as @xmath113 where the averaging is over the oscillation period , @xmath114 is the low- ( zero- ) frequency spectral density of the set current ( see appendix ) , and the time dependence comes from oscillating bias voltage @xmath74 . consequently , the low - frequency spectral densities of @xmath79 and @xmath80 fluctuations , and their mutual spectral density are @xmath115 where @xmath116 notice much simpler equations for @xmath117 and @xmath118 in the case @xmath89 : @xmath119 the change of the measured charge @xmath29 can in principle be monitored via the change of any quadrature @xmath79 or @xmath80 . the corresponding _ responses _ can be defined as the derivatives @xmath120 or @xmath121 , while the corresponding _ sensitivities _ ( minimal detectable charge @xmath122 for a small measurement bandwidth @xmath123 ) are @xmath124 numerical calculations show that in the usual case @xmath78 thus defined rf - set sensitivity for overtones ( @xmath88 ) can be comparable to the sensitivity using the carrier frequency ( @xmath125 ) . however , because of relatively small amplitude of reflected overtones in this case , their monitoring is impractical , and so we mainly consider monitoring of @xmath91 and @xmath92 which are referred below as @xmath126 and @xmath127 . monitoring both quadratures @xmath126 and @xmath127 simultaneously , one can improve the sensitivity compared with monitoring of only one quadrature . it is easy to show that the resulting sensitivity can be obtained as the optimization over angle @xmath128 of the sensitivity corresponding to monitoring the linear combination @xmath129 ( experimentally , this is just a phase shift in the homodyne detector ; notice that the contribution @xmath130 is noiseless by assumption ) . the optimum sensitivity is achieved at @xmath131 and the resulting sensitivity is @xmath132^{1/2 } , \label{dq1}\ ] ] that can be rewritten as @xcite @xmath133^{1/2 } , \label{dq2}\ ] ] where @xmath134 $ ] is the noise correlation factor [ @xmath135 is a real magnitude because we consider only low - frequency sensitivity ; for finite frequency @xmath135 in eq . ( [ dq1 ] ) should be replaced by @xmath136 . notice that the response optimization for @xmath137 monitoring is achieved at different phase shift : @xmath138 that leads to the response @xmath139^{1/2}$ ] . let us also consider the case when the reflected wave is monitored by simple rectification . assuming for simplicity the monitoring of only the first harmonic amplitude @xmath140^{1/2}$ ] ( overtones are filtered out ) , the rf - set sensitivity can be calculated as @xmath141^{1/2 } } { |(x - v_{in } ) dx / dq_s + y dy / dq_s| } \label{dq3}\ ] ] while the response is obviously @xmath142/a_1 $ ] . the formulas for monitoring of @xmath143th overtone are similar , except @xmath57 does not contribute to the amplitude @xmath144 . in the most interesting for practice case @xmath145 the magnitude of @xmath127 is small in comparison with @xmath126 for @xmath146 [ see eqs . ( [ x1s])([y1s ] ) ] , and the correlation factor @xmath147 also vanishes [ because @xmath148 and the set bias voltage contains mostly @xmath149-component see eq . ( [ vbsimple ] ) ] . then eq . ( [ dq2 ] ) for the best homodyne detection practically coincides with eq . ( [ dq3 ] ) for rectification and reduces to the sensitivity @xmath150 for monitoring of @xmath126 component only ( similarly , the formulas for response also practically coincide ) . because of that , the numerical results for rf - set response and sensitivity in the case @xmath89 will assume monitoring of @xmath126 quadrature . since the increments of the measured charge @xmath29 and the effective set charge @xmath37 are related by a constant factor , @xmath151 the rf - set response and sensitivity in respect to @xmath29 and in respect to @xmath37 differ by the factor @xmath152 . all our numerical results will be in terms of @xmath37 measurement , so for @xmath126 monitoring we will use the derivative @xmath153 as a measure of the rf - set response and the magnitude @xmath154 as a measure of the sensitivity . we have studied numerically the dependence of the rf - set response and sensitivity on various parameters , which include `` fixed '' parameters ( which can not be easily changed in an experiment ) and the parameters of the operation point . the fixed parameters are : effective junction capacitances of the set @xmath34 and @xmath35 ( we assume @xmath155 unless mentioned otherwise ) , resistances @xmath26 and @xmath27 ( we always assume @xmath156 ) , temperature @xmath157 , cable wave impedance @xmath5 ( we always assume @xmath158 ) , the tank circuit frequency @xmath159 , and the @xmath0-factor @xmath160 . the operating point parameters are the effective set background charge @xmath37 , dc bias voltage @xmath64 , amplitude @xmath57 of the incident wave , and its frequency @xmath75 ( in most cases we assume @xmath89 ) . we use the set parameters for normalization , so that a natural unit for temperature is @xmath161 ( where @xmath162 ) , the voltage unit is @xmath163 , the rf - set response @xmath153 ( or @xmath164 ) can be measured in units @xmath165 , and the unit for sensitivity @xmath166 is @xmath167 ( where @xmath168 ) . notice that all considered magnitudes have a simple scaling with @xmath169 ; however , there is no simple scaling with @xmath170 because of the dimensionless parameter @xmath171 . = 3.0 in = 3.0 in figure [ res - sens-2d](a ) shows the rf - set response @xmath153 on the plane of operating point parameters @xmath37 and @xmath57 for the case @xmath172 , @xmath173 ( i.e. @xmath174 ) , @xmath175 , @xmath89 , and @xmath176 ( the set is symmetric , @xmath155 , @xmath156 ) . figure [ res - sens-2d](c ) is similar , except @xmath177 . we do not show @xmath164 because it is practically vanishing . because of the same reason , the sensitivity @xmath178 shown in figs . [ res - sens-2d](b ) and [ res - sens-2d](d ) is calculated using only quadrature @xmath126 . both the response and sensitivity are obviously poor when the amplitude of oscillations at the set is below the coulomb blockade threshold @xmath179 [ this happens at @xmath180 where @xmath181 , and corresponds to the triangles in the lower left corners of figs . [ res - sens-2d](a)(d ) ] , and they are also poor when @xmath57 is much larger than this condition ( notice that better response corresponds to larger @xmath153 , while better sensitivity corresponds to smaller @xmath182 ) . even though the regions of relatively good response and sensitivity are similar , the maximum response and optimum sensitivity are achieved at quite different points in the @xmath183 plane . in particular , the amplitude @xmath57 of the incoming wave is significantly larger for maximum response , than for best sensitivity . in the present - day experiments , maximization of response [ figs . [ res - sens-2d](a ) and [ res - sens-2d](c ) ] is still of the major importance , because the noise from the next amplifying stage is still significant [ if a large constant noise had been added to @xmath184 in eq . ( [ dq0x ] ) , then the sensitivity would be mainly determined by the denominator @xmath153 ] . however , if the amplifier noise is small compared to the contribution from the set shot noise , then the best operation point should optimize the shot - noise - limited sensitivity [ figs . [ res - sens-2d](b ) and [ res - sens-2d](d ) ] . in the case of comparable contributions from two noises , we have a trade - off between two regimes . in this paper we will concentrate on the analysis of the maximum response ( mr ) mode and optimized sensitivity ( os ) mode , keeping in mind that experimentally optimal regime is somewhere in between , depending on the amplifier noise . = 3.0 in = 3.0 in figure [ v0dep ] shows dependence of the response @xmath153 and sensitivity @xmath178 on the dc bias voltage @xmath64 in the mr and os modes ( optimizations are over @xmath57 and @xmath37 ) . several curves on each plot are for different @xmath0-factors : @xmath186 , 30 , 50 , 70 , and 90 , while other parameters are similar to the parameters of fig . [ res - sens-2d ] . one can see that the best response in both mr and os regimes as well as the best sensitivity in os mode are achieved at @xmath175 . this is because both positive and negative branches of the symmetric set @xmath62 curve ( see fig . [ i - v ] in appendix ) contribute equally at @xmath175 , and therefore the signal is maximal . in the mr mode at @xmath187 the optimum background charge [ see fig . [ res - sens-2d](a ) ] is about @xmath188 ( so the coulomb blockade threshold @xmath179 is about @xmath189 ) while the optimum amplitude @xmath100 of the set bias @xmath190 oscillations is about @xmath191 ( these numbers have only weak dependence on @xmath0 , while the optimum @xmath57 obviously depends on @xmath0 quite significantly ) . when @xmath192 starts to increase , it becomes advantageous to increase @xmath37 ( so @xmath179 decreases ) while @xmath100 stays approximately constant , so that both positive and negative branches of the set @xmath62 still contribute to the response . however , since these branches can not both contribute in the optimal way , the response decreases with @xmath64 [ see fig . [ v0dep](a ) ] . for large enough @xmath64 it becomes preferable to use only one ( positive ) branch [ this regime corresponds to the lower maximum in fig . [ res - sens-2d](c ) , while the upper maximum corresponds to the two - branch regime ] ; then the optimal wave amplitude drops to @xmath193 and the optimal @xmath37 corresponds to @xmath179 slightly above @xmath64 ( by about @xmath194 ) . this change causes the kinks on the response curves in fig . [ v0dep](a ) and jumps down on the sensitivity curves in fig.[v0dep](c ) . notice that in the mr mode the sensitivity at @xmath195 is better than at @xmath175 . in the os regime the optimal amplitude is significantly less than in the mr regime . it depends mainly on temperature , and for parameters of fig . [ v0dep ] the amplitude @xmath100 at the set is typically between 0.08 and 0.1 in units of @xmath163 . the `` above blockade '' voltage @xmath196 is few times smaller than @xmath100 and is comparable to the temperature . the best sensitivity is achieved at @xmath175 when both branches participate , while at large enough @xmath64 when it becomes preferable to use only one branch , both sensitivity and response practically do not depend on @xmath64 [ see figs . [ v0dep](b ) and [ v0dep](d ) ] . = 2.7 in = 2.7 in the dependence on @xmath0-factor is summarized in fig . [ qdep ] , which shows the response and sensitivity in the mr and os modes for @xmath197 , @xmath173 , and @xmath89 . the results are presented for @xmath175 ( thick lines ) , which provides the best mr and os response and best os sensitivity , and also for @xmath198 ( thin lines ) , which is a typical value for the case when only one branch of the set @xmath62 is involved . one can see that the response @xmath153 grows linearly with @xmath0 at small @xmath0 [ see eq . ( [ resp - lin ] ) ] and reaches the maximum at @xmath0 around 50 ( this number is somewhat different for different regimes ; for example it is almost 70 for the thin solid line ) , which is close to the crude theoretical estimate @xmath199 for the impedance matching . however , unlike in the linear model , this maximum does not correspond to the exact impedance matching . for example , the impedance matching ( minimum of reflection ) occurs at @xmath200 for the upper curve in fig . [ qdep](a ) and at @xmath201 for the curve second from the top , while for two lower curves ( os mode ) it does not occur at all in a reasonable range of @xmath0 . in contrast to the response behavior , _ the rf - set sensitivity monotonically worsens with increase of _ @xmath0 . qualitatively , this happens because the noise @xmath184 has @xmath202 scaling [ see eq . ( [ sx1 ] ) ] , while the response @xmath153 has slower than @xmath0 dependence [ see eq . ( [ resp - lin ] ) ] . this simple analysis predicts the parabolic dependence @xmath203 , which crudely fits the curves in fig . [ qdep](b ) using @xmath204 about 1.5 - 3 times larger than @xmath171 , though the curves in fig . [ qdep](b ) actually have slower than parabolic dependences at large @xmath0 . comparing the mr and os modes at @xmath205 ( just some typical number ) and @xmath175 , we see that the mr regime provides about 40% larger response , while the os regime provides about twice better sensitivity . even though these numbers depend significantly on the temperature and also depend on @xmath0 and @xmath171 , they show that the results for the mr and os regimes are not too much different ( not by an order of magnitude ) . comparison of the cases @xmath175 and @xmath198 shows that for the mr and os response as well as for the os sensitivity there is no much difference between these two cases , and the relative difference decreases with @xmath0 . in contrast , the difference between mr sensitivity at @xmath175 and at @xmath198 grows with @xmath0 and can become significant . the low-@xmath0 limit of the os sensitivity for @xmath175 is well described by the formula @xcite ( which also assumes @xmath206 ) @xmath207 [ we do not use a shorter formula @xmath208 to emphasize natural normalizations ] , and similar limit for @xmath209 ( one - branch case ) is close to @xcite @xmath210 [ see fig . [ qdep](b ) ] . however , the theory of ref . @xcite which assumes @xmath211 , is not able to describe significant change of the sensitivity with @xmath0 in fig . [ qdep](b ) . one can see that this dependence is even more significant in the mr mode . = 3.0 in = 3.0 in the numerical results on temperature dependence are shown in fig . [ tdep ] for @xmath173 and @xmath89 . similar to fig . [ v0dep ] we show on four panels the response @xmath153 and corresponding sensitivity in the mr and os modes ( the optimal dc bias value @xmath175 is used ) . it is important to notice that in the mr mode both response and sensitivity almost do not depend on temperature at @xmath212 , and the rf - set performance is still ok at temperatures @xmath213 ( response and sensitivity change less than twice compared to zero - temperature case ) . in the os regime the response also has a very weak temperature dependence at @xmath214 ; however , the sensitivity is strongly temperature - dependent . the low-@xmath0 os sensitivity can be accurately described by eq . ( [ dq - an ] ) up to temperatures @xmath215 , [ see the lowest dotted line in fig . [ tdep](d ) ] , and the curves for large @xmath0-factors also follow the scaling @xmath216 at temperatures @xmath217 . in the orthodox theory this @xmath218 dependence is valid even at very small temperatures leading to infinitely good sensitivity ( @xmath219 ) ; however , in reality the neglected contribution of cotunneling processes becomes significant in the os regime at small @xmath157 that changes the formalism.@xcite comparing the results for mr and os modes , we notice that while they are significantly different at low temperatures , the difference decreases with temperature , so that at @xmath220 the mr and os results are already quite similar . = 3.0 in = 3.0 in figure [ rdep ] shows the response and sensitivity dependence on the total set junction resistance @xmath170 for @xmath197 , @xmath175 , and @xmath89 ( we show results only for @xmath221 , because at @xmath222 the theory is too inaccurate due to neglected cotunneling processes ) . in the case @xcite @xmath211 the response @xmath153 scales as @xmath223 [ see eq . ( [ x1 ] ) ] and the sensitivity @xmath178 scales as @xmath224 [ see eq . [ dq - an ] ] . correspondingly , the solid lines ( @xmath186 ) in figs . [ rdep](a ) and ( b ) are practically straight lines passing through the origin , and in figs . [ rdep](c ) and ( d ) the solid lines are practically horizontal [ the level determined by eq . ( [ dq - an ] ) is shown in fig . [ rdep](d ) by the lowest dotted line , which fits well the solid line ] . the dependence on @xmath170 becomes nontrivial when @xmath171 is comparable to @xmath202 . it is interesting to notice that the inverse response remains a practically linear function of @xmath170 even for large @xmath0 , as seen in figs . [ rdep](a ) and ( b ) ( the slope of the lines decreases with @xmath0 while the offset increases ) . in particular , this means that in the orthodox model the rf - set response continues to increase with the set resistance decrease even when the matching condition is overshot . decrease of @xmath170 close or beyond the matching condition , worsens the sensitivity in comparison with scaling @xmath225 [ see figs.[rdep](c ) and ( d ) ] . in the mr mode the sensitivity worsens with decrease of @xmath170 even in absolute units ( @xmath226 ) when the matching condition is sufficiently overshot , while in the os mode the sensitivity still improves with decrease of @xmath170 in absolute units , in spite of worsening compared to @xmath224 scaling . = 3.0 in = 3.0 in now let us discuss the effect of asymmetric effective capacitances of the set junctions , @xmath227 . even when the physical junction capacitances are equal , @xmath228 , the effective capacitances @xmath34 and @xmath35 can be significantly different when the gate capacitance @xmath28 is comparable to junction capacitances . in our model shown in fig . [ fig1 ] this happens due to asymmetry of signal source capacitances , @xmath229 [ see eqs . ( [ c1 ] ) and ( [ c2 ] ) ] ; in rf - set experiments this effect is called asymmetric rf biasing of the set.@xcite for the conventional set setup , the biasing asymmetry is not important at all , because of the formal set equivalence under transformation @xcite @xmath230 for arbitrary @xmath231 , that means that the asymmetric biasing ( different @xmath34 and @xmath35 ) can be simply corrected by the background charge shift . however , for the rf - set there is no simple equivalence because the bias voltage @xmath190 changes in time . ( effective capacitance asymmetry can still be controlled by addition of the extra rf signal to the set gate . ) therefore it is interesting to find if the asymmetric biasing ( @xmath227 ) is better or worse than the symmetric case ( @xmath155 ) . figure [ asymbias ] shows the @xmath64 dependence of the rf - set response and sensitivity in the mr and os modes for several ratios @xmath232 ( with fixed total capacitance @xmath233 ) . one can see that in the asymmetric cases the best mr response and os sensitivity are achieved at nonzero @xmath64 and are typically @xmath234 than the corresponding values for the symmetric case . therefore , the _ asymmetric biasing is actually preferable _ ( that eases the concern about this issue discussed in ref . @xcite ) ; however , the advantage is rather minor . in particular , the os sensitivity in the asymmetric case is still limited by eq . ( [ dq - an-2 ] ) ; therefore the possible improvement is less than the sensitivity decline due to high @xmath0-factor [ see fig . [ qdep](b ) ] , and even less because of the difference between eqs . ( [ dq - an ] ) and ( [ dq - an-2 ] ) . so far we have considered the resonant case @xmath89 . [ actually , the exact resonance is at the frequency @xmath235 which is very close to @xmath236 . ] in this case the quadrature component @xmath126 is much larger than @xmath127 ( at @xmath107 ) and therefore the rf - set response and sensitivity in respect to monitoring @xmath126 quadrature practically coincide with that for monitoring the reflected wave amplitude @xmath237 ( we denote with @xmath237 the amplitude @xmath238 of the first harmonic ) or monitoring the optimized phase - shifted combination @xmath239 . the detuning of @xmath75 from the resonant frequency @xmath236 leads to a significant magnitude of @xmath127 quadrature , and so to different results for different ways of reflected wave monitoring . for a small frequency detuning @xmath240 the simple linear analysis using eq . ( [ resp - lin ] ) predicts the non - monotonic shape of the @xmath126-response frequency dependence with zeros at @xmath241 and with the full width at half height ( fwhh ) equal to @xmath242 ; while for the amplitude monitoring it predicts fwhh of @xmath243 . ( [ dadr ] ) which assumes both @xmath3 and @xmath57 being frequency - independent , gives the response fwhh equal to @xmath244 and @xmath245 for @xmath126 and @xmath237 monitoring , respectively ] . in reality the situation is more complex because even neglecting the nonlinear contributions , the effective set resistance @xmath3 ( and therefore @xmath14 ) depends on @xmath74 amplitude , which depends itself on @xmath75 , @xmath3 , and @xmath57 . moreover , the optimized point in the @xmath183 plane is also frequency - dependent . = 3.0 in = 3.0 in figure [ freq - dep ] shows the numerical results for the frequency dependence ( in the vicinity of the resonance ) of the optimized response and sensitivity . each panel shows four curves , which correspond to monitoring the components @xmath126 ( thick solid line ) , @xmath127 ( dashed ) , optimized phase - shifted combination @xmath246 ( thin solid ) , and amplitude @xmath237 ( dotted line ) . at each frequency we perform optimizations of the response ( mr mode ) and sensitivity ( os mode ) over @xmath183 plane for all four monitored magnitudes , so the optimizations are different for different curves . the @xmath0-factor is equal to @xmath247 , which is close to the estimate of the impedance - matching value ( see also fig . [ qdep ] ) ; however , instead of the naively expected value @xmath248 for the loaded @xmath0-factor , it is @xmath249 in the mr mode since @xmath250 ( these values are calculated at @xmath89 using effective set resistance ) . the shape of the @xmath126-response dependence in panel ( a ) is close to the prediction of eq . ( [ resp - lin ] ) using @xmath251 , though the minima do not reach zero and the curve beyond the minima is shifted up . the mr @xmath127-response is practically zero at @xmath89 ( since @xmath127 vanishes ) ; however , it becomes comparable to the resonant @xmath126-response at some frequency detuning ; the overall shape is close to the prediction of eq . ( [ resp - lin ] ) , but the maxima are about 15% higher . the mr @xmath237-response curve is about 50% wider than the prediction of the linear theory ( using @xmath251 ) and is very close to the curve for monitoring optimal phase - shifted quadrature @xmath246 . as expected , the @xmath237- and @xmath246-responses are better than @xmath126- and @xmath127-responses at finite detuning ( for both mr and os modes ) . the sensitivity ( both os and mr ) for @xmath246 or @xmath237 monitoring at finite detuning is also better then for @xmath126 or @xmath127 monitoring . an interesting observation is that while the rf - set response decreases with detuning , the sensitivity slightly improves with detuning for @xmath246 and @xmath237 monitoring . this effect is similar to the sensitivity improvement with the decrease of @xmath0-factor ( see fig . [ qdep ] ) . in experiments the incoming rf wave is usually tuned close to the resonance with the tank circuit ; in this case the contribution of overtones in the reflected rf wave is small in spite of significant set @xmath62 nonlinearity ( the set nonlinearity has been recently used @xcite for rf mixing ) . however , if the @xmath143th overtone is in resonance , @xmath252 , then the reflected wave may have a significant contribution from this overtone , and the rf - set operation can be based on monitoring this overtone . @xcite the use of different frequencies for the incident and reflected waves may be advantageous for some applications . also , it may be useful to have the absence of the monitored reflected wave when the set is off ( no current ) , while in the conventional regime this case corresponds to the largest reflected power . one more possible advantage is somewhat easier control of the amplitude of the set bias voltage oscillations , since now it is more directly related to the incident amplitude @xmath57 , while in the usual regime the relation depends on @xmath14 [ see eq . ( [ vbsimple ] ) ] which varies with operating point . ( the disadvantage is that the incident amplitude @xmath57 should be much larger than in usual regime that may lead to heating problems . ) = 2.7 in = 2.7 in figure [ overtones ] shows the rf - set response and sensitivity for @xmath253 and @xmath254 , in respect to monitoring @xmath255 and @xmath256 , correspondingly ( the @xmath126-quadratures are small ) . we use @xmath175 in the case @xmath254 and @xmath257 in the case @xmath253 [ for @xmath175 there is no second overtone because of the @xmath62 curve symmetry even at nonzero @xmath37 see fig . [ i - v](a ) and eq . ( [ anbn ] ) ] . as seen from fig . [ overtones ] , the mr responses and os sensitivities in the two cases ( second and third resonant overtone ) are quite similar , that is related to a strong nonlinearity of the set @xmath62 curve . moreover , comparing figs . [ qdep ] and [ overtones ] ( the parameters are the same ) we see that the rf - set performance in the regime of a resonant overtone is comparable to the performance in the conventional regime @xmath258 ( the mr response and os sensitivity are worse by only about 1.5 times ) . combined with the advantages discussed above , this makes the regime of resonant overtone potentially useful in experiments . in this paper we have developed the formalism for the calculation of the response and shot - noise - limited sensitivity of the normal - metal rf - set , extending the theory of ref . @xcite to the case of arbitrary large @xmath0-factor of the tank circuit and arbitrary frequency of the incident rf wave . the formalism has been used to analyze numerically the dependence of the rf - set response and sensitivity on the operation parameters . we have mainly studied two operation modes , optimized over the rf wave amplitude and the set background charge , corresponding to either maximum response ( mr mode ) or optimum sensitivity ( os mode ) . the rf amplitude for the optimum sensitivity is typically significantly smaller than for maximum response . the mr mode is the best experimental regime when the preamplifier noise is more significant than the effect of the set noise , while the os mode is the best when the preamplifier noise is negligible . analyzing the performance dependence on the set dc bias voltage , we have found that the best response and sensitivity are achieved at zero dc bias , though finite biasing does not change the rf - set performance much as long as it is within the coulomb blockade range . we have found that the dependence of the rf - set response on the unloaded @xmath0-factor of the tank circuit has a maximum at @xmath0 comparable to the simple impedance matching estimate @xmath259 . in contrast , the rf - set sensitivity monotonically worsens with increase of @xmath0 ; the dependence is approximately parabolic and the sensitivity change can reach few times ( compared to the low-@xmath0 case ) if @xmath0 is chosen too large . this means that to improve the sensitivity in an experiment , it is better to `` undershoot '' the @xmath0-factor compared to the impedance matching case , than to `` overshoot '' it . studying the temperature dependence we have found that the rf - set response saturates approximately at temperatures @xmath260 , that translates into 200 mk for a typical capacitance value @xmath261 af . the optimized sensitivity continues to improve as @xmath218 at lower temperatures until it reaches the quantum limitation due to cotunneling @xcite ( not studied here ) . the `` orthodox '' rf - set performance improves with decrease of the set resistance ; however , for high @xmath0-factors the dependence is significantly slower than the low-@xmath0 scaling @xmath223 for the response and @xmath262 for the sensitivity . we have analyzed the effect of the asymmetric rf biasing of the set which leads to unequal effective capacitances of the set and found that such asymmetry does not worsen the rf - set performance ( even slightly improves it ) . this answers the concern about asymmetric rf biasing raised in ref . @xcite . we have analyzed the effect of the carrier frequency detuning from the resonance and found that the decrease of the rf - set response with detuning can crudely be described by the simple linear theory ( though difference in linewidth can reach 50% ) ; however , the estimate of the loaded @xmath0-factor determining the linewidth is not simple since it significantly depends of the operation point . even with the frequency detuning , the rf - set performance for monitoring the reflected wave amplitude ( by rectification ) is found to be very similar to the case of optimal homodyne detection ( one - channel mixing with the optimal phase ) . it is important that the mixing phase optimizing the sensitivity can be different from the phase optimizing the response . unexpectedly , in contrast to the response decrease with detuning , the optimized sensitivity slightly improves with detuning . we have also analyzed the operation regime for which an overtone of the incident rf wave is in resonance with the tank circuit , and found that the rf - set performance in this regime is comparable to the performance in the conventional regime . taking into account an advantage of the frequency separation between the incident and monitored waves , this operation mode may be experimentally useful . the theoretical proposal of the resonant overtone regime @xcite has been recently realized experimentally in the group of keith schwab . experimental rf - set sensitivities using the second and third resonant overtones have been found practically coinciding with the sensitivity in the conventional regime . @xcite now let us compare our theoretical results for the rf - set sensitivity with the experimental value of @xmath263 from ref . @xcite for the normal - metal case ( the sensitivity of the superconducting rf - set was significantly better : @xmath264 ) . using the experimental parameters @xmath265 af , @xmath266 k@xmath267 , assuming the temperature @xmath268 mk , and using fig . [ rdep ] , we get the mr sensitivity of @xmath269 for @xmath205 and @xmath270 for @xmath172 , while the os sensitivities are @xmath271 for @xmath205 and @xmath272 for @xmath172 [ notice that all these numbers are significantly higher than the low-@xmath0 os estimate @xmath273 using eq . ( [ dq - an ] ) ] . unfortunately , it is not possible to extract the unloaded @xmath0-factor value from ref . @xcite ; however , it seems to be between 30 and 50 . it is also not known if the rf - set operation point was closer to the mr or os mode ( we guess the response was more important ) . nevertheless , we see that the difference between the theory and experiment is few times . we guess that the difference is mainly due to the preamplifier noise . some contribution may also come from higher effective temperature than we assumed and from a non - optimal operation point . another contribution to the difference may come from the neglected here effect of cotunneling , which limits the sensitivity . however , a rough estimate @xcite of this limit @xmath274 gives the value of @xmath275 , so it is unlikely to be the major reason for the difference . we hope that the further experimental progress will bring the rf - set sensitivity really close to the theoretical limit . there are still many theoretical questions about the rf - set performance , not answered in this paper . for example , as seen from the above estimate , the account of cotunneling and study of the quantum operation of the rf - set is starting to be important for experiments . the development of the theory for superconducting rf - set is even more important since in the majority of experiments with rf - sets the superconducting state is used . there is still no rigorous theory of the frequency dependence of the rf - set sensitivity . it is also important to consider the backaction from the rf - set and analyze if it can in principle be used as a quantum detector with high quantum efficiency ( ideality ) . these problems are going to be the topics of further studies . useful discussions with m. blencowe , k. likharev , and k. schwab are gratefully acknowledged . the work was partially supported by nsa and arda under aro grant daad19 - 01 - 1 - 0491 and by the semiconductor research corporation grant 2000-nj-746 . the numerical calculations were partially performed on the ucr - igpp beowulf computer lupin . d. v. averin and k. k. likharev , in _ mesoscopic phenomena in solids _ , edited by b. l. altshuler , p. a. lee , and r. a. webb ( elsevier , amsterdam , 1991 ) , p. 173 . t. a. fulton and g. d. dolan , phys . lett . * 59 * , 109 ( 1987 ) . k. k. likharev , proc . ieee * 87 * , 606 ( 1999 ) . a. n. korotkov , int . j. electronics * 86 * , 511 ( 1999 ) . b. starmark , t. henning , t. claeson , p. delsing , and a. n. korotkov , j. appl . phys . * 86 * , 2132 ( 1999 ) . j. pettersson , p. wahlgren , p. delsing , d. b. haviland , t. claeson , n. rorsman , and h. zirath , phys . rev . b 53 , 13272 ( 1996 ) . e. h. visscher , j. lindeman , s. m. verbrugh , p. hadley , j. e. mooij , and w. van der vleuten , appl . * 68 * , 2014 ( 1996 ) . r. j. schoelkopf , p. wahlgren , a. a. kozhevnikov , p. delsing , and d. e. prober , science * 280 * , 1238 ( 1998 ) . j. clarke , proc . ieee * 77 * , 1208 ( 1989 ) . p. wahlgren , r j. schoelkopf , a. a. kozhevnikov , p. delsing , d. e. prober , and t. claeson , j. supercond . * 12 * , 741 ( 1999 ) . a. aassime , g. johansson , g. wendin , r. j. schoelkopf , and p. delsing , phys . lett . * 86 * , 3376 ( 2001 ) . a. aassime , d. gunnarsson , k. bladh , p. delsing , and r. schoelkopf , appl . phys . lett . * 79 * , 4031 ( 2001 ) . a. aassime , p. delsing , and t. claeson , nanotechnology * 12 * , 96 ( 2001 ) . t. r. stevenson , f. a. pellerano , c. m. stahle , k. aidala , and r. j. schoelkopf , appl . phys . lett . * 80 * , 3012 ( 2002 ) . k. segall , k. w. lehnert , t. r. stevenson , r. j. schoelkopf , p. wahlgren , a. aassime , and p. delsing , appl . . lett . * 81 * , 4859 ( 2002 ) . k. bladh , d. gunnarsson , g. johansson , a. kck , g. wendin , a. aassime , m. taslakov , and p. delsing , physica scripta * t102 * , 167 ( 2002 ) . lehnert , k. bladh , l. f. spietz , d. gunnarsson , d. i. schuster , p. delsing , and r. j. schoelkopf , phys . lett . * 90 * , 027002 ( 2003 ) . t. m. buehler , d. j. reilly , r. p. starret , s. kenyon , a. r. hamilton , a. s. dzurak , and r. g. clark , microel . * 67 * , 775 ( 2003 ) . t. m. buehler , d. j. reilly , r. p. starret , a. d. greentree , a. r. hamilton , a. s. dzurak , and r. g. clark , cond - mat/0304384 ; cond - mat/0302085 . t. fujisawa and y. hirayama , appl . * 77 * , 543 ( 2000 ) . h. d. cheong , t. fujisawa , t. hayashi , and y. hirayama , appl . . lett . * 81 * , 3257 ( 2002 ) . a. n. korotkov , d. v. averin , k. k. likharev , and s. a. vasenko , in _ single - electron tunneling and mesoscopic devices _ , edited by h. koch and h. lubbig ( springer , berlin 1992 ) , p. 45 . a. n. korotkov , phys . b * 49 * , 10381 ( 1994 ) . v. a. krupenin , d. e. presnov , a. b. zorin , and j. niemeyer , j. low temp . phys . * 118 * , 287 ( 2000 ) . l. roschier , r. tarkiainen , m. ahlskog , m. paalanen , and p. hakonen , appl . . lett . * 78 * , 3295 ( 2001 ) . p. m. echternach , c. p. williams , s. c. dultz , s. braunstein , and j. p. dowling , quant . comp . * 1 * , 143 ( 2001 ) . a. d. armour , m. p. blencowe , and k. c. schwab , phys . * 88 * , 148301 ( 2002 ) . e. k. irish and k. schwab , cond - mat/0301252 . r. g. knobel and a. n. cleland , nature * 424 * , 291 ( 2003 ) . a. n. korotkov and m. a. paalanen , appl . . lett . * 74 * , 4052 ( 1999 ) . m. p. blencowe and m. n. wybourne , appl . . lett . * 77 * , 3845 ( 2000 ) . y. zhang and m. p. blencowe , j. appl . * 91 * , 4249 ( 2002 ) . y. zhang and m. p. blencowe , j. appl . phys . * 92 * , 7550 ( 2002 ) . a. b. zorin , phys . * 86 * , 3388 ( 2001 ) . v. o. turin and a. n. korotkov , cond - mat/0305012 ; to be published in appl . lett . m. h. devoret and r. j. schoelkopf , nature * 406 * , 1039 ( 2000 ) . a. n. korotkov , r. h. chen , and k. k. likharev , j. appl . phys . * 78 * , 2520 ( 1995 ) . r. knobel , c. s. yung , and a. n. cleland , appl . * 81 * , 532 ( 2002 ) . k. schwab , private communication . the set current @xmath276 and its low frequency shot noise @xmath277 have been calculated using the orthodox theory of single - electron tunneling,@xcite assuming that a stationary state of the set is achieved at any moment of time ( i.e. @xmath278 ) . we assume normal metal ( non - superconducting ) case and calculate the tunneling rates @xmath279 of electron tunneling to ( @xmath280 ) or from ( @xmath281 ) the island through the first or second junction as@xcite @xmath282 } \ , , \\ & & w_j^\pm ( m ) = \frac{e^2}{c_\sigma } \left [ \mp ( -1)^j \frac{v_bc_1c_2}{ec_j } -\frac{1}{2 } \mp m \mp \frac{q_0}{e } \right ] , \end{aligned}\ ] ] where @xmath283 is the number of extra electrons on the set , @xmath284 denotes the junction , and @xmath157 is the temperature . the low - frequency spectral density @xmath277 of the set current [ fig . [ i - v](b ) ] can be calculated as @xcite @xmath289 ( \hat{\bf\gamma}^{-1})_{m'm } \nonumber \\ & & \hspace{0.3 cm } \times [ \gamma_1^+(m-1 ) \sigma_{st}(m-1 ) -\gamma_1 ^ -(m+1)\sigma_{st}(m+1 ) \nonumber \\ & & \hspace{0.3 cm } -(i / e)\sigma_{st}(m ) ] + 2e^2 \sum_m \sigma_{st}(m ) [ \gamma_1^+(m)+\gamma_1 ^ -(m ) ] , \label{s_i}\end{aligned}\ ] ] where @xmath290 is the three - diagonal matrix of the set charge evolution , @xmath291 ] $ ] . notice that eq . ( [ s_i ] ) can be easily used numerically even though the matrix @xmath290 is singular and therefore does not have a unique inverse ; eq . ( [ s_i ] ) is constructed in a way that the non - uniqueness is not important , and therefore the standard algorithm for solving a linear system of equations with three - diagonal matrix can be readily used . ( there is no problem of the matrix singularity at a finite frequency @xmath292 , when the inverse of the matrix @xmath293 should be calculated . ) actually , instead of using eq . ( [ s_i ] ) , we have used a somewhat faster algorithm for calculation of @xmath277 , described in section vii of ref .

the determination of the binding energy of nuclei starting from realistic @xmath1 potentials is one of the subjects which has received permanent attention from the early days when the brueckner - bethe - goldstone ( bbg ) equation introduced methods to overcome the strong repulsion of the nuclear forces at short distances . at present , several many body techniques compete to accurately determine the binding energy of nuclear matter starting from the realistic nn potentials . one of them is the correlated basis functions ( cbf ) @xcite , which follows the line of the hypernetted chain approach ( fhnc ) @xcite . another one follows the traditional bbg approach @xcite , and , although costly numerically , the variational monte carlo method ( mc ) has allowed to make , in principle , exact calculations , although limited to nuclei with small or medium value of a @xcite . methods like the selfconsistent treatment of the nucleon selfenergy have also introduced new advances in the field @xcite . the need for three body forces has also been emphasized and the present status is that it is difficult to be quantitative on the strength of these forces , and usually they are parameterized in order to adjust the precise value of the binding energy @xcite . it has also been noted in @xcite that short range correlations play an important role when considering these three body forces . a common feature of these approaches is that they start from the realistic nucleon nucleon interaction , obtained from fits to nn data and deuteron data . they use hence the free nn interaction as input . one of the important ingredients of this interaction is the one pion exchange ( ope ) . however , from detailed studies of the pion nuclear interaction it is well known that the pion properties in the nuclear medium are sizably renormalized @xcite . there is also the question of the intermediate range attraction , which is basic in the binding of nuclei . models for this interaction would contain @xmath0 exchange , uncorrelated two pion exchange and omega exchange @xcite . in as much as the pion properties are changed in the medium , so should the two pion exchange be modified . medium effects in the two - pion exchange have been investigated in early works like in ref . @xcite restricting themselves in this case to a subset of two pion exchange diagrams with no @xmath3-isobar intermediate states , by including pauli blocking in the intermediate nucleons . the medium modification of the two pion exchange got a new impulse after the models of the @xmath4 interaction in the medium showed large modifications @xcite , later on softened by the introduction of chiral constraints @xcite . the implementation of this medium modified @xmath4 interaction in the correlated two pion exchange @xmath1 potential increased appreciably the @xmath1 attraction in nuclear matter . this was partially reduced by the consideration of the chiral constraints in refs . @xcite . in these latter references the importance of short range correlations which modify the @xmath5 nucleus selfenergy was already discussed . the further use of medium modified vector meson masses led to improvements in the nuclear matter saturation curve . a new perspective into this problem has been made possible by studies of meson meson interaction within chiral unitary approaches @xcite which allow to improve the description of the correlated two pion exchange @xmath1 interaction @xcite , as an alternative to the conventional @xmath0-meson exchange interaction . this picture of the @xmath0 exchange was mandatory after extensive studies showing that the @xmath0 is not a genuine resonance , made up of @xmath6 but just the manifestation of a pion - pion resonance state created by the interaction of the pions , what is called a dynamically generated resonance . this shows up naturally within the context of chiral unitary approaches which use the input of the chiral lagrangians for the meson meson interaction and extends chiral perturbation theory to implement exactly unitarity in coupled channels @xcite . this means the @xmath0 exchange inside a nuclear medium will also be modified as direct consequence of the change of the pion properties . our aim in the present paper is to start from this new picture for the @xmath0 exchange , use also the standard approach for the uncorrelated two pion exchange and modify these in the nuclear medium to see what changes one finds from these sources . further improvements come from the consideration of short range correlations not only in the pion selfenergy but also in the vertex functions appearing in the model . the changes obtained are moderate , thanks to the simultaneous consideration of these nuclear short range effects in the calculation . in the absence of these , the renormalization of the @xmath1 interaction is huge . yet , even the moderate results obtained are large enough to motivate further calculations of the nuclear binding and other properties of matter . the paper is organized as follows . in sec . ii we provide those elements of the chiral lagrangian which are relevant for the present calculations and briefly discuss peculiarities of the finite baryon density . in sec . iii we consider the modification of the one pion exchange @xmath1 force in the nuclear medium and in sec . iv we discuss the propagation of two pions in the nuclear matter . v and vi are devoted to the in - medium two pion exchange in the scalar - isoscalar channel , both , correlated and uncorrelated . the technical details are relegated to the appendix . in this section we will briefly specify those elements of the effective chiral lagrangian in the meson - baryon sector which are relevant for the subsequent calculations . the effective chiral lagrangian is written as the sum of a purely mesonic lagrangian @xmath7 and the baryonic lagrangian @xmath8 @xmath9 both are organized in a derivative and quark mass expansion . the lowest order mesonic lagrangian @xmath10 is given by @xmath11 and contains the most general low - energy interactions of the pseudo - scalar meson octet . ( [ lpipi ] ) the symbol @xmath12 indicates the trace in flavor space , the goldstone fields are collected in a unitary matrix @xmath13 , @xmath14 mev is the pseudoscalar decay constant and the leading symmetry - breaking term @xmath15 is linear in the quark masses . for @xmath16 and in the isospin limit @xmath17 . the lowest order baryon octet - meson octet lagrangian reads @xmath18 \rangle\ ] ] where the brackets @xmath19 $ ] and @xmath20 denote commutators and anti - commutators , respectively . the covariant derivative of the @xmath21 baryon matrix @xmath22 is defined as @xmath23\ ] ] in the absence of external field eqs . ( [ barlagr ] ) and ( [ covarderiv ] ) involve other quantities @xmath24 the @xmath21 axial vector coupling constants are determined by neutron and hyperon @xmath25-decay . one finds @xmath26 , @xmath27 and the axial coupling constant is @xmath28 . in the @xmath16 limit the lagrangian simplifies to @xmath29 where @xmath30 is a two component dirac field @xmath31 . the @xmath32 vertex ( a ) , the contact @xmath33 ( b ) and pion pole ( c ) terms . ] the opep diagrams . the first graph is the vacuum contribution . the second and third diagrams of the upper and first two diagrams of the lower sets of diagrams correspond to the @xmath44 and @xmath45 rpa series . the last two diagrams account for the excitation of @xmath46 states and the contribution of the @xmath47-wave optical potential , respectively . the latter diagrams play a minor role and will not be further discussed . ] we start with the one - pion exchange @xmath1 potential ( opep ) . the typical diagrams modifying it are shown in fig . [ opepmedimgraphs ] where the propagation of exchanged pions is distorted by interactions with nucleons forming the fermi sea . the fermionic bubbles describe the decay of the pion in @xmath44 and @xmath45 states and take into account the conventional nuclear matter polarization effects . all these diagrams are responsible for the interaction of two nucleons in the particle - particle ladder channel with the in - medium virtual excitations . in fig . [ opepmedimgraphs ] the first graph is the vacuum contribution . the second and third diagrams correspond to the @xmath44 rpa series and the last diagrams accounts for the excitation of @xmath46 states and the contribution of the @xmath47-wave optical potential , respectively . the @xmath48-wave pion self energy is given by @xmath49^{-1}\ ] ] where @xmath50 is the landau - migdal parameter @xcite , @xmath51 is the nuclear matter density and @xmath52 is the lindhard function accounting for the direct and crossed contributions of @xmath44 and @xmath45 excitations with the normalization of the appendix of ref . @xcite . left panel : the opep in the vacuum ( solid curve ) and at normal @xmath53 @xmath54 nuclear matter density ( dashed curve ) . right panel : opep at several densities . , title="fig : " ] left panel : the opep in the vacuum ( solid curve ) and at normal @xmath53 @xmath54 nuclear matter density ( dashed curve ) . right panel : opep at several densities . , title="fig : " ] the opep in the momentum space takes the form @xmath55 where we have defined @xmath56 where @xmath57 is the pion propagator in the medium @xmath58 and @xmath59 stands for a monopole form factor @xmath60 with the cut off scale @xmath61 gev , @xmath62 is a momentum transfer and @xmath63 . note , that opep depends on the real part of the polarization operator only , since for @xmath64 one has @xmath65 . the well - known vacuum @xmath1 amplitude is recovered in eq . ( [ wopep ] ) by setting @xmath66 or in the limit @xmath67 . our results for @xmath68 are presented in fig . [ opepmedium ] ( left ) . the standard vacuum behavior is shown by the solid curve . the dashed curve represents the modified opep at normal nuclear matter density were we observe an additional strong increase of the strength associated with the attractive excitation of @xmath44 and @xmath69 collective states , with @xmath44 playing a dominant role . their individual contributions are shown by dot - dot - dashed and dot - dashed curves , respectively . the opep with short range correlations ( solid curves ) as a function of the nuclear matter density . the left and right panels are the central and tensor parts , respectively.,title="fig : " ] the opep with short range correlations ( solid curves ) as a function of the nuclear matter density . the left and right panels are the central and tensor parts , respectively.,title="fig : " ] the analytic properties of the in - medium pion propagator which we use here may be verified by using the dispersion representation for the green function @xmath70 in this case the opep of eq . ( [ wopep ] ) can be written in terms of the absorptive part of the pion propagator only @xmath71\end{aligned}\ ] ] the causality requires that both equations must produce the same result . indeed as one can see in fig . [ opepmedium ] the curves calculated with dispersive and absorptive parts of the pion propagator are practically indistinguishable . we would like to note that a strong modification of the opep at finite baryonic density observed here is not new and was predicted long time ago by migdal @xcite . there it was also shown that in - medium modified opep helps to explain the unnatural parity states in finite nuclei , for instance , the shift of @xmath72 state in closed shell nuclei . in the right panel of fig . [ opepmedium ] we show our combined plot for a few densities @xmath73 and @xmath74 . here , we also show the sensitivity of our results to the value of the landau - migdal parameter @xmath75 . we find that the increase of @xmath75 from 0.6 to 0.7 makes the opep softer . this fact suggests that the proper treatment of the @xmath1 short range correlations is important for understanding the in - medium properties of the opep . at this point we would like to mention that in a realistic calculation one will have to add strong repulsive forces at short distances . this can be done in a straightforward way using any of many body schemes discussed in the introduction . the correlations of this part of the interaction would effectively modulate the @xmath5 exchange interaction , introducing the correlation parameter @xmath75 @xcite . the denominator in eq . ( [ inmediumgf ] ) takes into account this effect between @xmath48-wave bubbles in the diagrams of fig . [ opepmedimgraphs ] , but not between the external nucleon and the contiguous bubble . to account for this we make the separation between the longitudinal and transverse parts of the pion effective interaction @xcite @xmath76 where @xmath77 is the cartesian component of the unit vector @xmath63 and @xmath78f({\mbox{\boldmath $ q$}})^2\ ] ] @xmath79 when we perform the sum of diagrams in fig . [ opepmedimgraphs ] then we get @xmath80\sigma_1^i \sigma^j_2 { \mbox{\boldmath $ \tau$}}_1 { \mbox{\boldmath $ \tau$}}_2 \nonumber \\ & = & w_0(q,\rho ) { \mbox{\boldmath $ \sigma$}}_1 { \mbox{\boldmath $ \sigma$}}_2 { \mbox{\boldmath $ \tau$}}_1 { \mbox{\boldmath $ \tau$}}_2 + w_2(q,\rho ) ( { \mbox{\boldmath $ \sigma$}}_1 \hat{{\mbox{\boldmath $ q$ } } } { \mbox{\boldmath $ \sigma$}}_2 \hat{{\mbox{\boldmath $ q$ } } } - \frac{1}{3}{\mbox{\boldmath $ \sigma$}}_1 { \mbox{\boldmath $ \sigma$}}_2 ) { \mbox{\boldmath $ \tau$}}_1 { \mbox{\boldmath $ \tau$}}_2 \end{aligned}\ ] ] @xmath81 \\ w_2(q,\rho)&= & \frac{\mathcal{v}_l(q)}{1-\mathcal{u}(q,\rho)\mathcal{v}_l(q ) } - \frac{\mathcal{v}_t(q)}{1-\mathcal{u}(q,\rho)\mathcal{v}_t(q)}\end{aligned}\ ] ] where we have explicitely separated the central and the tensor parts of the interaction . in fig . [ opepmediummodif ] we show the results for the central @xmath82 ( left panel ) and tensor @xmath83 ( right panel ) parts ( omitting the spin - isospin operators ) as a function of the baryonic density . note that in the limit @xmath84 , @xmath85 where @xmath86 is given by eq . [ wopep ] . in this section we turn to the dynamics of the two pion system in the nuclear medium . our main interest here is the propagation of two @xmath47-wave pion pairs . as it is well known in the @xmath47-wave scattering the use of the proper unitarization schemes lead to the generation of the @xmath0-meson . later on we will use this result for the in - medium @xmath1 interaction mediated by exchange of two correlated pions in the scalar - isoscalar @xmath0-meson channel . the unitary series representing @xmath4 scattering amplitude . ] for the @xmath87 scattering process , defined by the cartesian isospin indices @xmath88 , the use of the standard @xmath15pt procedure in expanding the @xmath89 of eq . ( [ lpipi ] ) to order @xmath90 results in the tree level contact interaction @xmath91 where @xmath92 and @xmath93 is the off - shell part of the invariant @xmath4 amplitude . at this order of the pion field expansion the isoscalar @xmath47-wave @xmath4 partial amplitude ( @xmath94 ) is obtained from the standard decomposition @xmath95 where @xmath96 are the legendre polynomials and @xmath97 accounts for the statistical factor occurring in states with identical particles : @xmath98 for @xmath43 in the unitary normalization of the states @xcite . the tree level scalar - isoscalar @xmath4 scattering amplitude is @xmath99 in eq . ( [ vpipi00 ] ) the off shell part depends on choice of @xmath13 and is equal to zero for on mass shell pions . following ref . @xcite and using the bethe - salpeter equation we unitarize the @xmath47-wave @xmath4 scattering amplitude ( see fig . [ unitseries ] ) @xmath100^{-1},\ ] ] where @xmath101 is a scalar two - pion loop function @xmath102 ( k^2-m_{\pi}^2+i0^+)}\ ] ] where @xmath103 and @xmath104 . the @xmath105 function is analytic with a cut along the positive real axis starting at the @xmath4 threshold . note that , eq . ( [ tpipiunitary ] ) contains a pole in the second riemann sheet corresponding to the @xmath0-meson with mass and width @xmath106 mev . renormalization of the in - medium @xmath4 amplitude including @xmath48-hole and @xmath3-hole excitations . ] in the nuclear medium the @xmath47-wave @xmath4 scattering amplitude and therefore the @xmath0-meson get renormalized and explicit calculations were done in refs . the diagrams at one loop level are shown in fig . [ renormpipiinmedium ] . for instance , the amplitudes corresponding to the insertion of fermion bubbles in the upper meson line are given by @xmath107 ^ 2 \\ v_{\mbox{\tiny ii}}^{up}(s ) & = & \frac{-i2}{3f_{\pi}^2 } \left ( \frac{d+f}{2 f_{\pi } } \right)^2 \int \frac{d^4 k}{(2 \pi)^4 } { \mbox{\boldmath $ k$}}^2 \mathcal{u}(k ) d_{\pi}(k ) d_{\pi}(p - k ) \big [ v_{on}(s ) + \frac{1}{3 f_{\pi}^2 } \sum_i \lambda_i \big ] \\ v _ { \mbox{\tiny iii}}^{up}(s ) & = & \frac{i}{9f_{\pi}^4 } \left ( \frac{d+f}{2 f_{\pi } } \right)^2 \int \frac{d^4 k}{(2 \pi)^4 } { \mbox{\boldmath $ k$}}^2 \mathcal{u}(k ) d_{\pi}(p - k)\end{aligned}\ ] ] as was shown in refs . @xcite , in the center of mass frame of the two pions the off shell part ( @xmath108 terms ) in @xmath109 cancels exactly the @xmath110 and @xmath111 terms . and one is left only with the diagrams of type ( i ) but with the on shell @xmath4 amplitude @xmath112 it is straightforward to iterate the @xmath44 and @xmath45 excitations in fig . [ renormpipiinmedium ] ( i ) and the loop function , @xmath113 , is given by @xmath114 where @xmath115 is in - medium modified scalar loop integral @xmath116 using the spectral representation for the in - medium pion propagators , eq . ( [ inmediumgf ] ) , we get for the loop function @xmath117 \int \limits_{x}^{\infty } d y \big[\mbox{im } \tilde{d}_{\pi}(y - x,{\mbox{\boldmath $ p$}}-{\mbox{\boldmath $ k$}})\big ] \nonumber \\ \times \frac{y}{(p_0+y)(p_0-y+i0^+ ) } \end{aligned}\ ] ] we refer to ref . @xcite where different aspects of @xmath47-wave @xmath4 scattering in the nuclear medium are discussed and also the behavior of the @xmath0-meson mass and width at finite baryonic density is addressed . but here , we would like to illustrate the impact of the nuclear medium on the @xmath4 system . for that consider the imaginary part of the loop function in the @xmath4 center of mass frame @xmath118 with @xmath119 . this situation is relevant for the in - medium @xmath4 scattering and contains the proper information about the dynamics of the pole position of @xmath0 at finite density . our results for the imaginary part of the scalar loop function for several densities are shown in fig . [ imgpipimedium](left panel ) . the solid curve correspond to the vacuum loop function which can be obtained from eq . ( [ gpipigen ] ) by substituting the imaginary part of in - medium pion propagators by their vacuum expressions @xmath120 which is proportional to the two - body phase space of two particles ( pions ) . as one can see in fig . [ imgpipimedium ] ( left ) the effect of the medium is remarkable due to the increase of available for pions phase space because of additional pion decay branches like @xmath44 , @xmath45 and @xmath46 . the imaginary part of the in - medium scalar loop function @xmath121 in the @xmath122 cm frame ( left panel ) with the normal dressing of the pion propagators including @xmath44 , @xmath45 and @xmath46 excitations . the right panel shows the loop function in the space like region @xmath123.,title="fig : " ] the imaginary part of the in - medium scalar loop function @xmath121 in the @xmath122 cm frame ( left panel ) with the normal dressing of the pion propagators including @xmath44 , @xmath45 and @xmath46 excitations . the right panel shows the loop function in the space like region @xmath123.,title="fig : " ] the kinematics relevant for the @xmath1 force , where two nucleons interact by exchange of mesons is defined by the moving reference frame where @xmath123 . in this case the @xmath4 interaction in the medium has to be generalized from the results in @xcite since there @xmath124 but @xmath125 . we repeat all the steps that led to cancellations in @xcite and find again the same cancellations as before but with some remnant @xmath126 dependent terms vanishing in the limit @xmath127 . to evaluate these terms we simplify the calculation assuming @xmath62 relatively small ( this is fine for momenta below the fermi momentum ) . concretely , we assume @xmath128 where @xmath129 is the cut off in the three momentum ( of the order of @xmath130 gev in @xcite ) that one uses to regularize the @xmath121 function . we also assume @xmath131 which implies @xmath132 as it is the case . and 3 ) we expand @xmath133 in terms of @xmath134 and vice - versa to relate different terms . after all this is done we find that the corrections can be taken into account by means of the change in the lindhard function @xmath135\ ] ] the expression in brackets in eq . ( [ umodif ] ) is @xmath136 for very small @xmath137 and also @xmath136 for @xmath138 . hence with very good approximation we can take the bracket equal to unity and thus there are no other corrections to be done to the result of @xcite except the obvious one of changing @xmath139 . for these value of @xmath140 the @xmath121 is only real , contrary to the case studied in @xcite . results for @xmath141 can be seen in fig . [ imgpipimedium ] ( right panel ) for different nuclear densities @xmath142 , @xmath143 and @xmath74 . we can see that the corrections are sizable particularly at small values of @xmath144 . the correlated two pion exchange ( crtpe ) in the scalar - isoscalar channel , the equivalent to a @xmath0 exchange in meson exchange models @xcite , or correlated two pion exchange in the dispersion relations in @xcite was studied in @xcite within the context of chiral lagrangians . one starts from the diagrams of fig . [ ctpevacuum.eps ] , where the @xmath145 scattering shows the off shell ambiguities . to avoid these ambiguities with isoscalar exchange it was stated in ref . @xcite that one must include the subset of diagrams of fig . [ setcans ] to find cancellations of the off shell @xmath4 isoscalar amplitude . this statement was rigorously verified in ref . @xcite , where it was shown that the consideration of these subset of chiral diagrams , fig . [ setcans ] , including the contact @xmath41 interactions , results in the cancellation of the off - shell part of the @xmath4 amplitude , and the on - shell part of the @xmath4 amplitude can be factorized out from the loop integrals . one step forward was given in @xcite , where iteration of the @xmath4 interaction , through the bethe - salpeter equation , was done by means of which a simple analytical expression was obtained for the correlated two pion exchange in the scalar - isoscalar channel @xmath146^{-1}\ ] ] where @xmath147 in the @xmath1 c.m . the vertex function @xmath148 for the triangle loop with two mesons and one baryon propagator , including @xmath149 and @xmath3 intermediate states , is evaluated in @xcite using a cut off in @xmath150 of about 1 gev . note that , the bracket in eq . ( [ vcrtpesigma ] ) contains a pole in the @xmath140-channel corresponding to the @xmath0-meson . this restores the relation to the @xmath0-meson exchange , which now enters the formalism as a dynamical resonance in the @xmath4 system @xcite ( see also related discussions in refs . @xcite ) . diagrams representing the correlated two pion exchange with @xmath1 , @xmath151 and @xmath152 intermediate states . ] the set of diagrams which cancel the off - mass shell part of the scalar - isoscalar correlated two - pion exchange . ] the in - medium diagrams involving the @xmath48-wave pion dressing and vertex corrections . the diagrams with the intermediate @xmath3 states are not shown . ] the diagrams responsible for the renormalization of crtpe in the nuclear medium are shown in fig . [ corrtpecrmedium ] . there , in analogy to what was done for the @xmath4 interaction we include @xmath5 selfenergy corrections as well as vertex corrections . we find that the cancellation of the off shell part of the vacuum @xmath0-exchange discussed in @xcite is also exact at finite baryonic density but for zero momentum transfer only ( see appendix b ) . the results of the derivation for @xmath153 can be summarized as follows : \1 ) the expressions for the triangle vertex functions @xmath154 and @xmath155 of ref . @xcite are obtained in the same way replacing the two free pion propagators by the renormalized ones . \2 ) the lindhard function entering the pion selfenergy is changed to @xmath156\ ] ] to account for the corrections obtained at @xmath153 . these corrections are negligible for small @xmath144 and for @xmath157 of the order of @xmath158 , hence moderate in all cases . \3 ) the final expression for the potential @xmath159 is given by eq . ( [ vcrtpesigma ] ) , which accounts for the @xmath4 rescattering , by substituting @xmath160 by @xmath161 of eq . ( [ gpipigen ] ) and taking the expression for the in - medium vertex function @xmath162 where @xmath163 \nonumber \\ \nonumber \\ & & \times \int \limits_0^{\infty } dy \frac{x+y+e({\mbox{\boldmath $ k$}})-m_n } { ( x+y)(x+e({\mbox{\boldmath $ k$}})-m_n)(y+e({\mbox{\boldmath $ k$}})-m_n ) } \big[\mbox{im } \ , \tilde{d}_{\pi}(y,{\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$}})\big]\end{aligned}\ ] ] @xmath164 \nonumber \\ \nonumber \\ & & \times \int \limits_0^{\infty } dy \frac{x+y+e_{\delta}({\mbox{\boldmath $ k$}})-m_n } { ( x+y)(x+e_{\delta}({\mbox{\boldmath $ k$}})-m_n ) ( y+e_{\delta}({\mbox{\boldmath $ k$}})-m_n ) } \big[\mbox{im } \ , \tilde{d}_{\pi}(y,{\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$}})\big ] \end{aligned}\ ] ] the coupling constants @xmath165 are defined by @xmath166 the momentum space unitary @xmath0 meson exchange @xmath1 potential at finite baryon density . left panel without short range correlations . right panel with short range correlations.,title="fig : " ] the momentum space unitary @xmath0 meson exchange @xmath1 potential at finite baryon density . left panel without short range correlations . right panel with short range correlations.,title="fig : " ] there is still one more correction to be done in order to account for short range correlations . so far we have replaced the free pion propagator by the renormalized one of eq . ( [ inmediumgf ] ) . however , since the interaction between the @xmath44 bubble and the external nucleons is affected by correlations one should take @xmath167 ( see eq . ( [ vl ] ) ) instead of @xmath168 ^ 2 d_{\pi}$ ] . thus we would get the series @xmath169 where @xmath170 and @xmath171 are the free and dressed pion propagators , respectively . hence , the expressions for @xmath154 and @xmath155 get modified by including inside the @xmath172 integral the factors @xmath173 our results for @xmath0-meson exchange in the momentum space are shown if fig . [ sigmaexchmediummod ] for both cases , no short range correlations ( left ) and with short range correlations ( right ) in this section we consider another sort of intermediate distance contributions to the @xmath1 force generated by the uncorrelated two pion exchange . the material presented here for the vacuum @xmath1 scattering is standard and we merely generalize it to the nuclear medium . the planar box diagrams involving the nucleon and @xmath3 intermediate states . ] in the perturbative expansion of the @xmath1 force we must to take into account the planar and crossed box diagrams shown in figs . [ planarmedium ] and [ crossedmedium ] , respectively . we will discuss the scalar - isoscalar part of this contributions only . the contribution of the isovector exchange is small and can be found , for instance , in ref . @xcite in vacuum , the expression for the planar box diagrams , fig . [ planarmedium ] , including the nucleon pole and @xmath151 and @xmath152 intermediate states reads @xmath174 \big[{\mbox{\boldmath $ \sigma$}}^{(1)}_n \cdot { \mbox{\boldmath $ k$}}\big]^{\dagger } \big[{\mbox{\boldmath $ \sigma$}}^{(2)}_m \cdot ( { \mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$}})\big ] \big[{\mbox{\boldmath $ \sigma$}}^{(2)}_m \cdot { \mbox{\boldmath $ k$}}\big]^{\dagger } \ , \big\{\mathcal{r}_{nm}^{(p)}({\mbox{\boldmath $ k$ } } , { \mbox{\boldmath $ q$ } } ) \ , i_{nm}^{(p)}\big\}\ ] ] in eq . ( [ planar ] ) the sum over @xmath175 is assumed and the spin transition operators are @xmath176 , @xmath177 . the coupling constants @xmath165 are defined in eq . ( [ kappaconstants ] ) . the isospin factors are given by @xmath178 for which we find @xmath179 the function @xmath180 in eq . ( [ planar ] ) contains the integration over the time - like component of the four vector @xmath181 . @xmath182 where @xmath170 and @xmath183 is the pion and nonrelativistic baryon propagators , respectively . @xmath183 is given by @xmath184 this integration can be carried out explicitely @xmath185[e_m - e]\big)(e_n+e_m-2 e ) \nonumber\\ & & + \big(\omega_1 + \omega_2\big ) \big(2\omega_1\omega_2+[e_n+e_m-2 e]^2\big)\big ] \nonumber\\ & & \times \frac{1}{\omega_1\omega_2 ( \omega_1 + \omega_2 ) } \left\{\frac{1 } { e_n+e_m- 2 e - i 0^+}\right\ } \nonumber\\ & & \times \frac{m_n}{e_n [ \omega_1 + e_n - e][\omega_2 + e_n - e ] } \nonumber\\ & & \times \frac{m_m}{e_m [ \omega_1 + e_m - e][\omega_2 + e_m - e ] } \end{aligned}\ ] ] recall that @xmath64 , @xmath186 , @xmath187 in @xmath1 c.m . frame and in eq . ( [ rnmfull1 ] ) @xmath188 are the on - shell energies of intermediate pions and baryons with c.m . energy of the initial nucleons @xmath189 . in the nuclear matter @xmath190 , and the expression for @xmath191 is obtained by using the dispersion representation for the in - medium pion propagator @xmath192 where @xmath193 is given by eq . ( [ rnmfull1 ] ) with substitution @xmath194 and @xmath195 . the crossed box diagrams with nucleon and @xmath3 intermediate states . ] for the crossed box diagrams , fig . [ crossedmedium ] , we have @xmath196 \big[{\mbox{\boldmath $ \sigma$}}^{(1)}_n \cdot { \mbox{\boldmath $ k$}}\big]^{\dagger } \big[{\mbox{\boldmath $ \sigma$}}^{(2)}_m \cdot { \mbox{\boldmath $ k$}}\big ] \big[{\mbox{\boldmath $ \sigma$}}^{(2)}_m \cdot({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$}})\big]^{\dagger } \ , \big\{\mathcal{r}_{nm}^{(c)}({\mbox{\boldmath $ k$ } } , { \mbox{\boldmath $ q$ } } ) \ , i_{nm}^{(c)}\big\}\ ] ] the isospin factors are the following @xmath197 and the vacuum expression for @xmath198 is given by @xmath199 the integration can be done analytically and our result reads @xmath200 \nonumber\\ & & \times \frac{1}{\omega_1\omega_2 ( \omega_1 + \omega_2 ) } \nonumber\\ & & \times \frac{m_n}{\tilde{e}_n [ \omega_1 + \tilde{e}_n - e][\omega_2 + \tilde{e}_n - e ] } \nonumber\\ & & \times \frac{m_m}{e_m [ \omega_1 + e_m - e][\omega_2 + e_m - e ] } \end{aligned}\ ] ] here @xmath201 , @xmath202 and @xmath203 are defined in eq . ( [ kin ] ) and @xmath204 the corresponding expression for the in - medium crossed box diagrams takes the form @xmath205 the momentum space uncorrelated two - pion exchange at finite baryonic density . the left panel correspond to normal pion dressing without sr correlations and right panel with sr correlations.,title="fig : " ] the momentum space uncorrelated two - pion exchange at finite baryonic density . the left panel correspond to normal pion dressing without sr correlations and right panel with sr correlations.,title="fig : " ] for the generic case of non - vanishing initial momenta the analytical structure of eq . ( [ rnmfull1 ] ) is driven by the term in figure brackets @xmath206 here we have to pay a special attention to the case where two intermediate @xmath1 states @xmath207 appear , because in time ordered perturbation theory these diagrams are generated by iterations of the opep in a lippmann - schwinger equation ( lse ) . considering the @xmath1 intermediate state only , the iterated tpe can be easily identified and comes from the nucleon pole in eq . ( [ rnmfull ] ) ( in the lower half of the complex plane ) corresponding to @xmath208 with @xmath209 @xmath210\ ] ] expanding eq . ( [ nucleonpole ] ) in powers of @xmath211 @xmath212 and inserting the leading order result in eq . ( [ planar ] ) one can get the second order term in the non - relativistic lippmann - schwinger equation . it is instructive to derive the contributions of the two remaining poles ( in lower half - plane ) from the pion propagators @xmath213 after the expansion of this result in powers of @xmath211 we get @xmath214 in this limit eq . ( [ pipoles ] ) cancels exactly the corresponding crossed box diagram with two intermediate nucleons in the isoscalar channel . indeed , considering the crossed box diagram with two intermediate nucleons , the leading term of @xmath215 in the @xmath211 expansion is given by @xmath216 note that for the isovector @xmath4 exchange because of the different sign of the @xmath217 term in @xmath218 , @xmath219 in eqs . ( [ ipl ] ) and ( [ icr ] ) these two contributions would add . the cancellation discussed above hold also in the nuclear medium . the result of the vacuum scalar - isoscalar @xmath1 force generated by the planar and crossed box diagrams , with the nucleon pole diagrams excluded , is shown in the left panel of fig . [ uncormainmedium ] by the solid curve . it is in agreement with ref . @xcite where it was shown that the consistent use of the cut off regularization in both the correlated two pion exchange , and uncorrelated two pion exchange , together with the contribution of a repulsive @xmath220-exchange lead to a scalar - isoscalar potential in good agreement with the argonne @xcite potential in the whole range of relevant distances . the corresponding results for the nuclear matter are shown in fig . [ uncormainmedium ] ( left ) for three densities @xmath142 , @xmath143 and @xmath74 . qualitatively the behavior is similar to the vacuum case but with a strong enhancement toward the small momentum transfer . as we have seen this feature is generic in present calculations . uncorrelated @xmath4 exchange becomes extremely attractive at intermediate distances . the spin sum over intermediate baryon states of the @xmath221 operator in eqs . ( [ planar ] ) and ( [ crossed ] ) gives in the scalar channel @xmath222 ^ 2\ ] ] where @xmath223 @xmath224 , @xmath225 @xmath226 and @xmath227 @xmath228 . now we again wish to take into account the short range correlations @xmath229 ^ 2 \nonumber \\ \longrightarrow \tilde{\mathcal{w}}_l(k+q ) \tilde{\mathcal{w}}_l(k ) \left [ ( \widehat{{\mbox{\boldmath $ k$}}+ { \mbox{\boldmath $ q$ } } } ) \cdot \hat{{\mbox{\boldmath $ k$ } } } \right]^2 \nonumber \\ + \left [ \tilde{\mathcal{w}}_l(k+q)\tilde{\mathcal{w}}_t(k ) + \tilde{\mathcal{w}}_l(k ) \tilde{\mathcal{w}}_t(k+q)\right ] \left\ { 1 - \left [ ( \widehat{{\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } } ) \cdot \hat{{\mbox{\boldmath $ k$}}}\right]^2 \right\ } \nonumber \\ + \tilde{\mathcal{w}}_t(k+q ) \tilde{\mathcal{w}}_t(k ) \left\ { 1 + \left [ ( \widehat{{\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } } ) \cdot \hat{{\mbox{\boldmath $ k$ } } } \right]^2 \right\ } \nonumber \\\end{aligned}\ ] ] where @xmath230 where @xmath231 is the lindhard function and @xmath167 and @xmath232 are given by eq . ( [ vl ] ) and ( [ vt ] ) , respectively . by comparing the results in fig . [ opepmedium ] and fig . [ opepmediummodif ] we observe that the effect of correlations has been essential and reduces drastically the medium effects found in fig . [ opepmedium ] without corrections . we observe that the medium corrections weaken the strength of the central part of the opep . on the other hand the effect of the medium corrections on the tensor part are more moderate . the @xmath0 exchange presents similar features as we can see in fig . [ sigmaexchmediummod ] . there we see that the medium effects in the absence of short range correlations are rather large and increase the strength of the interaction in about a factor two at @xmath233 . however , as soon as the correlations are taken into account the medium modifications are reduced drastically and at @xmath233 just reduce the strength in about 25 percent . the medium effects in the uncorrelated two pion exchange are also relatively small , with respect to the size of the interaction in the vacuum , see fig . 14 ( right ) . once again the consideration of the correlations has been essential and reducing the size of the medium effects . however , the fact that the strength of this interaction is bigger than that of the correlated two pion exchange makes the absolute correction relevant and of the same size as the corrections discussed before . the modification of the @xmath1 interaction at normal nuclear matter density in momentum space ( left panel ) and in configuration space ( right panel ) . shown are the central part of the opep ( solid curve ) , tensor part of the opep ( dashed curve ) , correlated @xmath0-meson exchange ( dot - dashed curve ) and uncorrelated @xmath4 exchange ( dot - dot - dashed curve).,title="fig : " ] the modification of the @xmath1 interaction at normal nuclear matter density in momentum space ( left panel ) and in configuration space ( right panel ) . shown are the central part of the opep ( solid curve ) , tensor part of the opep ( dashed curve ) , correlated @xmath0-meson exchange ( dot - dashed curve ) and uncorrelated @xmath4 exchange ( dot - dot - dashed curve).,title="fig : " ] since we have taken also the effect of correlations in the free part of the interaction and any realistic calculation of the binding energy of matter will explicitly account for these correlations , the use of our modified in - medium potentials would lead to double - counting if any of these methods is used . for this purpose the relevant results from this work should be the differences between the in - medium potential and the free one . in this case we start already with one bubble and the correlations account for the repulsion between the external leg and the one in the bubble . the two external legs still have to be correlated and this will be done with the use , for instance , of the ordinary bethe golstone equation . after this discussion of the medium effects in the different terms , we show the differences in momentum space between the medium and free parts of the different terms in fig . [ uncormainmediummm ] ( left ) . the results are shown at @xmath233 and we see that these corrections are moderate , but they could have a relevant role in the binding of matter . in order to have a qualitative idea of the relevance of these corrections to the potentials we rewrite them in coordinate space and show these results in fig . [ uncormainmediummm ] ( right ) . the effects seem sizeable at short distances , but this will be irrelevant in any realistic calculation of binding energies since the consideration of the short range correlations between the external legs will make this interaction inoperative . more interesting is the strength of the corrections around 1 fm , and there we see that all them are relatively small , of the order of 20 mev or less , the biggest one being the central part of the ope . however , given the size of the empirical scalar isoscalar attraction @xcite which is of the order of 20 mev at intermediate densities , the corrections found here are not negligible . it would be thus interesting to see the effects of the results obtained here in observables like the binding of nuclear matter and other properties , which we hope to stimulate with the results obtained here . in this paper we studied the modification of the one pion exchange , as well as two pion exchange potential inside a nuclear medium . for this purpose we separated the two pion exchange into an uncorrelated two pion exchange and the correlated two pion exchange . we study both in the scalar isoscalar channel , which is by large the most important one generated by this interaction . the correlated two pion exchange gave rise to the equivalent of the @xmath0 exchange in other models and we studied the medium modification to it . on the other hand , for the uncorrelated two pion exchange we have followed a traditional approach in which only terms with at least one intermediate @xmath3 state are considered . one of the important findings here was the effect played by the nn short range correlations which drastically moderated the medium corrections to the potential . in the absence of these , the corrections where unrealistically large . yet , even if relatively moderate , the medium corrections found in this paper are sizeable enough to have relevant repercussions in the binding and other properties of nuclear matter and we would like to encourage calculations in this direction . we would like to acknowledge a. ramos for a critical reading of the manuscript and useful suggestions . this work is partly supported by dgicyt contract number bfm2003 - 00856 , and the e.u . euridice network contract no . hprn - ct-2002 - 00311 . this research is part of the eu integrated infrastructure initiative hadron physics project under contract number rii3-ct-2004 - 506078 . the set of feynman diagrams shown in fig . [ 3pinnvertcharge ] appear in the construction of the @xmath40 transition amplitude . the corresponding @xmath41 vertex functions are given by contact terms appearing in the construction of the @xmath40 transition amplitude . ] i l_b & = & \frac{d+f}{2 } \frac{2 \sqrt{2}}{12 f_{\pi}^3 } \big [ { \mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ q$}}_2 + 2 { \mbox{\boldmath $ k$}}_2 - { \mbox{\boldmath $ k$}}_1 ) \big ] \\ i l_c & = & \pm \frac{d+f}{2 } \frac{2}{12 f_{\pi}^3 } \big [ { \mbox{\boldmath $ \sigma$}}(2 { \mbox{\boldmath $ q$}}_1 + { \mbox{\boldmath $ k$}}_1 + { \mbox{\boldmath $ k$}}_2 ) \big ] \\ i l_d & = & \frac{d+f}{2 } \frac{2 \sqrt{2}}{12 f_{\pi}^3 } \big [ { \mbox{\boldmath $ \sigma$}}(2{\mbox{\boldmath $ q$}}_1 + { \mbox{\boldmath $ k$}}_1 + { \mbox{\boldmath $ k$}}_2 ) \big ] \\ i l_e & = & \frac{d+f}{2 } \frac{2 \sqrt{2}}{12 f_{\pi}^3 } \big [ { \mbox{\boldmath $ \sigma$}}(2{\mbox{\boldmath $ q$}}_2 + { \mbox{\boldmath $ k$}}_1 + { \mbox{\boldmath $ k$}}_2 ) \big ] \\ i l_f & = & 0\end{aligned}\ ] ] in this appendix we demonstrate the cancellation of the off shell part of the correlated two pion exchange in the scalar isoscalar channel . for simplicity we consider the nucleon intermediate states only . the generalization to @xmath151 and @xmath152 is straightforward . the diagrams are shown in fig . [ corrtpecrmedium ] and corresponding amplitudes are given by @xmath235 \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[{\mbox{\boldmath $ k$}}^2 \mathcal{u}(k)\big ] s_f(p_1-k ) d^2(k ) d(k+q ) \\ -i t_a^{(1 ) } & = & 2 \left(\frac{d+f}{2 f_{\pi}}\right)^4 \frac{1}{f_{\pi}^2 } \big[v_{n\pi\pi}(t ) \big ] \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[{\mbox{\boldmath $ k$}}^2 \mathcal{u}(k)\big ] s_f(p_1-k ) d(k ) d(k+q ) \\ -i t_a^{(2 ) } & = & 2 \left(\frac{d+f}{2 f_{\pi}}\right)^4 \frac{1}{f_{\pi}^2 } \big[v_{n\pi\pi}(t ) \big ] \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[{\mbox{\boldmath $ k$}}^2 \mathcal{u}(k)\big ] s_f(p_1-k ) d^2(k ) \\ \nonumber\end{aligned}\ ] ] @xmath236 \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[({\mbox{\boldmath $ k$}}+ { \mbox{\boldmath $ q$}})^2 \mathcal{u}(k+q)\big ] s_f(p_1-k ) d(k ) d^2(k+q ) \\ -it_b^{(1 ) } & = & 2 \left(\frac{d+f}{2 f_{\pi}}\right)^4 \frac{1}{f_{\pi}^2 } \big[v_{n\pi\pi}(t ) \big ] \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[({\mbox{\boldmath $ k$}}+ { \mbox{\boldmath $ q$}})^2 \mathcal{u}(k+q)\big ] s_f(p_1-k ) d(k ) d(k+q ) \\ -it_b^{(2 ) } & = & 2 \left(\frac{d+f}{2 f_{\pi}}\right)^4 \frac{1}{f_{\pi}^2 } \big[v_{n\pi\pi}(t ) \big ] \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[({\mbox{\boldmath $ k$}}+ { \mbox{\boldmath $ q$}})^2 \mathcal{u}(k+q)\big ] s_f(p_1-k ) d^2(k+q ) \nonumber \\ -i t_c & = & -i t_a(k \to l , q \to - q , p_1 \to p_2 ) \nonumber \\ -i t_d & = & -i t_b(k \to l , q \to - q , p_1 \to p_2)\end{aligned}\ ] ] where @xmath237 is the triangle loop integral @xmath238 \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ l$}}\big ] s_f(p_2-l ) d(l - q ) d(l)\ ] ] and the on - mass - shell @xmath4 scattering amplitude is @xmath239 @xmath240 \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big [ ( 3 { \mbox{\boldmath $ k$}}{\mbox{\boldmath $ q$}}+ 2 { \mbox{\boldmath $ k$}}^2 ) \mathcal{u}(k)\big ] s_f(p_1-k ) d(k ) d(k+q ) \\ -i t_f & = & \left(\frac{d+f}{2 f_{\pi}}\right)^4 \frac{1}{f_{\pi}^2 } \big[v_{n\pi\pi}(t ) \big ] \\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big [ ( { \mbox{\boldmath $ q$}}^2 - 2 { \mbox{\boldmath $ k$}}^2 - { \mbox{\boldmath $ k$}}{\mbox{\boldmath $ q$ } } ) \mathcal{u}(k+q)\big ] s_f(p_1-k ) d(k ) d(k+q ) \nonumber \\ t_g & = & -i t_e(k \to l , q \to - q , p_1 \to p_2 ) \\ -i t_h & = & -i t_f(k \to l , q \to - q , p_1 \to p_2 ) \\ -i t_i & = & \left(\frac{d+f}{2 f_{\pi}}\right)^4 \frac{1}{f_{\pi}^2 } \big[v_{n\pi\pi}(t ) \big ] \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ q$}}-2{\mbox{\boldmath $ k$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[{\mbox{\boldmath $ k$}}^2 \mathcal{u}(k)\big ] s_f(p_1-k ) d^2(k ) \\ -i t_j & = & - \left(\frac{d+f}{2 f_{\pi}}\right)^4 \frac{1}{f_{\pi}^2 } \big[v_{n\pi\pi}(t ) \big ] \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ q$}}+2{\mbox{\boldmath $ k$ } } ) \big ] \big[{\mbox{\boldmath $ k$}}^2 \mathcal{u}(k)\big ] s_f(p_1+q - k ) d^2(k ) \\ -i t_k & = & -i t_i(k \to l , q \to - q , p_1 \to p_2 ) \\ -i t_l & = & -i t_j(k \to l , q \to - q , p_1 \to p_2 ) \end{aligned}\ ] ] in the present case @xmath64 and for @xmath241 the amplitude @xmath242 cancels exactly @xmath243 . the same cancelation is found for @xmath244 and @xmath245 . the sum of @xmath246 and @xmath247 is given by @xmath248 \nonumber\\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[{\mbox{\boldmath $ k$}}{\mbox{\boldmath $ q$}}\ , \mathcal{u}(k)\big ] s_f(p_1-k ) d(k ) d(k+q)\end{aligned}\ ] ] the sum of @xmath249 and @xmath250 takes the form @xmath251 \\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) { \mbox{\boldmath $ q$}}\mathcal{u}(k+q)\big ] s_f(p_1-k ) d(k ) d(k+q ) \nonumber\end{aligned}\ ] ] from this the sum of four diagrams is given by @xmath252 \\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$ } } ) \big ] \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big [ { \mbox{\boldmath $ q$}}^2 \mathcal{u}(k+q ) + { \mbox{\boldmath $ k$}}{\mbox{\boldmath $ q$}}\big\{\mathcal{u}(k+q)-\mathcal{u}(k)\big\ } \big ] s_f(p_1-k ) d(k ) d(k+q ) \nonumber\end{aligned}\ ] ] alternatively , the spin flip parts of eqs . ( [ mm ] ) and ( [ mn ] ) can be canceled if we change in eq . ( [ mn ] ) @xmath253 and using again the fact that @xmath64 and the properties of the lindhard function @xmath254 we get @xmath255 \\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[{\mbox{\boldmath $ \sigma$}}{\mbox{\boldmath $ k$}}\big ] \big[{\mbox{\boldmath $ \sigma$}}({\mbox{\boldmath $ k$}}+{\mbox{\boldmath $ q$}})\big ] \big [ { \mbox{\boldmath $ k$}}{\mbox{\boldmath $ q$}}\mathcal{u}(k)\big ] s_f(p_1-k ) d(k ) d(k+q ) \nonumber\end{aligned}\ ] ] and @xmath256 \\ & & \times \int \frac{d^4 k}{(2 \pi)^4 } \big[({\mbox{\boldmath $ k$}}^2 + { \mbox{\boldmath $ k$}}{\mbox{\boldmath $ q$}}){\mbox{\boldmath $ k$}}{\mbox{\boldmath $ q$}}\big ] \mathcal{u}(k ) s_f(p_1-k ) d(k ) d(k+q ) \nonumber\end{aligned}\ ] ] finally @xmath257 \\ & & \times \int \frac{d^4 l}{(2 \pi)^4 } \big[({\mbox{\boldmath $ l$}}{\mbox{\boldmath $ q$}}-{\mbox{\boldmath $ l$}}^2){\mbox{\boldmath $ l$}}{\mbox{\boldmath $ q$}}\big ] \mathcal{u}(l ) s_f(p_2-l ) d(l ) d(l - q ) \nonumber\end{aligned}\ ] ] as one can see in the limit @xmath241 we get an exact cancellation of the off mass shell part of the @xmath4 amplitude and only @xmath258 and @xmath259 are left . a. fabrocini , and s. fantoni , in _ first course on condensed matter _ , acif series , vol.viii , d. prosperi , s. rosati , and g. violini eds . , world scientific , singapore ( 1986)87 . s. fantoni and a. fabrocini , in _ microscopic quantum many - body theories and their applications _ edited by j. navarro and a. polls , lecture notes in physics , vol . 510 ( springer - verlag , berlin 1998 ) . e. oset , h. toki and w. weise , phys . * 83 * ( 1982 ) 281 . w. r. gibbs and b. f. gibson , ann . nucl . part . * 37 * ( 1987 ) 411 . t. e. o. ericson and w. weise , `` pions and nuclei '' , oxford university press , 1988 . c. m. chen , d. j. ernst and m. b. johnson , phys . c * 47 * ( 1993 ) 9 . j. nieves , e. oset and c. garcia - recio , nucl . phys . a * 554 * , 554 ( 1993 ) . m. nuseirat , m. a. k. lodhi and w. r. gibbs , phys . c * 58 * ( 1998 ) 314 . t. s. h. lee and r. p. redwine , ann . nucl . part . sci . * 52 * ( 2002 ) 23 . r. machleidt , k. holinde and c. elster , phys . * 149 * ( 1987 ) 1 . h. muther , a. faessler , m. r. anastasio , k. holinde and r. machleidt , phys . c * 22 * , 1744 ( 1980 ) . t. s. jensen and a. f. miranda , phys . c * 55 * , 1039 ( 1997 ) . o. jaekel , h. w. ortner , m. dillig and c. a. z. vasconcellos , nucl . a * 511 * , 733 ( 1990 ) . s. kamefuchi , l. oraifeartaigh and a. salam , nucl . * 28 * , 529 ( 1961 ) m. m. kaskulov and h. clement , phys . c * 70 * , 014002 ( 2004 ) . m. m. kaskulov and h. clement , phys . c * 70 * , 057001 ( 2004 ) ; h. a. clement , m. m. kaskulov and e. a. doroshkevich , int . j. mod . a * 20 * , 674 ( 2005 ) . d. jido , e. oset and j. e. palomar , nucl . a * 694 * , 525 ( 2001 ) .

searching for new materials with nontrivial topological properties is a forefront research field of condensed matter physics . following the theoretical prediction and experimental discovery of two - dimensional and three - dimensional topological insulators @xcite , the fully - gapped topological insulators and topological superconductors have been classified according to their symmetries @xcite . one focus of recent research is on a class of materials with discrete degeneracy points in the band structure @xcite . these dirac and weyl semimetals are analogies or generalizations of graphene , which is a prototypical two - dimensional ( 2d ) dirac semimetal @xcite . the three - dimensional ( 3d ) weyl semimetals realize the chiral anomaly which results in several novel transport properties , including the anomalous hall effect and the chiral magnetic effect @xcite . the interest in 3d weyl semimetal is soon extended to 2d dirac and weyl semimetals @xcite and 3d dirac and weyl superconductors @xcite . the 2d weyl superconductor , which is also a member of this new family of materials characterized by the presence of weyl fermions , has however till now received very little attention . this is partly because of the difficulty in realizing such a phase . on one hand , accidental band degeneracy is known to be very unlikely in 2d @xcite . in a superconductor , at least one particle band and one hole band ( which are related to each other by charge conjugation ) cross each other and thus invert in a certain region of the wave vector space , which makes possible the occurrence of nodal points as the parts of the crossing curve between the particle and hole bands that are not gapped out . however , by opening a full gap the superconductor can usually gain more condensation energy and so naturally occurring nodal superconductors are rare . on the other hand , the realization of weyl nodes requires non - degenerate fermi surface . this requires the breaking of at least one of the time - reversal symmetry and the inversion symmetry @xcite . while the former requires the persistence of magnetic order in the superconducting phase , the latter point to noncentrosymmetric materials . in bulk materials , the competition with other symmetry breaking tendencies makes the superconducting phase less favored as compared to the simple time - reversal invariant and inversion symmetric materials . therefore , the 2d weyl superconductor appears to be the most elusive member of the weyl materials . in this work , we propose that a 2d weyl superconductor can be created in a very inclusive system , a heterostructure consisting of a half - metal ( hm ) with rashba spin - orbital coupling ( rsoc ) and a spin - singlet superconductor . the combination of a single nondegenerate fermi surface @xcite , the transmutation between spin - singlet and spin - triplet pairing by the rsoc @xcite , and the possible existence of nodes in the order parameter of the spin - singlet pairing , make the hm layer of the proposed heterostructure a unique platform to realize the 2d weyl superconductor . an important tunability of the system , sketched in fig.1 , is the orientation of magnetization of the hm . earlier , the heterostructure with an @xmath1-wave superconductor was found to give a 2d weyl superconductivity phase ( swsc ) when the spontaneous magnetization of the hm lies along a mirror symmetric direction within the film plane @xcite . in the heterostructure with a @xmath0-wave superconductor ( such as the cuprates ) , on the other hand , a 2d weyl superconductor ( dwsc ) is realized for an arbitrary magnetization orientation ( the case with purely out - of - plane magnetization was also studied by sato and fujimoto in a slightly different setting @xcite ) . the bulk and edge state spectra and their evolution with the magnetization direction show distinct differences from those for the swsc . for example , besides the type - i weyl nodes which are realized in swsc , we also find type - ii weyl nodes in a dwsc . in weyl semimetal , type - i and type - ii weyl nodes are defined separately as weyl point with a point - like fermi surface and weyl point appearing at the contact of electron and hole pockets @xcite . here , we define type - i and type - ii weyl nodes in a superconductor analogously . namely , for type - i weyl node the constant energy surface ( not necessarily the fermi surface ) crossing the weyl point spans a single point , whereas for type - ii weyl node the constant energy surface crossing the weyl point forms an electron - like and a hole - like quasiparticle ` fermi surface ' and connect to each other at the weyl point . the transition between type - i and type - ii weyl nodes is achieved in a dwsc by varying the direction of the spontaneous magnetization of the hm . in addition , for a general magnetization orientation , the state studied consists simultaneously of normal and superfluid components . in terms of the zero - energy quasiparticle spectrum , the varying magnetization orientation drives the system to go through a series of lifshitz transitions , which is usually encountered in a metal in the normal phase @xcite . it is also shown to be qualitatively different from all known types of pairings , including the bcs pairing @xcite , the fflo pairing @xcite , and the breached pairing state ( sarma phase ) @xcite . the organization of this paper is as follows . in sec . ii , we define the model and parameters for the proposed heterostructure . in sec . iii , we report the energy spectra of the system . these include the energy spectra for bulk and a strip of the system , the spectral functions and density of states for the two edges of a strip , the clarification of the nature of proximity - induced superconducting state , and the lifshitz transitions in the quasiparticle spectrum . the transition between type - i and type - ii weyl nodes is also studied in this section . ( the analytical analyses are put to appendix a. ) in sec . iv , we study the symmetry and topological properties of the system . in sec . v , we show the energy spectra for the heterostructure with the hm replaced by an ordinary ferromagnetic material having a partially polarized fermi surface . finally , we summarize our results in sec . vi . [ fig1 ] -wave or @xmath0-wave spin - singlet superconductor and an insulating substrate , viewed laterally . @xmath2 and @xmath3 represent the magnitude and direction of the exchange field ( magnetization ) in the hm , which is assumed to lie on the @xmath4 plane.,title="fig:",width=283,height=113 ] + figure 1 is a schematic plot of the system that we will study . it consists of a hm ( or , highly polarized ferromagnetic metal ) thin film grown on an insulating substrate and capped with a spin - singlet superconductor with @xmath1-wave or @xmath0-wave pairing symmetry . to elucidate the principle more clearly , we consider the simplest model for the system . we describe the hm thin film by a one - orbital model defined on a square lattice . perfect interfaces between the hm and the substrate and the spin - singlet superconductor are assumed , so that we will ignore the effect of impurity and disorder . denoting the basis vector as @xmath5 $ ] , the model hamiltonian for the hm thin film , with a rsoc term induced by the formation of the heterostructure , is @xmath6 , where @xcite @xmath7 @xmath8 is the two - dimensional unit matrix , @xmath9 ( @xmath10 ) are pauli matrices in the spin subspace . @xmath12 is the hopping amplitude and we assume @xmath13 , @xmath14 is the chemical potential , @xmath15 is the strength of the rsoc . in writing down eq.(1 ) , we have assumed the magnetization to be fixed on the @xmath4 plane . in terms of @xmath2 ( @xmath16 ) and @xmath3 ( @xmath17 , with respect to the positive direction of the @xmath18 axis ) , which are separately the magnitude and polar angle of the magnetization ( see fig.(1 ) ) , we have @xmath19 , @xmath20 . to describe a hm with @xmath21 , we set @xmath2 to be of the same order of magnitude as @xmath12 , and is much larger than the externally induced @xmath15 . the chemical potential is tuned to make sure that it crosses only the lower spin - split band of @xmath22 , which amounts to @xmath23 . the proximity - induced superconductivity in the hm arising from coupling with an @xmath1-wave or @xmath0-wave spin - singlet superconductor is described by @xmath24 where @xmath25 @xcite . @xmath26 is a multiplication of the pairing amplitude @xmath27 and a symmetry factor @xmath28 . for the sake of simplicity , we take @xmath29 for @xmath1-wave superconductor and @xmath30 or @xmath31 for @xmath0-wave superconductor @xcite . the two latter cases can be realized by growing the same @xmath0-wave superconductor ( e.g. , cuprates ) on the top of the hm film along two special directions , respectively . the @xmath32-axis , which is the projection of the magnetization plane ( @xmath4 ) onto the plane of the films ( @xmath33 ) in the heterostructure , is along the antinodal direction of the @xmath0-wave pairing gap for @xmath30 and along the nodal direction of the @xmath0-wave pairing gap for @xmath31 . the heterostructure formed by growing the @xmath0-wave superconductor along more general ( less symmetric ) directions , while possible in principle , can be understood in terms of a combination of the results for these two special cases . in the nambu basis , @xmath34 $ ] , the full model is written as @xmath35 , where the bogoliubov - de gennes ( bdg ) hamiltonian with proximity - induced pairing term and rsoc is @xmath36 @xmath37 ( @xmath10 ) are pauli matrices in the nambu space . diagonalizing eq.(3 ) gives the four quasiparticle bands @xmath38 , with @xmath39=@xmath40 , @xmath41 , @xmath42 , @xmath43 and larger @xmath39 means higher energy . in the hm thin film , only the spin - polarized band crossing the chemical potential is important to the low - energy properties of the system . to facilitate the analysis , we construct the low - energy effective model by projecting to the subspace containing only information of the band contributing to the fermi surface . first , we diagonalize the model defined in eq.(1 ) by a unitary transformation @xmath44 . @xmath45 is a diagonal matrix with the two diagonal elements storing the eigenvalues of @xmath22 , @xmath46 where @xmath47 . notice that , @xmath48 ( @xmath47 ) are the electronic energy bands in the normal state ( i.e. , without the proximity - induced pairing term ) and should be distinguished from the quasiparticle bands @xmath38 ( @xmath49 - 2 , -1 , 1 , 2 ) defined as eigenvalues of eq.(3 ) . the unitary matrix @xmath50 containing the eigenvectors of @xmath22 is written as @xmath51 the normalized eigenvectors are chosen as @xmath52 where @xmath47 and @xmath53 . @xmath54 is defined in eq.(4 ) . in the eigenbasis , the pairing term is rewritten as @xmath55 for @xmath56 that we will focus on , only the @xmath57 band contributes to the low - energy properties . we thus retain only this band and the pairing term within this band to construct the low - energy effective model . in the reduced nambu space defined by the basis @xmath58 $ ] , the low - energy effective model is written as @xmath59 the intra - band pairing amplitude is @xmath60 \notag \\ & = & \frac{2\delta_{0}\eta(\mathbf{k})}{d_{-}(\mathbf{k})d_{-}(-\mathbf{k})}[\lambda^{\ast}_{\mathbf{k}}\tilde{m}_{z}(\mathbf{k})+\tilde{m}_{x}(\mathbf{k } ) ] \notag \\ & \equiv&\delta_{0}\eta(\mathbf{k})f(\mathbf{k}),\end{aligned}\ ] ] where we have introduced the abbreviations @xmath61 , @xmath62 , and @xmath63 . the effective paring is therefore determined by two symmetry factors , @xmath28 and @xmath64 . the eigenstates of @xmath65 will be used as approximations for the two low - energy quasiparticle bands , @xmath66 and @xmath67 , in the following analytical analysis . for the heterostructure with @xmath1-wave superconductor and in - plane exchange field ( @xmath68 or @xmath69 ) , the proximity - induced superconducting phase in the hm thin film is an swsc with a single pair of weyl nodes located on the mirror invariant line of the 2d brillouin zone ( bz ) @xcite . it has also been verified that , the swsc is a direct result of an emergent mirror reflection symmetry that exists only for an in - plane exchange field @xcite . the heterostructure with a @xmath0-wave superconductor , however , supports a dwsc phase independent of the @xmath3 angle of the exchange field . this is understandable from the intraband pairing amplitude in eq.(9 ) . namely , because of the presence of the @xmath70 factor , there are always point nodes inherited from those of the @xmath0-wave superconductor . the remaining wave - vector dependence of the effective pairing amplitude , contained in @xmath64 , is the same as that in the heterostructure with @xmath1-wave superconductor . in particular , @xmath71 @xcite . from the positions of the nodes resulting from the @xmath28 factor and from the @xmath64 factor , we know that as @xmath3 varies the number of pairs of weyl nodes vary between two and three for @xmath30 and remains to be two in the whole range of @xmath3 for @xmath31 . we study the energy spectra of the two cases separately in what follows . two kinds of topological phase transitions were identified in the heterostructure with @xmath1-wave superconductor , as we vary @xmath3 @xcite . one is between a fully gapped phase and a bulk gapless phase with ubiquitous local gap . another is between the bulk gapless phase with ubiquitous local gap and the swsc phase with a pair of weyl nodes connecting the two low - energy quasiparticle bands @xmath66 and @xmath67 . for the heterostructure with @xmath0-wave superconductor , the persistence of the weyl nodes for all @xmath3 angles makes exact analogies of the corresponding phase transitions nonexistent . however , we can still define several critical angles around which to expect some qualitative changes in the quasiparticle energy spectrum for the system in bulk and strip configurations . in this respect , for @xmath30 , we can define three critical angles in the range @xmath72 $ ] as follows . the critical angles are defined in terms of the quasiparticle energies of the four weyl nodes and those for two special fermi points . for the model parameters of interest to us ( @xmath73 ) , the positions of the reference wave vectors keep almost unchanged as @xmath3 varies . the four weyl nodes are at @xmath74 with @xmath75 determined by @xmath76.\ ] ] the two special fermi points are chosen as @xmath77 , with @xmath78 determined by @xmath79.\ ] ] the critical angle @xmath80 is determined by the condition @xmath81 , where @xmath82 is the sign function . the critical angle @xmath83 is determined by the condition @xmath84 . making use of the low - energy effective model defined by eqs.(8 ) and ( 9 ) , @xmath80 and @xmath83 are determined approximately by @xmath85 and @xmath86 the third critical angle is @xmath87 , at which the number of weyl nodes changes from two pairs to three pairs . for small @xmath88 , @xmath77 gives the approximate positions of the two additional weyl nodes . we proceed to show the numerical results for the energy spectra of the dwsc in both bulk and strip configurations , for several typical values of @xmath72 $ ] . the bulk energy spectra are shown along several lines with fixed values of @xmath89 , since the bulk quasiparticle spectrum is symmetric with respect to @xmath90 . we focus on a typical set of parameters as @xmath91 , @xmath92 , @xmath93 , @xmath94 , and @xmath95 . a relatively large value of @xmath27 is adopted to see the qualitative features of the spectra more clearly . [ fig2 ] , for a typical set of parameters , @xmath96 , @xmath94 , @xmath95 , @xmath93 . for the bulk system , the dispersions of the @xmath66 and @xmath67 bands are shown along four lines in the 2d bz with fixed @xmath90 values at 0 ( black lines ) , @xmath97 ( blue lines ) , @xmath98 ( red lines ) , and @xmath99 ( green lines ) . @xmath75 is defined by eq.(10 ) . the edge state spectra correspond to a strip with 1500 unit cells along the @xmath32 direction . @xmath100 for ( a ) and ( f ) , @xmath101 for ( b ) and ( g ) , @xmath102 for ( c ) and ( h ) , @xmath103 for ( d ) and ( i ) , @xmath104 for ( e ) and ( j).,title="fig:",width=321,height=566 ] figure 2 shows the quasiparticle energy spectra for @xmath30 in both the bulk [ 2(a ) to 2(e ) ] and strip [ 2(f ) to 2(j ) ] configurations . five values of @xmath3 , including @xmath105 [ 2(a ) and 2(f ) ] , @xmath80 [ 2(b ) and 2(g ) ] , @xmath83 [ 2(c ) and 2(h ) ] , @xmath106 [ 2(d ) and 2(i ) ] , and @xmath107 [ 2(e ) and 2(j ) ] are considered . for the bulk spectra , we plot for each @xmath3 the quasiparticle bands @xmath66 and @xmath67 along four lines in the 2d bz with correspondingly fixed values of @xmath90 , which are @xmath105 , @xmath97 , @xmath98 , and @xmath108 . from fig.2(b ) and fig.2(c ) , the critical angles @xmath80 and @xmath83 determined by eq.(12 ) and eq.(13 ) are very close to the exact values according to their definitions . as @xmath3 varies from @xmath105 to @xmath107 , the positions of the weyl nodes determined by the symmetry factor @xmath30 keep almost unchanged . at @xmath68 , it is clear from fig.2(e ) that , besides the four weyl nodes determined by @xmath109 that are always present , two additional nodes emerges along @xmath110 as a result of @xmath71 @xcite . in the strip configuration ( with two edges parallel to the @xmath111-axis ) , a pair of weyl nodes with the same @xmath112 ( equal to @xmath98 or @xmath113 ) and energy project to the same point of the edge bz . the breaking of translational symmetry along the @xmath32 direction gives a finite coupling between the two weyl nodes . it is thus natural to expect that the weyl nodes will be gapped out for a strip of finite width . besides , it is known that there are edge modes for @xmath29 . it is interesting to see whether the edge states persists and how they are modified by the ubiquitously existing weyl nodes . these questions are answered partly by the results in figs . 2(f ) to 2(j ) for strips with 1500 unit cells along the @xmath32 direction . clearly , a gap opens at the weyl nodes . the edge modes persist but the dispersion is changed from the crossing - like structure to a shape similar to an @xmath114 symbol but then connected again to the edge modes beyond the projection of the weyl nodes @xcite . interestingly , the two edge modes are nondegenerate . this means that there are no additional edge modes associated with the weyl nodes arising from the symmetry factor @xmath28 . for @xmath115 , a third pair of weyl nodes along @xmath110 emerges . same as the case for the swsc @xcite , the edge modes become majorana fermi lines connecting the two new weyl nodes . in this case , since the edge modes pass through the projection of the two pairs of bulk weyl nodes resulting from @xmath109 , the bulk nodes extending throughout the bulk of the system act as channels coupling the edge modes . therefore , close to the projection of the bulk weyl nodes at the edge bz , the edge modes are re - hybridized . for @xmath31 , there are always two pairs of weyl nodes for an arbitrary @xmath116 . one pair is along @xmath110 and the other pair is along @xmath117 . since weyl nodes exist along both of the two special high symmetry lines , there are no simple analogies to @xmath80 and @xmath83 defined for @xmath30 . nevertheless , we can still define two critical angles in a manner similar to the above discussions . for @xmath118 that we focus on in this work , the three special reference wave vectors for defining the two critical angles are chosen as @xmath119 , @xmath120 , and @xmath121 . @xmath75 and @xmath78 are determined by eq.(10 ) and eq.(11 ) , respectively . at the first critical angle @xmath122 , we have @xmath123 . at the second critical angle @xmath124 , we have @xmath125 . there are no simple analytical expressions for @xmath126 and @xmath127 . we obtain them numerically by the definition . the energy spectra of the system in the configurations of a bulk material and a strip are shown in figure 3 . the parameters as chosen as @xmath91 , @xmath92 , @xmath93 , @xmath94 , and @xmath95 . we show the results for five values of @xmath3 , including @xmath105 [ 3(a ) and 3(f ) ] , @xmath126 [ 3(b ) and 3(g ) ] , @xmath127 [ 3(c ) and 3(h ) ] , @xmath106 [ 3(d ) and 3(i ) ] , and @xmath107 [ 3(e ) and 3(j ) ] . for the bulk spectra , we plot for each @xmath3 the quasiparticle bands @xmath66 and @xmath67 along three lines in the 2d bz with correspondingly fixed values of @xmath90 , which are @xmath105 , @xmath98 , and @xmath128 . @xmath75 is still determined by eq.(10 ) . @xmath129 is determined by @xmath130,\ ] ] where @xmath131 . the persistence of two pairs of bulk weyl nodes and their locations , are confirmed by the results presented in figs.3(a ) to 3(e ) . a qualitative difference between the present case and those for the swsc and the dwsc with @xmath30 is , however , clear from the dispersions along lines with @xmath132 on figs.3(d ) and 3(e ) . namely , the weyl nodes located at @xmath133 change from the first type ( type - i ) to the second type ( type - ii ) as @xmath3 increases and becomes larger than a critical value . this change in the character of the weyl nodes has not been noticed in either the swsc or the dwsc with @xmath30 . the pair of nodes located at @xmath134 , on the other hand , keeps to be type - i as @xmath3 changes . from appendix a with more detailed analyses , the weyl nodes at @xmath134 are the only weyl nodes in the present heterostructure ( fig.1 ) that can become type - ii . this is the first case where type - ii weyl nodes are predicted in the presence of superconducting correlation . it is also the first example of type - ii weyl nodes in two - dimensional systems , as far as we know . [ fig3 ] , for a typical set of parameters , @xmath96 , @xmath94 , @xmath95 , @xmath93 . for the bulk system , the dispersions of the @xmath66 and @xmath67 bands are shown along three lines in the 2d bz with fixed @xmath90 values at 0 ( black lines ) , @xmath98 ( blue lines ) , and @xmath128 ( red lines ) . the edge state spectra correspond to a strip with 1500 unit cells along the @xmath32 direction . @xmath100 for ( a ) and ( f ) , @xmath122 for ( b ) and ( g ) , @xmath124 for ( c ) and ( h ) , @xmath103 for ( d ) and ( i ) , @xmath104 for ( e ) and ( j).,title="fig:",width=321,height=566 ] the energy spectra for a strip are shown in figs.3(f ) to 3(j ) , for five typical angles defined above . while the two weyl nodes at @xmath133 still project to the same point ( i.e. , @xmath117 ) of the edge bz , the two weyl nodes at @xmath134 remain well separated from each other . the broken translational invariance along the @xmath32 direction couples the two weyl nodes at @xmath133 and opens a gap in the nodes , similar to the situation for @xmath30 . the edge modes also hybridize and open a gap at @xmath117 which is where the two weyl nodes at @xmath133 projects . five values of @xmath3 are considered , including @xmath100 for ( a ) and ( b ) , @xmath101 for ( c ) and ( d ) , @xmath102 for ( e ) and ( f ) , @xmath103 for ( g ) and ( h ) , @xmath104 for ( i ) and ( j ) . the parameters are taken as @xmath96 , @xmath94 , @xmath95 , @xmath93 . the darker the color , the larger the spectral weight.,title="fig:",width=321,height=566 ] [ fig5 ] . five values of @xmath3 are considered , including @xmath100 for ( a ) and ( b ) , @xmath122 for ( c ) and ( d ) , @xmath124 for ( e ) and ( f ) , @xmath103 for ( g ) and ( h ) , @xmath104 for ( i ) and ( j ) . the parameters are taken as @xmath96 , @xmath94 , @xmath95 , @xmath93 . the darker the color , the larger the spectral weight.,title="fig:",width=321,height=566 ] the energy spectra of the edge states can be probed by angle - resolved photoemission spectra ( arpes ) and scanning tunneling spectroscopy ( sts ) . instead of the quasiparticle energy spectrum , the experimentally relevant quantity is the spectral function along the edge of a strip , which is defined as the imaginary part of the retarded green s function for the edge layer of the strip @xcite . we have calculated the retarded green s functions @xmath135 for the left ( @xmath136 ) and right ( @xmath137 ) edges of a strip with @xmath138 layers , in terms of standard iterative green s function method.@xcite as shown in figs.4 and 5 are results separately for @xmath30 and @xmath31 , corresponding to five @xmath3 values taken separately in figs.2 and 3 . while the correspondences between the spectral functions in figs.4 and 5 and the quasiparticle energy spectra in figs.2 and 3 are very clear , the spectral functions contain important new information on how the edge states are distributed on the two edges . in addition , the green s functions obtained in terms of the iterative green s function method can be regarded as corresponding to a very wide strip ( @xmath139 ) , there is thus no finite - size effect in the results . as a result , there is no gap opening in the spectral function at the weyl nodes at @xmath140 for @xmath30 and at @xmath117 for @xmath31 . [ fig6 ] ( a , b , c , d , e ) and for @xmath31 ( f , g , h , i , j ) . results for the left and right edges are assigned respectively with a label of @xmath136 and @xmath137 . in the calculation , the width of the strip ( @xmath138 ) is assumed as infinite so that there is no finite - size effect . parameters are taken as @xmath96 , @xmath94 , @xmath95 , @xmath93 . the angles used are ( a ) @xmath100 , ( b ) @xmath101 , ( c ) @xmath102 , ( d ) @xmath141 , ( e ) @xmath115 , ( f ) @xmath100 , ( g ) @xmath122 , ( h ) @xmath124 , ( i ) @xmath141 , and ( j ) @xmath115 . see main text for the definitions of the angles.,title="fig:",width=321,height=566 ] while the direct observation of the edge states in terms of arpes requires very high resolution or large proximity - induced pairing amplitude in the hm film , the sts spectrum can be used as a more sensitive probe to detect the edge states . for clean samples with ideal edges , the sts detects the density of states ( dos ) on the edge of the sample , defined as an integration of the spectral function over the edge momenta in the 1d edge bz @xcite . as shown in fig.6 are the results for both @xmath30 and @xmath31 . prominent fine structures in the dos are found , which show striking evolutions with @xmath3 and are related directly to the evolutions in the energy spectra of the strip . there are qualitative differences in the behavior of the dos between those for @xmath30 and those for @xmath31 . the behaviors of the two are also very different from the results for the heterostructure with an @xmath1-wave superconductor @xcite . the dos follows closely the evolution of the edge state spectrum . we emphasize here two most interesting features in the results . the first point is the striking differences in the dos at @xmath100 for the three choices of @xmath28 . in contrast to @xmath30 and @xmath29 which both have well - defined zero - energy edge states , the zero - energy edge states for @xmath31 merge with the bulk weyl nodes and therefore have vanishing weight on the edge . this explains the finite dos at @xmath142 for @xmath30 and @xmath29 @xcite and the vanishing zero - energy dos for @xmath31 . in addition , the edge states for @xmath30 are distorted and have two finite - energy critical points with zero slope [ figs . 4(a ) and 4(b ) ] . they in turn give the two low - energy van hove singularities in fig.6(a ) . the two low - energy van hove singularities in fig.6(f ) are also related to the presence of finite - energy critical points with zero slope in the edge states , as is clear from figs.5(a ) and 5(b ) . the second feature is the existence of a critical angle at which the edge states on one edge becomes completely ( the @xmath1-wave case ) or partly ( the two @xmath0-wave cases ) flat . in this aspect , the results for @xmath30 are similar to those for @xmath29 @xcite . the appearance of a nearly flat portion in the edge state dispersion shown in figs . 4(d ) and 4(f ) manifest as a prominent zero - energy dos peak in figs . 6(b ) and 6(c ) . for @xmath31 , a portion of the edge state dispersion for one edge also becomes flat at the transition point from the type - i weyl node to the type - ii weyl node , which would also give a zero - energy van hove singularity in the dos . from figs . 3(h ) and 5(e ) , @xmath127 is very close to but slightly smaller than this transition point . correspondingly , we get in fig.6(h ) a not fully developed zero - energy van hove singularity . in addition , the van hove singularity for @xmath31 occurs on the opposite edge as compared to those for @xmath30 and @xmath29 . this is understood from the fact that , for the edge states close to @xmath117 on the same edge ( e.g. , the left edge with @xmath136 ) and at @xmath100 , their velocity for @xmath31 is opposite in sign to those for @xmath30 and @xmath29 . here , we summarize the main differences between the energy spectra of the heterostructures with @xmath0-wave form factors [ @xmath30 and @xmath31 ] and those for the heterostructure with @xmath1-wave form factor [ @xmath29 ] @xcite . for the @xmath1-wave case , there are three different phases as we change the direction of the magnetization , the fully - gapped phase , the phase without a full gap but has a ubiquitous local gap , and the swsc phase with a pair of nodes between the two low - energy quasiparticle bands when the magnetization lies along an in - plane direction . for both @xmath30 and @xmath31 of the @xmath0-wave cases , the energy spectra always have nodes and thus describe 2d dwsc . in terms of the number of weyl nodes , there is a single pair of weyl nodes in the swsc . the two nodes have opposite nonzero quasiparticle energies . for @xmath30 , when a nonzero out - of - plane component of the magnetization exists , there are four weyl nodes . when the magnetization lies completely along an in - plane direction , an additional pair of weyl nodes emerge and so there are in total three pairs of weyl nodes . the quasiparticle energies for the nodes are zero for completely out - of - plane magnetization , but are nonzero when an in - plane component of the magnetization exists . for @xmath31 , there are always two pairs of weyl nodes in the quasiparticle energy spectrum . the energies of the pair of nodes along @xmath146 are fixed at @xmath147 for an arbitrary orientation of the magnetization . for the other pair of nodes along @xmath148 , the energies are zero for purely out - of - plane magnetization but is nonzero in the presence of a finite in - plane magnetization component . as regards the type of weyl nodes , only type - i weyl nodes are possible for swsc and dwsc with @xmath30 . for @xmath31 , the pair of nodes fixed at zero energy along @xmath146 change between type - i and type - ii as the direction of the magnetization varies . in this subsection , we make some comments on the nature of the phases described by eq.(3 ) , for a hm with spin - singlet pairing term and rsoc . for purely out - of - plane magnetization , the quasiparticle spectrum is similar to a conventional bcs superconductor : a full gap opens for the @xmath1-wave case which is effectively a spinless @xmath149 superconductor @xcite , whereas for the two @xmath0-wave cases the spectrum consists of four isolated nodes and is gapped everywhere else . as we tune the magnetization away from the out - of - plane direction , a nonzero in - plane component of the magnetization ( i.e. , @xmath150 ) cooperates with the rsoc to tilt the quasiparticle spectrum continuously along the @xmath112 direction . for the @xmath1-wave case , along with the transition from the fully - gapped phase to the phase without a full gap but has a ubiquitous local gap , the @xmath67 and @xmath66 quasiparticle bands begin to overlap in energy @xcite . the ` fermi surface ' of the quasiparticle bands , constituted by states with quasiparticle energy @xmath147 , consists of two disconnected pockets : one ` hole'-like pocket from @xmath66 and one ` electron'-like pocket from @xmath67 . for the two @xmath0-wave cases , while the spectra have nodes for all magnetization orientations , the energy overlap between the @xmath67 and @xmath66 quasiparticle bands also increases as the magnetization tilts continuously to the in - plane direction . for @xmath30 , the ` hole'-like and ` electron'-like pockets are always disconnected , similar to the case of @xmath29 . for @xmath31 , on the other hand , the ` hole'-like and ` electron'-like pockets begin to be connected through the two weyl nodes lying along the @xmath90 axis once they turn into type - ii . as shown in fig.7 is an illustration of the evolution of the @xmath147 fermi contour explained above , for five typical values of @xmath3 in the range @xmath151 $ ] . they are interpreted naturally as a series of lifshitz transitions @xcite . namely , as the magnetization orientation varies , the @xmath147 fermi contour undergoes transitions between four and three topologically distinctive configurations for @xmath30 and @xmath31 , respectively . similar analysis can be applied to the case of @xmath29 . it is important to note that the lifshitz transition is usually encountered and discussed in the normal phase of a metal , as an abrupt change in the fermi surface topology of the metal @xcite . its revelation in the presence of pairing indicates the essential difference between the state of the present system and other more well - known types of pairings . [ fig7 ] equal - energy contours for the quasiparticle bands . @xmath30 for ( a , b , c , d , e ) and @xmath31 for ( f , g , h , i , j ) . parameters are taken as @xmath96 , @xmath94 , @xmath95 , @xmath93 . the angles used are ( a ) @xmath100 , ( b ) @xmath101 , ( c ) @xmath102 , ( d ) @xmath141 , ( e ) @xmath115 , ( f ) @xmath100 , ( g ) @xmath122 , ( h ) @xmath124 , ( i ) @xmath141 , and ( j ) @xmath115 . see main text for the definitions of the angles . the numberings of the quasiparticle bands to which the pockets belong are indicated in the figures.,title="fig:",width=321,height=566 ] now , we compare the quasiparticle spectrum for the present state and those for several well - known types of pairings , including the bcs - like pairing @xcite , the fflo pairing @xcite , and the so - called breached pairing ( or , sarma state ) @xcite . the bcs pairing naturally occurs in a system with a centro - symmetric fermi surface . the pairing between states of equal and opposite momenta thus gives a spectrum in which the @xmath67 quasiparticle band is completely empty and the @xmath66 quasiparticle band is fully occupied . for our case , the pairing is also between states of equal and opposite momenta . but the energy spectrum deviates significantly from the conventional bcs spectrum that is symmetric with respect to @xmath147 for each @xmath152 . in the fflo pairing , the net momentum of a pair is nonzero . a simplest case where this phase might be found is when the fermi surface is asymmetric with respect to the center of the bz . in our case , while the fermi surface is asymmetric in the presence of an in - plane component of the exchange field , the net momentum of a cooper pair is always zero . finally , in the breached pairing state , the pairing gap between _ two _ bands of unequal fermi momenta opens at nonzero energy . for our system , in the presence of an in - plane component of the exchange field and away from @xmath117 , the pairing gap between the @xmath67 and @xmath66 bands also opens at nonzero energy and is similar to the breached pairing state . however , on one hand the present system consists of only one fermi surface in the normal phase , on the other hand the states on the fermi surface with @xmath153 are usually gapped at @xmath147 . only for @xmath31 and when the two nodes along @xmath146 have turned from type - i to type - ii , would the full fermi surface be gapless and the spectrum be most similar to the breached pairing state . however , the @xmath147 states in this case consist of both ` electron'-like states from @xmath67 and ` hole'-like states from @xmath66 , rather than from a single quasiparticle band in the case of a standard breached pairing state . from the above comparisons , we see that the appearance of the unique pairing state in the present system is caused by a synergic action of two factors : the bcs - like pairing correlation induced through proximity effect , and the asymmetry of the fermi surface in the hm . the first aspect of the model is a result of conservation of the parallel component of the momentum through the interface which is assumed to be perfectly translational invariant . the second factor results from a combination of the in - plane component of the magnetization and the rsoc . in our treatment of the pairing in the hm , we have assumed that it arises completely from the proximity coupling with the spin - singlet superconductor . if instead we assume that there is an intrinsic pairing tendency in the hm , then a self - consistent calculation is necessary in addition to the proximity effect . in this case , in the presence of an in - plane component of the magnetization which drives the fermi surface asymmetric with respect to the bz center , the pairing term obtained in a self - consistent manner could also possibly be the fflo pairing @xcite . in ref.@xcite focusing on the heterostructures with @xmath1-wave superconductor , we have characterized the phases and phase transitions driven by @xmath3 in terms of four topological quantities : a pfaffian @xmath154 invariant defined based on the particle - hole symmetry of the model , the tknn number defined in terms of the occupied quasiparticle states , the band - wise chern numbers associated with separate nondegenerate quasiparticle bands , and the zak phase which is a berry phase defined along one - dimensional subspace of the 2d bz @xcite . the band - wise chern numbers in the @xmath1-wave case are motivated from the ubiquitous local gaps in the quasiparticle spectrum , except for @xmath115 and @xmath155 which correspond to the 2d weyl superconductivity phase @xcite . in the case of heterostructure with @xmath0-wave superconductor , however , there are always nodes between the two low - energy quasiparticle bands @xmath66 and @xmath67 . therefore , the band - wise chern number can not be defined for the @xmath0wscs . in addition , the zero - energy nodes existent for out - of - plane magnetization for @xmath30 and always existent for @xmath31 also cause difficulty in calculating the tknn number . thus , we focus on the pfaffian @xmath154 invariant and the zak phase to characterize the topological properties of the phases in @xmath0wscs . before studying the topological invariants , we list the relevant symmetries ( or , broken symmetries ) of the model defined by eq.(3 ) . one important aspect of the model is the absence of time - reversal symmetry , @xmath156 , where @xmath157 denotes complex conjugation . once @xmath158 , no matter what the value of @xmath3 is , we have @xmath159 and thus the time - reversal symmetry is broken . besides , the inversion symmetry , which for the present single - orbital model is simply the operation bringing @xmath152 to @xmath160 , is broken by the rsoc since we have @xmath161 . however , the model defined by eq.(3 ) always has the particle - hole symmetry , @xmath162 . that is , @xmath163 holds for all three @xmath28 and for all values of @xmath3 . the next relevant symmetry is the mirror reflection symmetry . for the mirror reflection with respect to the plane perpendicular to the @xmath32 axis , the corresponding mirror reflection operator is @xmath164 . similarly , we have @xmath165 and @xmath166 , for mirror reflections with respect to the planes perpendicular to the @xmath111 axis and @xmath18 axis . for @xmath30 and @xmath29 , we have @xmath167 @xmath168 and @xmath169 for @xmath31 , by redefining the mirror reflection operator for @xmath32 and @xmath111 directions as @xmath170 and @xmath171 , it is easy to see that the bdg hamiltonian @xmath172 satisfies the same transformation properties as the other two pairing channels . from eqs.(16)-(18 ) , the model has the mirror symmetry along @xmath32 only when @xmath173 . no mirror symmetry along @xmath111 exists , once @xmath158 . when @xmath174 , @xmath175 acts effectively as an inversion symmetry of the model . this effective inversion symmetry ensures that each pair of weyl nodes having opposite wave vectors can share the same quasiparticle energy for @xmath100 or @xmath176 . finally , the model has the four - fold rotation symmetry around the @xmath18 axis when @xmath174 , namely @xmath177\tilde{s}_{z}=h(\mathbf{k}).\ ] ] @xmath178 . @xmath179 for @xmath29 . @xmath180 for @xmath30 and @xmath31 . as a result of this four - fold rotation symmetry , all four weyl nodes share the same quasiparticle spectrum for @xmath100 and @xmath176 . other symmetries ( e.g. , the so - called magnetic mirror reflection symmetries @xcite ) of the model , while on one hand can be defined as a combination of several of the above symmetries , on the other hand are not directly relevant to the following discussions . therefore , we do not list them here . from the particle - hole symmetry of the model defined by eq.(3 ) , the pfaffian @xmath181 invariant for the dwscs can be defined as @xcite @xmath182,\ ] ] where @xmath183 , the function @xmath82 gives the sign of a real number @xmath32 . for both @xmath30 and @xmath31 , we have @xmath184 . for the present parameters with large @xmath2 and a single nondegenerate fermi surface , we have @xmath185 , the phase is therefore always nontrivial ( @xmath186 ) from the pfaffian @xmath154 invariant . if the energy spectrum of the particle - hole symmetric bdg hamiltonian , for example eq.(3 ) with @xmath29 and @xmath100 , is fully gapped , the pfaffian @xmath181 invariant was shown by ghosh _ et al _ to give the parity of the chern number @xcite . in another work based on a model which can be considered as the @xmath100 case of eq.(3 ) with @xmath30 or @xmath31 , sato and fujimoto pointed out that although the spectrum is gapless , the parity of the chern number is still well - defined and determined by the same formula [ i.e. , eq.(20 ) ] for the pfaffian @xmath181 invariant @xcite . therefore , for both of the two choices of @xmath28 for the @xmath0wscs and for purely out - of - plane magnetization , the parity of the chern number is well - defined and is odd for the parameters considered in sec.iii . from the topological nature of @xmath187 , it is then natural to expect that the system described by eq.(3 ) retains a certain topological character as @xmath3 deviates from @xmath105 and @xmath176 . this is confirmed from the persistence of the edge states shown in sec.iii , except for fig.3(j ) where the edge modes are merged to the bulk bands . from the analysis of the zak phase in sec.iv c , however , there should be hidden edge states for the parameters of fig.3(j ) . the edge states for a general @xmath3 evolve continuously from those for @xmath100 . so , it is important to define the chirality of the edge states for @xmath100 . in the simplest case of a @xmath149 superconductor , the edge states are linearly dispersive and counter - propagating on the two opposing edges . the chirality of the edge state can be defined in terms of the sign of the velocity of the edge modes . the hm layer of the heterostructure with @xmath1-wave superconductor is equivalent to a @xmath149 superconductor , the chirality of the edge states can thus be defined in this manner @xcite . in the heterostructures with @xmath0-wave superconductors , however , the edge states is curvous and the velocity changes sign several times ( see figs . 2 and 3 ) . nonetheless , we can still define the chirality of the edge states in terms of the velocity of the edge modes in the following manner . on an edge with a unique edge state , we divide the full dispersion of it into several segments connected together through points with zero velocity . each segment is assigned a chirality , positive or negative , equal to the sign of the velocity in this section of the dispersion . the overall chirality is defined simply as a multiplication of the chiralities of all the segments . in this manner , it is straightforward to see from figs . 4 and 5 in sec . iii c ( and , fig . 5 in ref.@xcite for the @xmath1-wave case ) that the edge states on the same edge and for @xmath100 have the same chirality for all three @xmath28 . in addition , as @xmath3 increases , the chirality of the edge states on the left edge ( @xmath136 ) keeps unchanged , whereas the chirality of the edge states on the right edge ( @xmath137 ) change sign . the @xmath3-dependence of the ( generalized ) chirality of the edge states for all three pairings are thus qualitatively the same . the zak phase is a berry phase defined as a line integral over the berry connection along a 1d loop of the 2d bz @xcite . in calculation , one component ( i.e. , @xmath90 or @xmath112 ) of the wave vector is fixed as a parameter , integration over the wave vector along the orthogonal direction then traces a 1d loop , as a result of the periodicity of the bz . for example , by taking @xmath112 as a parameter , the zak phase for the @xmath39-th ( @xmath188 , @xmath41 , @xmath42 , @xmath43 ) quasiparticle band is defined as @xmath189 where @xmath190 is the @xmath32 component of the berry connection for the @xmath39-th quasiparticle band , @xmath191 . @xmath192 is the eigenvector of the @xmath39-th quasiparticle band . another set of zak phase , @xmath193 , is defined in the same manner by reversing the roles of @xmath90 and @xmath112 . when the effective 1d model is inversion symmetric , the zak phase can only take two values ( @xmath105 or @xmath176 ) , otherwise the zak phase can take any value between @xmath105 and @xmath194 @xcite . for the @xmath1wsc defined by @xmath29 and @xmath173 , the relevant symmetry is @xmath195 . the positions of the two weyl nodes of the @xmath1wsc can be captured by a jump in the zak phase between @xmath105 and @xmath176 @xcite . in addition , on an edge running along @xmath111 ( so that @xmath112 is still a good quantum number but @xmath90 is not ) , there are edge states in the segment of @xmath112 for which the zak phase is @xmath176 @xcite . from eqs.(16)-(18 ) , the @xmath0wscs with @xmath30 and @xmath31 have the same mirror symmetry ( or , modified mirror symmetry , for @xmath31 ) as @xmath29 . as a result , the cases in which the zak phase can take a well - defined quantized value of @xmath105 or @xmath176 are still associated with @xmath115 or @xmath196 , when the magnetization has only an in - plane component . in this case , the quantization of @xmath197 follows from eq.(16 ) . in addition , by combining the particle - hole symmetry @xmath198 and @xmath199 , we can see that the model has a ` inversion - like ' symmetry along @xmath112 for fixed @xmath90 . more exactly , we have @xmath200 therefore , the zak phase defined along @xmath112 loops should also take quantized values ( @xmath105 or @xmath176 ) . on the other hand , once the out - of - plane component of the magnetization ( @xmath201 ) is nonzero , no ` inversion - like ' symmetry can be defined along an arbitrary 1d loop with one wave vector fixed at a general value . a few exceptions to this conclusion are for @xmath174 and for those loops with the fixed wave vector taking time - reversal - invariant values ( i.e. , @xmath105 and @xmath202 ) . for these special loops , the 1d effective model has an ` inversion - like ' symmetry enforced by the symmetry of @xmath175 defined by eq.(18 ) . [ fig8 ] , at @xmath68 . @xmath30 for ( a ) , and @xmath31 for ( b ) . the parameters are the same as those used in sec.iii.,title="fig:",width=245,height=321 ] we have calculated explicitly the zak phases @xmath193 and @xmath197 for both @xmath30 and @xmath31 , at @xmath100 and @xmath68 . in agreement with the above expectation based on symmetry , @xmath193 and @xmath197 take quantized values of @xmath105 or @xmath176 for @xmath68 . for @xmath100 , the zak phases are quantized only when the fixed wave vector takes @xmath105 or @xmath202 . we show in figure @xmath203 the nontrivial results for the zak phase , which are @xmath197 for @xmath30 [ fig.8(a ) ] and @xmath193 for @xmath31 [ fig.8(b ) ] , both at @xmath68 . only the zak phases for the two low - energy quasiparticle bands ( namely , @xmath204 ) are shown since the results for the two high energy bands are always zero . fig.8(b ) does not contain results for @xmath110 since the integration loop for @xmath110 contain two nodes between @xmath66 and @xmath67 which makes the zak phase ill defined . in fig.8(a ) , results for the loops containing the nodes are also excluded . in addition , we have obtained ( at @xmath68 ) @xmath205 for @xmath30 and @xmath206 for @xmath31 . comparing fig.8(a ) with the zak phase calculated for the @xmath1wsc @xcite , it is clear that they are essentially the same . although the pairing amplitudes are taken at different values , the positions of the weyl nodes of the @xmath1wsc and the new pair of weyl nodes for the @xmath0wsc with @xmath30 are identical to each other , because they are determined by the vanishing of the effective pairing amplitude . and , though there are two additional pairs of weyl nodes from @xmath207 , the zak phase do not see them . this is also a very natural result , though , because each node contributes @xmath176 or @xmath105 to the zak phase , so the net change of the zak phase upon crossing the line with two weyl nodes is zero ( modulo @xmath194 ) . in addition , it is clear that the region with a nontrivial zak phase of @xmath176 is also characterized by the presence of edge modes connecting the two additional pairs of weyl nodes emerging from the emergent mirror symmetry of the model . on the other hand , the fact that @xmath205 for all @xmath90 is consistent with the fact that the change in the number of weyl nodes over which the integration loop sweeps is always an even number , zero or two . therefore , even if each node contributes @xmath176 , the total change in the zak phase defined modulo @xmath194 is still zero . the zak phase for @xmath31 is however more peculiar . namely , @xmath206 although the two weyl nodes along @xmath148 are separated from each other and thus the integration loop sweeps only a single weyl node . this result can be understood from the low - energy effective models for the two nodes along @xmath148 for @xmath31 : they are real symmetric [ see eqs . ( a11)-(a13 ) in appendix a ] and thus the wave function can be taken as purely real . as such , no nontrivial phase factor can be collected when we traverse the nodes . on the other hand , the low - energy effective models for the two weyl nodes of @xmath31 along @xmath146 ( which are type - ii for @xmath68 ) are complex . this explains why the zak phase @xmath193 shows a jump of @xmath176 when the integration path traverse one node . @xmath206 is also in agreement with the absence of well - defined edge states in fig.3(j ) for edges running along the @xmath111 direction . on the other hand , from the nontrivial wave vector dependence of @xmath193 , it is natural to expect the presence of edge states on the edges running along @xmath32-direction this in turn can account for the nontrivial result of the pfaffian @xmath154 invariant . however , since the bulk quasiparticle spectrum is completely gapless from the @xmath90 direction , the edge states , even if they exist , would merge to the bulk spectra and become undiscernible . however , it is still possible that the nontrivial zak phase @xmath193 can have some influence on the properties of the system . we leave the analysis over this interesting possibility to later works . finally , we consider the case of @xmath100 . consistent with the symmetry analysis , we have found that the zak phase @xmath197 [ @xmath193 ] for @xmath100 does not take quantized value for a general @xmath112 ( @xmath90 ) . on the other hand , the zak phase corresponding to wave vectors @xmath105 and @xmath202 do take quantized values ( @xmath105 or @xmath176 ) . these quantized zak phase can be utilized to define a new @xmath154 invariant , @xmath209 where @xmath210 or @xmath111 when considering an edge running along @xmath211 direction . @xmath212 and @xmath42 corresponds to the existence and nonexistence of well - defined edge state at @xmath213 . this connection can be verified easily by making a comparison between the energy spectra shown in sec.iii and the following numerical results for the zak phase , at @xmath100 : for @xmath30 , @xmath214 and @xmath215 . for @xmath31 , @xmath216 . for @xmath31 , this result is consistent with the fact that since for both edges running along @xmath32 and @xmath111 , there are always projections of two weyl nodes to the origin of the 1d bz , which act as bulk channels coupling the two edges together , no well - defined edge states can be defined at the center of the one dimensional edge bz . in the above discussions , we have focused on heterostructures in which the hm has a single nondegenerate fermi surface . it is interesting to ask what if the spin polarization is not complete , namely what if the two spin - split bands both contribute to the fermi surface . this is relevant since ferromagnetic materials with partially polarized fermi surface is more commonly found than hm . this problem is also relevant if we turn our attention from the hm layer to the layer of spin - singlet superconductor . it is clear that the same model can be used to describe this layer . the only qualitative difference is just in the magnitude of the exchange field , since the proximity - induced exchange field in the @xmath0-wave superconductor is in general not large enough to split completely one spin component off the fermi surface . consider a ferromagnetic metal with a fermi surface consisting of two sheets with opposite spin - polarization and different volumes . the two sheets with opposite spin polarizations would naturally be separated from each other . replacing the hm in the heterostructure studied above with such a partially polarized ferromagnet , we can expect to apply the previous analyses separately to each spin component . as such , we can expect to get two sets of low - energy effective models with similar properties but related separately to the two spin - split bands . in particular , for each band , there will be a pair of counter - propagating chiral majorana edge states . it is then interesting to ask whether or not the two sets of chiral edge modes are always separated from each other . or , in cases where they coexist in some parameter ranges , how would they interact with each other . ( a ) @xmath68 which corresponds to an @xmath1wsc , results for the two edges are the same . ( b ) and ( c ) are spectral functions on the left ( @xmath136 ) and right ( @xmath137 ) edges of the strip for @xmath100 . the parameters used are @xmath91 , @xmath92 , @xmath94 , @xmath217 , @xmath218 . the darker the color , the larger the spectral weight.,title="fig:",width=245,height=472 ] to answer the above question , we have studied the changes in the properties of the weyl superconductivity phase brought by increasing the chemical potential so that it crosses both of the two spin - polarized subbands of the ferromagnet . as shown in fig.9(a ) are the spectral functions for the edge layers ( results for the two edges are the same ) of a strip for @xmath29 and @xmath68 , namely the @xmath1wsc . while on one hand the edge modes for the two spin species keep largely independent of each other , a hybridization gap opens at ( @xmath117 , @xmath142 ) . besides , because the fermi sheets of the two spin channels open gaps independently , the edge modes in the non - overlapping regions of the pairing gaps of the two pin - split bands will be masked by the bulk states of the opposite - spin band . overall , the chiral edge modes appear in _ three disconnected _ regions of the local energy gap of the energy spectrum . in fig.10 , we show the spectral functions on the two edges of a strip of the @xmath0wsc at @xmath100 for @xmath30 [ 10(a ) and 10(b ) ] and @xmath31 [ 10(c ) and 10(d ) ] . the results for @xmath29 at @xmath100 are shown as figs.9(b ) and 9(c ) . for @xmath30 , we observe similar coexistence of two chiral edge states and the opening of a hybridization gap at ( @xmath117 , @xmath142 ) between the two edge states belonging separately to the two spin - subbands . for @xmath31 , the two copies of edge states simply superimpose together and no hybridization gap opens . in ref.@xcite , it was pointed out that the topological phase with a single majorana edge state can be obtained for vanishingly small exchange field when the chemical potential is at the band bottom ( @xmath219 ) . while that case still has a single band crossing the chemical potential , our present case has two bands contributing to the fermi surface . from the numerical results in fig.10 and comparing it with fig.9 for the @xmath1-wave case , we see that while the edge states exist when the fermi surface consists of two sheets , there is a pair of counter - propagating edge modes on each edge which usually hybridize and open a gap . the physical picture for the @xmath0-wave case is therefore of no essential difference from the @xmath1-wave case as regards the number of edge states and their hybridization . since there are even number of edge states on each edge , they are not topologically protected . this is consistent with the pfaffian @xmath154 invariant defined in eq.(20 ) , which gives the topologically trivial result of @xmath220 for the present set of parameters . the phase on the superconductor side of the heterostructure is therefore topologically trivial . topologically nontrivial physics are located in the hm layer . [ fig10 ] ( a , b ) and @xmath31 ( c , d ) . the angle is taken as @xmath100 . the parameters are chosen as @xmath91 , @xmath92 , @xmath94 , @xmath95 , @xmath218 . darker color means larger spectral function.,title="fig:",width=321,height=264 ] in summary , we have predicted the realization of two - dimensional weyl superconductivity phases in a heterostructure consisting of a half - metal ( or a highly spin - polarized ferromagnet ) clamped in between an insulating substrate and a spin - singlet superconductor . when the superconductor has @xmath0-wave symmetry ( with a symmetry factor of @xmath221 or @xmath222 ) , the two - dimensional weyl superconductivity phase is realized for an arbitrary magnetization direction in the half - metal . quasiparticle energy spectra are studied which exhibit interesting variations with the direction of the magnetization . in particular , we have predicted the transition between type - i and type - ii weyl nodes by changing the magnetization orientation . in addition , the evolution of the phase with the magnetization orientation is unveiled through the lifshitz transitions in the zero - energy quasiparticle spectrum . the phases are shown to be qualitatively different from several well - known types of pairings . we characterize the topology of phases by the pfaffian @xmath154 invariant and the zak phase . this work was supported by the texas center for superconductivity at the university of houston and the robert a. welch foundation ( grant no . l.h . would also like to acknowledge the support from the china scholarship council . the numerical calculations were performed at the center of advanced computing and data systems at the university of houston . in this section , we perform an analysis on the weyl nodes that are supported in the present system defined by fig . 1 and eq.(3 ) of the main text . in particular , as fig.3 has shown the transition from type - i weyl node to type - ii weyl node , it is interesting to investigate the possibility of realizing type - ii weyl nodes in other cases . from the analysis presented below , the type - ii weyl nodes found in fig.3 for @xmath31 are the only cases where type - ii weyl nodes can be realized in the proposed system . we also point out the possibility of realizing semi - weyl nodes . the low - energy effective model for a single weyl node can be written generally in terms of the relative momenta @xmath223 as @xcite @xmath224 where @xmath225 is a label for the specific weyl node under consideration , @xmath226 is the second - order unit matrix . the two eigenvalues of the model are @xmath227 @xmath228 describes a type - ii weyl node if and only if there exists a direction @xmath229 along which @xmath230 . otherwise , the weyl node is type - i . for the heterostructure with @xmath1-wave superconductor , the weyl superconductivity phase ( @xmath1wsc ) appears only when the magnetization lies along in - plane directions ( @xmath68 and @xmath231 ) . from the low - energy effective model of the weyl nodes for @xmath68 @xcite , the weyl nodes can become type - ii only when the strength of the rsoc ( @xmath15 ) is of the same order of magnitude as the hopping amplitude @xmath12 . since this is unrealistic , the weyl nodes in the @xmath1wsc can only be type - i . for the @xmath0wsc with @xmath30 , the two weyl nodes emerging for in - plane magnetization are qualitatively the same as the weyl nodes for @xmath1wsc . following the same analysis as above , they can only be type - i for realistic model parameters . the remaining four weyl nodes exist for all @xmath3 angles . their positions in the 2d bz are nearly @xmath3-independent and are @xmath232 ( @xmath47 , @xmath233 ) . @xmath98 is defined by eq.(10 ) . for a general @xmath3 , we have @xmath234 @xmath235 @xmath236 and @xmath237 the constant term in @xmath238 gives the energy of the weyl node but is irrelevant to the type of the weyl node . @xmath238 only gives a velocity along the @xmath239 direction , of the order no larger than @xmath15 . @xmath240 , on the other hand gives a velocity of the same order as @xmath12 along @xmath239 . as a result , the four weyl nodes @xmath241 ( @xmath47 , @xmath233 ) are type - i for realistic model parameters and all @xmath3 values . for the @xmath0wsc with @xmath31 , there are always four weyl nodes located approximately at @xmath242 ( @xmath47 , @xmath233 ) . to simplify notation , we introduce the following abbreviations : @xmath243 , @xmath244 , @xmath245 , @xmath246 , @xmath247 , @xmath248 . for a general @xmath3 , the effective low - energy model for the weyl nodes are @xmath249 @xmath250 @xmath251 and @xmath252 again , because @xmath238 only gives the quasiparticle a velocity along @xmath239 , only if the velocity along @xmath239 from @xmath253 is smaller than that from @xmath238 , can the weyl node become type - ii . note that @xmath254 , the effective model for the two weyl nodes @xmath255 becomes @xmath256 @xmath257 @xmath258 and @xmath259 . from the above effective model , the two weyl nodes at @xmath255 can only be type - ii when @xmath15 is of the same order as @xmath12 , which is unrealistic in the present setting . an interesting critical case for these two nodes is when the coefficient of @xmath260 in eq.(a12 ) becomes zero . when this happens , the linear in @xmath260 term vanishes , we have to add the term quadratic in @xmath260 . the nodes thus become linear in @xmath239 but quadratic in @xmath260 , and is thus semi - weyl @xcite . in a hm with large magnetic anisotropy energy , this critical point corresponding to a general @xmath3 angle should be very difficult to attain . if the magnetic anisotropy energy is small so that we can tune @xmath3 through the full @xmath194 range , it is still possible to turn the two weyl nodes at @xmath255 to be semi - weyl . the effective model for the other pair of weyl nodes , @xmath261 , are @xmath262 @xmath263 @xmath264 and @xmath265 . for experimentally relevant parameters satisfying @xmath266 , @xmath267 , and @xmath268 , the type - ii weyl node is clearly achievable . in the limiting cases of @xmath269 and @xmath270 , we can also get an analytical expression for the critical angles marking the transition between type - i and type - ii weyl nodes , @xmath271 overall , we have verified both numerically and analytically that a transition between type - i and type - ii weyl nodes can be achieved in our proposed experimental configuration , related to the pair of weyl nodes lying along @xmath146 for @xmath31 . finally , we reemphasize that the two cases , @xmath30 and @xmath31 , can be realized by growing the same @xmath0-wave superconductor on the hm thin film along two inequivalent special directions . guoqing chang , su - yang xu , daniel s. sanchez , shin - ming huang , chi - cheng lee , tay - rong chang , guang bian , hao zheng , ilya belopolski , nasser alidoust , horng - tay jeng , arun bansil , hsin lin , and m. zahid hasan , sci . adv . * 2 * , e1600295 ( 2016 ) .

for the past several decades since the discovery of pulsars , many authors have investigated the properties of dense matter which exist inside neutron stars and supernova cores ( see , e.g. , refs . these objects have been shown that they consist of a variety of material phases whose physical properties reflect in many astrophysical phenomena of these objects . especially , the properties of nuclear matter under extreme conditions , which is one of the essential topics for understanding the mechanism of collapse - driven supernovae @xcite , the structure of neutron star crusts @xcite and its relating phenomena , have been studied actively . this subject is also interesting as one of the fundamental problems of the complex fluids of nucleons . at subnuclear densities , nuclear matter exhibits the coexistence of a liquid phase with a gas phase due to the internucleon interaction which has an attractive part . at sufficiently low temperatures relevant to neutron star interiors , and sufficiently below the normal nuclear density , long - range coulomb interactions make the system divide periodically into gas and spherical liquid drops , adding a crystalline property to the liquid - gas coexistence . in the density region where nuclei are about to melt into uniform matter , it is expected that the energetically favorable configuration of the mixed phase possesses interesting spatial structures such as rodlike and slablike nuclei and rodlike and spherical bubbles , etc . , which are referred to as nuclear `` pasta '' . this picture was originally proposed by ravenhall et al . @xcite and hashimoto et al . @xcite independently . their predictions both based on free energy calculations with liquid drop models assuming some specific nuclear shapes . these works clarify that the most energetically stable nuclear shape is determined by a subtle balance between the nuclear surface and coulomb energies . detailed aspects of equilibrium phase diagrams , such as a series of nuclear shapes which can be realized as the energetically most favorable and the density range corresponding to the phases with nonspherical nuclei , vary with nuclear models @xcite . however , the realization of the `` pasta '' phases as energy minimum states can be seen in a wide range of nuclear models and the phase diagrams possess an universal basic feature that , with increasing density , the shape of the nuclear matter region changes like sphere @xmath4 cylinder @xmath4 slab @xmath4 cylindrical hole @xmath4 spherical hole @xmath4 uniform @xcite . this feature is also reproduced by the thomas - fermi calculations by several groups @xcite . the phases with these exotic nuclear structures , if they were realized in neutron star crusts or supernova cores , bring about many astrophysical consequences . as for those in neutron star phenomena , it is interesting to note the relevance of nonspherical nuclei to pulsar glitches and cooling of neutron stars . although the question whether the mechanism of pulsar glitches is depicted by vortex pinning model or star quake model has yet to be settled completely , the existence of nonspherical nuclei in neutron star matter ( nsm ) have significant effects in both cases . as for the former , while the force needed to pin vortices has yet to be clarified completely even for a bcc lattice of spherical nuclei mainly due to the uncertain properties of impurities and defects @xcite , the effect of spatial structure of normal nuclear matter on vortex dynamics can not be ignored . as for the latter , the existence of `` pasta '' phases with slablike and rodlike nuclei would change the elastic properties of inner crust matter from those of crystalline solid to those of liquid crystal as indicated by pethick and potekhin @xcite , which results in significant decrease of the maximum elastic energy that can be stored in the inner crust . the presence of nonspherical nuclei would also accelerate the cooling of the corresponding region of neutron stars by opening semileptonic weak processes which are unlikely to occur for spherical nuclei @xcite . `` pasta '' phases in supernova matter ( snm ) are expected to affect the neutrino transport and hydrodynamics in supernova cores . let us first note that the neutrino wavelengths , typically of order 20 fm , are comparable to or even greater than the internuclear spacing , leading to diffractive effects on the neutrino elastic scattering off such a periodic spatial structure of nuclear matter @xcite . these effects , induced by the internuclear coulombic correlations , would reduce the scattering rates and hence the lepton fraction @xmath5 . for the bcc lattice of spherical nuclei , such a reduction was examined by horowitz @xcite by calculating the associated static structure factor . it is also noteworthy that nonspherical nuclei and bubbles are elongated in specific direction . in such direction , the neutrino scattering processes are no longer coherent , in contrast to the case of roughly spherical nuclei whose finiteness in any direction yields constructive interference in the scattering . the final point to be mentioned is that the changes in the nuclear shape are accompanied by discontinuities in the adiabatic index , denoting how hard the equation of state of the material is . these discontinuities may influence the core hydrodynamics during the initial phase of the collapse @xcite . though the properties of `` pasta '' phases in equilibrium state have been investigated actively , the formation and the melting processes of these phases have not been discussed except for some limited cases which are based on perturbative approaches @xcite . it is important to adopt a microscopic and dynamical approach which allows arbitrary nuclear structures in order to understand these processes of nonspherical nuclei . at finite temperatures , it is considered that not only nuclear surface becomes obscure but also nuclei of various shapes may coexist . therefore , it is necessary to incorporate density fluctuations without any assumptions on nuclear shape to investigate the properties of `` pasta '' phases at finite temperatures . although the works done by williams and koonin @xcite and lassaut et al . @xcite do not assume nuclear structure , they can not incorporate fluctuations of nucleon distributions in a satisfying level because they are based on the thomas - fermi calculation , which is one - body approximation . in addition , only a single structure is contained in the simulation box in these works , there are thus possibilities that nuclear shape is strongly affected by boundary effect and some structures are prohibited implicitly . in the present work , we study the structure of cold dense matter at subnuclear densities in the framework of quantum molecular dynamics ( qmd ) @xcite , which is one of the molecular dynamics ( md ) approaches for nucleon many - body systems ( see , e.g. , ref . @xcite for review ) . md for nucleons including qmd , which is a microscopic and dynamical method without any assumptions on nuclear structure , is suitable for incorporating fluctuations of particle distributions . previously , we have reported the first results of our study on nuclear `` pasta '' by qmd , which demonstrated that the `` pasta '' phases can be formed in a dynamical way for matter with proton fractions @xmath6 and 0.5 @xcite . in this full paper , we present new results for astrophysically interesting neutron - rich matter of @xmath7 in addition to the cases of @xmath6 and 0.5 reported before . the plan of this paper is as follows . in section ii , we describe the framework of the qmd model used in the present study . we then show the results of our simulations in section iii and analyze the structure of matter obtained by the simulations using two - point correlation functions and minkowski functionals in section iv . in section v , we investigate the properties of the effective nuclear interaction used in this work in order to examine the validity of our results in terms of nuclear forces . astrophysical discussions are given in section vi . summary and conclusions are presented in section vii . we have various types of molecular dynamics methods for nucleons including representative ones such as fermionic molecular dynamics ( fmd ) @xcite , antisymmetrized molecular dynamics ( amd ) @xcite and qmd , etc . in the present work , we choose qmd among them balancing between calculation cost and accuracy . the typical length scale @xmath8 of inter - structure is @xmath9 fm and the density region of interest is just below the normal nuclear density @xmath10 . the required nucleon number @xmath11 in order to reproduce @xmath12 unit structures in the simulation box is about @xmath13 ( for slabs ) . it is thus desirable that we prepare nucleons of order 10000 if we try to reduce boundary effects down to a satisfactory level by reproducing several several unit structures in the box . while it is a hard task to treat such a large system with , for example , fmd and amd whose calculation costs scale as @xmath14 , it is feasible to do it with qmd whose calculation costs scale as @xmath15 . this difference comes from summations in the slater determinants in the trial wave functions of the former models . in qmd , on the other hand , the total @xmath11-nucleon wave function @xmath16 is assumed to be a direct product of single - nucleon states @xmath17 : @xmath18 the single - nucleon state is represented by a gaussian wave packet : @xmath19}\ , \label{packet}\ ] ] where @xmath20 and @xmath21 are the centers of position and momentum of the packet @xmath22 , respectively , and @xmath23 is a parameter related to the extension of the wave packet in the coordinate space . it is also noted that we mainly focus on the macroscopic structures ; the exchange effect would not be so important for them . this can be seen by comparing the typical values of the exchange energy for the macroscopic scale and of the energy difference between two successive phases with nonspherical nuclei . suppose there are two identical nucleons , @xmath24 and 2 , bound in different nuclei each other . the exchange energy between these particles is calculated as an exchange integral : @xmath25 where @xmath26 is the potential energy . an asymptotic form of the wave function is given by @xmath27 with @xmath28 and @xmath29 , where @xmath30 is the binding energy and @xmath31 is the nucleon mass . the exchange integral reads @xmath32 } \sim 5 \times 10^{-6 } { \rm mev}\ ] ] for the internuclear distance @xmath33 fm and @xmath34 mev , which is extremely smaller than the typical energy difference per nucleon between the `` pasta '' phases of order 0.1 kev ( for nsm , see fig . 4 in ref . @xcite ) - 10 kev ( for snm , see fig . 4 in ref . therefore , it is expected that qmd , which is less elaborate in treating the exchange effect , is not bad approximation for investigating the nuclear `` pasta '' . consequently , qmd has the advantages over the other models in the present study . in the future , we will have to confirm the validity of the results obtained by qmd using other more elaborate model such as amd or fmd to treat the exchange effect more precicely . however , this problem is beyond the scope of the present work . to simulate nuclear matter at subnuclear densities within the framework of qmd , we use a qmd model hamiltonian developed by maruyama et al . @xcite , which is constructed so as to reproduce bulk properties of nuclear matter and properties of finite nuclei . this model hamiltonian consists of the following six terms @xmath35 where @xmath36 is the kinetic energy , @xmath37 is the pauli potential introduced to reproduce the pauli principle effectively , @xmath38 is the skyrme potential which consists of an attractive two - body term and a repulsive three - body term , @xmath39 is the symmetry potential , @xmath40 is the momentum - dependent potential introduced as two fock terms of the yukawa interaction and @xmath41 is the coulomb potential . the expressions of these terms are given as @xmath42 } \ \delta_{\tau_i \tau_j } \delta_{\sigma_i \sigma_j}\ , \label{pauli}\\ v_{\rm skyrme } & = & { \alpha\over 2\rho_0}\sum_{i , j ( \neq i ) } \rho_{ij } + { \beta\over ( 1+\tau)\ \rho_0^{\tau } } \sum_i \left [ \sum_{j ( \neq i ) } \int { d^3{\bf r } \ \tilde{\rho_i}({\bf r } ) \ \tilde{\rho_j}({\bf r } ) } \right]^{\tau}\ , \label{skyrme}\\ v_{\rm sym } & = & { c_{\rm s}\over 2\rho_0 } \sum_{i , j(\neq i ) } \ , ( 1 - 2 | c_i - c_j | ) \ \label{sym}\\ v_{\rm md } & = & v_{\rm md}^{(1 ) } + v_{\rm md}^{(2 ) } \nonumber \\ & = & { c_{\rm ex}^{(1 ) } \over 2\rho_0 } \sum_{i , j(\neq i ) } { 1 \over 1+\left[{{\bf p}_i-{\bf p}_j \over \hbar \mu_1}\right]^2 } \ \rho_{ij } + { c_{\rm ex}^{(2 ) } \over 2\rho_0 } \sum_{i , j(\neq i ) } { 1 \over 1+\left[{{\bf p}_i-{\bf p}_j \over \hbar \mu_2}\right]^2 } \ \rho_{ij}\ , \label{md}\\ v_{\rm coulomb } & = & { e^2 \over 2}\sum_{i , j(\neq i ) } \left(\tau_{i}+\frac{1}{2}\right ) \ , \left(\tau_{j}+\frac{1}{2}\right ) \int\!\!\!\!\int d^3{\bf r}\,d^3{\bf r}^{\prime } { 1 \over|{\bf r}-{\bf r}^{\prime}| } \ , \rho_i({\bf r})\rho_j({\bf r}^{\prime})\ , \label{coulomb}\end{aligned}\ ] ] where @xmath43 means the overlap between the single - nucleon densities , @xmath44 and @xmath45 , for @xmath22-th and @xmath46-th nucleons given as @xmath47 @xmath48 is the nucleon spin and @xmath49 is the isospin ( @xmath50 for protons and @xmath51 for neutrons ) and @xmath52 and @xmath23 are model parameters determined to reproduce the properties of the ground states of the finite nuclei , especially heavier ones , and the saturation properties of nuclear matter @xcite . a parameter set used in this work is shown in table [ parameter ] . the single - nucleon densities @xmath44 and @xmath53 are given by @xmath54}\ , \quad \\ \tilde{\rho_i}({\bf r } ) & = & \frac{1}{(2\pi \tilde{l})^{3/2}}\ \exp{\left [ - \frac{({\bf r } - { \bf r}_i)^2}{2\tilde{l } } \right]}\ , \end{aligned}\ ] ] with @xmath55 the modified width @xmath56 in @xmath53 is introduced in the three - body term of skyrme interaction [ eq . ( [ skyrme ] ) ] to incorporate the effect of the repulsive density - dependent term . .effective interaction parameter set + ( incompressibility @xmath57=280 mev ; medium eos model in ref . @xcite ) [ cols="^,^,^,^ " , ] [ parameter ] we adopt qmd equations of motion with friction terms to simulate the dynamical relaxation : @xmath58 where the friction coefficients @xmath59 and @xmath60 are positive definite , which determine the relaxation time scale . the relaxation scheme given by eqs . ( [ qmdeom fric ] ) is referred to as the steepest descent method and it leads to the continuous decrease in @xmath61 as @xmath62 even though it is recognized that this method is not efficient , it is expected that the dynamics given by eqs . ( [ qmdeom fric ] ) with @xmath63 deviates slightly in a short period from the physically grounded dynamics given by qmd equations of motion without the friction terms [ equations without the second terms in the right - hand sides of eqs . ( [ qmdeom fric ] ) ] , which we would like to respect . we have performed qmd simulations of an infinite @xmath64 system with fixed proton fractions @xmath65 0.5 and 0.3 for various nucleon densities @xmath66 [ the density region is ( 0.05 - 1.0 ) @xmath67 . we set 2048 nucleons ( 1372 nucleons in some cases ) contained in a cubic box which is imposed by the periodic boundary condition . throughout this paper , the numbers of the protons ( neutrons ) with up - spin and with down - spin are equal . the relativistic degenerate electrons which ensure the charge neutrality are regarded as a uniform background and the coulomb interaction is calculated by the ewald method taking account of the gaussian charge distribution of each wave - packet ( see appendix [ ewald sum ] ) . this method enables us to efficiently sum up contributions of long - range interactions in a system with periodic boundary conditions . for nuclear interaction , we use the effective hamiltonian developed by maruyama et al . ( medium eos model ) @xcite whose expressions are given in the last section . we first prepare an uniform hot nucleon gas at @xmath68 mev as an initial condition equilibrated for @xmath69 fm/@xmath70 in advance . in order to realize the ground state of matter , we then cool it down slowly for @xmath71 fm/@xmath70 , keeping the nucleon density constant with the frictional relaxation method [ eqs . ( [ qmdeom fric ] ) ] , etc . @xcite until the temperature gets @xmath72 mev or less . note that no artificial fluctuations are given in the simulation . the qmd equations of motion with the friction terms given by eqs . ( [ qmdeom fric ] ) are solved using the fourth - order gear predictor - corrector method in conjunction with multiple time step algorithm @xcite . integration time steps @xmath73 are set to be adaptive in the range of @xmath74 fm/@xmath70 depending on the degree of convergence . at each step , the correcting operation is iterated until the error of position @xmath75 and the relative error of momentum @xmath76 become smaller than @xmath77 , where @xmath75 and @xmath76 are estimated as the maximum values of correction among all particles . we mainly use pcs(pentium iii ) equipped with mdgrape-2 , which accelerates calculations of momentum - independent forces including the long - range coulomb force . shown in figs . [ fig pasta sym ] and [ fig pasta x0.3 ] are the resultant nucleon distributions of cold matter at @xmath78 0.5 and 0.3 , respectively . we can see from these figures that the phases with rodlike and slablike nuclei , cylindrical and spherical bubbles , in addition to the phase with spherical nuclei are reproduced in both the cases of @xmath78 0.5 and 0.3 . we here would like to mention the reasons of discrepancies between the present result and the result obtained by maruyama et al . which says `` the nuclear shape may not have these simple symmetries '' @xcite . one of the most crucial reasons seems to be the difference in treatment of the coulomb interaction . in the present simulation , we calculate the long range coulomb interaction in a consistent way using the ewald method . for the system of interest where the thomas - fermi screening length is comparable to or larger than the size of nuclei , this treatment is more adequate than that which inctroduces an artificial cutoff distance as in ref . the other crucial reason is the difference in the relaxation time scales @xmath79 fm/@xmath70 ; we set @xmath80 in the present work , but maruyama et al . set @xmath81 several@xmath82 fm/@xmath70 @xcite . in our simulation , we can reproduce the bubble - phases [ see ( d ) and ( e ) of figs . [ fig pasta sym ] and [ fig pasta x0.3 ] ] with @xmath83 fm/@xmath70 and the nucleus - phases [ see ( b ) and ( c ) of figs . [ fig pasta sym ] and [ fig pasta x0.3 ] ] with @xmath84 fm/@xmath70 . however , the matter in the density region corresponding to a nucleus - phase is quenched in an amorphous - like state when @xmath85 fm/@xmath70 . in the present work , we take @xmath79 much larger than typical time scale @xmath86 fm/@xmath70 for nucleons to thermally diffuse in the distance of @xmath9 fm at @xmath87 and @xmath88 mev . this temperature is lower than the typical value of the liquid - gas phase transition temperature in the density region of interest , it is thus considered that our results are thermally relaxed in a satisfying level . phase diagrams of matter in the ground state are shown in figs . [ phase diagram 1](a ) and ( b ) for @xmath78 0.5 and 0.3 , respectively . as can be seen from these figures , the obtained phase diagrams basically reproduce the sequence of the energetically favored nuclear shapes predicted by simple discussions @xcite which only take account of the coulomb and surface effects ; this prediction is that the nuclear shape changes like sphere @xmath4 cylinder @xmath4 slab @xmath4 cylindrical hole @xmath4 spherical hole @xmath4 uniform , with increasing density . comparing figs . 3 ( a ) and ( b ) , we can see that the phase diagram shifts towards the lower density side with decreasing @xmath89 , which is due to the tendency that the saturation density is lowered as the neutron excess increases . it is remarkable that the density dependence of the nuclear shape , except for spherical nuclei and bubbles , is quite sensitive and phases with intermediate nuclear shapes which are not simple as shown in figs . 1 and 2 are observed in two density regions : one is between the cylinder phase and the slab phase , the other is between the slab phase and the cylindrical hole phase . we note that these phases are different from coexistence phases with nuclei of simple shapes , which will be referred to as `` intermediate phases '' . we have also performed qmd simulations of matter with proton fraction @xmath7 as a more realistic condition for the neutron star matter . in this case , we have to deal with a larger system than in the cases of @xmath0 and 0.3 because enough number of protons for reproducing several nuclei in a simulation box are required to obtain significant results ; protons play an important role in generating the long - range order due to their electric charge . we have investigated the neutron - rich matter at @xmath7 with 10976 nucleons , in which 1098 protons and 9878 neutrons are contained . following basically the same procedure that used for the cases of @xmath0 and 0.3 ( see section [ sim p - rich ] for detail ) , we tried to obtain the ground - state matter . however , in the present case , we quickly relax from the initial state at @xmath90 mev to the state at @xmath91 mev , at which matter is still uniform , with a nos - hoover - like thermostat @xcite which will be discussed in another paper @xcite . after the relaxation at @xmath91 mev for @xmath92 7000 fm/@xmath70 , we then cool down the system with a relaxation time scale @xmath84 fm/@xmath70 using the qmd equations of motion with friction terms ( [ qmdeom fric ] ) . these simulations are performed by fujitsu vpp 5000 equipped with mdgrape-2 . some resultant nucleon distributions are shown in figs . [ fig sphere x0.1 ] and [ fig cylinder x0.1 ] , which correspond to the sphere phase and the cylinder phase , respectively . as can be seen in fig . [ fig sphere x0.1 ] , dripped neutrons spread over the whole region in the simulation box , which lead to smaller density contrast compared with that for the cases of @xmath0 and 0.3 depicted in figs . [ fig pasta sym ] and [ fig pasta x0.3 ] , respectively . the results obtained for @xmath7 are summarized in the phase diagram shown in fig . [ phase diagram 2 ] . a striking feature is that the wide density region from @xmath93 to @xmath94 is occupied by an intermediate phase . the structure of matter seems to change rather continuously from that consists of branching rodlike nuclei connected to each other [ obtained in the lower density region of the intermediate phase denoted by ( c , u ) ] to that consists of branching bubbles connected to each other [ higher density region of the intermediate phase ( c , u ) ] . however , in the present neutron - rich case , `` pasta '' phase with slablike nuclei can not be obtained as far as we have investigated , which will be discussed at the end of the next section . it is also noted that the density at which matter turns into uniform is lower than those in the cases of @xmath0 and 0.3 , which is consistent with the tendency that matter becomes the more neutron - rich , the saturation density becomes the lower . to analyze the spatial distribution of nucleons , we calculate two - point correlation function @xmath95 for nucleon density field @xmath96 @xmath97 where @xmath11 stands for nucleons@xmath98 . @xmath95 is here defined as @xmath99 where @xmath100 denotes an average over the position @xmath101 and the direction of @xmath102 , and @xmath103 is the fluctuation of the density field @xmath104 given by @xmath105 with @xmath106 we construct the nucleon density distribution @xmath104 from a data set of the centers of position of the nucleons by the following procedure . we first set @xmath107 ( for @xmath0 and 0.3 ) or @xmath108 ( for @xmath7 ) grid points in the simulation box and then distribute particle numbers on each grid point using the cloud - in - cell method ( see , e.g. , ref . next , we carry out the smoothing procedure in the discrete fourier space with a gaussian smoothing function corresponding to the distribution of the wave packet given by eq . ( [ packet ] ) . the density distributions @xmath109 , where @xmath110 or 1 projects on particle type @xmath22 , in the discrete real space can be obtained by the inverse fourier transformation . the fourier transformations are performed using the fft algorithm . the resultant two - point correlation functions @xmath111 , @xmath112 and @xmath113 at various densities below @xmath114 at which matter becomes uniform at zero temperature for @xmath0 , 0.3 and 0.1 are plotted in figs . [ corr x0.5 ] , [ corr x0.3 ] and [ corr x0.1 ] , respectively . we can see the general tendency , which is common for the different values of @xmath89 , that the amplitude of @xmath115 decreases with increasing the density . it is noted that even though the change in the amplitude of @xmath115 is quite noticeable , the smallest zero - point @xmath116 of @xmath115 takes similar values at various densities especially for @xmath0 and 0.3 . this feature means that the typical length scales of the nuclear structures , i.e. , the internuclear distance and the nuclear radius , keep comparable at subnuclear densities from @xmath117 to @xmath114 , which is consistent with the results obtained by the previous works ( see , e.g. , refs . this behavior just below @xmath114 will be discussed further concerning a problem about the properties of the transition to uniform matter . we also note that a strong attractive force acting between a proton and a neutron leads to the good agreement of the zero - points of @xmath118 and @xmath119 even for @xmath6 and 0.1 as well as for @xmath0 although the zero - point @xmath120 of @xmath119 is @xmath121 fm ( @xmath122 fm ) larger than that of @xmath118 for @xmath6 ( 0.1 ) at each density . this shows that the phase of the density fluctuations of protons and neutrons correlate so strongly with each other at zero temperature that they almost coincide . as can be seen by comparing @xmath111 for different values of the proton fraction @xmath89 , the amplitude decreases and the value of the smallest zero - point @xmath120 increases with decreasing @xmath89 . this behavior means that , as matter becomes more neutron - rich , not only the nucleon density distribution gets smoother but also the spatial structure becomes larger . let us then examine @xmath95 for each value of the proton fraction . for symmetric matter ( @xmath0 ) , protons and neutrons are equivalent except for the mass difference and the coulomb interaction . therefore , the two - point correlation functions @xmath123 , @xmath118 and @xmath119 are almost the same at larger values of @xmath124 . at smaller values of @xmath125 , @xmath118 is slightly smaller than @xmath119 because the repulsive coulomb interaction among protons tends to reduce the proton density inhomogeneity especially in the smaller scale . for asymmetric matter ( @xmath6 and 0.1 ) , on the contrary , the amplitude of @xmath119 is much smaller than that of @xmath118 due to the dripped neutrons which distribute rather uniformly outside the nuclei . let us here examine the properties of the transition from the phase with spherical bubbles to uniform matter for @xmath0 . for this purpose , two - point correlation function of the nucleon density fluctuation is useful . in fig . [ corr2 melt ] , we thus plot the two - point correlation function of the nucleon density fluctuation @xmath111 for several densities around the melting density @xmath114 . to compute @xmath111 , we use a 1372-nucleon system cooled down until the temperature gets @xmath126 0.05 mev by qmd equations of motion ( [ qmdeom fric ] ) with friction terms . for uniform phase , @xmath111 should be zero except for the contribution of short - range correlation . the behavior of @xmath111 shows that @xmath114 lies between 0.7@xmath127 and @xmath128 at @xmath0 [ see also fig . [ phase diagram 1](a ) ] above which long - range correlation disappears . it is noted that the smallest value of @xmath116 at which @xmath129 keeps around 8 fm even at densities just below @xmath114 . this means that , for @xmath0 , the phase with spherical bubbles whose radii are around @xmath120 suddenly disappear rather than they shrink gradually and the system turns into uniform with increasing the density because the quantity @xmath120 nearly corresponds to the half wave - length of the inhomogeneous density profile . the discontinuous change in the density profile indicates that the transition between the phase with spherical bubbles and the uniform phase is of first order . this conclusion is also obtained in the previous calculations for which the spatial structure of the nuclear matter region and/or the shape of the density profile are assumed @xcite . in the present work , we have confirmed the first order nature by qmd simulations without these assumptions as several authors have done so without the assumptions by the thomas - fermi approximation @xcite . for the cases of @xmath6 and 0.1 , we could not see the significant sign of the first order nature of the transition between the mixed phase and the uniform phase because the amplitude of @xmath111 is quite small just below @xmath114 . further study is necessary to determine the properties of the transition for theses cases of asymmetric matter . to extract the morphological characteristics of the nuclear shape changes and the intermediate phases , we introduce the minkowski functionals ( see , e.g. , ref . @xcite and references therein ; a concise review is provided by ref . @xcite ) as geometrical and topological measures of the nuclear surface . let us consider a homogeneous body @xmath130 in the @xmath131-dimensional eucledian space , where @xmath132 is the class of such bodies . morphological measures are defined as functionals @xmath133 which satisfy the following three general properties : * _ motion invariance _ : the functional is independent of the position and the direction of the body , i.e. , @xmath134 where @xmath135 denotes any translations and rotations . * _ additivity _ : the functional of the union of two bodies should behave like a volume . the contribution of the overlapping region should be subtracted , i.e. , @xmath136 where @xmath137 . * _ continuity _ : if the body is approximated with pixels , the functional of the approximate body converges to that of the original body when the pixels get smaller , i.e. , @xmath138 where @xmath57 is a convex body and @xmath139 is a sequence of convex bodies . hadwiger s theorem in integral geometry states that there are just @xmath140 independent functionals which satisfy the above properties ; they are known as minkowski functionals . in three dimensional space , four minkowski functionals are related to the volume , the surface area , the integral mean curvature and the euler characteristic . in classifying nuclear shapes including those of the intermediate phases obtained in our simulations , the integral mean curvature and the euler characteristic are useful , which will be discussed later . both are described by surface integrals of the following local quantities : the mean curvature @xmath141 and the gaussian curvature @xmath142 , i.e. , @xmath143 and @xmath144 , where @xmath145 and @xmath146 are the principal curvatures and @xmath147 is the area element of the surface of the body @xmath57 . the euler characteristic @xmath148 is a purely topological quantity and is given by @xmath149 thus @xmath150 for the sphere and the spherical hole phases and the coexistence phase of spheres and cylinders , and @xmath151 for the other ideal `` pasta '' phases , i.e. the cylinder , the slab and the cylindrical hole phases which consist of infinitely long rods , infinitely extending slabs and infinitely long cylindrical holes , respectively . we introduce the area - averaged mean curvature , @xmath152 , and the euler characteristic density , @xmath153 , as normalized quantities , where @xmath154 is the volume of the whole space . we calculate the normalized minkowski functionals , i.e. , the volume fraction @xmath155 , the surface area density @xmath156 , the area - averaged mean curvature @xmath157 and the euler characteristic density @xmath158 for @xmath0 and 0.3 by the following procedure . as described in subsection [ 2pcf ] , we first construct proton and nucleon density distributions @xmath159 and @xmath160 , where @xmath161 or 1 is the isospin projection on the proton state . we set a threshold proton density @xmath162 and then calculate @xmath163 , where @xmath164 and @xmath165 are the volume and the surface area of the regions in which @xmath166 . we find out the value @xmath167 where @xmath168 and define the regions in which @xmath169 as nuclear regions . for spherical nuclei , for example , @xmath170 corresponds to a point of inflection of a radial density distribution . in the most phase - separating region , the values of @xmath171 distribute in the range of about @xmath172 @xmath173 in the both cases of @xmath174 and 0.3 , where @xmath171 is the threshold nucleon density corresponds to @xmath170 . we then calculate @xmath155 , @xmath175 , @xmath176 and @xmath148 for the identified nuclear surface . we evaluate @xmath155 by counting the number of pixels at which @xmath177 is higher than @xmath170 , @xmath175 by the triangle decomposition method , @xmath176 by the algorithm shown in ref . @xcite in conjunction with a calibration by correction of surface area , and @xmath148 by the algorithm of ref . @xcite and by that of counting deficit angles @xcite , which are confirmed that both of them give the same results . we have plotted the resultant @xmath66 dependence of @xmath155 , @xmath156 , @xmath157 and @xmath153 for the isodensity surface of @xmath178 in figs . [ fig minko 1 ] and [ fig minko 2 ] . in addition to the values of @xmath155 , @xmath156 and @xmath157 for the isodensity surface of @xmath178 , we have also investigated those for the isodensity surfaces of @xmath179 to examine the extent of the uncertainties of these quantities which stem from the arbitrariness in the definition of the nuclear surface . as shown in fig . [ fig minko 1 ] , these uncertainties are at most @xmath180 for @xmath155 and @xmath181 @xmath182 for @xmath156 . for @xmath157 , we could not observe remarkable differences from the values for @xmath178 ( they were smaller than 0.015 @xmath182 ) . we could not see these kinds of uncertainties in @xmath153 , except for the densities near below @xmath114 . as shown in fig . [ fig minko 1 ] , the volume fraction @xmath155 of the nuclear regions increases almost monotonically under @xmath114 in both cases of @xmath0 and 0.3 . this feature reflects the incompressible nature of nuclear matter . it is interesting to see the density dependence of the nuclear surface density @xmath156 because this quantity is directly related to the surface energy density , which is one of the key factors in determining the nuclear shape . figs . [ fig minko 1](c ) and [ fig minko 1](d ) show that , as the nucleon density increases , @xmath156 increases at a nearly constant rate until @xmath183 , and then its increasing rate becomes rather smaller around the density region of the slab phase , and finally it begins to decrease in the density region of the cylindrical hole phase or the spherical hole phase . this general behavior can be understood from the density dependence of the surface energy density obtained by simple arguments , which only allow for the nuclear surface and the coulomb effects ( see , e.g. , refs . @xcite ) . we also plot @xmath155 and @xmath156 for the nucleon density distribution @xmath184 as functions of the threshold density @xmath185 evaluated at various values of @xmath66 . the results for @xmath0 and 0.3 are shown in figs . [ fig minko thres x0.5 ] and [ fig minko thres x0.3 ] , respectively . features of nuclear shape changes can be seen in the behavior of the curves of @xmath156 . peaks in the higher @xmath185 region are attributed to nucleons in the nuclear matter regions and broad bumps in the lower @xmath185 region around 0.05 0.1@xmath127 observed for @xmath6 are due to the dripped neutrons outside nuclei . @xmath156 in the intermediate @xmath185 region in which its slope is nearly constant mainly comes from contribution of nuclear surfaces . as the nucleon density @xmath66 increases , the higher @xmath185 peak becomes more clear and the position of the center of the peak finally coincides with @xmath66 in the uniform phase . this feature shows that the nuclear matter regions become more uniform with increasing the density . when the dispersion of the internucleon distance is small , large surface area caused by the gaussian density distribution of each nucleon can be picked up with a single value of @xmath185 . as can be seen in fig . [ fig minko thres x0.3 ] , the lower @xmath185 bump , in turn , disappears with increasing @xmath66 . this is because , as @xmath66 increases , the dripped neutron gas becomes more inhomogeneous and tends to distribute close to the nuclear surface leading to a lower proton fraction in the nuclear matter regions . it is also noted that , as the nucleon density increases , the slope in the intermediate @xmath185 region changes from negative to positive at the density corresponding to the phase of slablike nuclei ( 0.4@xmath127 for @xmath0 and 0.35@xmath127 for @xmath6 ) , which is consistent with what is expected from the sign of @xmath157 for the nuclear surface [ see figs . [ fig minko 2](a ) and [ fig minko 2](b ) ] . let us then focus on @xmath157 and @xmath158 to classify the nuclear shape . the behavior of @xmath157 shows that it decreases almost monotonically from positive to negative with increasing @xmath66 until the matter turns into uniform . the densities corresponding to @xmath186 are about 0.4 and 0.35@xmath127 for @xmath0 and 0.3 , respectively ; these values are consistent with the density regions of the phase with slablike nuclei ( see fig . [ phase diagram 1 ] ) . as mentioned previously , @xmath153 is actually positive in the density regions corresponding to the phases with spherical nuclei , coexistence of spherical and cylindrical nuclei , and spherical holes because of the existence of isolated regions . as for those corresponding to the phases with cylindrical nuclei , planar nuclei and cylindrical holes , @xmath187 . the fact that the values of @xmath153 are not exactly zero for nucleon distributions shown as the slab phase in figs . [ fig pasta sym ] and [ fig pasta x0.3 ] reflects the imperfection of these `` slabs '' , which is due to the small nuclear parts connecting the neighboring slabs . however , we can say that the behaviors of @xmath153 plotted in figs . [ fig minko 2](c ) and [ fig minko 2](d ) show that @xmath153 is negative in the density region of the intermediate phases , even if we take into account the imperfection of the obtained nuclear shapes and the uncertainties of the definition of the nuclear surface . this means that the intermediate phases consist of nuclear surfaces which are saddle - like at each point on average and they consist of each highly connected nuclear and gas regions due to a lot of tunnels [ see eq . ( [ euler ] ) ] . using the quantities @xmath157 and @xmath158 , the sequence of the nuclear shapes with increasing the density can be described as follows : @xmath188 @xmath4 @xmath189 @xmath4 @xmath190 @xmath4 @xmath191 @xmath4 @xmath192 @xmath4 @xmath193 @xmath4 @xmath194 @xmath4 uniform . let us now consider the discrepancy from the results of previous works which do not assume nuclear structure @xcite ; the intermediate phases can not be seen in these works . we can give following two reasons for the discrepancy . * these previous calculations are based on the thomas - fermi approximation which can not sufficiently incorporate fluctuations of nucleon distributions . this shortcoming may result in favoring nuclei of smoothed simple shapes than in the real situation . * there is a strong possibility that some highly connected structures which have two or more substructures in a period are neglected in these works because only one structure is contained in a simulation box . it is not unnatural that the phases with highly connected nuclear and bubble regions are realized as the most energetically stable state @xcite . it is considered that , for example , a phase with perforated slablike nuclei , which has negative @xmath153 , could be more energetically stable than that with extremely thin slablike nuclei . the thin planar nucleus costs surface - surface energy which stems from the fact that nucleons bound in the nucleus feel its surfaces of both sides . the surface - surface energy brings about an extra energy increase in addition to the contribution of the surface energy . we have to examine the existence of the intermediate phases by more extensive simulations with larger nucleon numbers and with longer relaxation time scales in the future . in the case of @xmath7 , the criterion for identification of the isodensity surface corresponding to the nuclear surface using the second derivative of @xmath195 does not work at higher densities . we thus use another method to calculate the normalized minkowski functionals of the nuclear surface for @xmath7 . in fig . [ fig euler typical ] , we have plotted the @xmath185 dependence of the euler characteristic density @xmath158 at @xmath196 as an example . we can see that this curve consists of three components : the peaks of the lower @xmath185 region , the plateau region and the peaks of the higher @xmath185 region , which are due to dripped neutrons ( thus these peaks can not be observed for @xmath0 ) , nuclear surfaces and nucleons in nuclei , respectively . these components can also be seen at the other values of @xmath66 lower than @xmath114 . however , we have to mention that the higher the density becomes , the smaller the plateau region gets , which means that the density contrast between the dripped neutron gas region and the nuclear matter region becomes obscure . here , we take the mean values of the normalized minkowski functionals in the plateau region as those for the nuclear surface , which are plotted as crosses in fig . [ fig minko x0.1 ] . the error bars shown in this figure are the standard deviations of these quantities in the plateau region . consistency between this method and the one using the second derivative of @xmath195 has been confirmed for @xmath6 . [ fig minko x0.1 ] shows the resultant normalized minkowski functionals for the nuclear surface at various values of @xmath66 below @xmath114 . the qualitative behaviors of @xmath155 and @xmath157 for @xmath7 are the same as those for @xmath0 and 0.3 ; as @xmath66 increases , @xmath155 increases and @xmath157 decreases ( from positive to negative ) almost monotonically in the density region of 0.1@xmath197 . however , in the behaviors of @xmath156 and @xmath158 , qualitative differences can be observed between the present case and the cases of @xmath0 and 0.3 . as can be seen in fig . [ fig minko x0.1](b ) , @xmath156 increases almost linearly until just below @xmath114 . the absence of the phases with cylindrical bubbles and with spherical bubbles in the phase diagram of @xmath7 ( fig . [ phase diagram 2 ] ) is well characterized by the behavior of @xmath158 shown in fig . [ fig minko x0.1](d ) . in the cases of @xmath0 and 0.3 , @xmath158 increases from negative to positive with increasing density in the density region higher than that of the slab phase . however , for @xmath7 , we can not observe the tendency that @xmath158 starts increasing even at just below @xmath114 . as a result , @xmath158 keeps negative until matter becomes uniform . ( further discussion will be given in section [ int ] . ) let us then consider the phase with slablike nuclei , which have not been obtained in the simulations for @xmath7 . if it were realized by using a longer relaxation time scale , it is expected to be obtained at @xmath198 0.34 @xmath127 according to the behaviors of the @xmath157 and @xmath158 . however , we can not see any signs from fig . [ fig minko x0.1](b ) that @xmath156 stops increasing in this density region unlike the behaviors of @xmath156 in the cases of @xmath0 and 0.3 . here , we would like to mention that , according to the landau - peierls argument , thermal fluctuations are effective at destroying the long - range order of one - dimensional layered lattice of slablike nuclei rather than that of triangular lattice of rodlike nuclei and of bcc lattice of spherical ones . thus , the melting temperature of the planar phase would be lower than the other phases , which leads to a longer time scale for formation of the slablike nuclei by the thermal diffusion . therefore , a further investigation with a longer relaxation time scale is necessary to determine whether or not the phase with slablilke nuclei is really prohibited in such neutron - rich matter in the present model . in fig . [ fig minko thres x0.1 ] , we have also plotted @xmath155 and @xmath156 for the density distribution @xmath184 as functions of @xmath185 like figs . [ fig minko thres x0.5 ] and [ fig minko thres x0.3 ] . in comparison with fig . [ fig minko thres x0.3 ] , the contribution of the dripped neutrons is shown more clearly in this case . we can see that the peak in the lower @xmath185 region due to the dripped neutrons combines to the peak in the higher @xmath185 region . this behavior stems from the fact that a part of the dripped neutrons at lower densities are absorbed into nuclear matter region with increasing the density at fixed @xmath89 ; finally , all the neutrons are contained there in the uniform phase . we can also expect from the @xmath185 dependence of @xmath156 that the phase with slablike nuclei might be obtained around @xmath199 , where the slope of the plateau region is close to zero . let us here examine the effective nuclear interaction used in this work . structure of matter at subnuclear densities is affected by the properties of neutron - rich nuclei and of the pure neutron gas resulted from the nuclear interaction . key quantities are the energy per nucleon @xmath1 of the pure neutron matter , the proton chemical potential @xmath2 in the pure neutron matter , and the nuclear surface tension @xmath3 . there is a tendency , especially in the case of neutron star matter , that the higher @xmath1 , the density @xmath114 at which matter becomes uniform is lowered . this is because larger @xmath1 tends to favor uniform nuclear matter without dripped neutron gas regions than mixed phases with dripped neutron gas regions . in the neutron star matter , there is also a tendency that the lower @xmath2 , the smaller @xmath114 . this is because @xmath200 represents the degree to which the neutron gas outside the nuclei favors the presence of protons in itself . the quantity @xmath3 controls the size of the nuclei and bubbles , and hence the sum of the coulomb and surface energies . with increasing @xmath3 and so this energy sum , @xmath114 gets lowered . it is important to check whether the effective nuclear force given by eqs . ( [ hamiltonian])([coulomb ] ) yields unrealistic values of these quantities or not . if @xmath1 , @xmath201 and @xmath3 for the present model are unrealistically small in comparison with those for the other models , our results which have reproduced the `` pasta '' phases might be quite limited for the present model hamiltonian . in order to evaluate @xmath1 , we set 1372 neutrons in a simulation box imposed of the periodic boundary condition . this system is cooled down by the qmd equations of motion with friction terms [ see eqs . ( [ qmdeom fric ] ) ] until the temperature becomes @xmath126 1 kev . the resultant values of @xmath1 are plotted in fig . [ fig e_n ] . we note that our results for @xmath2020.2 , 0.6 and 1.0 @xmath127 ( the result for @xmath203 is not plotted in fig . [ fig e_n ] ) coincide with the results for zero - proton ratio plotted in fig . 9 of ref . @xcite . the values of @xmath1 for the present model behave like those for the skm skyrme interaction especially in the density region of @xmath204 @xmath173 ; they are close to the result of the variational chain summation obtained by akmal , pandharipande and ravenhall @xcite at @xmath205 . the steep rise in @xmath1 in the higher neutron density region ( @xmath206 @xmath173 ) compared to those obtained from the hartree - fock theory using various skyrme interactions would help neutron - rich matter , which have larger dripped neutron density of @xmath206 @xmath173 , to be uniform . therefore , we can say that this behavior of @xmath1 for the present qmd model hamiltonian suppresses the density region in which the `` pasta '' phases are the most energetically favorable in neutron star matter and in the case of @xmath7 in the present study . we also note that @xmath1 at lower neutron densities of @xmath207 @xmath173 is relatively small . this , in turn , would lead to increase the density at which matter with lower dripped neutron density ( e.g. , the case of @xmath6 in the present study ) turns into uniform . next , we calculate the proton chemical potential @xmath2 in the pure neutron matter . we use the cold neutron matter prepared for the above calculation of @xmath1 as an initial condition . we insert a proton into this pure neutron matter , and then minimize the total energy by the frictional relaxation method with fixing the positions and momenta of the other neutrons . the position of the inserted proton is chosen randomly in the simulation box , and its momentum is chosen randomly from @xmath208 mev@xmath209 . we evaluate @xmath2 as the difference in the total energy between that before the insertion of the proton and that after the optimization of the position and the momentum of the proton . in fig . [ fig mu_p0 ] , we plot @xmath2 for the present model hamiltonian . as can be seen from this figure , the result for the present model hamiltonian generally reproduce the data of the other results obtained from the hartree - fock theory using the various skyrme interactions at densities @xmath210 @xmath173 . at lower densities of @xmath211 @xmath173 , errors are quite large and data scatter significantly . this is because density fluctuations in pure neutron matter obtained by qmd would be unrealistically large at such low densities due to the fixed width of the wave packets in this model . however , it is noted that even in such a density region , our data are generally consistent with the other results mentioned above . finally , we turn to the surface tension , which affects energetically favorable nuclear shape most directly among the three quantities discussed here . we have calculated @xmath3 from the energy change induced by the change in the area of planar nuclear matter . we use nucleon gas with proton fractions @xmath7 , 0.2 , 0.3 and 0.5 composed of 1372 nucleons . in the calculation of @xmath3 , the coulomb interaction is excluded . in order to prepare slablike nuclear matter , we first cool down the above nucleon gas from @xmath90 mev to @xmath212 mev using eqs . ( [ qmdeom fric ] ) in a shallow trapping harmonic potential @xmath213 where @xmath214 mev @xmath215 . the simulation box here is imposed of the periodic boundary condition in the @xmath89- and @xmath216-directions , and is imposed of the open boundary condition in the @xmath217-direction . the box size @xmath218 in the @xmath89- and @xmath216-directions is set 20.26 fm . when the temperature reaches @xmath212 mev , we remove the trapping potential and , except for the case with @xmath0 , we change the boundary condition in the @xmath217-direction from open to periodic . the box size in the @xmath217-direction @xmath219 is chosen so as to at least all nucleons dripped outside the nuclear matter region can be contained in the box : @xmath220 fm ( @xmath7 ) , 79.12 fm ( @xmath221 ) , and 82.85 fm ( @xmath6 ) . after we relax the system for @xmath126 7000 fm/@xmath70 , we prepare three kinds of samples : one is nothing changed ( sample 1 ) and the others have the area of the @xmath222 side of the simulation box increased ( decreased ) by 1 @xmath223 with the total volume of the box kept constant [ sample 2 ( sample 3 ) ] . we further cool them down until @xmath224 kev . the resultant nucleon density profiles for the sample 1 of @xmath7 projected on the @xmath217-axis is shown in fig . [ fig profile ] . as can be seen from this figure , @xmath225 fm is much larger than the thickness of the slab @xmath226 fm , and thus the volume of the dripped neutron gas region is almost the same among the three kinds of samples because the volume change in the nuclear matter region is negligible . it is also noted that @xmath131 is much larger than the surface thickness @xmath227 fm , which ensures that the surfaces at both sides of the slab are separated well . therefore , we can say that the energy difference between the sample 2 and 3 is just due to the difference in the surface area of the planar nuclear matter . following the spirit of ravenhall , bennett and pethick ( hereafter rbp ) @xcite and by using the sample 1 , we define the proton fraction @xmath228 in the nuclear matter region as an averaged value for the region of the width of @xmath229 fm in the central part of the slab , where the proton and neutron density profiles ripple around constant values . we extract @xmath3 from the total energy @xmath230 of the sample 2 and @xmath231 of the sample 3 as follows : @xmath232 where @xmath233 is the area of the @xmath222 side of the sample 2 ( sample 3 ) given by @xmath234 @xmath235 ( @xmath236 @xmath235 ) . the factor 2 in the denominator represents , of course , the contribution of the two sides of the planar nuclear matter . as shown in fig . [ fig esurf ] , all the results plotted in this figure almost coincide with each other at @xmath237 , where uncertainty in the nuclear surface tension is rather small . these results deviate significantly at lower @xmath228 , and the values of @xmath3 of the present calculation lie between those obtained by baym , bethe and pethick ( hereafter bbp ) @xcite and by rbp @xcite , which is based on the hartree - fock calculation using a skyrme interaction , at @xmath238 . thus we can say that , in comparison with the result by rbp taken to be standard here , the contribution of @xmath3 of the present qmd model tends to favor uniform nuclear matter rather than the inhomogeneous `` pasta '' phases for lower @xmath228 of @xmath239 . now let us discuss the fact that the intermediate phases with a negative value of @xmath158 are obtained instead of the bubble phases in a wide density region for @xmath7 . in considering this problem , we should note the general tendency that @xmath201 increases and @xmath3 decreases as matter becomes neutron - rich ( see figs . [ fig mu_p0 ] and [ fig esurf ] ) . the density dependence of @xmath156 monotonically increases until @xmath114 as shown in fig . [ fig minko x0.1 ] would partly stem from the small value of @xmath3 . the apparently lower melting density @xmath114 in this case than in the cases of @xmath0 and 0.3 , even though @xmath3 is small , is due to a large @xmath201 , which increases typically of order 10 mev as @xmath240 increases . the small @xmath3 and the large @xmath201 in neutron - rich matter would help nuclear matter regions and neutron gas regions mix each other at the cost of small surface energy below @xmath114 . as a result , the structures with a negative @xmath158 could be favored . according to fig . [ fig mu_p0 ] , the quantity @xmath2 obtained for the present model hamiltonian is consistent with those for the other skyrme - hartree - fock calculations in the relevant region of @xmath241 @xmath173 . it is thus possible that a result which shows the structure of matter changes from negative @xmath158 to uniform without undergoing `` swiss cheese '' structure will be obtained by another calculation for neutron - rich matter using some framework without assuming nuclear shape . in closing this subsection , we summarize the consequences of the resultant @xmath1 , @xmath2 and @xmath3 for the present qmd hamiltonian . * for symmetric matter ( @xmath242 ) + according to @xmath3 at @xmath237 , the present model is consistent with the other results , and is an appropriate effective interaction for the study of the `` pasta '' phases at @xmath0 . * for neutron - rich cases + for neutron - rich cases such as @xmath243 , in which the dripped neutron density @xmath240 grows @xmath244 @xmath173 just before matter turns into uniform , the present qmd model can be taken as a conservative one in reproducing the `` pasta '' phases . its @xmath1 , @xmath2 and @xmath3 act to suppress the density region of the `` pasta '' phases compared to other skyrme - hartree - fock results . * for intermediate cases + at intermediate proton fraction of @xmath245 , @xmath1 of the present model acts to favor the inhomogeneous `` pasta''phases rather than the uniform phase and @xmath3 acts in the opposite way in comparison with other skyrme - hartree - fock results . here we would like to discuss astrophysical consequences of our results . pethick and potekhin have pointed out that elastic properties of `` pasta '' phases with rodlike and slablike nuclei are similar to those of liquid crystals , which stems from the similarity in the geometrical structures @xcite . it can also be said that the intermediate phases observed in the present work are `` spongelike '' ( or `` rubberlike '' for @xmath246 ) phases because they have both highly connected nuclear and bubble regions shown as @xmath247 . the elastic properties of the spongelike intermediate phases are qualitatively different from those of the liquid - crystal - like `` pasta '' phases because the former ones do not have any directions in which the restoring force does not act ; while the latter ones have . our results imply that the intermediate phases occupy a significant fraction of the density region in which nonspherical nuclei can be seen ( see figs . [ phase diagram 1 ] and [ phase diagram 2 ] ) . according to figs . [ phase diagram 1 ] and [ phase diagram 2 ] , we expect that the maximum elastic energy that can be stored in the neutron star crust and supernova inner core is higher than that in the case where all nonspherical nuclei have simple `` pasta '' structures . besides , the cylinder and the slab phases , which are liquid - crystal - like , lie between the spongelike intermediate phases or the crystalline solidlike phase , and the releasing of the strain energy would , in consequence , concentrate in the domain of these liquid - crystal - like phases . the above effects of the intermediate phases should be taken into account in considering the crust dynamics of starquakes and hydrodynamics of the core collapse , etc . if these phases exist in neutron star matter and supernova matter . in the context of pulsar glitch phenomena , the effects of the spongelike nuclei on the pinning rate and the creep velocity of superfluid neutron vortices also have yet to be investigated . for neutrino cooling of neutron stars , some version of the direct urca process which is suggested by lorenz et al . @xcite , that this might be allowed in the `` pasta '' phases , would be suppressed in the intermediate phases . this is due to the fact that the proton spectrum at the fermi surface is no longer continuous in the spongelike nuclei . an important topic which we would like to mention is about the effects of the intermediate phases on neutrino trapping in supernova cores . the nuclear parts connect over a wide region which is much larger than that characterized by the typical neutrino wave length @xmath248 fm . thus the neutrino scattering processes are no longer coherent in contrast to the case of the spherical nuclei , and this may , in consequence , reduce the diffusion time scale of neutrinos as in the case of `` pasta '' phases with simple structures . this reduction softens the supernova matter and would thus act to enhance the amount of the released gravitational energy . it would be interesting to estimate the neutrino opacity of the spongelike phases and the `` pasta '' phases . finally , we would like to mention the thermal fluctuations with long wavelengths leading to displacements of `` pasta '' nuclei , which can not be incorporated into the simulations using a finite - size box . even if we succeed in reproducing the phase with slablike nuclei in neutron - rich matter in the future study , we should remind the above effect of thermal fluctuations to consider the real situation of matter in inner crusts of neutron stars . following the discussions in refs . @xcite by a liquid - drop model , it is likely that the extension of slablike nuclei is limited to a finite length scale of @xmath249 fm in the temperature regions typical for neutron star crusts and supernova cores . we have performed qmd simulations for matter with fixed proton fractions @xmath0 , 0.3 and 0.1 at various densities below the normal nuclear density . our calculations without any assumptions on the nuclear shape demonstrate that the `` pasta '' phases with rodlike nuclei , with slablike nuclei , with cylindrical bubbles and with spherical bubbles can be formed dynamically from hot uniform matter within the time scale of @xmath80 fm/@xmath70 in the proton - rich cases of @xmath0 and 0.3 . we also demonstrate that the `` pasta '' phase with cylindrical nuclei can be formed dynamically within the time scale of @xmath84 fm/@xmath70 for the neutron - rich case of @xmath7 . our results imply the existence of at least the phase with cylindrical nuclei in neutron star crusts because they cool down keeping the local thermal equilibrium after proto - neutron stars are formed and their cooling time scale , which is macroscopic one , is much larger than the relaxation time scale of our simulations . in addition to these `` pasta '' phases with simple structures , our results obtained here also suggest the existence of intermediate phases which have complicated nuclear shapes . we have systematically analyzed the structure of matter with two - point correlation functions and with morphological measures `` minkowski functionals '' , and have demonstrated how structure changes with increasing density . making use of a topological quantity called euler characteristic , which is one of the minkowski functionals , it has been found that the intermediate phases can be characterized as those with negative euler characteristic . this means that the intermediate phases have `` spongelike '' ( or `` rubberlike '' for @xmath246 ) structures which have both highly connected nuclear and bubble regions . the elastic properties of the spongelike intermediate phases are qualitatively different from those of the liquid - crystal - like `` pasta '' phases . we have also investigated the properties of the effective qmd interaction used in the present work in order to examine the validity of our results . important quantities which affect the structure of matter are the energy per nucleon @xmath1 of the pure neutron matter , the proton chemical potential @xmath2 in pure neutron matter and the nuclear surface tension @xmath3 . the resultant these quantities show that the present qmd interaction has generally reasonable properties at subnuclear densities among other nuclear interactions . it is thus concluded that our results are not exceptional ones in terms of nuclear forces . our results which suggest the existence of the highly connected intermediate phases as well as the simple `` pasta '' phases provide a vivid picture that matter in neutron star inner crusts and supernova inner cores has a variety of material phases . the stellar region which we have tried to understand throughout this paper is relatively tiny , but there has quite rich properties which stem from the fancy structures of dense matter . g. w. is grateful to t. maruyama , k. iida , a. tohsaki , h. horiuchi , k. oyamatsu , h. takemoto , k. niita , s. chikazumi , i. kayo , c. hikage , t. buchert , n. itoh and k. kotake for helpful discussions and comments . g. w. also appreciate j. lee for checking the manuscript . this work was supported in part by the junior research associate program in riken through research grant no . j130026 and by grants - in - aid for scientific research provided by the ministry of education , culture , sports , science and technology through research grant ( s ) no . 14102004 , no . 14079202 and no . 14 - 7939 . in calculating a long - range interaction , such as the coulomb interaction , it is necessary to sum up all contributions of particles at a sufficiently far distance . the ewald sum is a familiar technique for efficiently computing the long - range contributions in a system with the periodic boundary condition ( see , e.g. , refs . @xcite ; recent mathematically careful discussion relating to the conditionally convergence of the coulomb energy can be seen , e.g. , in ref . the basic idea of the ewald sum is that the contributions of particles in a long distance in real space can be calculated as contributions in the neighborhood in fourier space : the contributions of particles in a short distance is summed up in real space and those of particles in a long distance is summed up in fourier space . let us consider a system consisting of charged particles which have a gaussian charge distribution and a uniform background charge which cancels the total charge of charged particles . these @xmath11 particles are assumed to be in a cubic simulation box with volume @xmath250 which is imposed by the periodic boundary condition . if every particle @xmath22 with total charge @xmath251 is surrounded by a gaussian charge distribution with total charge @xmath252 , the electrostatic interaction of particle @xmath22 turns into a screened short - range interaction . thus the total coulomb energy @xmath253 of this system can be decomposed into as follows ( see fig . [ fig ewald ] ) : @xmath254 where @xmath255 is the sum of the coulomb energy between an each unscreened charged particle @xmath22 and the other charged particles with screening charges , @xmath256 is that between an each unscreened charged particle @xmath22 and compensating charges which cancels the screening gaussian charges , and @xmath257 is the sum of spurious self interactions between an each charged particle @xmath22 and its compensating charge @xmath22 . charge densities of the real charged particles @xmath184 and of the screening charges @xmath258 can be written as @xmath259 where @xmath260 and @xmath261 are the reciprocals of the widths of charge distributions for a charged particle and a screening charge , respectively , and * n * denotes a position vector of a periodic image normalized by @xmath262 . thus a distribution of screened charges @xmath263 is @xmath264 the electrostatic potential @xmath265 due to @xmath266 can be obtained as because a solution of the poisson equation @xmath268\ ] ] is @xmath269 where erf(@xmath89 ) and erfc(@xmath89 ) are the error function and the complementary error function , respectively , and they are defined as @xmath270 and @xmath271 . here , we determine the constant @xmath272 so that the average value of @xmath273 in the simulation box @xmath154 be zero : @xmath274 thus , @xmath265 leads to next , we calculate @xmath256 , which is the sum of the coulomb energy between an unscreened charged particle @xmath22 and a charge density @xmath281 which consists of the compensating charges and the background charge : @xmath282 fourier transforming the charge distribution @xmath281 yields @xmath283 using @xmath284 and the poisson equation @xmath285 in the fourier form @xmath286 we can at once obtain the electrostatic potential @xmath287 due to the charge density @xmath288 : @xmath289 the term with @xmath290 is canceled due to charge neutrality . thus the coulomb energy @xmath256 is given by @xmath291 we have to subtract a sum of spurious self interactions @xmath292 between a charged particle @xmath22 and its compensating charge from @xmath256 . according to eq . ( [ solution ] ) , an electrostatic potential @xmath293 due to a gaussian compensating charge is @xmath294 , thus the self interaction of particle @xmath22 reads to @xmath295 and hence , @xmath296 finally , the total coulomb energy can be calculated by eqs . ( [ coul ] ) , ( [ u_short ] ) , ( [ u1 ] ) and ( [ u_self ] ) . the positive background charge does not appear explicitly because the average value of @xmath273 within the simulation box is set to be zero . the total coulomb energy per particle @xmath297 and the @xmath89 component of the coulomb force @xmath298 acting on a particle for various values of @xmath261 are plotted in fig . [ fig ewald ] . in this calculation , we use 1024 positive charged particles ( protons ) distributed randomly in a simulation box of @xmath299 fm ( i.e. @xmath300 ) , which is imposed of the periodic boundary condition . figs . [ fig ewald ] ( a ) and [ fig ewald ] ( b ) show the results for point charges and for gaussian charge distributions , respectively , which are calculated for the same configuration of the particle positions @xmath301 . the width of the gaussian charge distributions is set to be @xmath302 with @xmath303 @xmath235 , which corresponds to the width of the wave packets in the qmd model used in this work . we note that there are plateau regions of @xmath297 and @xmath298 whose values do not depend on @xmath261 . these constant values give results to be obtained . we note that the range of @xmath304 of the plateau regions become larger with increasing @xmath305 , where @xmath305 is the cutoff radius in the unit of @xmath306 for the summation in fourier space . as can be seen from fig . [ fig ewald ] , @xmath261 dependences of @xmath297 and @xmath298 for the present qmd model with a finite width of the gaussian charge distributions are weaker than those for the point charges . these features are also confirmed for different proton number densities of 0.2 and 0.3 @xmath127 . in our simulations , @xmath261 is set 13 or 14 , which are considered to be large enough to calculate the total coulomb energy per particle in accuracy less than @xmath307 kev , which is the typical value of the energy difference between successive `` pasta '' phases in neutron star matter obtained by previous works . bethe , g. borner and k. sato , astron . astrophys . * 7 * , 279 ( 1970 ) . g. baym , h. a. bethe and c. j. pethick , nucl . a175 * , 225 ( 1971 ) . k. sato , prog . * 54 * , 1325 ( 1975 ) . h. a. bethe , rev . phys . * 62 * , 801 ( 1990 ) . c. j. pethick and d. g. ravenhall , annu . nucl . part . * 45 * , 429 ( 1995 ) . d. g. ravenhall , c. j. pethick and j. r. wilson , phys . lett . * 50 * , 2066 ( 1983 ) . m. hashimoto , h. seki and m. yamada , prog . 71 * , 320 ( 1984 ) . lorenz , d.g . ravenhall and c.j . pethick , phys . lett . * 70 * , 379 ( 1993 ) . g. watanabe , k. iida and k. sato , nucl . * a676 * , 455 ( 2000 ) ; erratum , nucl . phys . a ( in press ) . g. watanabe , k. iida and k. sato , nucl . * a687 * , 512 ( 2001 ) ; erratum , nucl . a ( in press ) . r. d. williams and s. e. koonin , nucl . a435 * , 844 ( 1985 ) . m. lassaut , h. flocard , p. bonche , p.h . heenen and e. suraud , astron . astrophys . * 183 * , l3 ( 1987 ) . k. oyamatsu , nucl . phys . * a561 * , 431 ( 1993 ) . b. jones , mon . not . . soc . * 296 * , 217 ( 1998 ) ; mon . not . . soc . * 306 * , 327 ( 1998 ) ; phys . lett . * 83 * , 3589 ( 1999 ) . c. j. pethick and a.y . potekhin , phys . b427 * , 7 ( 1998 ) . c. j. horowitz , phys . d * 55 * ( 1997 ) 4577 . k. iida , g. watanabe and k. sato , prog . * 106 * , 551 ( 2001 ) . j. aichelin and h. stcker , phys . lett . * b176 * , 14 ( 1986 ) . h. feldmeier and j. schnack , rev . phys . * 72 * , 655 ( 2000 ) . g. watanabe , k. sato , k. yasuoka and t. ebisuzaki , phys . c * 66 * , 012801(r ) ( 2002 ) . h. feldmeier , nucl . phys . * a515 * , 147 ( 1990 ) . a. ono , h. horiuchi , t. maruyama and a. ohnishi , prog . theor . phys . * 87 * , 1185 ( 1992 ) ; phys . rev . lett . * 68 * , 2898 ( 1992 ) . t. maruyama , k. niita , k. oyamatsu , t. maruyama , s. chiba and a. iwamoto , phys . c * 57 * , 655 ( 1998 ) . in some cases , the nos - hoover - like thermostat which will be described in another paper @xcite is also used . in the present study , employment of this thermostat is almost limited in a higher temperature region where matter is completely uniform . g. watanabe , k. sato , k. yasuoka and t. ebisuzaki , in preparation . m. p. allen and d. j. tildesley , _ computer simulation of liquids _ ( clarendon , oxford , 1987 ) . t. maruyama , private communication . s. nos , j. chem . 81 * , 511 ( 1984 ) . w. g. hoover , phys . rev . a * 31 * , 1695 ( 1985 ) . r. w. hockney and j. w. eastwood , _ computer simulation using particles _ ( iop publishing , bristol , 1988 ) . r. ogasawara and k. sato , prog . * 68 * , 222 ( 1982 ) . k. michielsen and h. de raedt , phys . rep . * 347 * , 461 ( 2001 ) . j. schmalzing , s. engineer , t. buchert , v. sahni and s. shandarin , ( 1996 ) unpublished . j. r. gott iii , a. l. melott and m. dickinson , astrophys . j. * 306 * , 341 ( 1986 ) ; d. h. weinberg , publ . pacific * 100 * , 1373 ( 1988 ) . g. watanabe , phd thesis , univ . of tokyo ( 2003 ) . existence of some phase with a complicated structure is also suggested by p. magierski and p. h. heenen , phys . c * 65 * , 045804 ( 2002 ) . they focus on shell effects of dripped neutrons . non - integral values of the `` dimensionality '' @xmath131 in the early work done by ravenhall et al . @xcite also suggest the possibilty of complicated nuclear shapes . akmal , v. r. pandharipande and d. g. ravenhall , phys . c * 58 * , 1804 ( 1998 ) . b. friedman and v. r. pandharipande , nucl . a361 * , 501 ( 1981 ) . c. j. pethick , d. g. ravenhall and c. p. lorenz , nucl . phys . * a584 * , 675 ( 1995 ) . f. douchin and p. haensel , phys . lett . * b485 * , 107 ( 2000 ) . o. sjberg , nucl . a222 * , 161 ( 1974 ) . j. siemens and v. r. pandharipande , nucl . a173 * , 561 ( 1971 ) . d. g. ravenhall , c. d. bennett and c. j. pethick , phys . rev . lett . * 28 * , 978 ( 1972 ) . m. p. tosi , in : f. seitz and d. turnbull ( eds . ) , _ solid state physics : advances in research and applications _ ( academic press , new york , 1964 ) p. 1 . h. takemoto , t. ohyama and a. tohsaki , prog . . phys . * 109 * , 563 ( 2003 ) .

the initial stellar mass function ( imf ) has been observed directly in dozens of clusters in our galaxy and the large magellanic clouds , and inferred indirectly for our galaxy and other galaxies from metallicities , supernova rates , ionization levels , luminosities , dynamical masses , colors , and other tracers . there is some convergence in these observations to an imf that is approximately a power law above 1 m@xmath0 , having a slope between @xmath1 and @xmath2 on a @xmath3 histogram ( see reviews in massey 1998 and scalo 1998 ) , with a flattening below this down to @xmath4 m@xmath0 or lower . the power - law part is in rough agreement with the salpeter ( 1955 ) function , which has a slope of @xmath5 on such a plot , and also with the scalo ( 1986 ) determination , which revised the often - used miller & scalo ( 1979 ) function at intermediate and high mass ( the miller - scalo function is too steep ) . scalo ( 1998 ) recently emphasized the variations in the derived power law slopes from region to region , which amount to about @xmath6 . there are no obvious correlations between these variations and physical properties of the star - forming regions , including the cluster density , galactocentric radius , and metallicity ( scalo 1998 ) . there are also `` gaps '' in the power law parts of some cluster imfs , according to scalo ( e.g. , ngc 103 ; phelps & janes 1993 ) , and a `` ledge '' in the orion imf by hillenbrand ( 1997 ) . these features led scalo to suggest that physical effects might be involved , making imf variations larger than poisson statistics . an emphasis on the similarities , rather than the differences , in the imf has been made by massey and collaborators ( see review in massey 1998 ) . they find an imf power law slope that is fairly constant , around the salpeter value , for clusters with a factor of @xmath7 range in densities ( massey & hunter 1998 ; luhman & rieke 1998 ) and a factor of @xmath8 range in metallicities ( massey , johnson & degioia - eastwood 1995a ; see also the globular cluster study by bellazzini et al . 1995 ) . this slope was also found by kennicutt , tamblyn & congdon ( 1994 ) in galaxy - wide surveys , and by bresolin & kennicutt ( 1997 ) for individual hii regions in galaxies , using the equivalent widths of hydrogen emission lines , and it was found by greggio et al . ( 1993 ) , marconi et al . ( 1995 ) , holtzman , et al . ( 1997 ) , and grillmair , et al . ( 1998 ) from color magnitude diagrams of stars in the lmc and local dwarf galaxies . chemical evolution also tends to suggest an invariant imf , at least at intermediate and high mass , in the stellar halo ( nissen et al . 1994 ) and disk ( tsujimoto et al . 1997 ; wyse 1998 ) , in qso damped ly@xmath9 ( lu et al . 1996 ) and ly@xmath9 forest ( wyse 1998 ) lines , the intracluster medium ( renzini et al . 1993 ; wyse 1997 , 1998 ; however , see loewenstein & mushotzky 1996 ) , and elliptical galaxies ( see review in wyse 1998 ) . the observable in this case is the ratio of the iron to the oxygen abundance , which is always about the same . since this ratio scales with the ratio of intermediate mass to high mass stars , the imf must be about the same too . a different result was obtained by heikkil , johansson & olofsson ( 1998 ) , who conjectured that a low c@xmath10o / c@xmath11o ratio in the lmc results from a relatively low proportion of high - mass stars , as observed , for example , by massey et al . ( 1995b ) ; clearly the chemical and fractionation processes involved with this important isotope ratio have to be investigated further . gibson & matteucci ( 1997 ) also obtained a different result from evolution and metallicity models of elliptical galaxies , including infall and wind ejection of gas . they suggested the imf had a higher proportion of massive stars than the solar neighborhood , as it appears to in some starburst models ( discussed below ) . angeletti & giannone ( 1997 ) modeled elliptical galaxies and fit the mass - to - light and [ fe / h ] ratios with an imf that has a mass - dependent high mass slope , different from the salpeter value . the imf in local ob associations might differ from the salpeter value too . brown ( 1998 ) reviewed observations , including those with distances from the hipparcos satellite , which suggest a slightly steeper slope , @xmath12 to @xmath13 for local associations , but the stellar types in these studies were obtained from photometry , and massey ( 1998 ) discussed how such data can artificially steepen the imf without spectroscopic observations too . at the low mass end , there is no _ systematic _ indication yet of a turnover at masses lower than 0.1 m@xmath0 , but there are some regions where such a turnover has been reported , e.g. , in nearby field stars ( reid & gazis 1997a ) , orion ( hillenbrand 1997 ) , the pleiades ( reid 1998 ) , ngc 6231 ( sung , bessell , & lee 1998 ) , 30 dor ( nota 1998 ) , and the globular cluster ngc 6397 ( king et al . there are also regions where the imf appears to rise below 0.1 m@xmath0 , again for local stars ( mra et al . 1996 ; gould et al . 1997 ) . scalo ( 1998 ) suggests that the local field results at low mass contain uncertainties in the star formation rate , pre - main - sequence contribution , and mass - luminosity relation . the older cluster results are uncertain also because of evaporation of low mass members , although reid ( 1998 ) suggests evaporation is not important for the pleiades . in any case , scalo ( 1998 ) and reid ( 1998 ) both provide evidence that the low mass imf , below 1 m@xmath0 , varies slightly in shape from region to region , although no sensible correlations with properties of the environment have yet been noticed . variations in the low mass imf are also seen for globular clusters , without any obvious dependence on metallicity ( santiago et al . 1996 ; piotto et al . 1997 ; see review in cool 1998 ) , but there may be some dependence on the cluster evaporation rate and susceptibility to orbit shocking ( cool 1998 ) , since low mass stars leave the cluster most easily . the imf for the young open cluster mentioned above , ngc 6231 ( sung et al . 1998 ) , is somewhat peculiar , because the turnover mass is around 2.5 m@xmath0 , much higher than in ob associations and other clusters . however , ngc 6231 is a dense cluster @xmath8 my old , and a high turnover mass in the central regions might be expected from rapid thermalization leading to mass segregation ( e.g. , scalo 1986 ) . the low mass imf is also probed by observations of embedded clusters , where most of the stars are pre - main sequence ( zinnecker et al . 1993 ; fletcher & stahler 1994a , b ; lada & lada 1995 ; lada , alves & lada 1996 ; megeath 1996 ; horner , lada , & lada 1997 ) . pre - main - sequence ( pms ) clusters have not yet evolved kinematically , so differential evaporation and segregation of high and low mass members should not be as much of a problem as it is for older clusters . pms stars are also much brighter than they will be on the main sequence , so even stars less massive than the hydrogen burning limit can be observed in local clusters ( see review in lada , lada , & muench 1998 ) . lada et al . ( 1998 ) combined herbig s ( 1998 ) optical data on the cluster ic 348 , which was used to determine the history of star formation in the region , with their own k - band luminosity function to determine the low mass imf . they found that the imf in this cluster turns over below 0.25 m@xmath0 . recent searches for brown dwarfs ( zuckerman & becklin 1992 ; reid & gizis 1997b ; reid 1998 ) suggest these objects are so infrequent that the imf can not continue to rise at low mass . the number of possible brown dwarfs is consistent with an extension of the flattening seen between 1 m@xmath0 and 0.1 m@xmath0 ( reid 1998 ) . in spite of these approximate similarities in the imf from region to region , there have been two clear and consistent deviations that , if real , would have important consequences for star formation . one is the apparent shift toward relatively higher masses in regions of extreme activity , such as starburst galaxies ( e.g. rieke et al . 1980 ) , and the other is the apparent shift toward relatively lower masses in regions of extreme _ in_activity , the so - called `` field '' population , away from ob associations , as discussed by massey et al . ( 1995b ) . indeed , there may be a monotonic transition from low - mass weighted imfs to high - mass weighted imfs with increasing star formation rate , ism pressure , radiation field , etc . , perhaps because of a change in cloud properties , cloud destruction capabilities of young stars , and other things . the `` normal '' imf occupies what seems to be a broad range of star formation rates between these two limits , but this broad range is not so broad in terms of _ possible _ ism conditions . it may only be broad in the sense that it is common in main galaxy disks where some type of feedback between star formation and gas stability is maintained for long periods of time . reviews of imfs in starburst galaxies are in scalo ( 1990 ) , zinnecker ( 1996 ) , and leitherer ( 1998 ) . the first observation that suggested a `` top - heavy '' imf was that the luminosity of the starburst regions are so high , and the dynamical masses from the rotation curves so low , that there can not be a full complement of low mass stars , which dominate the mass , to go with the high mass stars , which dominate the light ( rieke et al . 1980 , 1993 ; kronberg , biermann , & schwab 1985 ; wright et al . 1988 ; see reviews in telesco 1988 , joseph 1991 , rieke 1991 , thuan 1991 ) . thus the imf contains a higher proportion of high mass stars than the solar neighborhood . doane & matthews ( 1993 ) reached the same conclusion for m82 based on many properties of galactic evolution models , including the remaining gas mass , total mass , star formation rate , and supernova rate . doyon , joseph , & wright ( 1994 ) and smith et al . ( 1995 ) found top - heavy imfs from spectroscopic line ratios for the starburst galaxies ngc 3256 and ugc 8387 , respectively . smith , herter & haynes ( 1998 ) found the same result from infrared excesses in 20 starburst galaxies . the distribution of stars was often not resolved in the early starburst studies . satyapal et al . ( 1995 , 1997 ) derived a smaller extinction - corrected k band luminosity for the inner 30 of the starburst galaxy m82 , and suggested the imf could have the salpeter slope all the way from 0.1 to 100 m@xmath0 with only 30% of the dynamical mass in the star formation event . ho & filippenko ( 1996a , b ) measured the velocity dispersions in two dense clusters in starburst galaxies , clusters that may evolve into globulars like those in the halo of our galaxy ( meurer 1995 ) . they found that the young globulars have large masses like old globulars , and that the young clusters would evolve into old clusters with a mass - to - light ratio like the old globulars after a hubble time . because of this , they suggested that the imf might be the same in the young starburst globulars as it is in old globular clusters , and in particular , that it contains low mass stars ( e.g. , de marchi & paresce 1997 ) . other considerations have led to revisions of the top - heavy models as well . devereux ( 1989 ) observed 20 nearby starbursts like m82 with infrared photometry , and found no evidence for an unusual imf in any of these ; he pointed out that a revision of extinction corrections for m82 could change the rieke et al . ( 1980 ) result too , as recently confirmed by satyapal et al . schaerer ( 1996 ) considered detailed evolutionary models of young stellar populations in starburst galaxies and also found that the imf was normal . calzetti ( 1997 ) modeled multiwavelength spectroscopy and broad - band infrared photometry of 19 starburst galaxies and found that the salpeter imf between 0.1 and 100 m@xmath0 fits the observations . stasiska & leitherer ( 1996 ) modeled the emission line spectra of giant hii region and starburst galaxies , having a factor - of - ten range in metallicities , and found a salpeter imf up to 100 m@xmath0 ; the lower mass limit could not be obtained , however . the proposed peculiarity of starburst imfs is therefore not certain , a conclusion reached by scalo ( 1990 ) as well . maybe only the extreme examples of starbursts require a high proportion of massive stars , while other starbursts , perhaps more `` ordinary , '' do not . the exact form of the peculiarity in a starburst imf is not known either . it could be `` truncated '' in the sense that the lower mass limit of the power law , where the power law slope begins to flatten , is higher than usual , perhaps as high as several m@xmath0 instead of several tenths of a m@xmath0 . such truncation may be observed already for 30 dor ( nota 1998 ) . or , it could have a power - law slope that is shallower than @xmath14 , or possibly , a higher upper mass limit . these differences are often hard to distinguish , even though all of them affect the mass - to - light ratio . fortunately , the exact form of the imf sometimes affects other things as well . for example , charlot et al . ( 1993 ) suggested that a truncated imf will produce a very red population of red giants , without the corresponding main sequence stars , after the turnoff age reaches the stellar lifetime at the truncation mass . wang & silk ( 1993 ) suggested that the truncated model gives an oxygen abundance that is too high when the star formation process is over . presumably these other constraints will help narrow down the possible choices . the extreme field studied by massey et al . ( 1995b ) is another example where the salpeter imf seems to break down . in regions of the lmc and our galaxy that are more than 30 parsecs from catalogued ob associations , massey et al . found o - type stars that probably did not drift there during their short lifetimes . massey et al . showed with an h - r diagram that there was an approximately steady rate of star formation there for the last 10 my years , and they derived a time - averaged imf . the result was very steep , with @xmath15 instead of @xmath16 for a salpeter function . there are several other peculiarities of the imf that may or may not have a specific physical origin . for example , it has been known for a long time ( larson 1982 ; myers & fuller 1993 ) that high mass clouds produce high mass stars , and low mass clouds produce low mass stars . two decades ago , this observation and others were related to the concept of a bimodel imf ( eggen 1976 , 1977 ; elmegreen 1978 ; gsten & mezger 1983 ; larson 1977 , 1986 ) , and to the impression that spiral arms preferentially formed massive stars ( mezger & smith 1977 ) , presumably because the molecular clouds in spiral arms are more massive , or somehow different from interarm clouds . today , we know that massive clouds also produce low mass stars in great abundance ( herbig & terndrup 1986 ) , and there are essentially no local regions that produce high mass stars exclusively . thus the `` observation '' that the maximum mass of a star correlates with the mass of the cloud could be only a statistical sampling effect ( larson 1982 ; elmegreen 1983 ; walter & boyd 1991 ; massey & hunter 1998 ) : high mass clouds produce more stars of all types , and so they are more likely to sample at the tail of the imf to produce a few massive stars . in the same sense , high mass clouds should also produce more dwarf stars than low mass clouds . an imf model based on random sampling from a hierarchical cloud ( elmegreen 1997a ) shows this effect clearly in figure [ fig : mvsm ] . each curve on the left is from a different cluster , with a number of stars in the clusters equal to 2500 , 5000 , 10000 , 25000 , 50000 , and 100000 . as the number of model stars increases , the mass of the largest star increases too , as shown on the right of the figure . the expected slope of the power law relation between cluster mass , @xmath17 , and maximum star mass , @xmath18 , is @xmath19 , i.e. , @xmath20 ; actually the calculation gives a slightly steeper slope because of storage limitations to the number of hierarchical levels allowed in the simulation ( these models use 8 levels with an average of 3 subclumps per clump at each level ) . this property of increasing maximum star mass with cloud mass should not depend on any specific imf model or on the star formation process . to be more precise about the stochastic interpretation of this effect , we can consider an imf @xmath21 , which gives the salpeter function when @xmath22 . the largest star , with mass @xmath18 , typically satisfies @xmath23 . this gives @xmath24 . because the total cluster mass is @xmath25 , the largest star is related to the total cluster mass as @xmath26.\ ] ] for @xmath27 and @xmath22 , this gives @xmath28 , and for @xmath29 @xmath30 , this gives @xmath31 . thus , a cluster requires several hundred m@xmath0 before a small o - type star is likely to appear , and a cluster with @xmath32 m@xmath0 is likely to contain @xmath8 modest o - type stars or a few very massive ones . this is the mass range for the largest open clusters in the milky way disk , indicating that when a few o stars appear , a cluster is likely to terminate its formation in an unbound state , presumably as a result of rapid cloud destruction , and make an expanding ob association ( elmegreen 1983 ) . larson ( 1991 , 1992 ) and khersonsky ( 1997 ) interpret this observation in the opposite way , saying that there is a physical effect which tends to produce higher mass stars in clouds of higher mass , and that the imf partly results from this physical effect . larson ( 1992 ) attributes the physical effect to a larger accretion length in larger clouds for material going into a star . if there is such an effect , and the observation is not entirely statistical as discussed above , then the superposition of imfs from clouds of different mass would be different from the imf in any one cloud . that is , clouds of low mass would produce an imf with a particular slope up to some maximum mass , and clouds of higher mass would produce an imf , perhaps with the same slope , up to higher masses . the sum of these two imfs would be an imf with greater slope than either separate cloud , because the summed imf contains low mass stars from both clouds , but only high mass stars from the more massive cloud . larson ( 1991 ) proposed his model by saying , in effect , that only one star forms in each cloud fragment at each level in the hierarchy , and that the star mass is always proportional to the square root of the fragment mass . then the sum over all fragments is the final imf , and this has a slope of @xmath33 , which is not too bad . stated this way , there is no problem with the model because each component in the sum is only a single star , not a whole cluster of stars with a separate imf . however , even in this model , if we add together the stellar populations from many separate clouds with different total masses , then the summed imf will be steeper than the imf in each . at this point , it is important to recall that the imf obtained from large - scale surveys of galaxies and their metallicities is about the same as the imf observed in individual clusters . thus the summed imf from many regions of star formation has to be about the same as the imf from each separate cluster . this would seem to suggest that the observed increase in maximum star mass with cloud mass is not the result of a physical process that specifically increases the upper stellar mass in larger clouds . a second peculiarity about star formation is that the high mass stars tend to form after the low mass stars ( herbig 1962a , b ; iben & talbot 1966 ) . this has been determined most recently for 30 dor ( massey & hunter 1998 ) . again one can think of physical reasons for this , such as a gradual warming of the cloud following the formation of low mass stars , and an increase in the jeans mass with temperature ( silk 1977 ; yoshii & saio 1985 ) , but statistical effects can explain it too . the decreasing nature of the imf at high mass implies that massive stars are likely to form only after a lot of low mass stars have already formed . for a constant star formation rate , the average time between the formation of stars in a logarithmic mass range centered on @xmath34 is proportional to the rate at which these stars form , which is inversely proportional to the relative number of the stars , or @xmath35\right)^{-1}\propto m^{x}$ ] . for a constant star formation rate , this is also the average time after star formation begins for the _ first _ appearance of a star with this mass . thus the most common , low - mass stars form first , followed by the intermediate and then the high mass stars . in terms of the proportion of all stars formed , and for a constant star formation rate in units of mass per year , the maximum stellar mass increases with time as @xmath36 for star formation rate @xmath37 . thus , a star of mass @xmath38 forms after the proportional time given by @xmath39 for the power law portion of the imf only , with @xmath40 m@xmath0 and @xmath22 , this becomes , for masses in m@xmath0 , @xmath41 for example , in a cluster with a maximum stellar mass of 30 @xmath42 , stars with 10 m@xmath0 form in the last 20% of the time . note that this increase in time with stellar mass is only for the first appearance of a star with that mass . the _ average _ time of appearance is independent of stellar mass in the stochastic interpretation . this suggests a way to check the stochastic model of birth orders . if the average time of appearance of all stars with mass @xmath34 is the average age of the cloud , and this is true for any @xmath34 , and if the first time of appearance of a star with mass @xmath34 increases with time , as described above , then the birth order is dominated by stochastic effects . figure [ fig : mt ] shows the relative birth order of stars in the random sampling imf model by elmegreen ( 1998 ) . each point marks the earliest time of formation of a model star with that mass , among all of the stars that formed in 11 clusters on the left and 6 clusters on the right . each cluster contains 200 to 500 stars and is independently generated . the plot shows a distribution of points that traces the inverse of the imf in log intervals . the fiducial line has a slope of @xmath43 for comparison . the increase at low mass occurs because the model uses a gaussian probability for failure to form a star at low mass , which is an arbitrary assumption at this time ( it causes the model imf to turn over at low mass ) . there are no quantitative measures yet of how the earliest appearance of a star with a particular mass increases with time relative to the birth order of all stars , so the predictions of theory can not yet be checked . in a real star - forming region , the star formation rate is probably not constant in time , so the ordinate in figure [ fig : mt ] should not be time , but star number in its proper birth order ; i.e. one should plot , for the first appearance of a star of mass @xmath34 , its mass versus the number of stars that formed before it . an interesting implication of the above two discussions is that _ massive stars must not be able to severely limit or halt star formation in their primordial clouds . _ if they did , then , considering their relatively late appearance , each cloud would produce stars with the same local imf up to some maximum stellar mass at which point the cloud is finally destroyed . this maximum stellar mass at the time of cloud destruction is presumably larger for more massive clouds , because it is harder to destroy a massive cloud than a low mass cloud . thus the summed imf from many regions of star formation would be steeper than the local imf in each region , for the reasons discussed above , and we could not explain why the integral imf in a galaxy is the same as the cluster imf . perhaps massive stars only shred the low density parts of these clouds , allowing ( or perhaps stimulating ) , most of the cores that have already formed in these regions to continue their evolution toward stars . in view of this , it seems that stars of all mass are probably equally likely to form in any cloud ( that contains at least this much mass ) , regardless of what other stars have formed before them . moreover , the maximum stellar mass is not physically limited , it is only statistically limited in the sense that extremely massive stars are rare . an exception to this conclusion may arise in the `` field '' population studied by massey et al . ( 1995b ) , in which the summed imf of presumed dispersed clusters appears to be steeper than the imf in each . maybe massive stars can more easily disrupt their clouds in low pressure `` field '' environments . on the other hand , maybe the massey et al . observation can be explained in other ways . a third peculiarity of the imf is that massive stars are often closer to a cluster center than low mass stars . this effect has been discussed for a long time ( e.g. , sagar et al . 1988 ) , but it was never certain whether the observations reflected a birth position , or just a relaxation effect after the cluster formed . thermal relaxation makes the high mass stars sink to the center of the cluster , on average , in a thermalization time scale . however , hillenbrand & hartmann ( 1998 ) and bonnell & davies ( 1998 ) have recently shown that this peculiar mass distribution , which is observed in the trapezium cluster , could not result from thermalization because the orion cluster is too young ( see also jones & walker 1988 ) . similar mass segregation has been observed in ngc 2157 ( fischer et al . thus it is an effect that should be explained by an imf theory , and it is likely to result from a physical process during star formation rather than stochastic sampling . there have been many attempts to explain this relative birth position ( see reviews in larson 1991 , and zinnecker , mccaughrean & wilking 1993 ) , including excess gas clump collisions in the center of a cloud ( larson 1990 ; stahler , palla , & ho 1998 ) , excess gas accretion onto protostars in the cloud core ( larson 1978 ; larson 1982 ; zinnecker 1982 ; bonnell et al . 1997 ) , enhanced gas drag for massive protostars ( larson 1990 , 1991 ) , late - time formation of high mass stars in a collapsing cloud ( murray & lin 1996 ) , and hierarchical cloud structure in which the massive trunk of a hierarchical tree for the gas cloud is closer to the center of the tree than the low - mass branches ( elmegreen 1998 ) . a combination of these effects might be involved . there are evidently four distinct physical effects that have to be explained by a theory of the imf : + ( 1 ) the power - law slope at intermediate to high stellar mass , with a value close to that found by salpeter ( 1955 ) ; + ( 2 ) the flattening at low mass for clusters and star - forming regions in our galaxy ; + ( 3 ) the increase in the transition mass between power - law and flat regions of the imf in starburst galaxies , + ( 4 ) the preferential birth of high mass stars close to cluster centers . there are apparently three additional observations that can be explained by stochastic sampling effects : + ( 1 ) the tendency for the largest mass star in a region to increase with the total number of stars ; + ( 2 ) the tendency for the largest mass stars to form last , and + ( 3 ) the seemingly random variations in the intermediate and high mass imf slopes from region to region , along with the appearance of gaps , ledges , and other peculiar departures from a power law . in addition to these observations , there are tentative indications that other imf features may be present , including : + ( 1 ) a turnover at low mass , and + ( 2 ) a steepening of the power - law part in field regions . + future observations should be able to illuminate these uncertain features . the theory of the imf depends on both the theory of star formation and the theory of cloud structure . because of the complexity of these issues , the imf may not be understood for a long time . here we review some recent developments . previous reviews were in cayrel ( 1990 ) and clarke ( 1998 ) . one of the key problems in understanding the origin of the imf is to determine the processes that limit the pre - stellar gas accretion onto a star and define the star s final mass . there have been several ideas on this . the final mass could result from the star s own ability to limit the accretion of new gas onto its surface , perhaps because of an intense proto - stellar wind ( larson 1982 ; shu , adams , & lizano 1987 ) , which is known to exist at this phase of evolution ( lada 1985 ) . this possibility has led to several theories of the imf . one , proposed by nakano , hasegawa , & norman ( 1995 ) , considered three mass - limiting agents : protostellar winds , ionization , and depletion of the gas clump in which the star forms . they concluded that under normal conditions , the protostellar wind would limit the accretion . then they calculated the fraction of the clump that would get into the star , finding that the final star mass scaled with the clump mass to the @xmath44 power . to get an imf slope of @xmath45 , which they got from the field - star imf in scalo ( 1986 ) , they required a very steep clump mass spectrum , @xmath46 in linear intervals , instead of the usual observation of @xmath47 with @xmath48 ( blitz 1993 ; kramer et al . they justified this steep clump spectrum by noting that if the linewidths found in a cs survey of orion clumps ( tatematsu et al . 1993 ) were assumed to be virial velocities , and the masses calculated accordingly , then the orion cs clumps would have such a spectrum . another imf model with wind - limited stellar masses was proposed by adams & fatuzzo ( 1996 ) . they conjectured that the stellar luminosity scales with the wind mass loss rate , which , at the time when the accretion stops , is proportional to the direct accretion rate onto the star , independent of what goes onto the disk . this gave them a relation between the stellar properties , i.e. , luminosity and mass , and the gas cloud properties , sound speed and angular rotation rate . the angular rotation rate entered because they had to determine the fraction of the accreting gas that goes into the star and not the disk . finally , they related the stellar mass to the total luminosity of the star , from the sum of the accretion - driven luminosity , which depends on mass , and the luminosity of a main sequence star at intermediate mass , from the standard mass - luminosity relation . the result is a relation between stellar mass and the cloud properties , i.e. , sound speed and rotation rate . taking the mass distribution for clumps with the observed slope @xmath9 discussed above , and the empirical relationship between clump mass and velocity dispersion from larson s ( 1981 ) laws , they got a distribution function for the clump velocity dispersion , which then led to a distribution function for the final star mass if the angular rate of all the clumps is the same . adams & fatuzzo ( 1996 ) also discussed more general equations giving the stellar mass from cloud properties , considering random variations in these cloud properties , and derived a log - normal final mass distribution , as in larson ( 1973 ) , elmegreen & mathieu ( 1983 ) , zinnecker ( 1984 ) , and elmegreen ( 1985 ) . silk ( 1995 ) also got a relation between cloud core mass and turbulent linewidth , different from adams & fatuzzo s , considering centrifugally supported cores with luminosities equal to the accretion luminosities . he considered an equality between the turbulent linewidth and the thermal velocity in the core , which entered into the luminosity through the temperature using the usual radiative transfer equation for a star and the rossland mean opacity . then , after substituting the empirical correlations between clump rotation rate and size , and between linewidth and size , he got the cloud core mass as a function of the turbulent linewidth . the distribution function for cloud core mass then followed from a distribution function for linewidth , which came from a theory for the time - dependent deceleration of wind - driven bubbles in a cloud . the star mass is then taken to be proportional to the cloud core mass . silk ( 1995 ) made several comments that protostellar outflows limit the mass of a star , but this assumption was not present in any of the theory in his paper , nor in the imf that resulted , particularly considering that a fixed fraction of the cloud core mass was assumed to go into the star . this was unlike the results of the wind - limited accretion models proposed by nakano , hasegawa , & norman ( 1995 ) and adams & fatuzzo ( 1996 ) . one of the most popular and persistent methods for determining the mass of a star has been with a combination of cloud fragmentation , accretion , and clump collisions . larson ( 1978 ) showed with three - dimensional n - body experiments that gas clouds fragment hierarchically by self - gravity , and the fragments accrete material in competition with each other . the resulting mass distribution was modeled after a fractal , which was a remarkably prescient concept in astronomy considering that mandelbrot ( 1977 ) began to popularize his fractal geometry only a year earlier . larson showed that the fractal mass function would have a slope of about @xmath49 , in reasonable agreement with observations . larson also reasoned that the lower mass limit to a star is the thermal jeans mass in the cloud because smaller condensations are not likely to form in the initial collapse . this experiment was an important departure from standard fragmentation models of the time , since it was previously believed , following hoyle ( 1953 ) and hunter ( 1962 ) , that fragmentation decreased the jeans mass , leading to ever more fragmentation . in larson s result , the number of clumps formed was about equal to the number of jeans masses in the initial cloud , with no subfragmentation into smaller pieces . this result was seconded by tohline ( 1980 ) , but investigated again by silk ( 1982 ) , who concluded that initial cloud fragments , particularly elongated fragments , formed by dynamical collapse could in fact fragment again ( see also bonnell & bastien 1993 , burkert et al . 1997 ) . silk ( 1977 ) considered a different picture of cloud fragmentation , using a thermal jeans mass that increased with time as a result of heating from the stars that already formed . his model of fragmentation was simpler than larson s , so there was no built - in , power - law mass spectrum from the fragmentation process itself . the power law in silk s model came from the time increase in the thermal jeans mass and the identification of this mass with the stellar mass at all times . yoshii & saio ( 1985 ) followed this model by considering the additional influence of coalescence among the opacity - limited fragments . they showed that the opaque fragments are usually so small that coalescence is unimportant ; thus fragmentation and heating alone determine the imf as a time sequence of increasing thermal jeans masses , as in the silk ( 1977 ) model . bastien ( 1981 ) obtained a different result by using the jeans length to determine the fragment collision cross section . the initial jeans length is much larger than the size of the opaque fragments considered by yoshii & saio ( 1985 ) , so bastien ( 1981 ) found that fragment collisions were important . subsequent work by lejeune & bastien ( 1986 ) , and allen & bastien ( 1995 , 1996 ) considered time - dependent coalescence , and concluded again that both fragmentation and coalescence were important , with coalescence dominating the formation of massive stars . price & podsiadlowski ( 1995 ) used protostar collisions in a different way . they proposed that stars grow by accretion at some more - or - less uniform rate , and this accretion stops when two protostars collide , disrupting the gas reservoirs from each . the final stellar mass was then determined by the product of the accretion rate and the time interval between collisions . the imf was built up over time as the final stellar mass decreased in the presence of an increasing collision rate that resulted from the continuous formation of more and more protostars . another coagulation model was proposed by murray & lin ( 1996 ) . they suggested that fragmentation driven by thermal instabilities leads to clumps that fall in the potential well of a cloud , after which they collide and coalesce with other clumps that have fallen too . a star forms when the clump mass exceeds the thermal jeans mass , although further coalescence can increase that mass afterwards . the power - law distribution of masses follows from the coagulation process , in a manner similar to that proposed by silk & takahashi ( 1979 ) . there is no assumption about wind - limited accretion here ; the star mass is identified with the core mass directly . an interesting aspect of their model is that they considered positionally correlated velocities , as in a turbulent fluid , but they showed this had no important effect on the model imf . numerical sph simulations of an evolving imf made of accreting , pseudo - star particle clumps was done by bonnell et al . the advantages of numerical solutions like this is that they can treat well the competition for gas among all the nucleated centers . they found that nucleating centers originally close to the center of the whole cloud grow faster and to larger masses because of the larger gas density there . a different class of theory considers only the statistical aspects of the imf . the hoyle ( 1953 ) picture of fragmentation led naturally to these ideas , because successive fragmentation produces hierarchical structure and power law mass spectra regardless of many physical details . larson ( 1973 ) considered a modification to these ideas by proposing that only part of each fragment undergoes further fragmentation , and got from this a log - normal mass distribution instead of a power - law . when miller & scalo ( 1979 ) proposed that the imf actually had a log - normal form , the statistical implications of this were developed further by elmegreen & mathieu ( 1983 ) , zinnecker ( 1984 ) , and elmegreen ( 1985 ) , who showed that a variety of processes , acting together , would combine to make a log - normal function , regardless of the distribution functions resulting from each separate process . this point was made again more recently by adams & fatuzzo ( 1996 ) . the most recent development in this area has been by elmegreen ( 1997a , 1998 ) , who considered the random selection of masses from a hierarchically structured cloud . these papers depart from the usual scenarios by asserting that hierarchical cloud structure is independent of star formation that it is set up long before star formation begins in the diffuse cloud stage , and then continuously reestablished during the molecular cloud stage as a result of supersonic turbulence compression . there is no gravitationally - driven fragmentation at all , and no significant clump coalescence . in fact , the clumps in this model need not even be constant objects , able to move around and coalesce ; they can be ever - changing gas compressions and wavepackets in the chaotic turbulent flow . this model is motivated by the pervasive appearance of hierarchical structure and spatially - correlated velocities in interstellar clouds . these features are reminiscent of structures and flows in laboratory turbulence ( e.g. , see reviews in sreenivasan 1991 , and falgarone & phillips 1991 ) . the basic point of the elmegreen ( 1997a , 1998 ) model is that essentially all local star formation processes that have been considered in detail have a time scale for evolution that varies approximately as the inverse square root of local density , including contraction or collapse from self - gravity , turbulence compression , magnetic diffusion in virialized cores , and coalescence at each level in the hierarchy . this means that in a model like this , stars will appear here and there , randomly , alone or with neighbors , at a rate that depends almost exclusively on the local square root of density . on a large scale , this implies that dense clouds in high pressure regions of galaxies will form all of their stars quickly , while low density clouds in low pressure regions will take much more time . on a small scale , within any one cloud , it means that different parts of the hierarchy form stars at different times . the density always increases at lower levels in a hierarchical structure , and this is where the clumps , contained by other clumps on larger scales , are small and have low mass . thus the low mass pieces in a cloud are likely to form stars first . as a result , there is slightly less mass for other stars that form later in the same or higher levels . this competition for mass tends to steepen the power in the power - law mass function by several tenths , i.e. , from @xmath49 to @xmath22 , as in the salpeter imf . perhaps a more important difference in this new model is the way in which cloud structure is measured . typically , the power law index for molecular and diffuse cloud structure is determined from the mass distribution of separate clumps , resolved in large - scale surveys by various telescopes and then separated into discrete objects either by eye or by computer algorithms that do about the same thing as the eye . these mass distributions are always much flatter than the imf , having slopes @xmath9 in the range from 1.5 to 1.8 , when the imf has a slope of about @xmath50 . to understand the imf , however , we have to try to view interstellar clouds from the perspective of a forming star . a star does not care about the clumps that our telescopes resolve and our eyes choose to label as discrete , but only about the general distribution and motion of gas in all forms . in pre - star - forming clouds , where this structure is first established , it is largely hierarchical , from scales that are much smaller than mm - wave telescopes can resolve , up to perhaps a galactic scale height . hierarchical means that most of the small clouds are contained inside larger clouds , and most small stellar groupings are inside larger stellar groupings ( see review in elmegreen & efremov 1998 ) . for example , efremov ( 1995 ) and colleagues have estimated that 90% of the ob associations in the milky way ( efremov & sitnik 1988 ) , m31 ( efremov , ivanov , & nikolov 1987 ; battinelli 1991 , 1992 ; magnier et al . 1993 ; battinelli , efremov & magnier 1996 ) , m33 ( ivanov 1987 , 1992 ) , and the lmc ( feitzinger & braunsfurth 1984 ) are inside larger star complexes . scalo ( 1985 ) has reviewed hierarchical structure for interstellar clouds . it appears as if the hierarchical embedding of interstellar and young stellar structures is nearly all - inclusive . emission line surveys leading to `` discrete '' clump masses do not consider this aspect of cloud structure . if two clumps are close together and part of a larger structure , the algorithms call them two separate objects , and do not tabulate the larger `` object '' that contains them . for this reason , all emission line or extinction surveys catalogue clouds that are within a factor of @xmath51 to 10 of the angular resolution , regardless of the cloud distance or the wavelength of the observation . this factor of @xmath52 for recognized clump size corresponds to a factor of @xmath53 for cloud mass ( which scales as the size squared or cubed in co surveys ) , so the mass functions look reasonably well sampled , but in fact only a small part of the cloud structure is included . everything smaller than the telescope resolution is not seen , and everything larger than about 3 to 10 resolution elements is subdivided into its component parts and not called a separate object . this is how every cloud or clump mass function has been evaluated since the beginning of this exercise ( i.e. , since before field & saslaw 1965 ) . the elmegreen ( 1997a , 1998 ) model takes a different point of view . it considers the structure on all scales and asks for the probability that any particular mass is chosen from anywhere in the whole hierarchy . this is presumably what happens as a result of the combination of physical processes that leads to star formation in a real cloud : because of the self - similar nature of turbulent flows , each level in the hierarchy looks the same as any other level ( for masses above the thermal jeans mass ) , so each choice of level for a star - formation event would have the same likelihood as any other choice , modulated only by the density - dependent local evolution rate discussed in the previous paragraph . aside from this variable rate , the instantaneous distribution of masses for all structures , and the instantaneous probability of selecting any particular mass , is proportional to @xmath54 for linear intervals in mass ( i.e. , @xmath55 in the notation of the clump spectra given above ) . such a distribution for hierarchical structure was also recognized by larson ( 1978 ) and fleck ( 1996 ) . the result gives a steeper instantaneous mass function than emission line or extinction surveys , but there is no conservation of mass as in these standard surveys ( i.e. , the sum of the masses of all possible structures is larger than the mass of the whole cloud , because nearly all of the structures are contained in other structures , and so are multiply counted ) . this way of viewing cloud structure seems to be closer to what a real star - forming process `` sees '' before it actually begins the sequence of events that makes a star . the density dependent rate of star formation then steepens the mass function from @xmath54 to @xmath56 . this result is then identified with the salpeter imf . for the purposes of understanding the imf , there are probably better ways to measure cloud structures than with clump - finding algorithms . the structure on the edges of clouds has been analyzed in terms of fractals , rather than clumps , for many years ( beech 1987 ; bazell & dsert 1988 ; scalo 1990 ; dickman , horvath , & margulis 1990 ; falgarone , phillips , & walker 1991 ; zimmermann & stutzki 1992 , 1993 ; henriksen , 1991 ; hetem & lepine 1993 ; vogelaar & wakker 1994 ; pfenniger & combes 1994 ) . for the structures inside clouds , stutzki et al . ( 1998 ) showed that fourier transform power spectra of emission line intensity scans across molecular clouds gave power laws , and concluded that the internal structure was scale - invariant , as in a fractal . they reproduced this structure with a random fractal - generation model . other models that made cloud fractals were in hetem & lepine ( 1993 ) and elmegreen ( 1997b ) . the concept that the cloud mass distribution function is the result of fractal structure , presumably generated by turbulence , began with papers by fleck ( 1996 ) , elmegreen & falgarone ( 1996 ) , and stutzki et al . this seems to be a more reasonable explanation than the older collisional - build - up models of clump structure , particularly since off - center supersonic collisions between clumps should not be sticky ( scalo & pumphrey 1982 ; kimura & tosa 1996 ; fujimoto , & kumai 1997 ) . the probability of selecting structures for star formation in hierarchical clouds seems to show up more directly in the distribution of masses for open clusters . when a bound cluster forms , a high fraction of the gas mass has to go into stars ( see verschueren 1990 and references therein ) , so the cluster mass distribution should reflect the mass distribution of cloud structures pretty well . in this case , the result is obviously independent of telescope resolution or cloud - clump recognition bias , because clusters are seen optically , even at great distances . also , the range of masses for clusters can be reasonably large , exceeding a factor of 100 , so a power law can be measured fairly well if it exists . cluster selection suffers from various other biases , however , such as extinction , age limitations , and a loss of low - mass clusters with increasing distance . nevertheless , there are some studies that get around these biases , and they confirm the expected result . in two samples of nearby clusters , each with calibrated masses , battinelli et al . ( 1994 ) found power law mass functions with slopes of @xmath57 and @xmath58 . these clusters are close enough to the sun to be relatively free of distance and extinction effects . also , in the lmc , where @xmath59 clusters are catalogued ( bica et al . 1996 ) , calibrated photometrically for age ( girardi et al 1995 ) , and all at about the same distance , elmegreen & efremov ( 1997 ) found @xmath60 mass functions for discrete age groups ( e.g. , @xmath61 yrs ; dividing the clusters into age groups is important when there are no direct measures of cluster mass , because the conversion from luminosity to mass depends on age ) . the imf should be steeper than the cluster mass function because of the density - dependent formation rate of individual stars , and the competition for mass in the imf . these effects are not important for clusters . open clusters take a high fraction of all the gas mass originally available to them ( or else they would not have ended up bound ) , and the rate at which they do this is not important for their final masses . in the case of stars , new objects forming at one level steal mass away from the objects that form later at a higher level . but clusters do not do this : those which form inside other clusters in the general hierarchy can just merge together to make a single cluster of higher mass , appropriate to the higher level in the hierarchy . and if they do not merge , then they stay with their original masses , which are appropriate for the levels in which they form . other imf models based on fractal or hierarchical cloud structure were developed earlier by henriksen ( 1986 , 1991 ) and larson ( 1992 ) . henriksen used the size , @xmath62 , distribution function of structures in a fractal of dimension @xmath63 , given as @xmath64 by mandelbrot ( 1983 ) , and assumed that the density @xmath65 varied with size too , not as in a fractal ( which would be @xmath66 ) but as actually observed in self - gravitating clouds , namely @xmath67 . this gave a mass function for cloud structure that agreed well with observations if @xmath68 . note that if henriksen used the density dependence for a fractal , in a self - consistent manner with the size distribution , then the resulting mass function would have been independent of the fractal dimension , @xmath60 , as discussed above for hierarchical clouds in general . larson ( 1992 ) started again with the size distribution for a fractal , and assumed that final stellar mass is directly proportional to the linear size of the cloud structure , suggesting that such structures were filamentary anyway . then the slope of the imf in linear intervals became @xmath69 for @xmath70 , as implied by the fractal dimension of cloud perimeters , namely @xmath71 , added to 1 to account for the higher dimension of the surface . this imf , with a slope of @xmath72 , is significantly steeper than the salpeter imf , with a slope of @xmath73 , but larson suggested that maybe stars form in subparts of clouds where @xmath63 is smaller than 2.3 , or perhaps larger structures have larger temperatures , which would break the assumed linear relationship between star mass and scale size . these models differ from the elmegreen ( 1997a , 1998 ) model , even though both employ `` fractal '' cloud structure , because the latter uses general cloud shapes , not necessarily filaments or sheets , random sampling from all levels in the hierarchy of structures ( giving @xmath60 from sampling alone ) , a density - dependent rate that steepens the imf by preferred sampling at lower mass ( i.e. , steepening the result to @xmath74 ) , and a competition for mass , which steepens the imf further , to @xmath75 , as in the salpeter imf . with these different assumptions , the imf in the elmegreen model hardly depends on the fractal dimension at all ; it enters into the exponent of the imf power law as approximately @xmath76 , which is a very small increment for @xmath63 in the likely range of 2 to 3 . understanding the lower mass limit to star formation is an important part of any imf theory . various assumptions about this have already been mentioned in the review of theories presented above . three views on this subject seem to pervade most of them . in the wind - limited accretion models , the assumption is made that accretion onto a dense core continues until a protostellar wind develops , which presumably follows the onset of deuterium burning ( shu et al . thus low mass objects do not stay that way very long , they just keep accreting until they become active protostars . in most of the older fragmentation - coalescence models , the minimum mass is the mass of a gravitationally bound fragment that is optically thick . in the various scale - free models , based on gravitationally - driven fragmentation or turbulence , the minimum mass is usually the thermal jeans mass , particularly in the larson ( 1992 ) and elmegreen ( 1997a , 1998 ) models . this assumption is made because , without further fragmentation during the collapse itself , the thermal jeans mass is the smallest structure that can ever become self - gravitating , even in the presence of turbulence and magnetic fields . the first and the third of these mass limits equals about the observed limit , @xmath4 m@xmath0 . in the first case , it is self - defining : objects accrete until they become star - like , and then they stop accreting , so naturally the lowest possible mass is the mass of a star . this concept has the powerful advantage over the others that the basic star mass , and perhaps even the imf itself , can be the same everywhere , independent of cloud conditions . in the third case , the minimum theoretical mass just happens to equal the observed minimum in standard cloud conditions . an expression for this mass is the bonner - ebert critical mass , @xmath77 here we have normalized the result to standard cloud conditions , taking the pressure inside a cloud equal to the typical self - gravitating core pressure , not the boundary pressure ( which is sometimes lower by a factor of @xmath78 ) . our use of this form for the thermal jeans mass is motivated by the near constancy of @xmath79 and total @xmath80 in molecular clouds , according to the scaling relations ( larson 1981 ) . this is much more sensible than writing the thermal jeans mass in terms of @xmath79 and density , for example , when the density varies by several orders of magnitude inside a cloud . our use of the total pressure for @xmath80 also allows for a decrease of jeans mass in compressive flows , as in hunter & fleck ( 1982 ) . nevertheless , another theory of the imf in which all of the stellar masses have the local thermal jeans mass has been advanced by padoan , nordlund , & jones ( 1997 ) , with variations in mass entirely the result of variations in the density in a turbulent fluid . in this model , the high mass stars form in the low density regions , because the jeans mass is high there for a given temperature . the second minimum fragment mass discussed above , from the opacity limit of a self - gravitating condensation , is typically much smaller than the minimum star mass , so coalescence has to bring the star mass up to the observed value in these models . larson has looked for signatures of the thermal jeans mass in several ways , predicting that it should vary with cloud temperature ( larson 1985 , 1992 ) . in larson ( 1992 ) , he suggested that such variations were observed by myers ( 1991 ) in different molecular clouds . larson ( 1995 ) also suggested that the length scale at a break in the separation distribution for t tauri stars was the jeans length ( see also simon 1997 ; but see bate , clarke , & mccaughrean 1998 and nakajima et al . 1998 for other interpretations . ) elmegreen ( 1997a , 1998 ) considered @xmath81 from a different point of view , stating that most star - forming regions in normal galaxies actually do have nearly constant values of @xmath82 , with small variations in this quantity possibly leading to the observed variations in cluster imfs at low mass ( see section [ sect : intro ] above ) . this constancy is a result of the numerator in @xmath81 being approximately proportional to the cooling rate per unit mass of molecular gas ( neufeld , lepp , & melnick 1995 ) , and the denominator being approximately proportional to the local surface density of stars and gas in the galaxy , which determine the heating rate per unit mass from cosmic rays and background starlight . as long as cooling roughly balances background heating , this ratio will be about constant . other , more local variations in @xmath79 and @xmath80 tend to cancel out when combined as the quantity @xmath82 . for example , when @xmath80 goes up in a triggered region next to a previous generation of star - formation , @xmath79 generally goes up too because of the enhanced radiation field . conversely , when @xmath80 is low , as in low - mass clouds like taurus , then @xmath79 is low as well . elmegreen ( 1997a , 1998 ) also points out that the lower mass limit for stars actually _ does _ change , perhaps by a factor of 10 in extreme starburst regions ( see sect . [ sect : intro ] above ) where the average @xmath79 can be much larger than in local molecular clouds , e.g. , 100 k instead of 10 k ( cf . aalto et al . 1995 ) , and the pressure much larger too ( e.g. , @xmath83 ) . but this variation should be rare in normal galaxies because of feedback processes in the interstellar medium that tend to connect young stellar activity , interstellar pressure , and gas temperature . if we consider @xmath79 to be a measure of the rate of cooling per unit cloud mass , then @xmath79 in a star - forming region will be high when there is an unusually large embedded or nearby luminosity of stars per unit gas mass . this occurs when the efficiency for star formation is high in a high mass cluster ( the cluster has to be fairly high mass to contain the luminous stars that dominate the total radiation field ) . a high efficiency in a low mass cluster would not necessarily increase @xmath81 because low mass clusters are not very luminous per unit mass . a low efficiency in a high mass cluster , leading to an ob association , might not change @xmath81 much either , because the radiation luminosity per unit gas mass is not particularly large there either . thus the formation of globular clusters in regions with only moderately large pressures , not the enormous pressures typical of early galactic halos , might be elevate @xmath81 , but this conclusion is very uncertain . figure [ fig : mj ] shows two numerical experiments that made model imfs from random selections in a hierarchically structured gas cloud , using a rate of selection proportional to the square root of the gas density , as discussed above ( elmegreen 1997a ) . the models differ only in the value of @xmath81 , which enters into the probability for failure to form a star , taken to be @xmath84 in these cases ( @xmath81 in the figure is written in terms of the computer - program mass , which is scaled to a physical mass after multiplication by @xmath85 . the imfs are identical except for the larger value of the lower mass limit in the case with high @xmath81 . the two main contenders for the lower mass limit of star formation , wind - regulation and the thermal jeans mass , have obvious differences in predicted behaviors of the shape of the imf at low mass . if this shape varies a lot , with some clusters having clearly higher lower mass limits than other clusters , then wind self - regulation might be ruled out . moreover , if cluster imfs in regions with large turnover masses also have larger pre - stellar @xmath81 , then the thermal jeans mass would seem to be important . curiously , both of these models , the wind - regulated imf , and the @xmath81fractal cloud model , have their strong points where the other is weakest . the wind - regulated imf model seems to have difficulty explaining the nearly universal slope of the imf without some sensitivity to cloud parameters , yet it seems certain that wind self - regulation must play some role in determining the mass that gets onto a star , at least near the lower mass limit for star formation . similarly , models that rely on @xmath81 have an uncomfortable susceptibility to changes in cloud properties near the lower mass limit , but get the salpeter slope above this limit with essentially no free parameters . perhaps a combination of the two models would be better . there are several ways this could occur . first , the fractal model supposes that a star s mass can be identified with the mass of some structure in a turbulent cloud , but this is obviously too simple there is a lot that can happen during star formation that will vary the fraction of the cloud mass that gets into a star . if this fraction is randomly distributed , then the same imf slope and scatter about that slope result , as shown in connection with figure [ fig : scatter ] below , which is discussed in more detail in elmegreen ( 1998 ) . if it is not randomly distributed , then the power - law slope could change . but the entire power - law slope is not likely to come from wind - regulation , that would give the slope too great a sensitivity to cloud properties and the randomness of wind - clearing . this means that , perhaps to within a factor of 2 to 5 , star mass should really be identified with the mass of the clump in which it forms , i.e. , that bigger stars really do form in bigger clumps ( casoli et al . 1986 ; myers , ladd , & fuller 1991 ; myers & fuller 1993 ; pound & blitz 1995 ; motte , andr , & neri 1998 ; not to be confused with the statistical sampling statement discussed in section [ sect : intro ] , that bigger stars form in bigger _ whole _ clouds ) . this factor of 2 to 5 may contain all the detailed physics of wind clearing . protostellar winds should also play some role in regulating the smallest stars that form . according to the models by shu and collaborators , as reviewed in shu et al . ( 1987 ) , the onset of the wind is triggered by deuterium burning in a young , pre - stellar object . if the object is too small , there will be no wind , and presumably accretion will continue until deuterium burning begins . this concept combines well with the arguments based on a thermal jean mass . for example , if @xmath81 is much smaller than the deuterium burning limit , then a large number of small brown dwarfs or jupiters should form , and these could either leave the cloud in that state , forever drifting as brown dwarfs , or they could accrete more mass over time and end up finally able to start a wind . in the latter case , the effective lower mass limit to the imf would be the deuterium burning mass , not @xmath81 , and the larger masses would follow the fractal mass distribution as before , giving the salpeter function . on the other hand , if there are a lot of free brown dwarfs , which does not appear to be the case in the solar vicinity ( cf . [ sect : intro ] ) , and if there are also a lot of regions where @xmath81 is low , which does not appear to be the case locally either , then a model with accretion up to a wind - limited mass would not seem to work . if @xmath81 is much larger than the deuterium burning limit , then large , self - gravitating clumps will form in the cloud , but the rapid onset of a wind in these clouds might limit the gas mass that actually gets into the star to only a small fraction of what is available . then all of the excess mass between the deuterium limit and m@xmath86 would get dispersed back into the cloud for recycling into other stars , and the lower mass limit would be the deuterium limit . in this second case , the mass fraction that goes into each star will be small , but if it is about the same fraction for all clumps , then the higher mass stars will still map out the fractal properties of the cloud and produce the salpeter imf . again the details of how the imf gets established in a cloud depend strongly on unknown processes during the final collapse and dispersal phase of star formation , but the basic shape of the imf could be relatively independent of these details . there have been some fundamental changes lately in how we view molecular clouds and star formation , and these changes inevitably affect the various models for the imf . for example , we are beginning to think that star formation is relatively rapid in cloud cores , occurring in only one or two crossing times regardless of scale ( elmegreen & efremov 1996 ; efremov & elmegreen 1998 ) . this means that we can not generally wait for magnetic diffusion to occur as a cloud core slowly accretes across field lines . diffusion typically takes about 10 free fall times for cosmic ray ionization ( shu et al . 1987 ) , longer for ionization by starlight ( myers & khersonsky 1995 ) , and longer still if the gas is clumpy ( elmegreen & combes 1992 ) . indeed , nakano ( 1998 ) suggested that all star - forming cloud cores are magnetically _ super_critical ; i.e. , they would collapse dynamically across the field lines if turbulent motions were not present . there are two important implications of this change in thinking , if correct . first , there would no longer be a strong motivation for imf theories that consider bimodel star formation in the sense that low mass stars come from subcritical cores , and high mass stars come from magnetically supercritical cores ( lizano & shu 1988 ) . the imf is so remarkably uniform anyway , it does not seem possible to have widely different modes of star formation at high and low mass . any change from low mass dominance to high mass dominance in the imf could more easily be accommodated by an upward shift in the thermal jeans mass ( cf . section [ sect : theorylm ] ) . secondly , this time scale for star formation may be too short to allow multiple interactions between protostellar clumps . this would then rule out a broad class of coalescence and slow accretion models . such coalescence is unlikely anyway for clumps that move supersonically relative to each other ; they will fragment or disperse upon collision , rather than stick together . clump collisions may trigger star formation when they occur ( bhattal et al . 1998 ) , but multiple collisions are probably not the source of the cloud or star mass distributions . another observation that suggests the same thing is the extremely high stellar densities in young globular clusters , which form today in starburst regions . as suggested in section [ sect : intro ] , these globulars look like they will evolve into objects similar to our milky way s halo globulars , and the old milky way globulars have an apparently normal imf . if coalescence is important in local `` normal '' embedded clusters , it would seem to be devastating in globular clusters , where the stellar density can be @xmath87 times higher . conversely , if these globulars have normal imfs , then protostellar coalescence must be unimportant nearly everywhere . a similar constraint would come from the imf in small regions , where only a few stars form . there , protostar interactions are not likely to be important either . a second major shift in our view of molecular clouds is that most astronomers now think that pre - stellar structure comes from turbulence , not gravitationally - driven fragmentation . this is because the same character of structure is observed in both self - gravitating and non - self - gravitating clouds , and both types of clouds have correlated motions reminiscent of laboratory turbulence . a related point is that the cloud cores in which clusters form are not obviously collapsing as a whole : there are no inverse p - cygni profiles indicative of collapse motions for whole cores , as there are in some individual protostars . the stars are apparently forming inside more - or - less stable cloud cores . once again , the general collapse models seem untenable . the presence of supersonic turbulence also seems to rule out models involving thermal instabilities . turbulent motions in cluster - forming cores are generally much larger than thermal , and so the dominant forces that structure the gas are the turbulent forces , not the thermal . this means that turbulence causes the observed structures in pre - stellar clouds , not thermal instabilities . thermal processes are probably important on the smallest scales , but these are at and below @xmath81 . turbulence in pre - stellar clouds also appears to generate structure that is much lower in mass than even the smallest stars , perhaps as low as @xmath88 m@xmath0 ( heithausen et al . 1998 ; kramer et al . this means that star formation occurs in the middle range of all cloud structures , not at the bottom end . it also means that , while turbulence may generate a cascade of structures down to smaller and smaller scales , the end result of this cascade is not the formation of a star . turbulence makes cloud structures independent of star formation . self - gravity , sonic wave generation , magnetic reconnection , and many other aspects of interstellar cloud dynamics are likely to be important too by the time star formation begins . the remarkable near - uniformity of the imf in regions with a wide range of stellar densities , cluster masses , metallicities , ages , and galactic types strongly suggests there is a _ unifying process in star formation _ that makes the relative probability of forming various masses somewhat independent of physical parameters . for example , the imf seems independent of how star formation begins , i.e. , whether the region is triggered by some external pressure or forms quiescently ( e.g. see parker et al . 1992 and the contrary opinion by oey & massey 1995 ) , independent of the general cloud shape ( shell , layer , filament , etc . ) , magnetic field strength , ionization level , presence of other stars , and so on . this is remarkable because these physical variables do control the _ rate _ of star formation in some regions , and they certainly control _ where and when _ star formation begins . they just do not appear to influence the imf . the near - uniformity of the _ slope _ of the power law also implies several other things . ( 1 ) there is a _ universal aspect of cloud structure and evolution that produces the power - law portion of the imf independently of the temperature , pressure , and metallicity . _ this means that the power - law probably does not depend on the protostellar collapse process , e.g. , on the fragmentation or collapse of isothermal clumps , the accretion of gas in thermal equilibrium , or the thermal jeans mass . for all of these processes , thermal temperature and magnetic diffusion are probably important at some stage ( see reviews in nakano 1984 ; mouschovias 1991 ; shu , adams , & lizano 1987 ) . ( 2 ) the imf power - law is also likely to be a _ highly reduced average over many physical processes that either all give about the same result , or are combined so finely in every region that the same proportion of each process is always present . _ the variations in slope of the imf from region to region ( scalo 1998 ) seem at first to suggest something different , that the imf is not in fact uniform but depends sensitively on physical conditions . but are these variations statistically significant ? rarely do studies of cluster imfs contain more than several hundred stars . figure [ fig : fluctuations ] shows sample imfs from the model of random selection in a hierarchical cloud ( elmegreen 1998 ) . the panel in the upper left shows imf slopes for 100 different models , each with a different number of stars and therefore different upper mass limit . in all cases , the range of stars chosen for the power - law slope determination is such that there are 200 stars in the fit , starting from the histogram bin that is twenty filled bins ( a factor of 3 in mass ) away from the most massive star , and going to lower mass bins . this starting point for the fit ensures that there are enough stars to get an accurate slope , and it coincides with an astronomer s decision to avoid fitting the imf to the scarce few highest - mass stars in the cluster . we pick a different number of total stars for each imf , and use only the highest mass parts of each model , because this is what an astronomer would do as well . generally an observer catalogues only the brightest stars in a cluster and misses the low mass members . if a cluster has an imf made only from low mass stars , then that cluster is generally so small that there are no high mass stars at all . evidently the imf slope varies a lot around the salpeter value , shown in the figure by the dashed line at @xmath22 . this average slope is nearly independent of physical cloud properties in this model ( see the discussion in section [ sect : theory ] ) . the other three panels in figure [ fig : fluctuations ] show particular imfs and the mass ranges and fits used for the `` x '' marks in the top left panel . figure [ fig : scatter ] shows the rms deviations in the imf slopes around the mean values for power - law fits that include different numbers of stars in the same 100 imfs ( from elmegreen 1998 ) . each plotted value , connected by a smooth line , represents the rms deviation for the 100 models , calculated in the mass range from 1 to 10 m@xmath0 ( the characteristic physical mass in the model comes from @xmath81 , which is taken to be 0.35m@xmath0 from equation [ eq : mj ] ) . the rms deviation around the salpeter slope decreases systematically as more and more stars are included in the fit . the deviation can be as high as @xmath6 when only 80 stars are used , decreasing to @xmath4 when 1000 stars are used . this implies that an imf calculated with , say , 150 stars , and having a slope of @xmath89 instead of the salpeter slope of @xmath14 , is in fact statistically consistent with this salpeter slope in a `` universal '' imf . the fluctuations found by scalo ( 1998 ) in his review correspond to @xmath6 around the value @xmath90 . these fluctuations could be statistical , considering the number of stars included in typical imf surveys and the likely presence of measurement and star selection errors in the real data . the dashed line in figure [ fig : scatter ] is for another 100 models , independent of the first , and evaluated with a large ( @xmath91 ) and random fluctuation in the ratio of the star mass to the clump mass . such variations do not affect the average imf slope or the statistical fluctuations around it . the imf has been observed directly and indirectly for many years , with many different results , but there seems to be some convergence now in the basic form of the imf , and also some tantalizing indications that this form changes a little when the physical conditions for star formation change a lot . the main observational results were summarized in section [ sect : sum1 ] . the theory of the imf seems to be all over the map , even in recent years . this variation reflects our ignorance in the processes that determine a star s mass and in the processes that cause the complex spatial structures seen in interstellar clouds . the common aspects in many of these models , including clump and protostellar interactions , wind erosion of protostellar gas concentrations , minimum masses from the lack of self - gravity , and hierarchical cloud structure , are all likely to play some role in generating the imf . the worry is that the imf is such a highly reduced average over many physical processes that each process does not a have clear signature in the final result . in view of the observations , we should perhaps consider `` acceptable '' those models that do not have much sensitivity to either the local aspects of cloud structure , or the large scale aspects of time and place in the universe . this drives the recent appeal towards models based on common turbulence in one way or another , considering that turbulent structures are likely to be robust . unfortunately , we do not understand compressible mhd turbulence much either . there are several key observations that would help . one is the determination of star formation time scales , always in units of the cloud or clump crossing time . there are three important timescales : the rise time of the star formation rate in a cloud , the duration of star formation , and the decay time scale . if the rise time is less than a crossing time , and the duration only one or two crossing times , then imf models based on multiple clump interactions would seem to be ruled out . there is also a need to know the mass dependence of the mass fraction of clump gas that goes into a star . this might be determined from the ratio of luminosity to gas mass for type 0 sources , plotted as a function of gas mass for the dense cores where the stars are actually forming . theory should tell us what gas structures and evolutionary timescales are expected from self - gravitating mhd turbulence in the supersonic - subalfvnic regime of molecular clouds . if turbulence only makes moving waves and wavepackets , and star formation actually occurs in such regions , then star formation has to be relatively quick in terms of the local crossing time . it may be that in the absence of strong self - gravity , turbulent clumps are transient and amorphous , but when self - gravity becomes important , the clumps get some integrity and persist for a relatively long time .

high precision nucleon - nucleon ( nn ) interactions such as av18 @xcite , cdbonn @xcite , nijm i , ii and 93 @xcite accurately describe the nn data set up to about 350 mev . when these interactions are used to predict binding energies of three - nucleon ( 3n ) systems they underestimate the experimental bindings of @xmath5 and @xmath6 by about 0.5 - 1 mev @xcite . this missing binding energy can be cured by introducing a three - nucleon force ( 3nf ) into the nuclear hamiltonian @xcite . the study of elastic nucleon - deuteron ( nd ) scattering and nucleon - induced deuteron breakup also revealed a number of cases where the nonrelativistic description based on pairwise interactions is insufficient to explain the data . generally , the studied discrepancies between a theory based on nn interactions only and experiment become larger with increasing energy of the 3n system . adding a 3nf that includes long - range @xmath7 exchange to the pairwise interactions leads in some cases to a better description of the data . the parameters of such a 3nf must be separately adjusted to the experimental binding of @xmath5 and @xmath6 @xcite for each nn interaction . the elastic nd angular distribution in the region of its minimum and at backward angles is the best studied example @xcite . the clear discrepancy in these angular regions at energies below @xmath8 mev nucleon laboratory energy between a theory based on nn interactions only and the cross - section data can be removed by adding modern 3nfs based on chiral effective field theory @xcite to the nuclear hamiltonian . at energies higher than @xmath8 mev current 3nfs @xcite only partially improve the description of cross section data and the remaining discrepancies , which increase with energy , indicate the possibility of relativistic effects @xcite . the need for a relativistic description of 3n scattering was also raised when precise measurements of the total cross section for neutron - deuteron ( nd ) scattering @xcite were analyzed within the framework of nonrelativistic faddeev calculations @xcite . nn interactions alone were insufficient to describe the data above @xmath8 mev . in few - body models off - shell effects , relativistic effects , and three - body force contributions can not be cleanly separated . this is because different two - body interactions that give the same two - body @xmath0 matrix are related by a unitary scattering equivalence @xcite . to maintain this equivalence at the three - body level requires additional three - body interactions @xcite in one of the hamiltonians . since relativistic two - body models are fit to the same data as the corresponding nonrelativistic models , there is a similar on - shell two - body scattering equivalence . while the relativistic and nonrelativistic three - body predictions will be different one can in principle make up the difference ( in a chosen frame ) with a suitable three - body interaction . so while it is possible to simulate `` relativistic effects '' with a three - body interaction , a poincar invariant treatment of the dynamics provides the most direct way to model the consequences of imposing poincar invariance and s - matrix cluster properties . in this paper we investigate one particular representation of the poincar invariant three - body problem . we compare the predictions of relativistic and nonrelativistic three - body calculations where the input two - body interactions give the same two - body s matrix , have the same internal two - body wave functions , and have a kinematic three - dimensional euclidean symmetry . the relativistic and nonrelativistic models differ in how these interactions appear in the three body - problem . in addition , the internal and single particle variables are related by galilean boosts in the nonrelativistic case and lorentz boosts in the relativistic case . in the nonrelativistic case the two - body interactions for each pair are added to the center of mass kinetic energy operator . in the relativistic case , the non - linear relation between the two and three - body mass operators must be respected in order to obtain a scattering matrix that clusters into a product of the identity and the input two - body @xmath0-matrix @xcite . this non - linear dependence of the three - body invariant mass operator on the two - body interaction has dynamical consequences for the three - body system which complicates the structure of the faddeev kernel . in @xcite we used a poincar invariant formulation of the 3n scattering problem . a technique for constructing the relativistic nucleon - nucleon interaction from a standard high - precision interaction was given in @xcite . we used the same technique to construct the transition operators that appear in the kernel of the relativistic faddeev equation . application to a 3n bound state supported the relativistic effects previously found in @xcite . realistic nn interactions are fit by properly transforming experimental data to the center of momentum frame and fitting @xmath0-matrix elements computed using the nonrelativistic schrdinger equation to this data . while the same data could be precisely fit using @xmath0-matrix elements computed from a relativistic schrdinger equation , this has not been done @xcite with the same precision used to construct realistic interactions . in @xcite , instead , an analytical scale transformation of momenta was used to relate nn interactions in the nonrelativistic and relativistic schrdinger equations in such a way , that the two - body scattering matrix elements are identified as @xmath9 as functions of the center of momentum energy @xcite . in this work we use an alternative procedure @xcite that generates a relativistic nucleon - nucleon interaction with the property that the relativistic and nonrelativistic two - body @xmath0 matrices satisfy @xmath10 , where @xmath11 is the experimental two - body @xmath0 matrix , and @xmath12 is the cm momentum of one of the particles . when high - precision potentials are determined @xcite by properly lorentz transforming scattering data from the laboratory frame to the center of momentum frame the lorentz invariant scalar product @xmath13 is used to relate the laboratory beam energy @xmath14 to the c.m . momentum @xmath12 . the potential is determined by comparing the transformed experimental scattering observables to the scattering observables computed using the nonrelativistic lippmann - schwinger equation , identifying the @xmath12 appearing in the lippmann - schwinger equation with the @xmath12 computed from the invariant @xmath15 . with this procedure the resulting interactions are constructed so the s - matrix elements in the relativistic and nonrelativistic cases are identified as functions of @xmath12 rather than cm energy . though the difference in the two approaches leads to a small mismatch in the relativistic and nonrelativistic momentum , the interactions generated by the analytic scale transformation provide a useful first step to investigate the effects of the non - linear relation between the three - body mass operator and two - body interactions . in our initial studies @xcite the interaction generated by the analytic scale transformation was used to study the changes in elastic nd scattering and breakup observables when the nonrelativistic form of the kinetic energy is replaced by the relativistic one and a proper treatment of the dynamics is included . we found that the elastic scattering cross section is only slightly influenced by relativity . only at backward angles and higher energies the elastic cross sections are increased by relativity . due to the selectivity of the breakup reaction , however , regions of phase - space were found at higher energies of the incoming nucleon where relativity leads to a characteristic pattern by which relativity changes the nonrelativistic breakup cross section . namely , in this region of phase - space fixing the angle of the first detected nucleon and changing the angle of the second nucleon provides variations of the nonrelativistic cross section by relativity , increasing or decreasing it by a factor up to @xmath16 . for spin observables the implemented relativistic features lead only to small effects . recently an interesting estimate of `` relativistic corrections '' has been performed and its effect on low - energy nd analyzing power @xmath2 has been estimated using the plane - wave impulse approximation @xcite . the estimate is based on a perturbative realization of the poincar lie algebra to leading order in @xmath17 @xcite . the calculations are done in the plane - wave impulse approximation . a large increase by @xmath18 of the @xmath2 maximum at laboratory energy @xmath19 mev has been found . the authors comment that their estimates are both exploratory and incomplete . in addition to the absence of final state interaction , there are a number of other important differences with an exact formulation of this problem . such a large effect , which would significantly reduce the discrepancy between theory and data in the region of the @xmath2 maximum , calls for a relativistic study in an exactly poincar invariant treatment of the three nucleon dynamics . our previous study @xcite , performed without inclusion of wigner rotations , is too limited for spin observables . therefore , in order to make definite conclusions for @xmath2 we perform a complete dynamical calculation including the effects of wigner rotations . we focus on that issue and do not include 3nf s . the paper is organized as follows . in section ii we discuss the construction of our poincar invariant dynamical model . this includes a discussion of how high - precision interactions are used to construct the three - body mass operator ( rest hamiltonian ) . in section iii we discuss spin observables in poincar invariant quantum theory . in sec . iv we discuss the formulation of the faddeev equation to construct scattering observables for this three - nucleon mass operator . this includes an exact treatment of the faddeev kernel which avoids the approximations used in @xcite . we solve the relativistic 3n faddeev equation with and without wigner spin rotations . we show and discuss results for the neutron analyzing power @xmath2 . v contains a summary and conclusions . in a quantum theory the principle of special relativity requires that the probabilities computed for equivalent experiments done in different inertial coordinate systems are identical . since inertial coordinate systems are related by poincar transformations , it follows @xcite that equivalent states in different inertial coordinate systems are related by a unitary representation , @xmath20 , of the poincar group . this emphasis on the invariance of experimental measurements in different inertial frames is different than the covariance requirements that are historically motivated by the way symmetries are realized in classical wave equations . since any representation of the poincar group can be decomposed into a direct integral of irreducible representations , one way to construct a poincar invariant dynamics is to build it out of irreducible representations . the transformation properties of irreducible representations are well known and completely determined by group theoretical considerations . the dynamics is contained in the spectrum of the physical mass and spin operators which determines the values and multiplicities of the casimir invariants that appear in this decomposition . our construction begins with one - particle representations , which are irreducible representations . the particle s mass and spin fix the eigenvalues of the two casimir invariants of the poincar group . for computations it is necessary to choose a basis for the irreducible representation space . this is done by choosing a maximal set of commuting hermitian functions of the infinitesimal poincar generators . in addition to the mass and spin , it is possible to find four additional mutually commuting non - invariant functions of the generators . there is a second set of four operators that are conjugate to the non - invariant commuting observables . these operators change the eigenvalues and determine the spectrum of the commuting observables . all ten generators can be expressed as functions of these eight non - invariant operators and the two casimir operators . the irreducible representation space , @xmath21 , is the space of square integrable functions of the eigenvalues of the four commuting operators @xcite . our choice of basis for irreducible representation spaces is the simultaneous eigenstates of the linear momentum @xmath22 and the 3-component of the canonical spin , @xmath23 , which is the observable corresponding to the spin measured in the particle s rest frame if the particle is transformed to its rest frame with a rotationless lorentz transformation . in this basis the irreducible unitary representation of the poincar group is @xcite : @xmath24 \label{b.1}\ ] ] where @xmath25 , @xmath26 , and @xmath27 is the rotationless lorentz transformation that takes a particle of mass @xmath28 at rest to momentum @xmath22 . the quantity @xmath29 is the standard rotationless - boost wigner rotation . the representation ( [ b.1 ] ) is unitary for states with a @xmath30 normalization in the momentum . the important observation is that _ all _ mass @xmath31 spin @xmath32 irreducible representations of the poincar group in the @xmath33 basis have this form . the two or three - nucleon hilbert space is the tensor product of two or three single - nucleon irreducible representation spaces ; @xmath34 or @xmath35 . on each of these spaces @xmath36 define kinematic representations of the poincar group . these representations are reducible and do not contain any dynamics . we build dynamical irreducible representations by adding suitable interactions to the mass casimir operator of non - interacting irreducible representations . the first step needed to introduce interactions is then to decompose these non - interacting tensor product representations into a direct integral of irreducible representations . this is accomplished with poincar group clebsch - gordan coefficients in our chosen @xmath37 basis . the poincar group clebsch - gordan coefficients are the expansion coefficients of linear combination of tensor product states that transform irreducibly . the desired non - interacting irreducible states are computed by ( 1 ) constructing rest eigenstates of the two - body system , ( 2 ) decomposing them into irreducible representations under @xmath38 rotations , ( 3 ) boosting the result to an arbitrary frame . the resulting clebsch - gordan coefficients in this basis are @xcite@xcite@xcite @xmath39 @xmath40 @xmath41 d^{j_2}_{\mu_2 ' \mu_2}[r_w(b(\mathbf{p}/m_{120 } ) , k_2 ] = \label{b.3}\ ] ] @xmath42 @xmath43 d^{j_2}_{\mu_2 ' \mu_2}[r_w(b(\mathbf{p}/m_{120 } ) , k_2 ] . \label{b.4}\ ] ] in these expressions @xmath44 @xmath45 and the two - body invariant mass @xmath46 is replaced by the continuous variable @xmath47 . the quantum numbers @xmath48 and @xmath49 are kinematically invariant quantities that distinguish multiple copies of representations with the same mass ( @xmath50 ) and spin . for a two nucleon - system they have the same spectrum as the orbital and spin angular momentum operators in a partial wave representation of the nonrelativistic basis . an irreducible representation is also obtained by changing the order of the spin couplings in ( [ b.3]-[b.4 ] ) where the orbital angular momentum is first coupled to one of the spins , @xmath51 , and then the result is coupled to the second spin . @xmath52 . this representation is constructed by making the replacements in @xmath53 @xmath54 in equation ( [ b.3 ] ) or ( [ b.4 ] ) . in this representation the degeneracy parameters @xmath55 are replaced by @xmath56 . when we construct three - body irreducible representations by successive pairwise coupling we use the coupling ( [ b.3 ] ) in the first clebsch - gordan coefficient and the coupling ( [ b.8 ] ) in the second clebsch - gordan coefficient . this allows us to identify the quantum numbers of the relativistic irreducible basis with the quantum numbers that we used in previous nonrelativistic calculations ( @xcite ) . three - particle irreducible representations for systems of non - interacting particles can be constructed by successive pairwise coupling of irreducible representations : @xmath57 @xmath58 for three nucleon scattering or bound state problems it is sufficient and convenient to work in the three - body center of momentum frame . this simplifies the coefficients ; the wigner rotations in the second clebsch - gordan coefficient in ( [ b.9 ] ) become the identity and the normalization factor @xmath59 . both of these factors are non - trivial in the first coefficient . the form of these coefficients for the case that particle 1 is a spectator ( for details see @xcite ) and using a short hand notation , is @xmath60 d^{1\over{2 } } _ { \mu_3 ' \mu_3 } [ r_w ( b(-\mathbf{q}_1/m_{023 } ) ) ~ ) , { k}_3(~ \mathbf{q}_2,\mathbf{q}_3 ) ~ ) ] . \label{b.10}\end{aligned}\ ] ] where @xmath61 @xmath62 the important property of the states @xmath63 and @xmath64 is that they transform irreducibly . the mass and spin are given by @xmath65 @xmath66 respectively . in order to use these representations to construct dynamical representations an interaction is added to the two- or three - body invariant mass operator of the form @xmath67 for @xmath68 or @xmath69 for @xmath70 . diagonalizing @xmath71 or @xmath72 in the noninteracting irreducible basis gives simultaneous eigenstates of @xmath73 for @xmath68 and of @xmath74 for @xmath70 . in both the two and three - body case these eigenstates , @xmath75 and @xmath76 , where @xmath77 and @xmath78 are the eigenvalues of @xmath79 and @xmath80 , are complete on the two and three - body hilbert spaces respectively . the dynamical representation of the poincar group is defined by requiring that these eigenstates transform like ( [ b.1 ] ) with the mass being replaced by the mass eigenvalues @xmath81 or @xmath77 . this representation is unitary and defines the dynamics of the system . with our choice of irreducible basis , @xmath82 , the resulting irreducible representations of the poincar group have a mass independent representation of the three - dimensional euclidean subgroup , which dirac @xcite called an `` instant - form dynamics '' . for the three - nucleon case there remains the problem of how to construct realistic @xmath83-interactions . for two - body interactions the relation @xmath84 implies that the square of the two - body mass operator has a simple relation to the nonrelativistic rest hamiltonian with a `` realistic '' nn interaction @xcite provided one identifies the spectrally equivalent relative momenta , @xmath12 . in the relativistic case @xmath85 @xmath86 while in the nonrelativistic case @xmath87 : @xmath88 where @xmath89 is a rotationless lorentz boost and @xmath90 is the corresponding galilean boost . the kato - birman invariance principle @xcite implies that the mller wave operators satisfy @xmath91 where @xmath92 is any piecewise differentiable function of bounded variation with positive derivative @xcite . the functions @xmath93 and @xmath94 satisfy the conditions of kato - birman theorem . using equation ( [ b.23 ] ) along with the kinematic euclidean invariance of the hamiltonians @xmath95 and @xmath96 gives the following relation between the two - body scattering wave operators @xmath97 on the other hand the identification ( [ b.20 ] ) along with the reparametrization @xmath98 gives @xmath99 writing both wave operators as direct integrals over @xmath100 leads to the identifications @xmath101 and @xmath102 the identification of the relativistic and nonrelativistic wave operators as functions of @xmath12 ensures that the relativistic two - body model is fit to the same two - body @xmath0-matrix ( experimental data ) as the nonrelativistic model provided the interactions are related by ( [ b.20 ] ) . the identification of the wave operators also implies the identity of the scattering wave functions as a function of @xmath103 . the identity of the bound state wave functions is also due to the relation ( [ b.20 ] ) . in our calculations we use the interaction @xmath104 defined by @xmath105 which we construct @xcite from the @xmath83 interaction in ( [ b.20 ] ) by iterating @xmath106 in the irreducible plane - wave basis . because @xmath79 and @xmath107 have the same eigenvectors , the @xmath103 dependence of the wave functions constructed from @xmath79 are also identical to the corresponding nonrelativistic wave functions . the equation ( [ b.28 ] ) in the irreducible plane - wave basis has the form @xmath108 @xmath109 the iteration converges quickly for realistic interactions @xcite . while a mathematical proof of convergence of the iteration of ( [ b.29 ] ) is lacking , the results of the iterations are easily tested because the resulting @xmath110 must have the same eigenfunctions as the nonrelativistic two - body hamiltonian . we applied this approach using the cd bonn potential as the nonrelativistic interaction @xmath111 . in our previous studies @xcite we used the momentum transformation of ref . @xcite and in addition restricted to leading order terms in @xmath112 and @xmath113 expansion only @xmath114 we checked that in most cases this simple approximation leads to practically the same results as the exact approach applied in the present study . once the two - body mass operator is constructed , the three - body mass operator for the interacting @xmath115 pair is the well - defined non - linear function of the two - body mass : @xmath116 where @xmath117 is embedded in the three body hilbert space so it commutes with @xmath118 . if this is interpreted as the rest energy operator , the interacting pair and spectator energies are additive in the rest frame . this implies that the @xmath0 matrix clusters properly in the rest frame - while the invariance of @xmath0 in all frames ensures that this property extends to all inertial coordinate systems . pairwise interactions in the three - body system are defined by @xmath119 @xmath120 for any pair of particles these interactions commute with kinematic momentum and spin , and are independent on the momentum and magnetic quantum numbers . this ensures that the sum of the interactions has the general form ( [ b.19 ] ) . a generalization of the method used in ( [ b.28]-[b.29 ] ) can be used to construct @xmath121 by iterating @xmath122 specifically @xmath123 @xmath124 @xmath125 \label{b.34}\ ] ] this iteration also converges and is used to construct interactions @xmath126 for each pair of particles . the three - body mass ( rest energy ) operator is @xmath127 where @xmath128 is the three - body kinematic invariant mass : @xmath129 our relativistic faddeev equation is based on this mass operator ( [ b.35 ] ) with two - body interactions constructed from the cd bonn interaction using equations ( [ b.28 ] ) , ( [ b.29 ] ) , ( [ b.33 ] ) and ( [ b.34 ] ) . finally , just like in the nonrelativistic case , there is a natural order of coupling of the irreducible representations for computing each pairwise interaction . the change of basis relating different orders of coupling is needed for the implementation of the faddeev equation as an integral equation . the required basis change only changes the invariant degeneracy quantum numbers associated with each order of coupling . @xmath130 \label{b.37}\ ] ] the invariants @xmath131 $ ] are racah coefficients for the poincar group . they are constructed using four poincar clebsch - gordan coefficients . we compute this quantity using the balian - brezin method @xcite , where the variables associated with one order of coupling are expressed in terms of the variables associated with another order of coupling . the invariant coefficient , @xmath131 $ ] , can be computed by evaluating the expression at zero momentum , averaging over the magnetic quantum numbers , and evaluating the resulting expression at any kinematically allowed set of momenta @xcite . these racah coefficients contain wigner rotations and jacobians which do not appear in the nonrelativistic permutation operators . explicit expressions are given in appendix [ a2 ] . in a poincar invariant quantum theory or relativistic quantum field theory the spin of a particle can be defined as the angular momentum that is measured in the particle s rest frame . for any non - zero momentum , @xmath22 , different lorentz transformations can be used to transform a particle at rest to frame where the particle has momentum @xmath22 . because the commutator of two rotationless boost generators @xmath132 = -i \epsilon_{jkl } j_j \label{s.1}\ ] ] is a rotation generator , the spin of the particle with momentum @xmath22 depends on the choice of lorentz transformation that transforms the particle s momentum from zero to @xmath22 . in order to have an unambiguous definition of the spin it is necessary to choose a standard set of @xmath22 dependent lorentz boosts , @xmath133 , that transform a particle of mass @xmath28 at rest to momentum @xmath22 . then the spin of the particle can be unambiguously defined as the value of the spin measured in the particles rest frame if it is transformed to the rest frame using the standard lorentz transformation . with this definition , if two spins are equal in one frame they are equal in all frames . the choice of standard boost is not unique because if @xmath134 is any @xmath22 dependent rotation and @xmath135 then both @xmath136 and @xmath137 both transform @xmath22 to zero . each choice of boost leads to a different spin operator , corresponding to a different prescription for measuring the spin in an arbitrary frame . the rotation @xmath138 that relates different boosts is called a generalized melosh rotation @xcite@xcite . the spin operator @xmath139 associated with a boost @xmath140 is defined as the following function of the poincar generators : @xmath141 where @xmath142 is the pauli - lubanski vector , @xmath143 is the relativistic angular momentum tensor , and @xmath144 is the boost matrix with the parameters @xmath145 and @xmath28 replaced by the mass and momentum operators . while this quantity has the appearance of a four vector , it is not because of the operator dependence of the arguments of the boost . instead , under lorentz transformations the spin transforms like @xmath146 where @xmath147 is the wigner rotation associated with the boost @xmath140 . it is a consequence of the poincar commutation relations that the components of any of these spin observables satisfy @xmath38 commutation relations @xmath148_{- } = i \epsilon_{lmn } j_{xn } \label{s.7}\ ] ] for any `` @xmath149 '' . the operator @xmath150 is independent of the choice of boost because the generalized melosh rotations leave the scalar product of two vectors unchanged . the spin defined with the textbook rotationless or canonical boost is called the canonical spin . it is natural to ask how these different types of spins are measured in the laboratory . spins of isolated elementary or composite particles are measured in the laboratory through their response to classical electromagnetic fields . in the one - photon exchange approximation the photon couples to matrix elements of a covariant current operator . imposing poincar covariance , current conservation , and discrete symmetries allows one to express all current matrix elements in terms of an independent set of matrix elements , which have a 1 - 1 correspondence with invariant form factors . all conventional form factors can be expressed in terms of breit frame matrix elements with canonical spin and a quantization axis parallel to the breit frame momentum transfer . in the @xmath151 representation canonical boosts are represented by positive hermitian matrices . they have the general form @xmath152 where @xmath153 is the rapidity of the lorentz transformation and @xmath154 are the pauli matrices . in this paper all of our spins are canonical spins . the @xmath155 representation of canonical boosts are the standard rotationless boosts . given a definition of the form factors in terms of independent current matrix elements in a given basis , it is also possible to express them in terms of current matrix elements in any other standard frame using any other basis @xcite . for example the expression in terms of breit frame canonical spin matrix elements , can be replaced by a different independent set of laboratory frame helicity spin matrix elements . in quantum field theory the choice of boost is built into conventions used to define the dirac spinors . the relation of the invariant form factors to current matrix elements with different choices of spin determines the relationship between different spin observables and experiment . the spin degrees of freedom of the asymptotic incoming or outgoing particles are most conveniently expressed in terms of traces of density matrices , which is a reflection of the fact that realistic initial and/or final states are generally not pure states . scattering spin observables in cross sections are formally defined by @xcite @xmath156 where @xmath157 is the invariant scattering amplitude for the reaction under consideration . the connection between the invariant scattering amplitude defined in the particle data book @xcite and the transition amplitudes constructed by solving our formulation of the relativistic faddeev equations is given in @xcite . the quantities @xmath0 have the form @xmath158 where the index runs over all @xmath159 or @xmath160 initial or final sets of magnetic quantum numbers , @xmath161 are a basis for @xmath162 or @xmath163 matrices that are orthonormal with respect to the trace norm and @xmath164 are constant coefficients @xcite . if the initial and final asymptotic states are represented in a canonical spin basis , then the magnetic quantum numbers that appear in the invariant amplitudes @xmath157 transform with wigner rotations under lorentz transformations . the result is that the spin observable @xmath165 will not be invariant unless the matrices @xmath161 or coefficients @xmath164 are defined to transform in a manner that leaves the observable invariant . any spin observable can be made lorentz invariant , by defining the invariant observable as its value in a given frame if it is transformed to the frame with a specific lorentz boost . this can be used to get an invariant definition of the vector or tensor polarizations . in this paper invariant spin observables are defined to be the values of the observable in the laboratory frame ( rest frame of the target ) . to evaluate the corresponding spin observable it is only necessary to evaluate the expression @xmath166 for the values of the invariant amplitudes with laboratory kinematics . this observable is equal its value evaluated in other frames using the formula @xmath167 where the invariant amplitudes are evaluated in the other frame and @xmath168 and @xmath169 are products of wigner @xmath168-functions of the wigner rotations associated with the boost from the lab frame to the other frame . our specific interest in this paper is the observable @xmath2 with polarized incoming nucleon . the convention used to define @xmath2 is the madison convention , where the laboratory frame scattering plane is in the @xmath170 plane . the observable @xmath2 , is defined as @xmath171 because the wigner rotation for canonical boost along the direction of a particle s momentum is the identity , @xmath172 is unchanged so @xmath2 can also be evaluated in the c.m . frame without making any compensating wigner rotations . the nucleon - deuteron scattering with neutron and protons interacting through a nn interaction @xmath173 alone is described in terms of a breakup operator @xmath157 satisfying the faddeev - type integral equation @xcite @xmath174 the two - nucleon ( 2n ) @xmath175-matrix @xmath175 results from solving the lippmann - schwinger equation with the interaction @xmath173 . the permutation operator @xmath176 is given in terms of the transposition @xmath177 which interchanges nucleons i and j. the incoming state @xmath178 describes the free nucleon - deuteron motion with relative momentum @xmath179 and the deuteron state vector @xmath180 . finally @xmath181 is resolvent of the three - body center of mass kinetic energy . the elastic nd scattering transition operator u is given in terms of @xmath157 by @xcite @xmath182 this is our standard nonrelativistic formulation , which is equivalent to the nonrelativistic 3n schrdinger equation plus boundary conditions . the formal structure of these equations in the relativistic case remains the same but the ingredients change . as explained in @xcite the relativistic 3n rest hamiltonian ( mass operator ) has the same form as the nonrelativistic one , only the momentum dependence of the kinetic energy changes and the relation of the pair interactions in the three - body problem to the pair interactions in the two - body problem changes . consequently all the formal steps leading to eqs.([eq1a ] ) and ( [ eq1c ] ) remain the same . the free relativistic invariant mass of three identical nucleons in their c.m . system has the form ( [ b.36 ] ) while the free two - body mass operator has the form ( [ b.7 ] ) . as introduced in @xcite the pair forces in the relativistic 3n @xmath183 mass operator are given by ( [ b.32 ] ) where @xmath184 reduces to the interaction @xmath104 for @xmath185 . the transition matrix that appears in the kernel of the faddeev equation ( [ eq1a ] ) is obtained by solving the `` lippmann - schwinger equation '' , which must be solved as a function of @xmath186 @xmath187 the input two - body interactions are computed by solving equation ( [ b.33 ] ) and ( [ b.34 ] ) . the new relativistic ingredients in eqs.([eq1a ] ) and ( [ eq1c ] ) will therefore be the @xmath175-operator ( [ eq2a ] ) ( expressed in partial waves ) and the resolvent of the 3n invariant mass @xmath188 where @xmath128 is given by eq . ( [ b.36 ] ) and @xmath189 is the total 3n c.m . energy expressed in terms of the initial neutron momentum @xmath179 relative to the deuteron @xmath190 and @xmath191 is the deuteron rest mass . currently eq.([eq1a ] ) in its nonrelativistic form is numerically solved for any nn interaction using a momentum space partial - wave decomposition . details are presented in ref . this turns eq.([eq1a ] ) into a coupled set of two - dimensional integral equations . as shown in @xcite , in the relativistic case we can keep the same formal structure , though the permutation operators are replaced by the corresponding racah coefficients ( [ b.37 ] ) for the poincar group . these coefficients include both jacobians and wigner rotations that do not appear in the nonrelativistic permutation operators @xcite . these coefficients are computed in the appendix [ a2 ] , using methods that we have applied to compute the nonrelativistic permutation operators . in the nonrelativistic case the partial wave projected momentum space basis is @xmath192 where p and q are the magnitudes of standard jacobi momenta ( see @xcite ) , obtained by transforming single particle momenta to the rest frame of a two or three - body system using galilean boosts , and @xmath193 two - body quantum numbers with obvious meaning , @xmath194 refer to the third nucleon ( described by the momentum q ) , @xmath195 is the total 3n angular momentum and the rest are isospin quantum numbers . in the relativistic case this basis is replaced by the irreducible plane wave states defined in ( [ b.10 ] ) . the basis states ( [ b.10 ] ) are used for the evaluation of the partial wave representation of the permutation operator p with wigner rotations of spin states for nucleons @xmath196 and @xmath197 included . in the relativistic case we adopt the following short - hand notation for the irreducible three - body states , which also includes isospin quantum numbers coupled in the same order : @xmath198 equipped with that , projecting eq.([eq1a ] ) onto the basis states @xmath199 one encounters like in the nonrelativistic notation @xcite @xmath200 this is evaluated by inserting the complete basis of states @xmath201 and using ( [ b.10 ] ) . it can be expressed in a form which resembles closely the one appearing in the nonrelativistic regime @xcite @xmath202 where all ingredients are given in the appendix [ a2 ] . due to short - range nature of the nn interaction it can be considered negligible beyond a certain value @xmath203 of the total angular momentum in the two - nucleon subsystem . generally with increasing energy @xmath203 will also increase . for @xmath204 we put the t - matrix to be zero , which yields a finite number of coupled channels for each total angular momentum j and total parity @xmath205 of the 3n system . to achieve converged results at our energies we used all partial wave states with total angular momenta of the 2n subsystem up to @xmath206 and took into account all total angular momenta of the 3n system up to @xmath207 . this leads to a system of up to 143 coupled integral equations in two continuous variables for a given @xmath195 and parity . the subject of the present study is to investigate the influence of relativity on the nd elastic scattering nucleon analyzing power @xmath2 at low energies . we define the invariant observable @xmath2 to be the value measured in the laboratory frame ( target at rest ) . to this aim we solved faddeev equations at a number of the incoming neutron lab energies @xmath208 mev , @xmath209 mev , and @xmath210 mev . to check the energy dependence of the effect we added two additional energies @xmath211 mev and @xmath212 mev . in order to see the importance of specific relativistic features we solved the equation in the relativistic case with and without wigner rotations . this allowed us to see which effects , dynamical corrections , which are induced by the momentum dependence of the two - body force together with kinematical relativistic effects coming from the use of the poincar jacobi variables , or wigner rotations , dominate for @xmath2 . figs.[fig1 ] and [ fig2 ] illustrate the results . when only dynamical effects are taken into account ( wigner rotations neglected ) then at low energies of the incoming neutron the relativistic and nonrelativistic predictions are practically the same with the exception of the angular region close to the maximum of @xmath2 where the relativistic prediction is @xmath213 below the nonrelativistic one ( see fig.[fig1 ] ) . this small effect disappears at higher energies ( see fig.[fig2 ] ) . including wigner rotations lowers significantly the values of @xmath2 in a large region of angles around the maximum . the changes in the maximum are up to @xmath214 . again , when the energy of the incoming neutron increases nonrelativistic and relativistic predictions are practically identical . the large changes of @xmath2 occur in a region of energies where this observable is extremely sensitive to changes in @xmath215 , @xmath216 , and @xmath217 nn force components @xcite . at energies where this sensitivity dies out also the relativistic effects for @xmath2 become negligible . this allows to conclude that the large effects seen for @xmath2 at low energies are due to amplification of changes of the @xmath218 contributions due to relativity by a large sensitivity of @xmath2 to @xmath219-waves . in a recent study @xcite the changes of @xmath2 due to relativity by @xmath220 at @xmath221 mev have been reported . very probably the opposite sign of the effect found in that study can be attributed to the impulse approximation used when calculating @xmath2 . for @xmath222 elastic scattering at various lab energies @xmath223 of the incoming neutron . the dotted line is the result of the nonrelativistic faddeev calculation with the cd bonn potential . the relativistic predictions without and with wigner spin rotations are shown by the dashed and solid lines , respectively . the nd data at @xmath224 mev and @xmath209 mev are from ref . @xcite and at @xmath210 mev are from ref . @xcite . ] for @xmath222 elastic scattering at @xmath225 mev and @xmath212 mev . for description of lines see fig.[fig1 ] . all theoretical predictions are practically overlapping . the pd data at @xmath226 mev are from ref . @xcite and at @xmath212 mev from @xcite . ] we numerically solved the 3n faddeev equation for nd scattering including relativistic kinematics , dynamical relativistic effects and wigner rotations at the neutron lab energies @xmath227 , @xmath209 , @xmath210 , @xmath226 , and @xmath212 mev . as dynamical input we took the nonrelativistic nn potential cd bonn and generated in the 2n c.m . system an exactly on - shell equivalent relativistic interaction @xmath104 , by solving numerically nonlinear quadratic equation relating matrix elements of the nonrelativistic and relativistic potentials . we checked that the approximate procedure using an analytical scale transformation of momenta applied in our previous studies provides practically the same results as the present exact approach . in addition a similar nonlinear equation ( eq.([b.33 ] ) ) was used to generate the momentum dependent two - body interaction embedded in the @xmath197-particle hilbert space . we found that at low energies the effects of wigner rotations are most important for the analyzing power . they lower the maximum of @xmath2 by up to @xmath214 . the dynamical relativistic effects are of minor importance for @xmath2 and provide small changes of @xmath2 in a region close to its maximum . they lower @xmath2 by only @xmath3 . the relativistic effects disappear at higher energies . wigner rotations are negligible for the cross section and all other spin observables in elastic nd scattering with exception of four low energy spin correlations @xmath228 , @xmath229 , @xmath230 , @xmath231 and four low energy spin transfer coefficients : from deuteron to deuteron @xmath232 and @xmath233 , and from deuteron to neutron @xmath234 and @xmath235 . very probably this can be traced back as in the case of @xmath2 to their sensitivity to @xmath236-waves . these results shed new light on the low - energy analyzing power puzzle . it is known that the existing discrepancies between @xmath2 data and theoretical predictions based on nn potentials only can not be removed when current three - nucleon force , mostly of @xmath237-exchange character @xcite are included in the nuclear hamiltonian . this indicated that additional 3n forces should be added to the @xmath237-exchange type forces . such forces provided by @xmath238pt in nnlo and nnnlo orders are expected to provide the solution for the @xmath2 puzzle @xcite . it seems that in view of the present result they must increase the maximum of @xmath2 stronger than expected up to now . however , @xmath2 is a very sensitive observable and our approach , using the irreducible @xmath239 basis , is only one of many possible basis choices that lead to poincar invariant dynamical theories that are two - body scattering equivalent , which may give different three - body predictions . this work was partially supported by the helmholtz association through funds provided to the virtual institute `` spin and strong qcd''(vh - vi-231 ) . it was also performed in part under the auspices of the u. s. department of energy , office of nuclear physics , contract no . de - fg02 - 86er40286 with the university of iowa . the numerical calculations were performed on the ibm regatta p690 + of the nic in jlich , germany . in our calculations @xmath240 matrices are used to represent wigner rotations and lorentz transformations . in this way spin algebra is reduced to working with @xmath241 complex matrices , and there is no need for representations in terms of euler angles . the relevant relations are give below . the coordinates of a four - vector @xmath242 can be labeled by the @xmath241 hermitian matrix @xmath219 : @xmath243 where @xmath244 are the identity and the three pauli matrices . the components of @xmath242 can be extracted from the matrix @xmath219 using : @xmath245 because @xmath246 and @xmath247 for real @xmath242 , it follows that any linear transformation that preserves the hermiticity and determinant of @xmath219 is a real lorentz transformation . it is easy to show that if @xmath248 is a complex @xmath249 matrix with @xmath250 then the transformation @xmath251 has both of these properties : @xmath252 the most general @xmath241 matrix with determinant 1 can be written as @xmath253 where @xmath254 . if @xmath255 then @xmath256 is unitary and corresponds to an @xmath38 rotation through an angle @xmath257 about the @xmath258 axis . if @xmath259 then @xmath248 is a positive hermitian matrix that corresponds to a rotationless ( canonical ) lorentz boost with rapidity @xmath260 in the direction @xmath261 the rotation @xmath262 is given by @xmath263 the axis of rotation can be extracted using @xmath264 and the angle of rotation can be extracted from @xmath265 the rotationless boost that transforms a particle of mass @xmath28 at rest to total momentum @xmath22 can be labeled by the final four velocity @xmath266 : @xmath267 where @xmath268 is the rapidity of the lorentz boost which is related to @xmath269 by @xmath270 and @xmath271 @xmath272 satisfies @xmath273 where @xmath274 the rotationless boost above is called the canonical boost . the inverse transformation is obtained by reversing the sign of @xmath269 or @xmath22 . the wigner @xmath168 functions are homogeneous polynomials of degree @xmath275 in the coefficients of the @xmath240 matrices @xmath276 : @xmath277^{1/2 } \over ( j+\mu-\alpha)!\alpha ! ( \alpha - \mu + \nu ) ! ( j -\nu -\alpha ) ! } a_{11}^{j+\mu-\alpha } a_{12}^{\alpha } a_{21}^{\alpha -\mu+\nu } a_{22}^{j-\nu-\alpha } \label{a2.16}\ ] ] these are representations of both @xmath240 and @xmath38 . the @xmath278 representation is just the matrix @xmath248 . using eq . ( [ b.10 ] ) twice for the bra state @xmath279 and the ket state @xmath280 one gets for the matrix element of the permutation operator in our partial wave basis : @xmath281 ~ \cr & ~ & d^{{1\over{2}}~*}_{m_3 \mu_3 } [ r_w(b ( - \mathbf{q}/2\omega_m(\mathbf{k } ) ) , ( - \mathbf{k}(~ \mathbf{q}~ ' , -\mathbf{q } - \mathbf{q}~'~)~,\omega_m(\mathbf{k } ) ) ] ~ \cr & ~ & d^{{1\over{2}}}_{m_3 \mu_2 ' } [ r_w(b ( -\mathbf{q}~'~/2\omega_m(\mathbf{k } ) ) , ( \mathbf{k}(~ -\mathbf{q } -\mathbf{q}~ ' , \mathbf{q}~ ) ~ , \omega_m(\mathbf{k } ) ) ] ~ \cr & ~ & d^{{1\over{2}}}_{m_1 \mu_3 ' } [ r_w(b ( -\mathbf{q}~'~/2\omega_m(\mathbf{k } ) ) , ( -\mathbf{k}(~ -\mathbf{q } -\mathbf{q}~ ' , \mathbf{q}~ ) ~ , \omega_m(\mathbf{k } ) ) ] ~ \cr & ~ & _ 1 < ( ~ ( t{1\over{2 } } ) t ~ \vert ~ ( t'{1\over{2 } } ) t ~ > _ 2 , \label{eq17}\end{aligned}\ ] ] proceeding as in @xcite one gets the following expression for the matrix element of the permutation operator p : @xmath287 with @xmath288 and @xmath289 { d^{1/2}_{m_3 \mu_2'}}[r_w(b(\mathbf{q}'/2\omega_m(\mathbf{k } ' ) ) , ( \mathbf{k}',\omega_m(\mathbf{k } ' ) ) ] ) \cr & ~ & \sum_{\mu_3 ' } ( { 1\over{2 } } , \mu_2 ' , { 1\over{2 } } , \mu_3 ' \vert s ' , \mu_2 ' + \mu_3 ' ) \cr & ~ & \sum_{m_1 } ( \lambda 0 { 1\over{2 } } , m_1 , \vert i , m_1 ) { d^{1/2}_{m_1 \mu_3'}}[r_w(b(-\mathbf{q}'/2\omega_m(\mathbf{k } ' ) ) , ( -\mathbf{k}',\omega_m(\mathbf{k } ' ) ) ] \cr & ~ & \sum_{\mu } ( l , \mu-\mu_2-\mu_3 , s , \mu_2+\mu_3 \vert j , \mu ) ( -)^{\mu-\mu_2-\mu_3}y_{l -(\mu-\mu_2-\mu_3}(\hat{\mathbf{p } } ) ( j , \mu , i , m_1 \vert j , \mu+m_1 ) \cr & ~ & \sum_{\mu ' } ( l ' , \mu'-\mu_2'-\mu_3 ' , s ' , \mu_2'+\mu_3 ' \vert j ' , \mu ' ) y_{l ' \mu'-\mu_2'-\mu_3'}(\hat{\mathbf{p } } ' ) ( j ' , \mu ' , i ' , \mu+m_1-\mu ' \vert j , \mu+m_1 ) \cr & ~ & \sum_{m_2 } ( \lambda ' , \mu+m_1-\mu'-m_2 , { 1\over{2 } } , m_2 \vert i ' , \mu+m_1-\mu ' ) \cr & & ~~~~ { d^{1/2}_{m_2 \mu_2}}^*[r_w(b(-\mathbf{q}/2\omega_m(\mathbf{k } ) ) , ( \mathbf{k},\omega_m(\mathbf{k } ) ) ] y_{\lambda ' \mu+m_1-\mu'-m_2}(\hat{\mathbf{q } } ' ) . \label{eq23}\end{aligned}\ ] ] we use standard notation @xmath290 . it is assumed that z - axis is along @xmath269 and the momentum @xmath291 lies in the x - z plane what leads to the following components of the @xmath269 , @xmath291 , @xmath103 , and @xmath292 vectors @xmath293 \cr \mathbf{q}~ ' & = & [ q'\sqrt{1-x^2 } , 0 , q'x ] \cr \mathbf{k } & = & [ q'\sqrt{1-x^2 } , 0 , q'x + { 1\over { 2}}q(1+y_1(q , q',x ) ) ] \cr \mathbf{k } ' & = & [ -{1\over{2}}q'(1+y_2(q , q',x))\sqrt{1-x^2 } , 0 , -q-{1\over { 2}}q'(1+y_2(q , q',x))x ] \label{eq24}\end{aligned}\ ] ] though the direct evaluation of euler angles as arguments of the wigner d - functions could be used like in @xcite , here we use the @xmath240 representations of lorentz transfomations discussed in appendix [ a1 ] . this leads to @xmath294 & = & b ( \mathbf{p}_2/m ) b ( - \mathbf{q}/m_0 ) b ( \mathbf{k}/m)\nonumber\\ & = & \sqrt{\frac{e_0 + m_0}{2 m_0 } } \sqrt{\frac{\omega(k ) + m}{\omega(p_2 ) + m}}\nonumber\\ & - & \frac{\mathbf{k } \cdot \mathbf{q } } { \sqrt{2 m_0 ( e_0 + m_0 ) ( \omega(k ) + m ) ( \omega ( p_2 ) + m)}}\nonumber\\ & + & i \mathbf{k } \times \mathbf{q } \cdot \bsig \frac{1}{\sqrt{2 m_0 ( e_0 + m_0 ) ( \omega(k ) + m ) ( \omega ( p_2 ) + m)}}\end{aligned}\ ] ] where @xmath295

oscillatory exchange coupling between metallic magnetic layers across a non - magnetic spacer has been intensively studied over the last few years . one of the main issues in this field has been the determination of the oscillation periods of the coupling as a function of the spacer thickness . the physical mechanisms proposed for explaining the so - called oscillatory interlayer coupling include the quantum well theory ( qwt ) of edwards _ et al._@xcite and an extension of the ruderman - kittel - kasuya - yosida ( rkky ) theory adapted to the multilayer geometry due to bruno and chappert@xcite . both theoretical approaches seem to agree that the origin of the oscillation periods is in the structure of the spacer fermi surface ( fs ) . in fact , in the rkky theory the periods are associated with wave vectors directly obtained from the fs . more specifically , they are given by vectors perpendicular to the layer planes which span two points of the fs with anti - parallel fermi velocities@xcite . there has been a general belief that all oscillation periods are given by the rkky theory and therefore that such a simple picture is always valid . indeed , it has been shown that in certain simple models @xcite the periods predicted by the qwt coincide with the rkky ones and are given by the extremal dimensions of the spacer fs in the direction perpendicular to the layers . in another case @xcite , where the lattice lacks reflection symmetry about the layer planes , the correspondence between the quantum well and the rkky periods is rather subtle but still obtains . in the models mentioned above , harmonics of the rkky periods are obtained but no new fundamental periods . furthermore , dalbuquerque e castro _ et al_. @xcite showed analytically for a very general model that rkky theory holds in the limit of small exchange splitting in the ferromagnetic material . however , using a more general approach@xcite based on the existence of quantum well states@xcite , fundamental periods not predicted by rkky theory have been recently obtained for the particular case of a fcc lattice along the ( 110 ) direction @xcite . in that communication , within the single - band tight - binding model with nearest neighbour interaction , it is shown that in addition to the rkky predictions , a number of distinct periods is involved in the coupling . those new periods arise from non - negligible higher order terms which rkky , as a second order perturbative approach , does not account for . the resemblance of the fs considered with those of the noble metals suggests possible real effects in systems having these materials as spacers . moreover , the occurrence of non - rkky behaviour within the simplest single - band model indicates that it should also happen in the more elaborate multi - band case . in fact , non - magnetic transition metals with their multi - sheet fermi surfaces are suitable candidates for presenting non - rkky periods . furthermore , in most cases investigated so far , each oscillation period comes from distinct @xmath0-points of the 2-dimensional brillouin zone ( bz ) . even when multi - periodic oscillations occur , one value of @xmath1 yields only one period . the existence of more than one period for a single value of @xmath1 , which may arise in a number of cases , has not been thoroughly investigated and the combination of such periods may lead to new features of the oscillations . for the reasons above , a general understanding of the mechanisms determining all oscillation periods is needed . the purpose of this work is to generalize the method previously used to calculate the oscillation periods of the interlayer coupling . the relation between the periods and the spacer fs is clarified and in addition to the determination of the usual periods , the general conditions for occurrence of non - rkky behaviour are also established . furthermore , here we present a rather detailed set of results in order to illustrate better the different stages of the calculations . the sequence of the paper is as follows . in section [ general ] we present a general method for obtaining the oscillation periods of the interlayer coupling , where in addition to the determination of the usual periods , the existence of non - rkky behaviour is proved to be possible . similarly to previous theoretical approaches@xcite , an expression for the coupling in the asymptotic limit of large spacer thickness is derived . here however , the occurrence of multi - periodicity for a single value of @xmath1 becomes evident at the same time that the relation between the oscillation periods and the spacer fs is clarified . in section [ fcc110 ] , in order to illustrate the method , we look at the particular case of the fcc lattice along the ( 110 ) direction which unmistakably shows the relevance of terms beyond rkky approximation . finally , in the last section we conclude and present the general conditions for occurrence of non - rkky oscillations . based on the formalism introduced by dalbuquerque e castro _ et al._@xcite , where the coupling is expressed in terms of the one - electron green s functions of the multilayered system , we here present a general method for determining all the oscillation periods involved in the coupling . for the sake of simplicity , we consider a system consisting of two parallel magnetic planes embedded in an infinite non - magnetic material . we label the two ferromagnetic planes @xmath2 and @xmath3 , so that the number of non - magnetic spacer planes is @xmath4 . as far as the coupling as a function of the spacer thickness is concerned the number of magnetic planes does not affect the periods @xcite , influencing only phase and amplitude of the oscillations . since the periods are our main concern , the restriction to two single magnetic planes does not pose any limitations on the results here obtained . the expression for the coupling @xmath5 , namely the difference in the thermodynamical potential @xmath6 between the ferromagnetic and antiferromagnetic configurations , is given by@xcite @xmath7 where the function @xmath8 is given by @xmath9 in the equations above @xmath10 is the propagator between planes @xmath2 and @xmath3 for electrons with spin @xmath11 in the ferromagnetic configuration of the system , @xmath12 is the fermi function , @xmath13 is a diagonal matrix in orbital indices representing the strength of the local exchange potential in the ferromagnetic layers , and the summation over @xmath1 is restricted to the two - dimensional brillouin zone ( bz ) . it is worth stressing that all matrices above are in orbital indices . we assume for simplicity that the non - magnetic part of the potential in the ferromagnetic planes is identical to the spacer on - site potential . this is not an essential assumption and can be lifted without major effects on the results . in fact , it has been shown that the presence of a spin - independent square well does not influence the periods of the interlayer coupling , affecting only the phase and amplitude of the oscillation @xcite . as such , the exchange splitting of the magnetic material acts as two localized perturbations on the planes @xmath2 and @xmath3 and in this situation the propagators @xmath14 can be written as a function of the bulk spacer propagators @xmath15 and @xmath16 , _ i.e. _ , @xmath17 within the particular case of the single - band model the explicit dependence of @xmath18 upon the off - diagonal propagators @xmath15 and @xmath16 is , for each @xmath1 and @xmath19 , given by @xmath20 where @xmath21 is a general matrix element of the bulk spacer green s function and @xmath22 . an analogous expression can be obtained for @xmath23 . note that whereas eq.([ggeneral ] ) involves matrices in orbital indices , all terms in eq.([g ] ) are scalars . returning to the general case and for reasons of future comparison , it is worth presenting the rkky - limit expression for the bilinear coupling @xmath24@xcite . it is given by eq.([j ] ) with @xmath8 replaced by @xmath25 we recall that the expansion of @xmath26 in powers of @xmath27 , being @xmath28 the angle between the magnetization vectors in the magnetic planes , defines the bilinear and biquadratic couplings @xmath29 and @xmath30 , respectively @xcite . moreover , it is found that higher order terms than @xmath31 are negligible so that @xmath32 @xcite . it is clear from eqs.([fwq ] ) , ( [ ggeneral ] ) and ( [ frkky ] ) that in either case the spacer thickness dependence of the interlayer coupling is entirely determined by the off - diagonal propagators @xmath33 and @xmath16 . the investigation of the oscillation periods thus requires an analysis of @xmath16 as a function of @xmath3 . for fixed values of @xmath1 , the green s functions @xmath34 are equivalent to 1-dimensional propagators with @xmath0-dependent hoppings and in - plane energies . we recall that @xmath34 is a matrix over orbital indices and therefore a general matrix element @xmath35 represents the propagation of an electron with orbital @xmath36 from plane @xmath37 into an orbital @xmath38 at plane @xmath39 . it is given by @xmath40 where @xmath41 is the wave vector perpendicular to the layers , @xmath42 is the interplane distance , @xmath43 describes the bulk spacer band structure , being @xmath44 the band index , and @xmath45 is the projection of the eigenstate @xmath44 into the orbital @xmath36 for a given @xmath46 . the integral in eq.([integral ] ) can be evaluated for @xmath47 by simply extending @xmath46 to the complex plane and changing the contour integration from a straight line on the real axis to the boundaries of a semi - infinite rectangle in the upper - half plane whose base lies on the real axis between @xmath48 and @xmath49 . for the case @xmath50 , the extension is to a rectangle in the lower - half plane . in either case the integrand vanishes as @xmath51 and because @xmath52 and @xmath53 are equivalent wave vectors , the integrals along the vertical sides of the rectangle cancel each other , simplifying the problem to the calculation of the residues associated with the poles of the integrand inside the closed contour . the poles are in turn given by the values of @xmath46 satisfying the condition @xmath54 , where @xmath55 . the poles are clearly dependent on the band structure . due to the in - plane symmetry of the system , the general tight - binding hamiltonian of the homogeneous system is written as @xmath56 where @xmath57 are operators describing the electron hopping between planes @xmath58 and @xmath59 . the upper limit of the sum depends on the maximum number of planes @xmath60 connected through the electron hoppings as well as on the lattice structure . as we shall see , in the fcc lattice along the ( 110 ) direction for instance , even with only hoppings between nearest atoms being included , electrons may hop between nearest and next nearest planes . for a fixed value of @xmath61 , the wave vectors @xmath46 satisfying the equation @xmath62 are the solutions of a polynomial equation in @xmath63 , _ i.e. _ , @xmath64 where @xmath65 is a function of the @xmath0-dependent hoppings and in - plane energies . note that the solutions of the equation above are exponentials @xmath63 and not wave vectors @xmath46 themselves . we label the poles @xmath66 , where @xmath67 numbers the poles of a given band index @xmath44 . since @xmath68 poles appear in pairs and may be real or complex . we see from a simple taylor expansion of @xmath69 around the real poles that , for a given pair , which pole contributes to the integral depends on the sign of the derivative @xmath70 . for a given band @xmath44 which crosses the energy @xmath71 twice , yielding two pairs of real poles , the contributory poles @xmath72 and @xmath73 have different signs . in other words , when @xmath72 is in the range [ @xmath74 , @xmath73 is in the range [ @xmath75 or vice versa . it is interesting to look at the physical significance of the real poles . being obtained from the band structure @xmath43 , the poles @xmath66 are just the wave vectors with which electrons of energy @xmath19 having in - plane energies @xmath76 and hoppings @xmath77 propagate across the spacer . thus , for @xmath78 , where @xmath79 is the fermi energy , the real poles are just the perpendicular coordinates of the fs for a fixed value of @xmath0 . as discussed later , complex poles contribute significantly only in the case of thin spacers . by adding the residues associated with the contributory poles we obtain an analytical expression for @xmath21 , given by @xmath80 where the matrix elements of the matrix @xmath81 are @xmath82 _ { q_\perp = q_j^s}\right\}^{-1}\,\ , . \label{aa}\ ] ] the coefficients @xmath83 depend neither on @xmath37 nor on @xmath39 and the only dependence on the distance between the planes is inside the argument of the exponentials . in addition , the exponentials are independent of the orbital indices , which means that all matrix elements oscillate with the same periods . note that in calculating eq.([aa ] ) we have assumed that the coefficients @xmath84 are not ill - behaved , that is , they have neither singularities nor branch points inside the integration contour . in this way , we just need to evaluate the coefficients at the poles @xmath66 . nevertheless , in a general case the coefficients @xmath84 are not well behaved . although they do not have singularities like in the denominator of eq.([integral ] ) , branch points and cuts do exist . the contributions from the branch points must be taken into account otherwise a simple summation of the residues will not give the correct result of the integration . this is a very important point that , if overlooked , may lead to erroneous results . in fact , in a paper by bruno @xcite where the coupling is expressed in terms of transmission and reflection coefficients of electrons across quantum barriers , a similar integral to that in eq.([integral ] ) arises . in that communication no particular attention was given to the analytic properties of the coefficients @xmath85 and the result above was erroneously concluded to be exact for arbitrary values of @xmath37 and @xmath39 . it is important to stress that eqs.([glm ] ) and ( [ aa ] ) are not correct in general . the equations above become exact for all @xmath37 and @xmath39 only within the single - band model , where there are no coefficients introducing branch points . this is clearly seen in fig.[g05 ] where the real and imaginary parts of the function @xmath86 of a linear chain with hoppings up to four nearest neighbours within the single - band model are displayed . the parameters used were @xmath87 , @xmath88 , @xmath89 , @xmath90 and @xmath91 , where @xmath92 is the in - plane energy and @xmath93 are the hoppings to the @xmath37-th neighbour planes . the analytical results following eq.([glm ] ) , represented by the lines , are in absolute agreement with the brute - force calculation of eq.([integral ] ) , represented by the symbols . in the general case there are also contributions to the @xmath46 integral arising from branch cuts associated with the coefficients @xmath84 . since in general these cuts do not intersect the real axis , the factor @xmath94 in the integrand of eq.([integral ] ) ensures that the contribution is negligible for large @xmath95 . thus , failure to treat the branch cuts correctly does not introduce errors in the asymptotic limit of large @xmath95 . it is clear that @xmath21 is an oscillatory function of the distance between the planes and it oscillates with different periods , each one associated with its respective wave vector @xmath66 . in addition , the combination of exponentials allows a quasi - periodic behaviour of @xmath21 because , in general , the wave vectors @xmath66 are incommensurate . this fact , as we shall see , may have a striking effect on the oscillatory coupling . another point to be mentioned is the possibility of evaluating @xmath21 in eq.([glm ] ) for non - integer values of @xmath37 and @xmath39 , which allows the evaluation of the coupling for continuous values of spacer thickness . we now mention the case of complex poles . since they have non - zero imaginary parts , their contribution for large separation between the planes is again negligible . thus only real wave vectors contribute to the oscillations for large @xmath96 . bearing in mind that the principal contribution to the coupling comes from the energy region around the fermi level@xcite , it is clear that at least one real wave vector will be involved in the expression for @xmath97 . thus the sum over the band index @xmath44 in eq.([glm ] ) must be only over the bands which cross the fermi level . having investigated the function @xmath21 as a function of the distance between the planes , we recall that @xmath8 is expected to oscillate with the same periods . the obvious way of representing the function @xmath8 is through a simple fourier expansion , which nevertheless can not be done due to the quasi - periodicity of @xmath34 . in other words , because the wave vectors @xmath66 are in general incommensurate , @xmath33 and @xmath16 may have an overall non - periodic behaviour . to overcome this problem we make use of the multiple fourier series , recently used to deal with the problem of magnetic thickness dependence of the interlayer coupling@xcite . the procedure consists in replacing the spacer thickness @xmath3 multiplying the wave vectors @xmath66 in each exponential with fictitious spacer thicknesses @xmath98 , such that eq.([glm ] ) becomes @xmath99 the new function @xmath100 is now periodic in each variable @xmath98 separately , and so is its corresponding @xmath101 . bearing in mind that @xmath100 does not correspond to the physically real case , which is reproduced by making @xmath102 for every @xmath67 and @xmath44 , we evaluate the fourier series for each variable @xmath98 separately , eventually making them all equal @xmath3 . @xmath8 then becomes @xmath103 where @xmath104 are integers , @xmath105 are the fourier coefficients given by @xmath106 and @xmath107 are the periods associated with the wave vectors @xmath66 . the number of indices involved in the fourier coefficients is the same as the number of poles because each @xmath66 has been associated with a fictitious spacer thickness @xmath98 . in spite of the somewhat congested notation , in practice there are only a few indices in the coefficients due to the restricted number of real solutions crossing the fermi level . in fact , within the single - band model the calculation of the fourier coefficients can be extremely simplified , as shown in appendix [ fourier ] . the analysis of such a simplified model , although primarily illustrative , reflects some properties common to the more general case . two of these properties are @xmath108 and @xmath109 for odd values of @xmath110 . we now focus on the rkky - limit expression for the coupling . this is the limit of small ferromagnetic exchange splitting . it is clear from eq.([frkky ] ) that unlike the coupling @xmath111 , the rkky expression involves neither logarithmic functions nor combinations of @xmath112 and @xmath113 . in fact , @xmath114 simply contains a single product of the propagators @xmath16 and @xmath15 . bearing eq.([frkky ] ) in mind we define @xmath115/2 i$ ] and from the expression for an arbitrary element @xmath21 in eq.([glm ] ) we write @xmath116 clearly , the above equation shows that the integrand @xmath114 can be manipulated and put in a fourier expansion form , similarly to @xmath8 in eq.([f ] ) . however , the main difference is that in this case the number of terms involved in the expansion is limited , related to the maximum number of poles in eq.([glm ] ) , whereas the sum in eq.([f ] ) runs over an infinite number of terms . such difference is due to the fact that the rkky - limit expression results from a perturbative approach corresponding to the lowest order term of an expansion . all the terms included in @xmath114 are also included in the full expression @xmath8 . the higher order terms correspond to oscillations that are either harmonics of the fundamental rkky behaviour or effectively new oscillation periods . this point confirms the rkky theory as a limit of the more general approach used here . we also point out that in this case , the argument of the exponentials has the form @xmath117 , involving only two wave vectors and with the constraint @xmath118 , being @xmath119 and @xmath120 . this is a consequence of @xmath114 being the product of only two propagators , namely , @xmath16 and @xmath15 . in the more general case of @xmath8 , the infinite possibilities of combining the mentioned propagators does not pose such a restriction . at this point we raise the question of whether the rkky periods of the interlayer coupling remain valid as the exchange splitting of the ferromagnetic layers , associated with @xmath13 , is increased . the question can be reformulated as whether the higher order terms in @xmath8 , apart from the rkky harmonics , do contribute to the coupling . the answer to this question is in the evaluation of the integrals in eq.([j ] ) . having written @xmath8 ( and @xmath114 ) in terms of a fourier expansion , we can now make use of the stationary phase method@xcite , which yields an analytical expression for the coupling , asymptotically exact for thick spacers ( @xmath121 ) . inserting eq.([f ] ) into ( [ j ] ) , we have @xmath122 we recall that the coefficients @xmath105 and the wave vectors @xmath66 depend on both integration variables @xmath1 and @xmath19 . the stationary phase method shows that in the asymptotic limit of large spacer thickness the only contributions to the integral over @xmath1 arise from the neihbourhood of points which make the arguments of the exponentials stationary . we call these points @xmath123 , where the superscript @xmath124 labels the different vectors . furthermore , the integral over @xmath19 is also non - negligible only around the fermi energy @xmath79 . hence , the wave vectors that effectively contribute to the coupling must satisfy the following condition , @xmath125 \ , = 0 \,\ , , \label{grad}\ ] ] where @xmath126 is the two - dimensional gradient in @xmath0-space . in other words , the equation above gives the necessary conditions for a constructive interference between the electrons across the spacer . the otherwise destructive interference does not contribute to the coupling . eq.([grad ] ) simply selects the effective periods with which the coupling oscillates . note that for @xmath78 , each wave vector @xmath127 represents a surface in the reciprocal space and when put together , the surfaces @xmath66 reproduce the spacer fermi surface . following this argument and bearing in mind that non - real values of @xmath66 strongly damp the oscillations , we stress that only real solutions must be taken into account . the possible stationary solutions of eq.([grad ] ) reflect the lattice structure and are usually located at points of high symmetry , even though they may occur at general points of the 2-dimensional bz . in such limit we obtain an analytical expression for the coupling given by@xcite @xmath128 @xmath129 } \,\ , , \label{tspa}\ ] ] where @xmath130 , @xmath131 , and @xmath132 . here we neglect the energy dependence of the fourier coefficients which in some cases @xcite can be important . the factor @xmath133 takes the value @xmath134 for a minimum , @xmath135 for a maximum and @xmath136 for a saddle point of the surface @xmath137 . the sum over @xmath124 regards all different stationary solutions and the inner sum over @xmath138 runs over the values allowed by the stationary phase condition . the coupling oscillates with periods @xmath139 and their weights depend on geometric factors as well as on the fourier coefficients . an important issue is the relationship between the fermi surface and the wave vectors that effectively contribute to the coupling . the lowest order combinations amongst the possible wave vectors @xmath127 bear a direct correspondence with the fs . indeed , this correspondence agrees with the rkky geometrical picture for selecting the oscillation periods of the interlayer coupling . however , in the general case of a higher order combination of wave vectors the rrky relation no longer holds , and the effective wave vectors are not directly related to the spacer fs . this is a very important result which shows that the geometrical criteria used for selecting the oscillation periods of the interlayer coupling are incomplete . even though the wave vectors @xmath66 are directly related to the fs , the general wave vectors @xmath140 are not . furthermore , note that if @xmath141 has only one oscillation period , the full expansion in eq . ( [ jspa ] ) contains merely harmonics of the fundamental rkky periods . in other words , if the fs is cut in a single point by the line drawn from the origin perpendicular to the layers for a fixed @xmath1 , the rkky periods and its harmonics are the only ones present in the coupling . in fact , it explains the agreement between the periods predicted by the rkky and qwt in the models previously mentioned@xcite . for a multi - periodic @xmath142 though , where the fs is intersected in more than one point corresponding to incommensurate wave vectors , the higher order combinations of the wave vectors correspond neither to a rkky period nor to any of its harmonics , but to _ fundamentally new periods_. multi - periodicity of the off - diagonal propagator is a necessary but not sufficient condition . the argument above simply shows that non - rkky behaviour is possible in presence of multi - periodic @xmath143 . to find out whether it really occurs one has to solve eq.([grad ] ) and check whether there are real wave vectors @xmath66 for values of @xmath144, ,@xmath145 other than the ones obtained by the rkky theory . in summary , the determination of the oscillation periods results from the analysis of the bulk spacer off - diagonal propagators where the relevant real wave vectors , under the stationary phase condition , are obtained from the band structure . the relation between the spacer fs and the oscillation periods is the same as the geometrical rkky picture for the lowest order terms . nevertheless , for higher order contributions a direct correspondence with the fs is not evident and the stationary phase condition displayed in eq.([grad ] ) determines the periods with which the coupling oscillates . we have shown that oscillation periods other than the ones predicted by the rkky theory may arise . to illustrate the method and unambiguously prove the existence and relevance of periods other than rkky ones , we next evaluate the particular case of the fcc lattice along the ( 110 ) direction within the single - band tight - binding model . having presented the general method for determining the periods of the interlayer coupling , here we apply it to the particular case of a fcc(110 ) system within the single - band nearest - neighbour tight - binding model , which unmistakably shows the existence of non - rkky periods . moreover , in the light of the qwt this system is interesting because of the difficulty in determining analytically the energies of the resonances and size quantized states in such a structure@xcite . for a single band the notation introduced in the previous section becomes simpler due to the omission of the band index @xmath44 . in addition , all the matrices over orbital indices become scalars . we label the nearest neighbour hopping @xmath146 and with no loss of generality assume the on - site potential @xmath147 . the bulk density of states of the fcc lattice is displayed in fig.[dos ] . it is worth pointing out that there are different contributions to the interlayer coupling , depending on the position of the fermi energy inside the band . three distinct regions are observed and their boundaries are shown by the vertical dashed lines in fig.[dos ] . in addition to the traditional polyhedral bz , we make use of an auxiliary prismatic bz due to the 2-dimensional symmetry of the multilayers@xcite . both zones are equivalent and enclose the same volume in reciprocal space . the energy boundaries indicated correspond to fermi energies such that the fs just touches the two distinct brillouin zones . more specifically , for @xmath148 the corresponding fs touches the prismatic bz , whereas @xmath149 is the exact fermi energy where the fs opens the necks which are familiar from the noble metals @xcite . as previously mentioned , even within the nearest neighbour tight - binding model , in this case electrons are capable of hopping between nearest and next nearest planes . the bulk band structure is then given by @xmath150 we assume the @xmath151 direction perpendicular to the layers . the 2-dimensional bz over which the stationary solutions must be searched for is a rectangle defined by @xmath152 and @xmath153 , where @xmath154 and @xmath155 are the components of @xmath1 . the explicit dependence of @xmath156 on @xmath1 is given by @xmath157 , @xmath158 and @xmath159 . following the steps of the previous section , @xmath160 gives the poles which yield the periods of @xmath16 . eq.([e ] ) can be rewritten in terms of a second degree polynomium in @xmath161 whose solutions @xmath162 and @xmath163 are given by @xmath164 where @xmath165 and @xmath67 is either 1 or 2 . nevertheless , the poles @xmath162 and @xmath163 are in the @xmath46-complex plane . by mapping the relationship between the two complex functions we can uniquely relate one value of @xmath46 inside the integration contour to the corresponding value of @xmath161 and vice versa , as shown in fig.[map ] . the equation above shows that @xmath166 and @xmath167 have opposite imaginary parts and therefore , following the maps in fig.[map ] , @xmath162 and @xmath163 have opposite real parts . this confirms what has been said in the previous section concerning the position of two different wave vectors for a given band @xmath44 . in addition , looking at eq.([cosq ] ) with @xmath78 it is clear that non - rkky behaviour can not arise in the bottom region of the band ( @xmath168 ) . this is due to the fact that in such a region @xmath162 and @xmath163 can not both be real . in fact , the corresponding fs in this region is spherical - like . however , in the range @xmath169 , corresponding to the top regions inside the band , both @xmath162 and @xmath163 may be real indicating that non - rkky periods are possible . having obtained the poles , from eqs.([glm ] ) and ( [ aa ] ) we have @xmath170 where @xmath171\sqrt{1 - \cos^2(q_j d)}\right\}^{-1 } \,\,.\ ] ] in the present case @xmath172 . eq.([g0n2 ] ) shows that for values of @xmath1 and @xmath19 , for which @xmath162 and @xmath163 are real , @xmath16 oscillates with the superposition of two periods , @xmath173 and @xmath174 , which are in general incommensurate . in those cases @xmath16 exhibits a quasi - periodic dependence on @xmath3 . because there are only two poles in this case , the function @xmath8 , following eq.([f ] ) , becomes @xmath175 where the fourier coefficients are given by eq.([coefs ] ) . within the single - band model , the fourier coefficients can be calculated in an alternative way , shown in appendix [ fourier ] . following the appendix , we recall that the lowest order fourier coefficients correspond to the rkky - limit @xmath114 . they are @xmath176 where @xmath177 , @xmath178 and @xmath179 . more specifically , @xmath180 , @xmath181 , @xmath182 , and the property @xmath183 define all the six coefficients present in the rkky expression , where @xmath184 and @xmath185 are defined after eq.([g ] ) . beyond the lowest order coefficients are the terms expected to contribute to the non - rkky behaviour . it is then illustrative to present one of these terms , namely @xmath186 , @xmath187 \,\,.\ ] ] the stationary phase condition in eq.([grad ] ) is rewritten as @xmath188 the condition above imposes that both gradients must be colinear and moreover , the ratio between their moduli must be a rational number . the determination of the stationary points @xmath189 yielding the periods @xmath190 is rather tedious but straightforward . as previously mentioned , the solutions are often located at highly symmetric points of the bz but more general solutions may occur . nevertheless , because the rkky periods as well as the most important higher order terms come only from those high - symmetry points , we shall firstly focus on these solutions . across the whole energy band , four stationary points are found on the corners of the 2-dimensional irreducible bz . they are @xmath191 , @xmath192 , @xmath193 and @xmath194 . in the first region @xmath195 is the only period , which is actually present along the entire band . the graphical relation between the spacer fs and the oscillation period is displayed in fig.[fsef_3 ] , where the fs for an arbitrary energy ( @xmath196 ) is shown together with the wave vector @xmath197 and the cross section of the prismatic bz for @xmath198 . the cross section is defined by @xmath152 and @xmath199 . note that @xmath200 corresponds to the extremal wave vector spanning the fs , in according to the rkky picture . clearly , higher order periods @xmath201 for @xmath202 are merely harmonics of the fundamental @xmath203 . in the second region an additional period @xmath204 arises . similarly , the vector @xmath205 yields the fundamental period and possible higher combinations @xmath206 add no new terms but harmonics . fig.[fsef_1 ] shows a cross section of the fs ( @xmath207 ) for @xmath208 . the wave vectors @xmath209 and @xmath210 are highlighted and it is clear that @xmath211 satisfies the rkky criteria . at first glance one could think , accordingly to fig.[fsef_1 ] , that @xmath212 and @xmath213 are also contributory wave vectors for the same reasons as in the first region . nevertheless , by looking at the @xmath155-dependence of the fs we find that those vectors are not solutions of eq.([grad2 ] ) . the most interesting case occurs in the top region , where the fs has necks resembling those of the noble metals . in addition to @xmath195 at @xmath214 , other stationary solutions occur at @xmath215 and @xmath216 . at @xmath217 , analogously to the second region , the solution comes from @xmath218 and does not introduce new periods . at @xmath219 though , both gradients in eq.([grad2 ] ) vanish simultaneously allowing any values of @xmath220 and @xmath221 to satisfy the stationary phase condition . therefore , the coupling oscillates with as many periods @xmath222 as there are non - zero coefficients @xmath223 . the lowest order @xmath224 , @xmath225 and @xmath226 correspond to the rkky periods and the geometrical relation between the wave vectors @xmath227 , @xmath228 and the fs in fig.[fsef1 ] confirms this point . for the present model we find that @xmath229 . thus , the oscillation periods @xmath224 and @xmath225 can not be distinguished just by looking at discrete integer values of @xmath3 . in addition , @xmath230 , which leads the third energy region to contain three non - equivalent rkky periods . we recall that in fcc cu @xmath79 lies in the third energy region with fs necks . a quantitative description of the oscillation periods for cu can be obtained within the present framework by going beyond nearest neighbours and using the tight - binding parameters of halse@xcite . then @xmath226 and @xmath231 become distinct periods and , by taking into account the equivalence of @xmath224 and @xmath225 for a discrete lattice , we find exactly the four rkky periods of bruno and chappert @xcite . however , because in general the wave vectors @xmath227 and @xmath232 are incommensurate , higher order periods are not simple harmonics of the rkky ones , unlike in the two lower regions . in fact , the period @xmath233 does not correspond to any harmonics of the fundamental rkky periods . furthermore , even though the wave vectors @xmath227 and @xmath232 separately have clear correspondences with the fs , as shown in fig.[fsef1 ] , the wave vector @xmath234 has not . from this argument it becomes clear that higher order periods do not satisfy the geometrical picture of the rkky theory , having instead an indirect relation between the wave vectors and the fs . fig.[periods ] summarizes the results obtained here by displaying the oscillation periods as a function of the fermi energy @xmath79 . the full lines labeled from 1 to 5 are the fundamental rkky periods @xmath195 , @xmath204 , @xmath235 , @xmath225 and @xmath226 , respectively , and the dot - dashed line in the top region represents only one of the non - rkky periods , namely @xmath233 . recalling the relation @xmath236 , we see that the oscillation periods @xmath224 and @xmath225 can not be distinguished just by looking at discrete integer values of @xmath3 . still focusing on the top energy region , fig.[result ] exhibits the bilinear coupling @xmath32 at temperature @xmath237 as a function of the spacer thickness for @xmath238 and @xmath239 , where @xmath240 is the spacer band - width . this value of @xmath79 is chosen so that @xmath241 is well separated from the rkky periods , as may be seen in fig.[periods ] . the full line in fig.[result ] corresponds to the result obtained from eqs.([j ] ) and ( [ fwq ] ) , whereas the dashed line is the rkky approximation , following eqs.([j ] ) and ( [ frkky ] ) , scaled down by a factor of 8 for reasons of comparison . despite the agreement in the dominant long period , the difference in the fine structure of the oscillations reflects contributions beyond the fundamental rkky periods . the evaluation of the stationary phase approximation in this case unmistakably proves the existence and relevance of the non - rkky periods . fig.[coefsfig ] displays the absolute value of the ratio between some fourier coefficients @xmath223 and @xmath242 , the largest coefficient of a fundamental rkky period , as a function of @xmath13 . the full line corresponds to the fourier coefficient associated with the new period @xmath233 whereas the dashed and dot - dashed lines correspond to @xmath243 and @xmath244 , respectively , which are simple harmonics of the fundamental rkky periods . as expected for very small exchange splittings , the magnitudes of higher order coefficients relative to that of the fundamental rkky ones are negligible@xcite , indicating that non - rkky behaviour becomes more important as the exchange splitting is increased . an interesting consequence is that , even though the periods are determined from the spacer fs , their relevances depend on the nature of the magnetic materials involved . note that for @xmath245 the ratio @xmath246 . this indicates that for such a value of @xmath13 , which is exactly the same as in fig . [ result ] , the fourier coefficients are of the same order of magnitude ; the only reason why both periods do not have comparable weights is due to geometrical factors determined by the curvatures of the respective surfaces . in fact , the curvature of the surface @xmath247 vanishes at @xmath215 , making it impossible to evaluate the stationary phase approximation for the period @xmath226 even though it undoubtedly indicates a strong contribution . taking into account the geometrical curvatures when possible , we evaluate the stationary phase approximation and show in fig.[spa ] the separate contribution of some periods to the exchange coupling . the full line is the new period @xmath233 whereas the dashed line and the the dot - dashed line represent @xmath195 and @xmath224 , respectively . clearly , @xmath233 is as important as the rkky periods , except for the long dominant one . it is worth pointing out that the rate of decay of the oscillation amplitude associated with the new periods is exactly the same as for the lowest order ones , falling off as @xmath248 for @xmath249 . concerning the actual amplitude of the oscillations , two points should be highlighted . the strength of all periods depends rather critically on the matching of the bands in real materials , which is not well reproduced by our single - orbital model , and the omission of the energy dependence of the fourier coefficients could be serious for cu - based systems @xcite . the solutions treated above all come from the high - symmetry points @xmath250 , @xmath251 , @xmath219 and @xmath217 . note that fundamentally new periods arise but always in the presence of rkky ones . they result from crossed mixtures of the fundamental wave vectors , which in turn yield the rkky periods . nevertheless in general , as previously stated , solutions of eq.([grad2 ] ) away from high - symmetry points are found and we point out that their contributions are not necessarily associated with the existence of rkky periods . in fact , it can be shown in the second energy region that energy - dependent solutions exist along the line @xmath252 with corresponding wave vectors @xmath253 and @xmath254 which do not yield any possible rkky period . this indicates there are colinear gradients along that line satisfying eq.([grad2 ] ) for different values of @xmath220 and @xmath221 even though no rkky stationary solutions are obtained . these solutions are found to be far less significant in their contribution to the exchange coupling than the highly symmetric ones . in summary , we have shown that the rkky theory is not capable of predicting all the oscillation periods of the interlayer coupling , as generally believed . the actual relation between the periods and the spacer fs is more subtle than the rkky geometrical picture . even though the lowest order periods indeed correspond to the rkky predictions , non - negligible higher order terms arise and do not bear any direct correspondence to the fs . instead , they correspond to indirect combinations of fundamental wave vectors , which in turn , may or may not yield rkky periods . the appearance of non - rkky oscillations clearly requires the multi - periodicity of the bulk spacer off - diagonal propagator as a function of the spacer thickness , although this is not a sufficient condition . in addition , real solutions of the stationary phase condition must exist in order to have a constructive interference between the electrons across the system . the otherwise destructive interference damps the oscillation not contributing to the coupling . finally , if such behaviour can be found even within the single - band model , for which the fs is simple and has a single sheet , we may expect the occurrence of non - rkky periods in systems with more elaborate fs . surfaces with more than one sheet , such as non - magnetic transition metals for instance , are possible candidates for presenting such periods . in those cases , the interpretation of the results in terms of the rkky theory may be misleading . alternatively , within the single - band model , the fourier coefficients @xmath176 in section [ fcc110 ] can be calculated in the following way . bearing in mind the equivalence between @xmath172 , eq.([g ] ) can be concisely written as @xmath255 and analogously , @xmath256 the product of eqs.([gona ] ) and ( [ gnoa ] ) is @xmath257 writing the logarithmic function as a series , we have from eq.([fwq ] ) @xmath258^\ell \right\ } \,\ , , \label{fprime}\ ] ] where @xmath259 . we recall that the summations above have indices with different limits . whereas the external summation index @xmath37 runs over the integers starting from @xmath260 onwards , the inner indices @xmath67 and @xmath261 have zero as their initial values . eq.([fprime ] ) can be manipulated and rewritten as @xmath262 for a fixed @xmath37 there are twice as many inner summations , running from @xmath263 and @xmath264 as far as @xmath265 and @xmath266 . we define the integer @xmath267 as the exponent of the function @xmath268 in eq.([fprimea ] ) , i.e. , @xmath269 . note that @xmath267 runs over the positive integers greater than zero . squaring @xmath16 in eq.([g0n2 ] ) and inserting it into eq.([fprimea ] ) we obtain the function @xmath8 in terms of a fourier expansion . clearly , the exponential @xmath270 arises from terms of @xmath271 such that @xmath272 . the number of terms in turn depends on the number of possible combinations between the implicit indices @xmath67 , @xmath261 and @xmath37 , and the larger @xmath267 is , the more combinations are possible . for @xmath273 for instance , the only possible combination of the indices is @xmath274 and @xmath260 . the terms coming from this contribution correspond to the rkky - limit expression in eq.([frkky ] ) , which indicates @xmath114 as a subset of the full expression @xmath8 and confirms the rkky theory as a limit of the more general approach used here @xcite . for @xmath275 , more combinations are possible . in addition to @xmath276 and @xmath277 , four other combinations come from @xmath260 and @xmath278 . note that the lowest order coefficients are easily calculated through this alternative method but as the exponent @xmath267 is increased , more terms arise and increase the difficulty . nevertheless , because we are searching for periods other than the rkky ones , the lowest order terms beyond the rkky - limit are expected to be the important corrections to the coupling . finally , from the discussion above , two properties of the fourier coefficients are straightforwardly obtained . they are @xmath108 and @xmath109 unless @xmath110 is even . these properties are also valid in the general case of section [ general ] . d. m. edwards , j. mathon , r. b. muniz and m. s. phan , phys . rev . lett . * 67 * , 493 ( 1991 ) . d. m. edwards , j. mathon , r. b. muniz and m. s. phan , j. phys . : condens . matter * 3 * 4941 ( 1991 ) j. mathon , murielle villeret and d. m. edwards , j. phys . : condens . matter * 4 * , 9873 ( 1992 ) p. bruno and c. chappert , phys . * 67 * , 1602 ( 1991 ) p. bruno and c. chappert , phys . b * 46 * , 261 ( 1992 ) j. dalbuquerque e castro , m. s. ferreira and r. b. muniz , phys . rev . b * 49 * , 16062 ( 1994 ) m. s. ferreira , j. dalbuquerque e castro , d. m. edwards and j. mathon , j.m.m.m . * 154 * , l1 ( 1996 ) j. dalbuquerque e castro , j. mathon , murielle villeret and d. m. edwards , phys . rev . b * 51 * , 12876 ( 1995 ) m. s. ferreira , j. dalbuquerque e castro , r. b. muniz and d. m. edwards , j. phys . : condens . matter * 6 * , l619 ( 1994 ) d. m. edwards , j. mathon and r. b. muniz , phys . b * 50 * 16066 ( 1994 ) b. heinrich _ et al_. , phys . b * 44 * , 9348 ( 1991 ) a. v. kozlov and w. a. harrison , phys . rev . b * 48 * 12334 ( 1993 ) m. r. halse , trans . london a * 265 * , 507 ( 1969 ) n. w. ashcroft and n. d. mermin , solid state physics ( 1976 ) d. m. edwards , a.m. robinson and j. mathon , j.m.m.m . * 140 - 144 * , 157 ( 1995 ) j. dalbuquerque e castro , j. mathon , murielle villeret and a. umerski , phys . b * 53 * , 13306 ( 1996 ) p. bruno , phys . rev . b * 52 * , 411 ( 1995 )

single particle tracking microscopy provides the position time series @xmath0 of individual particle trajectories in a medium @xcite . the information garnered by single particle tracking yields insights into the mechanisms and forces , that drive or constrain the motion of the particle . an early example of systematic single particle tracking is given by the work of jean perrin on diffusive motion @xcite . due to the relatively short individual trajectories , perrin used an ensemble average over many trajectories to obtain meaningful statistics . a few years later , nordlund conceived a method to record much longer time series @xcite , allowing him to evaluate individual trajectories in terms of the time average and thus to avoid averages over not perfectly identical particles . today , single particle tracking has become a standard tool to characterise the microscopic rheological properties of a medium @xcite , or to probe the active motion of biomolecular motors @xcite . particularly in biological cells and complex fluids single particle trajectory methods have become instrumental in uncovering deviations from normal brownian motion of passively moving particles . ) as function of the lag time @xmath1 . left : normal diffusion growing like @xmath2 ( 1 ) , restricted ( confined ) diffusion with a turnover from @xmath3 to @xmath4 ( 2 ) , drift diffusion @xmath5 ( 3 ) . right : ergodic subdiffusion @xmath6 ( 1 ) , restricted ergodic subdiffusion turning over from @xmath6 to @xmath4 ( 2 ) , non - ergodic subdiffusion @xmath3 ( 1 ) , restricted non - ergodic subdiffusion turning over from @xmath3 to @xmath7 ( 2 ) , superdiffusion @xmath8 ( 3 ) . here , @xmath9 . note the double - logarithmic scale.,width=332 ] classical diffusion patterns are sketched in the left panel of fig . [ diffusion_modes ] . accordingly , one may observe free diffusion , leading to a linear growth with time of the second moment [ line 1 in fig . [ diffusion_modes ] ] . brownian motion may also be restricted ( corralled , confined ) . confinement in a cell , for instance , could be due to the cell walls . in that case the second moment initially grows linearly with time and eventually saturates to its thermal value equalling the second moment of the corresponding boltzmann distribution [ line 2 ] . in the presence of a drift the second moment grows with the square of time [ line 3 ] . such results are typical for simple fluids . in more complex environments , different patterns may be observed , as displayed on the right of fig . [ diffusion_modes ] . here , subdiffusion may occur , for which the second moment grows slower than linearly with time [ line 1 ] . restricted subdiffusion would depart from this behaviour to reach a plateau [ line 2 ] . driven motion may lead to a superdiffusive power - law form of the second moment with an exponent between 1 and 2 [ line 3 ] . however , as we demonstrate below , subdiffusion may also be non - ergodic , and the associated time averaged second moment may grow _ linearly _ with time [ line 1 ] . similarly strange behaviour may be observed for restricted non - ergodic subdiffusion , which exhibits a _ power - law _ growth , not a saturation to a plateau [ line 2 ] . non - ergodic processes come along with a significant scatter between individual trajectories . this is an effect of the ageing nature of the process that persists for long measurement times . in the following we discuss in detail the behaviour of passive subdiffusive motion in terms of time and ensemble averages and address the peculiarities , that may arise for non - ergodic systems . non - ergodic behaviour of the above sense is indeed observed experimentally . [ christine ] shows the time averaged mean squared displacement for lipid granules in a living fission yeast cell . the motion is recorded by indirect tracking in an optical tweezers setup . initially the granule is located in the bottom of the laser trap potential such that the granule moves freely . eventually the granule ventures away from the centre of the trap and experiences the hookean trap force . as demonstrated in a detailed analysis the granule motion indeed exhibits weak ergodicity breaking , giving rise to the characteristic turnover from an initially linear scaling @xmath2 with the lag time @xmath1 , to the power - law regime @xmath10 @xcite . moreover a pronounced trajectory - to - trajectory scatter is observed , again typical for systems with weak ergodicity breaking . the initially linear scaling @xmath11 turns over to the power - law regime @xmath12 , induced by the restoring force exerted on the granules by the laser trap @xcite . note also the characteristic scatter between individual trajectories . inset : average behaviour of the shown trajectories . in both graphs the unit of @xmath13 is volt@xmath14 , i.e. , the direct output voltage of the quadrant photodiode . the voltage is directly proportional to the distance from the centre of the laser trap.,width=332 ] free diffusion is typically quantified in terms of the second moment . the mean squared displacement @xmath15 is obtained as the spatial average over the probability density function @xmath16 to find the particle at position @xmath17 at time @xmath18 . the quantity ( [ eamsd ] ) therefore corresponds to the ensemble averaged second moment of the particle position , denoted by angular brackets , @xmath19 . in particular , the time @xmath18 enters into eq . ( [ eamsd ] ) only as a parameter . conversely , single particle trajectories @xmath0 are usually evaluated in terms of the time averaged mean squared displacement defined as @xmath20 where we use an overline @xmath21 to symbolise the time average . here @xmath1 is the so - called lag time constituting a time window swept along the time series , and @xmath22 is the overall measurement time . the time averaged mean squared displacement thus compares the particle positions along the trajectory as separated by the time difference @xmath1 . in an ergodic system the time average of a certain quantity obtained from sufficiently long time series is equal to the corresponding ensemble mean @xcite . for instance , for the mean squared displacement ergodicity would imply @xmath23 brownian motion is ergodic , as well as certain stationary processes leading to anomalous diffusion , such as fractional brownian motion considered below . there exist , however , non - ergodic processes , which are intimately connected to ageing properties . in what follows we discuss two prominent models for anomalous diffusion , the stationary fractional brownian motion and the ageing continuous time random walk , and analyse in detail the features of the associated time averaged mean squared displacement . generally , anomalous diffusion denotes deviations from the classical linear dependence of the mean squared displacement , @xmath24 . such anomalies include ultraslow diffusion of the form @xmath25 @xcite . in contrast , anomalous diffusion processes may become faster than ballistic , for instance for systems with correlated jump lengths or in systems governed by generalised langevin equations @xcite . here we are interested in anomalous diffusion with power - law dependence on time @xcite , @xmath26 for which the anomalous diffusion exponent belongs to the subdiffusive range @xmath9 , such that the limit @xmath27 corresponds to brownian motion . the proportionality factor @xmath28 in eq . ( [ aneamsd ] ) is the anomalous diffusion coefficient of physical dimension @xmath29 . the embedding spatial dimension is @xmath30 , and eq . ( [ aneamsd ] ) includes the complete gamma function @xmath31 . subdiffusion of the form ( [ aneamsd ] ) with @xmath9 occurs in the following biologically relevant systems . fluorescently labelled mrna in _ e.coli _ bacteria cells was observed to follow @xmath32 with @xmath33 @xcite . this result is consistent with more recent findings according to which free rna tracers in living cells exhibit @xmath34 , while dna loci show @xmath35 @xcite . telomeres in the nucleus of mammalian cells were reported to follow anomalous diffusion with @xmath36 at shorter times and value @xmath37 at intermediate times , before a turnover to normal diffusion occurs @xcite . also larger tracer particles show anomalous diffusion , such as adeno - associated viruses of radius @xmath38 nm in a cell with @xmath39 @xcite and endogenous lipid granules , of typical size of few hundred nm with @xmath40 @xcite . it should be noted that this subdiffusion observed by single particle tracking microscopy is consistent with results from other techniques , such as fluorescence correlation spectroscopy @xcite or dynamic light scattering @xcite . subdiffusion of biopolymers larger than some 10 kd in living cells is due to molecular crowding , the excluded volume effect in the superdense cellular environment @xcite . larger tracer particles also experience subdiffusion due to interaction with the semiflexible cytoskeleton @xcite . knowledge of the time or ensemble averaged mean squared displacement of an anomalous diffusion process is insufficient to fully characterise the underlying stochastic mechanism , as the associated probability density @xmath16 is no longer necessarily gaussian , and therefore no longer specified by the first and second moments , only @xcite . this property is in contrast to the universal gaussian nature of brownian motion which is effected by the central limit theorem . at the same time the very nature of the anomalous diffusion process may result in decisively different behaviours for diffusional mixing and diffusion - limited reactions @xcite . in biological cells this would imply significant differences for signalling and regulatory processes . for a better understanding of the dynamics in biological cells and other complex fluids knowledge of the underlying stochastic mechanism is therefore imperative . here we discuss the properties of continuous time random walk ( ctrw ) processes with diverging characteristic waiting time with respect to their time averaged behaviour , expanding on our earlier work @xcite . we show that for free motion the lag time dependence of the time averaged mean squared displacement , @xmath41 , is insensitive to the anomalous diffusion exponent @xmath42 for ctrw processes . in contrast , for confined ctrw subdiffusion a universal scaling behaviour emerges , @xmath43 , with dynamic exponent @xmath44 [ curves ( 1 ) and ( 2 ) in fig . [ diffusion_modes ] ] . subdiffusion governed by fractional brownian motion ( fbm ) leads to the scaling @xmath45 of the time averaged mean squared displacement , turning over to a saturation plateau under confinement , @xmath46 [ curves ( 1 ) and ( 2 ) in fig . [ diffusion_modes ] ] . we particularly emphasise the irreproducible nature of the time averaged quantities and their associated scatter around the ensemble mean for ctrw subdiffusion processes . this randomness of the time averages is captured by the distribution function of the amplitude of the time average . we show that even for relatively short trajectories this distribution is a good characteristic for the underlying process . in contrast for fbm processes the scatter typical of many single particle experiments is not found in the long time limit . in the remainder of this work , for simplicity we restrict the discussion to the one - dimensional case ( @xmath47 ) . generalisation to higher dimensions is straightforward . the article is structured as follows . we start with a brief introduction to ctrw and fbm . we then consider the cases of unbounded motion and confined anomalous diffusion in the subsequent two sections . the distribution of the time averages will be presented thereafter . finally , we discuss the velocity autocorrelation functions for subdiffusive ctrw and fbm processes , before presenting a concluding discussion . although both ctrw and fbm give rise to an ensemble averaged mean squared displacement of the form ( [ eamsd ] ) , they are fundamentally different processes , as outlined here . we note in passing that also in the random motion on a fractal support subdiffusion arises @xcite . we will not pursue this type of anomalous diffusion in the following . ctrw theory dates back to the work of montroll and weiss @xcite , and was championed in the analysis of charge carrier motion in amorphous semiconductors by scher and montroll @xcite . ctrw has become a standard statistical tool to describe processes ranging from particle motion in actin networks @xcite to the tracer motion in groundwater @xcite . each jump of a ctrw process is characterised by a random jump length and a random waiting time elapsing before the subsequent jump . at each jump the jump length and waiting time are chosen independently . for a subdiffusive process we assume that the variance of the jump lengths is given by a finite value @xmath48 , and we consider the unbiased case @xmath49 . on a lattice of spacing @xmath50 , we would have @xmath51 . in contrast , the waiting times @xmath52 are drawn from the probability density @xmath53 for large @xmath52 , with @xmath9 . this form of @xmath54 is scale - free , that is , the average waiting time @xmath55 diverges , causing effects such as ageing @xcite and weak ergodicity breaking @xcite . in eq . ( [ wtd ] ) , the quantity @xmath56 is a scaling factor . the anomalous diffusion constant in this case becomes @xcite @xmath57 the scale - freeness of @xmath58 allows individual waiting times @xmath52 to become quite large . no matter how long a given time series is chosen , single @xmath52 values may become of the order of the length of the entire time span covered in the trajectory . an example is shown in fig . [ timeseries ] : the stalling events with large waiting times @xmath52 are quite distinct . for values of the anomalous diffusion exponent @xmath42 that are closer to 1 the stalling is less pronounced . physically , the power - law form of the waiting time distribution @xmath58 may be related to comb models @xcite or random energy landscapes @xcite ; for more details , compare refs . . for brownian motion ( top ) , ctrw ( middle ) , and fbm ( bottom ) the anomalous diffusion exponent for fbm and ctrw is @xmath59 . for ctrw the stalling events are outstanding , while in the case of fbm strong antipersistence occurs . note that the stalling events and the antipersistence are less pronounced for larger values of @xmath42 . typical experimental data include additional noise , such that the appearance of measured data will not display the ideal behaviour shown here.,width=332 ] fbm is a gaussian process with stationary increments . its position is defined in terms of the langevin equation @xmath60 or , alternatively , @xmath61 the motion is driven by stationary , fractional gaussian noise @xmath62 with zero mean @xmath63 and long - ranged noise correlation @xcite @xmath64 contrasting the uncorrelated noise for normal diffusion @xmath27 : @xmath65 . in eq . ( [ fbnoise ] ) the anomalous diffusion exponent is connected to the traditionally used hurst exponent by @xmath66 , and we introduce the abbreviation @xmath67 for consistency with standard fbm notation . for subdiffusion the fractional gaussian noise is anticorrelated , decaying like @xmath68 . this implies that a given step is likely to go into the direction opposite to the previous step . the corresponding oscillating behaviour is seen in fig . [ timeseries ] . the position autocorrelation function of fbm becomes @xmath69 so that at equal times @xmath70 we recover the mean squared displacement ( [ aneamsd ] ) . if the fractional gaussian noise is not considered external , but the validity of the fluctuation - dissipation theorem is imposed , one obtains the generalised langevin equation ( gle ) @xcite , @xmath71 where we write @xmath72 instead of the exponent @xmath42 in eq . ( [ fbnoise ] ) . in the gle ( [ gle ] ) , we defined the coupling coefficient @xmath73}$ ] according to the fluctuation - dissipation theorem , where @xmath74 is the boltzmann constant and @xmath75 the absolute temperature . in what follows we only consider the overdamped limit , in which the inertia term @xmath76 can be neglected . the gle then gives rise to the form @xmath77 of the mean squared displacement . in contrast to fbm , that is , the gle leads to subdiffusion for persistent noise with @xmath78 , while @xmath79 yields superdiffusion . fbm and the related gle are used to describe processes such as long term storage capacity of water reservoirs @xcite , climate fluctuations @xcite , economical market dynamics @xcite , single file diffusion @xcite , and elastic models @xcite . motion of this type has also been associated with the relative motion of aminoacids in proteins @xcite , and the free diffusion of biopolymers under molecular crowding conditions @xcite . let us begin with considering anomalous diffusion on an infinite domain and without drift . to find an analytical expression for the time averaged mean squared displacement ( [ tamsd ] ) we note that even a brownian process recorded over a finite time span @xmath22 will show fluctuations in the number of jumps performed during @xmath22 . to average out these trajectory - to - trajectory fluctuations we introduce the ensemble mean , @xmath80 we can then express the integrand in terms of the variance of the jump lengths , @xmath48 , and the average number of jumps @xmath81 in the time interval @xmath82 , as follows : @xmath83 for a regular random walk on average every jump occurs after the waiting time @xmath55 . thus @xmath84 , and @xmath85 where we defined the diffusion constant @xmath86 $ ] . in the brownian limit the time averaged mean squared displacement ( [ btamsd ] ) in terms of the lag time @xmath1 takes on exactly the same form as the ensemble averaged mean squared displacement ( [ eamsd ] ) as function of time @xmath18 . this is not surprising , as brownian motion is an ergodic process . for sufficient duration @xmath22 of the time records any time average converges to the corresponding ensemble average , and thus the ensemble average in expression ( [ eatamsd ] ) is no longer necessary . the number of jumps of a ctrw process with waiting time distribution of the form ( [ wtd ] ) on average grows sublinearly with time , @xmath87 $ ] @xcite . this time evolution translates into the mean squared displacement ( [ eamsd ] ) with the anomalous diffusion coefficient ( [ an_coff ] ) . combining the time dependence of @xmath88 for ctrw subdiffusion with the definition ( [ eatamsd ] ) we obtain the following result for the time averaged mean squared displacement , @xmath89 in the limit @xmath90 @xcite . this follows from expansion of the relation @xmath91 . the noteworthy feature of this result is that the linear lag time dependence of normal diffusion remains completely unaffected by the anomalous nature of the stochastic process . only the dependence on the overall measurement time @xmath22 witnesses the underlying subdiffusion . ( [ ctrwtamsd ] ) can be understood if we notice that for normal diffusion we have @xmath92 [ since according to einstein @xmath93 where @xmath94 is the mean time between jump events . now , for the subdiffusive case @xmath55 diverges and must be replaced by @xmath95 , which explains the term @xmath96 . [ ctrw_tamsd ] shows 20 trajectories of a ctrw process with @xmath97 , displaying the general trend @xmath98 . , shown for 20 individual trajectories . the overall measurement time is @xmath99 ( a.u . ) . note the local changes of the slope in the trajectories as well as the complete stalling in the lowest curve , all bearing witness to the scale - free nature of the underlying waiting time distribution @xmath100.,width=332 ] rewriting the result ( [ ctrwtamsd ] ) in the form @xmath101 we see that the effective diffusion coefficient @xmath102 decays as function of the measurement time @xmath22 . the longer the system evolves after its initial preparation , the less mobile it appears , consistent with the ageing property of ctrw subdiffusion . for instance , in the picture of the trapping events , in the course of time deeper and deeper traps may be encountered by the diffusing particle , such that it gets more and more stuck . this increasing immobility is mirrored in the form @xmath102 . note that eqs . ( [ ctrwtamsd ] ) and ( [ rew ] ) suggest that from measuring the @xmath1 dependence of an anomalous diffusion process one might draw the erroneous conclusion that normal diffusion were observed . the disparity between the time averaged mean squared displacement ( [ ctrwtamsd ] ) and its ensemble averaged counterpart ( [ aneamsd ] ) demonstrates the weak ergodicity breaking characteristic of a process with diverging time scale @xcite . single or few , long waiting times are also responsible for the pronounced deviations between individual realisations shown in fig . [ ctrw_tamsd ] . this apparent irreproducibility of the time averaged mean squared displacement is again intimately coupled to the weak ergodicity breaking nature of the ctrw subdiffusion process . we will discuss this feature more quantitatively in terms of the distribution @xmath103 as function of the relative deviation @xmath104 of the time averaged mean squared displacement around its ensemble mean in section [ scatter ] . in contrast to the above behaviour of ctrw subdiffusion processes , fbm is ergodic . indeed , by help of the position autocorrelation ( [ fbm_auto ] ) the time averaged mean squared displacement ( [ eatamsd ] ) becomes @xcite @xmath105 exactly matching the ensemble averaged mean squared displacement , @xmath106 . however , ergodicity is reached algebraically slowly @xcite , see also below . in fig . [ fbm_tamsd ] the comparatively minute scatter between individual trajectories supports the ergodic behaviour of fbm processes . the somewhat larger deviations at longer lag times are due to worsening statistics when @xmath107 . . the scatter between individual trajectories is minute , mirroring the ergodicity of this process.,width=332 ] since the motion of tracer particles is typically confined , for example , by the cell walls or internal membranes , we now consider the important case of anomalous diffusion in a bounded domain . in a finite interval @xmath108 $ ] the mean squared displacement of a brownian particle initially released well away from the boundaries will grow linearly in time and eventually turn over to a stationary plateau of magnitude @xmath109 . similarly , if the particle evolves in the confinement of an external harmonic potential of the form @xmath110 , the thermal value @xmath111 $ ] will eventually be attained . how is this behaviour modified for a ctrw subdiffusion process ? under the influence of an arbitrary external potential @xmath112 defining the force @xmath113 , ctrw subdiffusion can be described by the fractional fokker - planck equation @xcite , or , equivalently , in terms of the following coupled langevin equations @xcite : @xmath114 here the position @xmath115 is expressed in terms of the parameter @xmath116 ( the internal time ) , and driven by the white gaussian noise @xmath117 . thus , eq . ( [ le1 ] ) defines standard brownian motion @xmath118 , where @xmath119 is the diffusivity for the normal diffusion process in internal time @xmath116 . laboratory time @xmath18 is introduced by the so - called subordination through the process @xmath120 , given by the probability density function @xcite @xmath121^{1/\alpha } \frac{t}{s^{1 + 1/\alpha}}\right),\ ] ] where @xmath122 is a one - sided lvy stable probability density with laplace transform @xmath123 . thus , eq . ( [ le2 ] ) transforms the brownian process @xmath118 with diffusivity @xmath119 into the subdiffusive motion with generalised diffusivity @xmath28 . on the basis of this scheme , the ensemble averaged position - position correlation in an arbitrary confining potential @xmath124 becomes @xmath125 valid in the limit @xmath126)^{1/\alpha}$ ] , where @xmath127 is the smallest non - zero eigenvalue of the corresponding fokker - planck operator @xcite . ( [ ctrw_conf_ea_auto ] ) demonstrates that , despite the confinement , the process is non - stationary . in result ( [ ctrw_conf_ea_auto ] ) we used the incomplete beta function @xmath128 and defined the first and second boltzmann moments , whose general definition is @xmath129 the partition function reads @xmath130 at large time separation , @xmath131 , the position autocorrelation decays algebraically , @xmath132 towards the value @xmath133 . the corresponding limiting behaviour of the incomplete beta function reads @xmath134 inserting expression ( [ ctrw_conf_ea_auto ] ) into the definition of the time averaged mean squared displacement ( [ eatamsd ] ) we obtain @xcite @xmath135dt.\end{aligned}\ ] ] then , with relation ( [ bet_lim ] ) we arrive at the scaling behaviour @xmath136 valid in the limits @xmath137 and @xmath138 ) ^{1/\alpha}$ ] . result ( [ conf_tamsd ] ) is quite remarkable : instead of the naively assumed saturation toward the stationary plateau a power - law growth @xmath139 is observed . only when the lag time @xmath1 approaches the overall measurement time @xmath22 the singularity in expression ( [ eatamsd ] ) causes a dip toward the plateau of the ensemble averaged mean squared displacement . additionally eqs . ( [ ctrw_conf_ea_auto ] ) and ( [ conf_tamsd ] ) are universal in the sense that the exact form of the confining potential solely enters into the prefactor through the first two moments of the boltzmann distribution corresponding to the confining potential @xmath124 . we note that the asymptotic scaling @xmath140 is consistent with the numerical analysis presented in ref . @xcite . [ confined_ctrw ] depicts the ensemble mean of the time averaged mean squared displacement ( [ eatamsd ] ) for two types of confinement , an harmonic potential and a box potential . the particle is initially placed at the bottom of the potential well and in the middle of the box potential , respectively . this is why at short lag times the particle exhibits the linear scaling @xmath141 typical for unconfined motion , before turning over to the confinement - induced scaling @xmath7 . note that this plot is fit - free , i.e. , the theoretically calculated asymptotic behaviours nicely fall on top of the simulations results . for an harmonic binding potential @xmath142 ( @xmath143 ) and a particle in a box of length @xmath144 ( @xmath145 ) , with @xmath146 , measurement time @xmath147 ( a.u . ) , @xmath148 , and @xmath149 . without fitting , the lines show the analytic results for the transition from @xmath150 for short lag times according to eq . ( [ ctrwtamsd ] ) ( @xmath151 and @xmath152 ) to @xmath7 for long lag times , eq . ( [ conf_tamsd ] ) ( ) . for long @xmath1 , @xmath153 exhibits universal behaviour , in the sense that the curve does not depend on the external field.,width=332 ] as shown in fig . [ conf_scatter ] individual trajectories still exhibit the pronounced scatter typical for ctrw subdiffusion . visually the scatter does not change between the unbiased initial motion and the confinement - dominated part of the process . this is due to the fact that the scatter is caused by the scale - freeness of the waiting times . in our process waiting times and jump lengths are decoupled , corresponding to the subordination property of ctrw subdiffusion @xcite . . the extent of the deviations between the trajectories does not change qualitatively between the initial unbiased motion and the later confinement - dominated regime . the squares ( @xmath143 ) represent simulations results for the ensemble average @xmath154 . same parameters as in fig . [ confined_ctrw].,width=332 ] what happens if we introduce a cutoff in the power - law of the waiting time distribution of the form @xmath155,\ ] ] in which a characteristic time scale @xmath156 is introduced , terminating the power - law scaling ? for the waiting time density ( [ wtd_cutoff ] ) the characteristic waiting time @xmath157 becomes finite . at sufficiently short times one would expect this process to still exhibit the features of ctrw subdiffusion , while at times @xmath158 the process should converge to regular brownian motion with a gaussian propagator . in fig . [ ctrw_cutoff_conf ] we demonstrate that for a suitable choice of the cutoff time @xmath156 the behaviour of the time averaged mean squared displacement @xmath13 under confinement preserves the characteristic non - ergodic features of ctrw subdiffusion , i.e. , the turnover from the initial scaling @xmath3 to the confinement - dominated scaling @xmath159 . below we will show that at the same time the distribution of the time average around its ensemble mean is significantly altered . ) , in the box @xmath160 $ ] with reflecting boundary conditions . the generic behaviour of initial and final scaling , @xmath161 and @xmath7 remains unaltered . we chose @xmath162 , @xmath163 , @xmath164 and overall measurement time @xmath165 ( a.u.).,width=332 ] ergodicity remains unaffected for fbm , that is confined to an interval of size @xmath108 $ ] . namely , convergence of @xmath13 is observed toward the stationary value @xmath109 @xcite . for fbm under the influence of an harmonic potential @xmath110 the position - position correlation can be calculated exactly @xcite . inserting into the ensemble mean ( [ eatamsd ] ) of the time averaged mean squared displacement one can show that the initial behaviour @xmath166 turns over to the stationary value @xmath167 $ ] . in general , the time averaged mean squared displacement for confined fbm converges to a constant : @xmath168 the deviations of the time averaged mean squared displacement @xmath169 between individual trajectories can be quantified in terms of the probability density function @xmath170 of the dimensionless ratio @xmath171 of the time averaged mean squared displacement over its ensemble mean . for an ergodic process this distribution is necessarily sharp , @xmath172 for long measurement time @xmath22 . deviations from this form are expected for non - ergodic processes such as ctrw subdiffusion , but also for relatively short trajectories . we here address both effects and demonstrate that the distribution @xmath170 is a quite reliable means to distinguish different stochastic processes even when the recorded time series are fairly short . for ctrw subdiffusion with power - law waiting time distribution ( [ wtd ] ) the distribution assumes the form @xcite @xmath173 for @xmath174 , where @xmath122 is again a one - sided lvy stable distribution , whose laplace transform is @xmath175 . special cases include the brownian limit ( [ phi1 ] ) for @xmath176 and the gaussian shape @xmath177 for @xmath146 . in fact , we observe an exponentially fast decay of @xmath170 with large @xmath178 for all @xmath9 , compare appendix [ app1 ] . for short trajectories the basic shape of the distribution @xmath170 is surprisingly well preserved @xcite . this is demonstrated in fig . [ ctrw_scatter ] , in which the characteristic asymmetric shape with respect to the ergodic value @xmath179 for the case @xmath97 is reproduced for a process with @xmath180 , even for a lag time as long as @xmath181 . no significant dependence on the confinement is observed . also for larger values of @xmath42 the finite value for small @xmath178 is similarly reproduced for short trajectories @xcite , pointing at the quite remarkably reliability on the shape of the scatter distribution @xmath170 for ctrw subdiffusion . as we show below , the distribution for fbm processes with its zero value at @xmath182 can be clearly distinguished from the ctrw form . of the time averaged mean squared displacement , with @xmath183 for a ctrw process with @xmath97 , @xmath163 , and @xmath180 ( a.u . ) . the filled and open symbols , respectively , represent the unconfined case and confined motion on the interval @xmath184 $ ] . we see quite good agreement with the expected gaussian limiting distribution of the scatter , eq . ( [ phi_gauss ] ) , centred around @xmath182 . in the inset we show the case for the largest measured lag time , @xmath185 , also in good agreement with the predicted shape of the distribution.,width=332 ] a useful parameter to quantify the violation of ergodicity is the ergodicity breaking parameter @xcite @xmath186 eb varies from @xmath187 for @xmath188 monotonically to @xmath189 in the brownian limit @xmath27 . for the special case @xmath146 one finds @xmath190 , while for @xmath191 , @xmath192 . the scatter distribution is sensitive to few extreme events . when these are lacking , the distribution should be significantly different from the form discussed for ctrw subdiffusion with diverging characteristic time scale . indeed , as shown in fig . [ cutoff_scatter ] for the cutoff waiting time distribution defined through eq . ( [ wtd_cutoff ] ) , @xmath170 drops down to zero around @xmath182 and assumes an almost gaussian shape around the ergodic value @xmath179 . note that the simulations were performed with the same parameters as for fig . [ ctrw_cutoff_conf ] , in which the breaking of ergodicity still persists , despite the cutoff . this demonstrates that different quantities have different sensitivity to the cutoff . for ctrw subdiffusion with cutoff , as defined in eq . ( [ wtd_cutoff ] ) . the observed form is clearly different from the much more asymmetric shape in fig . [ ctrw_scatter ] . the value of @xmath170 reaches zero for small values of @xmath178 . an almost gaussian shape centred around the ergodic value @xmath179 is observed . same parameters as in fig . [ ctrw_cutoff_conf].,width=332 ] fbm is based on long - ranged correlations . however , we can obtain an approximate expression for the scatter distribution @xcite @xmath193 where @xmath194 is an intrinsic time scale . expression ( [ fbmscatter ] ) is valid for sufficiently small lag times @xmath1 , for which the correlations are neglected @xcite . note that this distribution is independent of @xmath42 and centred around the ergodic value @xmath179 . in particular , in the long measurement time limit @xmath174 , the distribution converges to the sharp form @xmath195 . also this behaviour is surprisingly well preserved for short trajectories , as demonstrated in fig . [ fbm_scatter ] . of the time averaged mean squared displacement , with @xmath196 for fbm subdiffusion with @xmath97 and @xmath180 ( a.u . ) . the filled and open symbols , respectively , represent the cases for unconfined and confined motion . for small values of the lag time @xmath1 we see quite good agreement with the expected gaussian limiting distribution of the scatter , eq . ( [ fbmscatter ] ) , that is centred around @xmath179 . for larger values of @xmath1 deviations are observed , however , even for @xmath185 the value around @xmath182 is consistently zero . a similar behaviour is found for larger values of @xmath42 @xcite . note that when @xmath174 the process is ergodic and @xmath197 approaches a @xmath198 function.,width=332 ] for fbm one can define a quantity similar to the ergodicity breaking parameter ( [ eb ] ) . namely , without taking the long measurement time limit , we obtain the normalised variance of the time averaged mean squared displacement @xmath199 it turns out that the convergence to ergodicity is algebraically slow , and for subdiffusion @xmath9 one obtains a decay , which is inversely proportional to @xmath22 @xcite : @xmath200 the coefficient @xmath201 here is defined as @xcite @xmath202 it increases continuously from zero at @xmath203 to @xmath204 . the quantity @xmath205 is connected to the probability density ( [ fbmscatter ] ) by @xmath206 , obtained from approximating that the values of @xmath178 may range in the interval @xmath207 , which appears reasonable given the sharp decay of @xmath197 to 0 at @xmath182 for fbm . apart from the fact that @xmath205 explicitly depends on @xmath42 while the approximation leading to eq . ( [ fbmscatter ] ) loses the @xmath42 dependence , both quantities decay @xmath208 , and the internal time scale @xmath194 takes on a role similar to the lag time @xmath1 . a typical quantity accessible from experimental data is the velocity autocorrelation function , which is defined through @xmath209 here the velocity is defined as the difference quotient @xmath210 $ ] . the velocity autocorrelation ( [ autov00 ] ) was suggested as a tool to distinguish between different subdiffusion models @xcite . we show here that for confined processes the shape of the velocity autocorrelation function does not allow for a significant distinction between subdiffusive ctrw and fbm . note that the calculation of @xmath211 amounts to determine four two - point correlation functions of the type @xmath212 . for unbounded ctrw process starting with initial condition @xmath213 , the position correlation function is given by the following expression @xcite , @xmath214^\alpha . \label{autov03}\ ] ] this result for free ctrw is due to the fact that the jump lengths in the interval @xmath215 for @xmath216 have zero mean . with eq . ( [ autov03 ] ) we find @xmath217 0 & \tau\geq\epsilon \end{array}\right . \label{autov04}\ ] ] the velocity autocorrelation function for an unbounded ctrw process does not yield negative values due to the absence of correlations for different jumps . in this case the velocity autocorrelations can be easily distinguished from the behaviour of fbm processes , see below . note that @xmath211 in eq . ( [ autov04 ] ) is non - analytic at @xmath218 , an observation that is confirmed in simulations . the behaviour of the velocity autocorrelations for ctrw subdiffusion are displayed in fig . [ velcorr1 ] . for a confined ctrw process the situation is quite different . to explore the behaviour of the corresponding velocity autocorrelation function we use the general result for the correlation function of confined ctrw , eq . ( [ ctrw_conf_ea_auto ] ) . for simplicity we assume a symmetric potential such that @xmath219 . with the initial condition @xmath213 we obtain @xmath220 figs . [ velcor2 ] and [ velcor3 ] ( for the absolute value ) display excellent agreement of eq . ( [ autov05 ] ) with simulations of ctrw subdiffusion for a lattice of size 10 with reflecting boundary conditions , and @xmath146 . we observe that for confined motion the ctrw velocity autocorrelation function indeed attains negative values and has a minimum on @xmath218 , as the confinement effectively induces correlations . for long @xmath52 the velocity autocorrelation function decays to zero ( from the negative side ) as the power - law @xmath221 valid for @xmath222 . for free fbm the position correlation function ( [ fbm_auto ] ) together with the definition ( [ autov00 ] ) produce the result @xmath223 this function yields negative values for sufficiently long @xmath52 , and its minimum value @xmath224 is assumed at @xmath218 . for long @xmath52 it decays toward zero from the negative side in the power - law form @xmath225 valid for @xmath222 . for confined fbm with @xmath97 . the simulations were performed on the interval @xmath184 $ ] , and @xmath226 ( a.u . ) . for comparison , the velocity autocorrelation for free fbm is drawn based on eq . ( [ autov01]).,width=332 ] we see that both the velocity autocorrelation function of the confined ctrw and the one for free fbm acquire negative values and decay as power - laws for large @xmath52 , see below . further both obtain a sharp minimum for @xmath218 . confined fbm behaves similarly to free fbm , see fig . [ fbm_box ] . from our analysis it becomes clear that the shape of the velocity autocorrelation function may not be a good diagnosis tool to distinguish between subdiffusive ctrw and fbm processes . we note that for ctrw subdiffusion the time averaged correlation functions will be random variables , exactly like the time averaged mean squared displacement . the physical mechanisms leading to subdiffusion in biological cells or other complex systems are varied . the anomalous diffusion of inert biopolymers larger than some 10kd in living cells is due to molecular crowding , these are excluded volume effects in the viscoelastic , superdense cellular environment @xcite . potentially similar effects occur in single file diffusion : a tracer particle diffusing in a one dimensional channel and interacting with other brownian particles will slow down its random motion , and subdiffusion with @xmath146 is observed @xcite . what is the resulting motion like ? consider a particle , that gets repeatedly trapped during its random walk . such trapping may give rise to a broad distribution of waiting times , as embodied in the ctrw model . such broad distribution of trapping times could be due to chemical binding of the tracer particle to its environment on varied time scales , or due to active gelation / de - gelation of the crowding particles . long - tailed trapping time distributions were observed for @xmath227m - sized particles due to interaction with a semiflexible actin mesh @xcite . there , the anomalous diffusion exponent @xmath42 depends on the size of the tracer particles versus the typical actin mesh size . alternatively , motion patterns of fbm type may be due to coupling to the viscoelastic crowding environment @xcite . yet other approaches are based on polymer dynamics . de gennes reptation model of a polymer diffusing in a tube formed by foreign chains in a melt yields subdiffusion with @xmath228 , and was used as a possible explanation of the short time dynamics of telomere motion @xcite . one may speculate whether the widely observed subdiffusion of biomolecules and passive tracers is a mere coincidence of nature that should be attributed to the dense environment in the cell and/or to the wide distribution of obstacles and reactions in the cell . or , maybe , subdiffusion is by itself a goal which was obtained via evolution ? there exist claims that sub - diffusion is useful in certain cellular search strategies @xcite . one may also view subdiffusion to be a sufficiently slow process to help maintaining the organisation of the genome in the nucleus of the cell without the need for physical compartments @xcite . since reactions in many cases are controlled by diffusion , the emergence of slower - than - normal diffusion has far - reaching implications to signalling and regulatory processes in the cell . the standard theories of diffusion - controlled reactions under such circumstances must be replaced by subdiffusion - controlled reaction models . while these issues are clearly important , we here focused more on the characterisation of trajectories of single particles in the cell , in particular , as a quantitative diagnosis method to probe the nature of the stochastic motion in cells and other complex media . single particle tracking microscopy is a widely used technique . it allows one to locally probe complex systems in the liquid phase _ in situ_. in particular it has become one of the standard tools in biophysical , colloidal , polymeric , and gel - like environments . the motion of tracers in these systems is often anomalous . typically the experimentally recorded time series are evaluated in terms of the time averaged mean squared displacement @xmath13 . here we collected the behaviour of @xmath229 for the two most prominent anomalous stochastic processes , the continuous time random walk and fractional brownian motion . for ctrw subdiffusion , connected with ageing and weak ergodicity breaking , in an unconfined system the time averaged mean squared displacement scales linearly in time , @xmath13 [ eq . ( [ ctrwtamsd ] ) ] , renouncing the anomalous nature of the process . only the dependence on the overall measurement time @xmath22 pays tribute to the underlying subdiffusion . this behaviour contrasts the anomalous scaling @xmath230 of the ensemble averaged mean squared displacement . under confinement no plateau value is observed as in the ensemble average , instead , a power - law of the form @xmath43 , eq . ( [ conf_tamsd ] ) , is found . interestingly , this characteristic behaviour is approximately preserved when an appropriate cutoff in the waiting time is introduced . conversely the distribution of the time averaged mean squared displacement around its ensemble mean becomes almost gaussian in presence of the cutoff , while it has an exponential decay and a finite value at @xmath182 when the system is non - ergodic and exhibits ageing . fbm , in contrast , is ergodic , although ergodicity is reached algebraically slowly . free fbm subdiffusion shows @xmath231 , eq . ( [ fbm_free ] ) , while under confinement the plateau value of the ensemble average is reached , eq . ( [ fbm_cnf ] ) . for finite trajectories the distribution of the time averaged mean squared displacement is approximately gaussian for short lag times . our analysis demonstrates how different the two stochastic processes are , despite sharing the same form of the ensemble averaged mean squared displacement . at the same time the velocity autocorrelation of confined ctrw subdiffusion is hardly distinguishable from that of fbm subdiffusion . given a recorded time series of anomalous diffusion from experiment it is important to know more precisely which stochastic process is responsible for the observed behaviour , in particular , with respect to diffusion - limited reactions , general transport behaviour , and related processes such as gene regulation . the analysis presented here , along with complementary tools discussed in refs . @xcite , will be instrumental in the classification of anomalous diffusion behaviour . we thank c. selhuber - unkel and l. oddershede for providing the data used in fig . [ christine ] , and for many useful discussions . partial financial support from the deutsche forschungsgemeinschaft and the israel science foundation is gratefully acknowledged . using the theory of fox @xmath232-functions we obtain the exact form for the distribution @xmath170 of the dimensionless variable @xmath178 @xcite : @xmath233.\ ] ] this function has the series expansion @xmath234^n}{n!\gamma(-\alpha n)},\ ] ] and the asymptotic form @xmath235^{1/(1-\alpha)}\right)\end{aligned}\ ] ] valid at @xmath236 . s. b. yuste , g. oshanin , k. lindenberg , o. bnichou , and j. klafter , phys . e * 78 * , 021105 ( 2008 ) ; e. abad . s. b. yuste , and k. lindenberg , phys . e * 81 * , 031115 ( 2010 ) ; i. m. sokolov , s. b. yuste , j. j. ruiz - lorenzo , and k. lindenberg , phys . e * 79 * , 051113 ( 2009 ) ; d. froemberg and i. m. sokolov , phys . lett . * 100 * , 108304 ( 2008 ) . bouchaud , j. phys . ( paris ) i , * 2 * , 1705 ( 1992 ) ; g. bel and e. barkai , phys . rev . lett . * 94 * , 240602 ( 2005 ) ; a. rebenshtok and e. barkai , _ ibid . _ * 99 * , 210601 ( 2007 ) ; m. a. lomholt , i. m. zaid , and r. metzler , _ ibid . _ , 98 , 200603 ( 2007 ) ; i. m. zaid , m. a. lomholt , and r. metzler , biophys . j. * 97 * , 710 ( 2009 ) ; f. d. stefani , j. p. hoogenboom , and e. barkai , physics today * 62*(2 ) , 34 ( 2009 ) ; g. margolin and e. barkai , phys . 94 * , 080601 ( 2005 ) ; g. aquino , p. grigolini , and b. j. west , epl * 80 * , 10002 ( 2007 ) . r. metzler , e. barkai , and j. klafter , phys . lett . * 82 * , 3563 ( 1999 ) ; europhys . lett . * 46 * , 431 ( 1999 ) . e. barkai , r. metzler , and j. klafter , phys . e * 61 * , 132 ( 2000 ) ; r. metzler , j. klafter , and i. m. sokolov , _ ibid . _ * 58 * , 1621 ( 1998 ) . a. n. kolmogorov , dokl . 26 * , 115 ( 1940 ) ; b. b. mandelbrot and j. w. van ness , siam rev . * 1 * , 422 ( 1968 ) ; h. qian , fractional brownian motion and fractional gaussian noise . in g. rangarajan and m.z . ding ( eds ) , _ processes with long - range correlations _ ( springer , lecture notes in physics , vol.621 ) , pp.22 - 33 . t. e. harris , j. appl . * 2 * , 323 ( 1965 ) ; l. lizana and t. ambjrnsson , phys . . lett . * 100 * , 200601 ( 2008 ) ; l. lizana , t. ambjrnsson , a. taloni , e. barkai , and m. a. lomholt , phys . e * 81 * , 051118 ( 2010 ) ; e. barkai and r. silbey , phys . * 102 * , 050602 ( 2009 ) . m. magdziarz , a. weron , k. burnecki , and j. klafter , phys . * 103 * , 180602 ( 2009 ) . compare also m. magdziarz and j. klafter , phys . e , at press ; k. burnecki and a. weron , phys . e * 82 * , 021130 ( 2010 ) . h. c. fogedby , phys . e * 50 * , 1657 ( 1994 ) ; a. baule and r. friedrich , phys . e * 71 * , 026101 ( 2005 ) ; m. magdziarz , a. weron , and j. klafter , phys . rev . * 101 * , 210601 ( 2008 ) ; m. magdziarz , a. weron , and k. weron , phys . e * 75 * , 016708 ( 2007 ) ; e. heinsalu , m. patriarca , i. goychuk , g. schmid , and p. hnggi , phys . rev . e * 73 * , 046133 ( 2006 ) . a. m. mathai and r. k. saxena , the h - function with applications in statistics and other disciplines ( wiley eastern ltd . , new delhi , 1978 ) ; a. m. mathai , r. k. saxena , and h. j. haubold , the h - function , theory and applications ( springer , new york , 2010 ) .

in previous papers ( kato 2004 , 2008a ) , we proposed a resonant excitation mechanism of disk oscillations in deformed disks . the purpose is to propose an excitation mechanism of quasi - periodic oscillations ( qpos ) observed in neutron - star and black - hole x - ray binaries ( e.g. , kato and fukue 2007 , kato 2008b ) . in this resonant excitation model of disk oscillations , a deformation of disks from an axially - symmetric equilibrium state is essential . the deformation to be considered is a warp or an eccentric deformation of disks in the equatorial plane . an outline of the model is as follows . a non - linear coupling between a disk oscillation ( hereafter we call it the original oscillation ) and a deformed part of the disks ( warp or eccentric deformation ) brings about some forced disk oscillations ( we call them intermediate oscillations ) . the intermediate oscillations make a resonant coupling with the unperturbed disk at particular radii of the disk . after this resonant coupling , the intermediate oscillations feedback to the original oscillation by a nonlinear coupling with the deformed part of the disk . since this nonlinear feedback process involves a resonance , the original oscillation is amplified or dampened . kato ( 2004 , 2008a ) examined this nonlinear feedback processes , and derived an excitation criterion and growth ( damping ) rate of the resonant oscillations . it is of importance to note that in keplerian disks this resonant excitation process works only when the disks are general relativistic . that is , a non - monotonic radial distribution of epicyclic frequency is necessary for appearance of resonance and for trapping of oscillations . in these studies the intermediate disk oscillations were assumed to be local near to the resonant radius in the sense that their radial wavelength is shorter than the characteristic radial length of disks . that is , the radial derivative , @xmath0 , operated to the wave quantities was taken to be @xmath1 . further , @xmath2 was assumed to be constant . this treatment of @xmath3 const . , however , was inadequate around the resonant point . if this inadequate treatment is properly corrected , we find that unlike in the previous papers , the resonant point is not the place where a local dispersion relation of the intermediate oscillations is satisfied with a constant @xmath2 , but the radii of lindblad resonance for intermediate oscillations ( when resonance occurs by horizontal motions ) or the radii where the intermediate oscillations are trapped in the vertical direction ( when resonance occurs by vertical motions ) . in this paper , we modify the analyses in our previous papers so that the above - mentioned inadequate treatment is corrected . we apply the mathematical techniques used by meyer - vernet and sicardy ( 1987 ) in their study of resonant disk - satellite interaction . the results show that the stability criterion derived in the previous papers are unchanged . the expression for growth rate obtained in the previous papers is found to be still applicable , but in this paper the expression is changed in a form more suitable to numerical calculations . we first summarize in section 2 the basic equations and relations necessary to the study of the present resonant excitation problem , although they are given in the previous papers . in section 3 , the stability criterion and growth rate are derived to the case of pressure - less disks as it is instructive , and to the case of disks with pressure in section 4 . the final section is devoted to a brief discussion on a meaning of the instability criterion . we briefly summarize here basic equations and relations [ see kato ( 2008a ) for details ] that are necessary in the subsequent sections . in the present problem , the general relativity is essential . for simplicity , however , formulation in this paper is done in the framework of a pseudo - newtonian using the gravitational potential introduced by paczy ' nski and wiita ( 1980 ) , assuming that the central object has no rotation . we adopt a lagrangian formulation by lynden - bell and ostriker ( 1967 ) . the unperturbed disk is in a steady state with a steady flow @xmath4 . by using a displacement vector , @xmath5 , a weakly nonlinear hydrodynamical equation describing adiabatic , non - self - gravitating perturbations is written as , after generalizing the linear equation derived by lynden - bell and ostriker ( 1967 ) , @xmath6 where @xmath7 is a linear hermitian operator with respect to @xmath5 ( lynden - bell and ostriker 1967 ) and is @xmath8 \nonumber \\ -p_0\nabla({\rm div}\mbox{\boldmath $ \xi$ } ) -\nabla[(\mbox{\boldmath $ \xi$}\cdot\nabla)p_0 ] + ( \mbox{\boldmath $ \xi$}\cdot\nabla)(\nabla p_0 ) , \label{2.4}\end{aligned}\ ] ] and @xmath9 and @xmath10 are the density and pressure in the unperturbed state , and @xmath11 is the barotropic index specifying the linear part of the relation between lagrangian variations @xmath12 and @xmath13 , i.e. , @xmath14 . since , the self - gravity of the disk gas has been neglected , the gravitational potential , @xmath15 , is a given function and there is no eulerian perturbation of it . at the stage of equation ( [ 2.3 ] ) , there is no restriction on the form of @xmath4 , but in the main part of this paper , @xmath4 is taken to be a pure cylindrical rotation . the right - hand side of wave equation ( [ 2.3 ] ) represents the weakly nonlinear terms . derivation of a detailed expression for @xmath16 is boring , because of the presence of many terms . however , after lengthy manipulations we can summarize them into a relatively simple form in the case of @xmath17 : @xmath18 using cartesican cordinates . when @xmath19 , some terms should be added on the right - hand side , which are shown in equation ( 82 ) of kato ( 2008a ) . an important characteristics of @xmath16 is that we have commutative relations ( kato 2008a ) for an arbitrary set of @xmath20 , @xmath21 , and @xmath22 : @xmath23 as shown later , the presence of these commutative relations is the reason why we have a simple stability criterion . we suppose that this commutative relation is a general property of conservative systems . we now assume that the disks are deformed from an axisymmetric steady state by some external or internal cause . the deformation is assumed , for simplicity , to be time - independent , and has the azimuthal wavenumber @xmath24 . the lagrangian displacement associated with the deformation , @xmath25 , is denoted by , @xmath5 [ see equation ( [ wave ] ) ] , and @xmath26 [ see equation ( [ intermediate ] ) ] are expressed in complex forms . when their nonlinear couplings are calculated , we must be careful since the real displacement associated with deformations , for example , is not @xmath27 itself , but the real part of @xmath27 given here . ] @xmath28 as the deformation , a warp will be the most probable , but it is not only one candidate of possible deformations . a plane - symmetric one - armed spiral deformation is one of another possible candidates that can excite disk oscillations . in both cases we have @xmath29 . our purpose here is to examine how the behavior of disk oscillations are affected by the disk deformation . as shown below , some of oscillation modes are resonantly excited on deformed disks through nonlinear coupling with disk deformation . the nonlinear coupling processes are schematically shown in figure 1 of kato ( 2004 ) . the displacement vector associated with a disk oscillation , @xmath5 , is assumed to have frequency @xmath30 and azimuthal wavenumber @xmath31 . separating the time and azimuthal dependences from @xmath32 , we introduce @xmath33 as @xmath34 \hat { \mbox{\boldmath $ \xi$}}(r , z ) . \label{wave}\ ] ] the first step of the nonlinear interaction between the disk oscillation characterized by @xmath35 and the deformation characterized by @xmath36 introduces two kinds of intermediate oscillations with azimuthal wavenumber being @xmath37 and @xmath38 . let us denote the displacement vector associated with these intermediate oscillations by @xmath39 { \hat{\mbox{\boldmath $ \xi$}}}_\pm^{\rm int}(r , z ) , \label{intermediate}\ ] ] where @xmath40 or @xmath38 . here , @xmath41 represents the intermediate oscillations resulting from the coupling between @xmath42 and @xmath43 , while @xmath44 does those resulting from the coupling between @xmath42 and @xmath45 , where the asterisk represents the complex conjugate . that is , @xmath41 and @xmath44 are described , respectively , by @xmath46 , \label{2.7}\ ] ] @xmath47 . \label{2.8}\ ] ] next , the second stage of the nonlinear coupling is considered , which is a feedback process returning to the original oscillation , @xmath48 , by the intermediate oscillations @xmath49 interacting with @xmath50 . in the case where the coupling occurs through @xmath51 , the feedback is described by @xmath52 . \label{2.9}\ ] ] on the other hand , in the case where the feedback occurs through @xmath44 , the equation corresponding to equation ( [ 2.9 ] ) is @xmath53 . \label{2.10}\ ] ] an important point to be noted here is that as a result of this feedback process , the original disk oscillation is amplified or dampened , i.e. , the frequency , @xmath30 , can be no longer real , since in the feedback processes a resonance is involved . how much is the imaginary part of @xmath30 ? this can be examined from equations ( [ 2.9 ] ) and ( [ 2.10 ] ) using the fact that the operators @xmath54 and @xmath55 are hermitian ( lynden - bell and ostriker 1967 ) . after some calculations the imaginary part of frequency , @xmath56 , can be formally expressed as ( e.g. , kato 2008a ) @xmath57 where @xmath58 denotes the cases of the coupling through @xmath59 and @xmath60 , respectively , and @xmath61 dv , \label{2.16}\ ] ] @xmath62 dv , \label{2.18}\ ] ] and @xmath63 { \hat{\mbox{\boldmath $ \xi$}}}dv , \label{2.12}\ ] ] where @xmath64 is the frequency of the original oscillation before the mode couplings . the above expressions for @xmath65 and @xmath66 have physical meanings that @xmath65 is the rate at which work is done on the original oscillations by the nonlinear resonant processes , and @xmath66 is the wave energy of the original oscillations [ see equation ( 93 ) by kato ( 2001 ) ] . it is important to change the above expressions for @xmath65 by using the commutative relation ( [ commutative ] ) as @xmath67dv , \label{work+}\ ] ] @xmath68dv . \label{work-}\ ] ] hereafter , we restrict our attention to oscillations in geometrically thin disks . the steady unperturbed disks are axially - symmetric and have no motion except for rotation . here , cylindrical coordinates ( @xmath69 , @xmath70 , @xmath71 ) are employed , in which the @xmath71-axis is perpendicular to the disk plane and the origin of the coordinates is at the disk center . the unperturbed flow is then described as @xmath72 , where @xmath73 is the angular velocity of disk rotation . we further assume , for simplicity , that the disk is isothermal in the vertical direction . in the vertically isothermal disks , the unperturbed density in disks , @xmath74 , is stratified as ( e.g. , kato et al . 1998 ) @xmath75 , \label{rho}\ ] ] where @xmath76 is the density on the equatorial plane , and @xmath77 is the half - thickness of the disk and is related to the vertical epicyclic frequency , @xmath78 , by @xmath79 where @xmath80 is the isothermal acoustic speed . the vertical epicyclic frequency , @xmath78 , is equal to the angular velocity of the keplerian rotation , @xmath81 , in the case of the central object being non - rotating , and practically equal to @xmath82 , since the disk is assumed to be geometrically thin . hereafter , however , we use @xmath78 without using @xmath81 or @xmath82 , so that we can trace back the effects of @xmath78 on the final results . concerning oscillations superposed on such disks , we consider oscillations whose radial wavelength is moderately short so that the radial variations of physical quantities in the unperturbed state [ including the radial variation of @xmath83 are neglected compared with the radial variation of wave quantities . furthermore , we assume that the oscillations also occur isothermally . then , we can neglect the term of @xmath84 $ ] in the operator @xmath55 in equation ( [ 2.4 ] ) . if the above simplifications and approximations are adopted , the @xmath69- and @xmath71- dependences of @xmath85 are approximately separated . that is , the @xmath71-dependence of @xmath85 is obtained by solving an eigen - value problem , and the eigen - functions are found to be an orthogonal set of hermite polynomials ( okazaki et al . 1987 ) . restricting our attention to one of such oscillations , we write @xmath85 as @xmath86 @xmath87 @xmath88 where @xmath89 is the hermite polynomial of argument @xmath90 , and @xmath91 characterizes the number of node(s ) of oscillations in the vertical direction . it is noted that the number of node(s ) of @xmath92 in the vertical direction is smaller than those of @xmath93 and @xmath94 by one , as shown in equation ( [ xi - z ] ) . however , the subscript @xmath95 ( not @xmath96 ) is attached to @xmath97 as @xmath98 in order to emphasize that @xmath99 , @xmath100 , and @xmath101 are a set of solutions . under these approximations , we express the @xmath69- , @xmath70- , and @xmath71- components of the homogeneous parts of wave equation ( [ 2.3 ] ) as @xmath102\breve{\xi}_{r , n } -i2\omega(\omega - m\omega)\breve{\xi}_{\varphi , n } + \omega^2_\bot h{d \breve{\xi}_{z , n}\over dr } = 0 , \label{4.1}\ ] ] @xmath103 @xmath104\breve{\xi}_{z , n } -n\omega^2_\bot h{d \breve{\xi}_{r , n}\over dr } = 0 . \label{4.3}\ ] ] nonlinear coupling between @xmath105 and @xmath106 introduce intermediate oscillations @xmath107 . the azimuthal wavenumber of the intermediate oscillations are @xmath37 or @xmath38 . their @xmath71-dependence are characterized by @xmath108 or @xmath96 when @xmath109 since @xmath110 and by @xmath95 when @xmath111 . to consider these various coupling cases separately , we write @xmath112 in the forms of @xmath113 @xmath114 @xmath115 where the subscript @xmath116 represents @xmath108 or @xmath96 or @xmath95 . nonlinear coupling terms are also separated into terms proportional to exp@xmath117 $ ] and @xmath118 . that is , in the case of coupling through @xmath119 , we write the coupling terms as @xmath120_r = \rho_0\sum_{{\tilde m},\tilde{n } } \breve{a}_{r,+,\tilde{n}}(r ) { \rm exp}[i(\omega t-{\tilde m}\varphi ) ] { \cal h}_{\tilde { n}}(z / h)+\cdot\cdot\cdot \label{5.7}\ ] ] @xmath120_\varphi = \rho_0\sum_{{\tilde m},\tilde{n } } \breve{a}_{\varphi,+,\tilde{n}}(r ) { \rm exp}[i(\omega t-{\tilde m}\varphi ) ] { \cal h}_{\tilde { n}}(z / h)+\cdot\cdot\cdot \label{5.7'}\ ] ] @xmath120_z = \rho_0\sum_{{\tilde m},\tilde{n } } \breve{a}_{z,+,\tilde{n}}(r ) { \rm exp}[i(\omega t-{\tilde m}\varphi ) ] { \cal h}_{\tilde { n}-1}(z / h)+\cdot\cdot\cdot , \label{5.7''}\ ] ] where @xmath121 denotes terms orthogonal both to @xmath122 and @xmath123 , and the subscript @xmath124 is added to @xmath125 s in order to emphasize that they are related to the @xmath70-dependence of @xmath126 $ ] . in a similar way , in the case of coupling through the intermediate oscillations of @xmath127 , @xmath128 $ ] can be expressed in forms similar to equations ( [ 5.7 ] ) ( [ 5.7 ] ) , introducing @xmath129 , @xmath130 , @xmath131 . the equations describing intermediate oscillations are then written as @xmath132 \breve{\xi}^{\rm int}_{r,\pm,{\tilde n } } -i2\omega(\omega-{\tilde m}\omega ) \breve{\xi}^{\rm int}_{\varphi,\pm,{\tilde n } } \nonumber \\ + \omega^2_\bot h{\partial \breve{\xi}^{\rm int}_{z,\pm,{\tilde n } } \over\partial r } = \breve{a}_{r,\pm,\tilde{n } } , \label{4.1'}\end{aligned}\ ] ] @xmath133 @xmath134 \breve{\xi}^{\rm int}_{z,\pm , { \tilde n } } -{\tilde n}\omega^2_\bot h{d \breve{\xi}^{\rm int}_{r,\pm,{\tilde n}}\over dr } = \breve{a}_{z , \pm,\tilde{n } } , \label{4.3'}\ ] ] where @xmath135 represents @xmath37 or @xmath38 , and @xmath136 does @xmath108 or @xmath96 when @xmath109 and @xmath137 when @xmath111 . from equations ( [ 4.1 ] ) and ( [ 4.2 ] ) we can eliminate @xmath138 to give @xmath139\breve{\xi}^{\rm int}_{r,\pm , { \tilde n } } -c_{\rm s}^2{d^2 \breve{\xi}^{\rm int}_{r,\pm , { \tilde n } } \over dr^2 } + \omega^2_\bot h { d\breve{\xi}^{\rm int}_{z,\pm , { \tilde n } } \over dr } = \breve{a}_{r,\pm,{\tilde n } } -i{2\omega\over\omega-{\tilde m}\omega}\breve{a}_{\varphi,\pm,{\tilde n}}. \label{4.1''}\ ] ] hereafter , we use equations ( [ 4.1 ] ) ( [ 4.1 ] ) as the basic equations describing the intermediate oscillations . finally , we should notice that in the case of isothermal disks described above , @xmath140 and @xmath141 given by equations ( [ work+ ] ) and ( [ work- ] ) are written in the forms of @xmath142 \cdot \biggr[{\breve \xi}_{r,+}^{\rm int}{\breve a}_{r,+}^ * + { \breve \xi}_{\varphi,+}^{\rm int}{\breve a}_{\varphi,+}^ * + \frac{1}{{\tilde n}}{\breve \xi}_{z , + } { \breve a}_{z,+}^*\biggr]dx , \label{w+}\ ] ] and @xmath143 \cdot \biggr[{\breve \xi}_{r,-}^{\rm int}{\breve a}_{r,-}^ * + { \breve \xi}_{\varphi,-}^{\rm int}{\breve a}_{\varphi,-}^ * + \frac{1}{{\tilde n}}{\breve \xi}_{z , -}{\breve a}_{z,-}^*\biggr]dx , \label{w-}\ ] ] where @xmath144 $ ] comes from the part of volume integration in the @xmath70- and @xmath71-directions . furthermore , the wave energy , @xmath66 , given by equation ( [ 2.12 ] ) can be expressed as @xmath145 where @xmath146 is a dimensionless quantity given by is different from that given in the previous paper ( kato 2008a ) , since the previous one is not suitable to represent the sign of wave energy . ] @xmath147 and the subscript c denotes the values at resonant radius , which is defined by @xmath148 or @xmath149 ( see below ) . it is noted that in the case of @xmath150 , @xmath151 is zero . before examining general cases of disks with pressure , we study here the limiting case of pressure - less disks , i.e. , @xmath152 and @xmath153 . equations ( [ 4.1 ] ) and ( [ 4.3 ] ) show that @xmath154 ( and @xmath155 ) and @xmath156 respond resonantly , respectively , at the radii where @xmath157 holds and at the radii where @xmath158 holds . we define here @xmath159 and @xmath160 by @xmath161 we call the resonance at @xmath148 horizontal resonances ( lindblad resonances ) and the resonance at @xmath149 vertical resonances . we consider these two resonances separately . as a typical example in infinitesimally thin disks , we consider here the case where all motions including those associated with disk deformations are in the equatorial plane , i.e. , @xmath150 , @xmath111 , and @xmath162 . , @xmath163 , and @xmath136 in general without specifying to particular values . ] at the radii of @xmath148 , @xmath154 and @xmath155 resonantly respond to @xmath125 s as [ see equations ( [ 4.1 ] ) and ( [ 4.2 ] ) ] @xmath164 @xmath165 . \label{3.2}\ ] ] the quantity @xmath156 can be taken to be zero , since it does not respond resonantly and has no contribution to the final results . to examine the resonant interaction at the radii of @xmath148 , we introduce a small imaginary part of @xmath30 , i.e. , @xmath56 , as @xmath166 , where the imaginary part , @xmath56 , is tentatively assumed to be negative so that causality is satisfied . later , the results obtained by @xmath167 is extended analytically to the whole region of @xmath56 as usually done in stability analyses . hereafter , the real part , @xmath168 , is written by @xmath30 or @xmath64 , without confusion . near to the resonance , @xmath169 is expanded as @xmath170 , \label{3.4}\ ] ] where the subscript c represents the values at the resonant radii , and @xmath171 is @xmath172_{\rm c}. \label{3.5}\ ] ] let us now consider an integration of @xmath173 , @xmath174 being an arbitrary smooth real function , along the radial direction including the resonant region . the imaginary part of the integration comes from the path near to the pole . thus , we have @xmath175 this can be summarized as @xmath176 where sign@xmath177 is the sign of @xmath178 at the resonant radii . let us now estimate the rate of work done by the resonance on the original oscillations by using equation ( [ 3.6 ] ) . in the case of coupling through @xmath179 , the work is given by equations ( [ w+ ] ) and ( [ w- ] ) . since @xmath180 and @xmath181 are given by equations ( [ 3.1 ] ) and ( [ 3.2 ] ) , respectively , we have from equation ( [ w+ ] ) and ( [ w- ] ) @xmath182 where we have used @xmath183 , and subscript @xmath77 is attached to @xmath56 in order to emphasize that this is a case of horizontal resonances . hence , the growth rate , @xmath184 [ see equation ( [ growth ] ) ] , can be expressed as @xmath185 it is noted that the above expression for growth rate is valid even when @xmath186 . equations ( [ 3.7 ] ) ( [ 5.30 ] ) formally hold even in the case with pressure , as shown in the next section . the vertical resonance might be unrealistic in pressure - less disks , but we briefly summarize the formal results here , since the expression for growth rate is applicable even in the case of disks with pressure , as shown in the next section . we now expand @xmath187 around @xmath149 as @xmath188 , \label{5.32}\ ] ] where @xmath189_{\rm c } , \label{5.33}\ ] ] and @xmath190 is the resonant radius of @xmath149 . in the above two cases of the horizontal and vertical resonances , the resonant radii are different each other . however , the same notation , @xmath190 , is used here and hereafter without confusion . after this preparation , we calculate the rate of work done on oscillations by the same procedures as those in the case of horizontal resonance . the results show that the growth rate is given by @xmath191 equations describing the intermediate oscillations [ equations ( [ 4.1 ] ) ( [ 4.3 ] ) or equations ( [ 4.1 ] ) and ( [ 4.3 ] ) ] are now solved without assuming @xmath152 and @xmath153 . since we are interested in behaviors of @xmath192 in a region close to the resonant point of @xmath148 or @xmath149 , all the coefficients of @xmath192 in equations ( [ 4.1 ] ) ( [ 4.3 ] ) are assumed to be constant except for @xmath169 or @xmath187 . that is , in the case of the horizontal resonance ( lindblad resonance ) , @xmath193 , and @xmath169 is taken as [ see equations ( [ 3.4 ] ) and ( [ 3.5 ] ) ] @xmath194 where @xmath195 is the radial distance from the resonant radius , @xmath190 , i.e. , @xmath196 . in the case of the vertical resonance , on the other hand , @xmath197 , and @xmath187 is taken as [ see equations ( [ 5.32 ] ) and ( [ 5.33 ] ) ] @xmath198 where @xmath196 . the resonant radii in the horizontal and vertical resonances are different each other , but the same notation , @xmath190 , has been adopted here without confusion . in disks with pressure , different from the case of pressure - less disks , the resonant region is not narrow . that is , the resonant region is widened by the effects of pressure ( see below ) , and thus the coupling terms , @xmath199 , @xmath200 , and @xmath201 , can be no longer regarded to be spatially constant in general . considering these situations , we introduce fourier transforms of @xmath202 and @xmath203 as @xmath204 @xmath205 after these preparations , to solve the set of equations ( [ 4.1 ] ) ( [ 4.3 ] ) , we apply a method of fourier transform adopted by meyer - vernet and sicardy ( 1987 ) in their study of resonant disk - satellite interaction . here and hereafter , such subscripts as @xmath58 and @xmath136 to be attached to @xmath192 , @xmath206 , @xmath207 , and @xmath208 are sometimes omitted in order to avoid unnecessary complications , unless they are explicitly necessary . in the present case , the fourier transform of equations ( [ 4.1 ] ) and ( [ 4.3 ] ) give , respectively , @xmath209 and @xmath210 where @xmath211 . eliminating @xmath212 from equation ( [ 4.4 ] ) and ( [ 4.5 ] ) , we have an inhomogeneous differential equation of @xmath213 with respect to @xmath2 : @xmath214 where @xmath215 is defined by @xmath216 let us first consider the case of @xmath217 . equation ( [ 4.6 ] ) is then solved to lead @xmath218 , \label{4.8}\ ] ] where @xmath219dk ' . \label{4.8a}\ ] ] in determining the integration range of equation ( [ 4.8a ] ) , we have used a boundary condition allowing @xmath220 for @xmath221 . this boundary condition is required , since @xmath213 is the fourier transform of @xmath222 . by using the inverse fourier transform : @xmath223 we can express @xmath222 in an integration form as @xmath224dk , \label{xi - r - again}\ ] ] where @xmath225 is defined by @xmath226 since @xmath227 and @xmath228 are related to @xmath229 by [ see equations ( [ 4.2 ] ) and ( [ 4.3 ] ) ] @xmath230 the work integral , @xmath231 , in the case of the coupling through @xmath232 , is written as , using equation ( [ w+ ] ) , @xmath233 d_{(+)}(k ) { \rm exp}\biggr[-\frac{\epsilon}{j_1'}k+i\biggr(kx-\frac{1}{3}\alpha_{\rm p}^3k^3\biggr ) \biggr ] \nonumber \\ & & \times\biggr[{\breve a}_{r,+}^*(x)+i\frac{2\omega}{\omega-{\tilde m } \omega}{\breve a}_{\varphi,+}^*(x ) + \frac{1}{j_2}ik\omega_\bot^2h{\breve a}_{z,+}^*(x)\biggr ] , \label{w+2}\end{aligned}\ ] ] where real parts on the right - hand side of equation ( [ w+2 ] ) have been omitted . in the case of @xmath234 , the expression for @xmath235 is slightly changed from equation ( [ 4.8 ] ) as @xmath236 , \label{4.8'}\ ] ] where @xmath237 is given by @xmath238dk ' , \label{4.8a'}\ ] ] where @xmath239 is defined by equation ( [ 4.7 ] ) . the integration range in equation ( [ 4.8a ] ) is changed from equation ( [ 4.8a ] ) so that the boundary condition @xmath240 is satisfied at @xmath241 . by performing the same procedures as those in the case of @xmath217 , we can derive an expression for @xmath231 . an expression for @xmath231 in this case is the same as equation ( [ w+2 ] ) , except that @xmath242 is now changed to @xmath243 , i.e. , @xmath233 d_{(-)}(k ) { \rm exp}\biggr[-\frac{\epsilon}{j_1'}k+i\biggr(kx-\frac{1}{3}\alpha_{\rm p}^3k^3\biggr ) \biggr ] \nonumber \\ & & \times\biggr[{\breve a}_{r,+}^*(x)+i\frac{2\omega}{\omega-{\tilde m } \omega}{\breve a}_{\varphi,+}^*(x ) + \frac{1}{j_2}ik\omega_\bot^2h{\breve a}_{z,+}^*(x)\biggr ] , \label{w+2a}\end{aligned}\ ] ] expressions of the work integral in the case of the coupling through @xmath244 can be derived by similar procedures as those in the case of the coupling through @xmath245 . in this case , the subscript + attached to @xmath246 s in equations ( [ w+2 ] ) and ( [ w+2a ] ) is changed to @xmath247 and @xmath248 . first , we consider the case where the coupling terms , @xmath199 , @xmath200 , and @xmath201 , can be regarded to be spatially constant in the resonant region . ( this is really realized in a particular case , as mentioned in the next section . ) an another purpose of examining the case of constant coupling terms is to have a rough image on the width of resonant region in disks with pressure . if @xmath249 , @xmath250 , and @xmath251 are spatially constant , their fourier transforms give @xmath252 in this case , @xmath253 given by equation ( [ 4.8a ] ) is reduced simply to @xmath254 , \label{d}\ ] ] where @xmath255 is the unit - step function , i.e. , @xmath256 for @xmath257 , while @xmath258 for @xmath259 . since @xmath253 does not involve @xmath195 , the integration with respect to @xmath195 in equation ( [ w+2 ] ) leads to @xmath260 . hence , performing the integration with respect to @xmath2 in equation ( [ w+2 ] ) , we have @xmath261 in the case of @xmath217 , where @xmath40 and @xmath262 , depending on the type of couplings and the form of deformation . in the case of @xmath234 , @xmath213 is given by equation ( [ 4.8 ] ) . by performing the same procedures as those in the case of @xmath217 , we can derive an expression for @xmath231 . the final expression for @xmath231 in this case has the opposite sign from equation ( [ w+3 ] ) . inequality @xmath217 implies that @xmath263 , since @xmath184 should be taken tentatively to be negative as mentioned before , while inequality @xmath264 means @xmath263 . considering this difference , we can summarize @xmath231 in the above two cases of @xmath217 and @xmath234 as @xmath265 where @xmath40 . in the case of coupling through @xmath266 , we have , after the similar processes as the above , @xmath267 where @xmath248 . finally , from @xmath268 given above and the wave energy , @xmath66 , given by equation ( [ 2.12 ] ) , we find that the growth rate , equation ( [ 5.28 ] ) , is written as @xmath269 which is identical to equation ( [ 5.30 ] ) , and also equal to the growth rate derived in the previous paper ( kato 2008a ) , as will be mentioned at the end of this section . the arguments in subsection 4.1.1 show that in the case where @xmath249 , @xmath250 , and @xmath251 are radially constant , the radial structure of resonant region is determined by [ see equation ( [ w+2 ] ) and notice that @xmath253 has the step - function ] @xmath270,\ ] ] where the term with @xmath271 can be practically neglected . the @xmath195-dependence of this integral has been examined by meyer - vernet and sicardy ( 1987 ) [ see equation ( 41 ) and fig , 5 in their paper ] . they show that the half width of the resonant region is @xmath272 , which is on the order of @xmath273 , i.e. , @xmath274 , in the present case . in general , @xmath249 , @xmath250 , and @xmath251 radially oscillate with the wavelength determined by the product of @xmath275 and @xmath276 . hence , in some cases the scale - length of the variation of @xmath277 s will be shorter than @xmath274 , although it depends on where the resonance occurs ( see the next section ) . to examine general cases , numerical calculations are needed . however , in order to have a rough image of the effects of spatial variation of @xmath249 , @xmath250 , and @xmath251 on the work integral @xmath268 , we consider here a limiting case where @xmath249 , @xmath250 , and @xmath251 oscillate periodically in the radial direction . that is , we assume @xmath278 where @xmath279 s and @xmath280 are the amplitudes and wavenumber of the spatial variations . then , the fourier transform defined by equation ( [ 4.2b ] ) show that @xmath281 let us consider the case of @xmath217 . in this case , @xmath253 given by equation ( [ 4.8a ] ) is reduced to [ cf . , equation ( [ d ] ) ] @xmath282 { \rm exp}\biggr[\frac{\epsilon}{j_1'}k_0+i\frac{\alpha_{\rm p}^3}{3}k_0 ^ 3\biggr ] . \label{newd}\ ] ] since @xmath249 , @xmath250 , and @xmath251 are given by equation ( [ periodic ] ) , performance of integration with respect to @xmath195 in equation ( [ w+2 ] ) gives @xmath283 . next , we perform the integration with respect to @xmath2 . since @xmath253 is now given by equation ( [ newd ] ) , we have @xmath284 similar arguments can be made in the case of @xmath234 . in this case , @xmath231 has the sign opposite to equation ( [ w+4 ] ) . expressions for @xmath231 in two cases of @xmath217 and @xmath234 can be combined into a single form as before . similarly , we can consider the case where the coupling occurs through @xmath285 . all of these results can be summarized as @xmath286 hence , the growth rate is written as @xmath287 in the case of the vertical resonance , from equations ( [ 4.1 ] ) and ( [ 4.3 ] ) , we have , respectively , @xmath288 and @xmath289 eliminating @xmath290 from the above two equations , we have an inhomogeneous differential equation with respect to @xmath291 as @xmath292 , \label{v4}\ ] ] where @xmath293 in deriving equation ( [ v4 ] ) , in order to avoid unnecessary complications , we have neglected @xmath294 compared with @xmath169 , since the magnitude of @xmath295 contributing to the present formulations is smaller than unity , i.e. , @xmath296 . in the case of @xmath297 , equation ( [ v4 ] ) is solved as @xmath298 , \label{v6}\ ] ] where the coefficient @xmath299 is given by @xmath300 \nonumber \\ & & \times { \rm exp } \biggr[\frac{\epsilon}{j_2'}k'+i{1\over 3}\alpha_{\rm p , v}^3k'^3\biggr]dk ' . \label{v7}\end{aligned}\ ] ] by using @xmath212 given above , we can write @xmath290 and @xmath301 as [ see equations ( [ 4.1 ] ) and ( [ 4.2 ] ) ] , @xmath302 and @xmath303 substitution of these relations into the work integral ( [ w+ ] ) gives @xmath304 d_{{\rm v},(+)}(k ) { \rm exp}\biggr[-\frac{\epsilon}{j_2'}k+ i\biggr(kx-\frac{1}{3}\alpha_{\rm p , v}^3k^3\biggr)\biggr ] \nonumber \\ & & \times \biggr[-i\frac{ik\omega_\bot^2 h}{j_1 } \biggr({\bar a}_r^*+i\frac{2\omega}{\omega-{\tilde m}\omega}{\bar a}_\varphi^*\biggr ) + \frac{1}{{\tilde n}}{\bar a}_z^*\biggr ] . \label{v10}\end{aligned}\ ] ] in the case of @xmath305 , @xmath306 has a similar form , but we should use , instead of @xmath299 , @xmath307 , which is given by @xmath308 \nonumber \\ & & \times { \rm exp } \biggr[\frac{\epsilon}{j_2'}k'+i{1\over 3}\alpha_{\rm p , v}^3k'^3\biggr]dk ' . \label{v7'}\end{aligned}\ ] ] as in the subsection 4.1.1 , we consider two cases where i ) @xmath251 is constant in the resonant region and ii ) @xmath251 is spatially periodic as @xmath309 . in the former case , summing the two cases of @xmath310 and @xmath311 , we have @xmath312 and in the latter case we have @xmath313 in the coupling through @xmath314 , the work integral , @xmath315 , has the same expression as equations ( [ v11a ] ) and ( [ v11 ] ) , except that the subscript @xmath124 is now changed to @xmath247 . after the above considerations , we summarize the growth rate , @xmath316 , in the former case of @xmath201 being constant as @xmath317 in the case where @xmath251 changes as @xmath318 , we have @xmath319 finally , we note here that the growth rates obtained in the previous papers are the same as those obtained in subsection 4.1.1 , although the final expression in the previous papers are different from those in subsection 4.1.1 . in the case of horizontal resonance , for example , @xmath320 in kato ( 2008a ) is equal to @xmath321 in the present paper . furthermore , if @xmath322 and @xmath323 are eliminated from equations ( 46 ) ( 48 ) in kato ( 2008a ) , we have @xmath324 where @xmath325 is @xmath326 . this expression for @xmath229 shows that @xmath327 defined by equation ( 52 ) in kato ( 2008a ) is written as @xmath328 then , substituting @xmath329 and equation ( [ * * ] ) with @xmath330 into equation ( 70 ) in kato ( 2008a ) , we obtain an expression for growth rate in the case of horizontal resonances , which is identical with equation ( [ 4.19 ] ) in subsection 4.1.1 . similar arguments easily show that the growth rate given by kato ( 2008a ) for vertical resonances , i.e. , equation ( 71 ) in kato ( 2008a ) , is identical with that in section 4.2 , i.e. , equation ( [ v11 ] ) . in this paper we have derived a stability criterion on resonant excitation of disk oscillations in a deformed disk . deformation of disks that can lead to excitation of disk oscillations are a warp and an eccentric deformation of disk plane , i.e. , @xmath331 , although this is not discussed in this paper ( see kato 2004 , 2008a ) . in previous papers ( kato 2004 , 2008a ) , the above resonant excitation problem was examined and a condition of excitation of disk oscillations was derived . from mathematical point of view , however , the treatment of disk oscillations around the resonant point was not suitable in the previous papers . in this paper , we corrected the inadequate points . the results of the correction show that the stability condition and the growth rate derived in the previous papers are kept qualitatively unchanged . in this paper , however , the growth rate of oscillations is written in a form applicable more easily to numerical calculations of growth rate . in disks with pressure , the resonant region is widened by pressure effects . based on arguments by meyer - vernet and sicardy ( 1987 ) , we showed in section 4 that the half - width of the resonant region is @xmath332 ( or @xmath333 ) , which is shorter than @xmath69 but longer than @xmath77 . the coupling terms comes from nonlinear products of original oscillation , @xmath334 , and disk deformation , @xmath335 . if the disk deformation is assumed to be global , the scale of spatial variation of the coupling terms is mainly governed by the spatial behavior of the original oscillation , @xmath334 . hence , in general , the width of resonant region is comparable with the wavelength of the original oscillations and may not be narrow enough to regard the coupling terms @xmath125 s as constant in the region . hence , in section 4 , we examined two limiting cases where i ) the coupling terms are constant in the resonant region and ii ) they vary sinusoidally in the region . analyses in section 4 show that in these two limiting cases there are no essential differences in growth rate , especially no difference in the criterion of stability . to examine what happens in intermediate cases of the above two limiting ones , however , numerical calculations will be needed . it is noted that the case where the coupling terms , @xmath203 , are roughly constant in the resonant region is realized when resonance occurs near to the boundary between the propagation and evanescent regions of the the original oscillations . this is because near to the boundary , the radial wavelength of the oscillations is long . applications of the present disk oscillation model to high - frequency qpos by kato and fukue ( 2006 ) and kato ( 2008a ) are concerned in such cases . the another limiting case where the coupling terms vary in the radial direction will be realized when resonance occurs in the propagation region far apart from the boundary between the propagation and evanescent regions . there are two kinds of resonances , i.e. , horizontal and vertical resonances . in both resonances , the condition of excitation of oscillations is that the wave energy of the original oscillation , @xmath146 , and @xmath178 at the resonant radius have the opposite signs , i.e. , @xmath336 . for the resonance to occur efficiently , the resonant radius must be within the region where both original and intermediate oscillations exists predominantly . hence , it will be reasonable to suppose that the sign of @xmath177 roughly represents the sign of the energy of the intermediate oscillation . [ notice that the sign of wave energy of the intermediate oscillation is determined by the sign of @xmath178 in the region where the wave exists predominantly , see equations ( [ 5.28 ] ) and ( [ 5.29 ] ) ] . in other words , the condition of @xmath337 is roughly equal to the condition that the original and intermediate oscillations have energies of the opposite signs . this instability criterion thus can be understand as a result of energy exchange between the original and intermediate oscillations at the resonant point ( ferreira and ogilvie 2008 ) . the resonant instability condition , i.e. , @xmath338 , however , is not exactly equal to the condition of both oscillaions ( original and intermediate one ) having the opposite signs . this suggests that the energy exchange at the resonant point is not only between the above two oscillations , but the disk rotation also plays a part and gains or losses its energy . in the case of @xmath339 , the original oscillation is damped . roughly speaking , in this case , both oscillations ( original and intermediate ones ) have the same signs in their wave energy . hence , in this case , the main energy flow will be between disk rotation and the oscillations so that the oscillations are damped . it is noticed that in vertical resonance we always have @xmath339 in the case where a resonant radius can appear in disk plane and the oscillations are always damped ( see kato 2004 , 2008a ) . finally , it is noted that the present resonant excitation mechanism has been applied to quasi - periodic oscillations observed in black - hole and neutron - star x - ray binaries , for example , by kato and fukue ( 2006 ) , kato ( 2008a ) and kato et al .